The Essential Can Du

THE
ESSENTIAL CANDU
A Textbook on the CANDU Nuclear Power Plant Technology
The Essential CANDU
A textbook on the CANDU Nuclear Power Plant Technology
i
Blank page
ii
The Essential CANDU
A textbook on the CANDU Nuclear Power Plant Technology
CANDU (short for CANada Deuterium Uranium)

www.unene.ca/education/candu-textbook
Editor-in-Chief: Wm. J. Garland, Professor Emeritus, McMaster University, Hamilton, Ontario
iii
Copyright ©2014 by UNENE
All rights reserved.
Fair Use: The textbook is provided free of charge in pdf format and is intended for education
and training purposes provided full attribution is given. It is provided in good faith as an aid to
CANDU awareness, education and training.
Citations: Since the contents of this web-based textbook are subject to change without notice, it
is important that the date of retrieval be included in the citation. To cite the textbook, the
following form is suggested:
The Essential CANDU, A Textbook on the CANDU Nuclear Power Plant Technology, Edi-
tor-in-Chief Wm. J. Garland, <chapter, page, etc, as appropriate>, University Network of
Excellence in Nuclear Engineering (UNENE), ISBN 0-9730040. Retrieved from
https://www.unene.ca/education/candu-textbook on <insert date of retrieval>.
Reproduction or translation of any part of this work beyond that permitted by The Canadian
Copyright Act without the permission of the copyright owner is unlawful. Requests for permis-
sion or further information should be addressed to
UNENE,
c/o Department of Engineering Physics, Bldg. JHE A315
McMaster University
Hamilton, Ontario, CANADA L8S 4L7
Tel: (905) 525-9140 ext. 20168
URL: http://www.unene.ca
Email: unene@mcmaster.ca
Library of Congress Cataloging in Publication data:

UNENE
The Essential CANDU.
Includes bibliographical references.

Nuclear reactors – Design and construction
Nuclear Engineering
Nuclear fission
ISBN 0-9730040
First Edition
Printed in Canada
iv
Dedication
This textbook is dedicated to the students of CANDU everywhere.
Disclaimer
UNENE assumes no liability for use or misuse of the textbook contents. While due care is taken
to ensure the contents are accurate, UNENE does not make any warranty, expressed or implied,
or assume any legal liability or responsibility for the accuracy, completeness, or usefulness of
any information contained herein.
Table of Contents
Volume 1
Preface
Acknowledgments
Prologue - CANDU in Context – Dr. William J. Garland
Chapter 1 - Introduction to Nuclear Reactors - Dr. Robin Chaplin
Chapter 2 - Genealogy of CANDU Reactors - Dr. Robin Chaplin
Chapter 3 - Nuclear Processes and Neutron Physics - Dr. Guy Marleau
Chapter 4 - Reactor Statics - Dr. Benjamin Rouben and Dr. Eleodor Nichita
Chapter 5 - Reactor Dynamics - Dr. Eleodor Nichita and Dr. Benjamin Rouben
Chapter 6 - Thermalhydraulic Design - Dr. Nikola K. Popov
Chapter 7 - Thermalhydraulic Analysis - Dr. William J. Garland
Chapter 8 - Nuclear Plant Systems - Dr. Robin Chaplin
Chapter 9 - Nuclear Plant Operation - Dr. Robin Chaplin
Chapter 10 - Instrumentation and Control - Dr. G. Alan Hepburn
Chapter 11 - Electrical Systems - Dr. Jin Jiang
Author Biographies
Abbreviations
Volume 2
Preface
Acknowledgments
Chapter 12 - Radiation Protection and Environmental Safety - Dr. Edward Waller
Chapter 13 - Reactor Safety Design and Safety Analysis - Dr. Victor G. Snell
Chapter 14 - Nuclear Plant Materials and Corrosion - Dr. Derek H. Lister and Dr. William G. Cook
Chapter 15 - Chemistry in CANDU Process Systems - Dr. William G. Cook and Dr. Derek H. Lister
Chapter 16 - Regulatory Requirements and Licensing – Dr. Victor G. Snell and Dr Nikola K. Popov
Chapter 17 - Fuel - Mr. Mukesh Tayal and Mr. Milan Gacesa
Chapter 18 - Fuel Cycles - Mr. Mukesh Tayal and Mr. Milan Gacesa
Chapter 19 - Storage and Disposal of Irradiated Fuel - Mr. Milan Gacesa and Mr. Mukesh Tayal
Chapter 20 - Fuel Handling an Storage- Ms. Diane Damario
Chapter 21 - CANDU In-Core Fuel Management - Dr. Benjamin Rouben
Author Biographies
Abbreviations
v
Preface
There has long been a need for a CANDU1 nuclear power plant textbook suitable for students,
educators, trainers and working professionals at a target level of senior undergraduate year
university engineering and science. The aim is to provide a concise, and consistently told, root
storyline that will enable those new to CANDU, whether a student, a manager, a teacher /
trainer, a journalist or a discipline area specialist, to learn about CANDU as an overall system
and to delve into any specific area as desired. This supports the agile deployment of personnel
into the industry and redeployment throughout the industry. Early on it was decided that the
textbook would be public domain available online in pdf form under the UNENE banner along
with associated prerequisite material and further reading. This web material can be copied to
CDs for distribution or printed in book form, yet is easily updated and expanded. It is to be a
living document appearing in the form of editions for revision tracking purposes.
The Essential CANDU is a textbook about nuclear science and engineering pertaining to CANDU
reactors using CANDU as the reference design. This distinguishes it from comparable textbooks
based on PWR reactors. This is not a product description of CANDU per se. The book will
emphasize theory over other aspects of nuclear science.
This is a textbook about CANDU nuclear reactor and power plant engineering, not just the
reactor physics of the core. While it is true that the reactor is certainly the distinguishing
feature of the plant, it is hardly the whole story. Within the Nuclear Steam Supply System
(NSSS) there are some 225 systems, but only a few can be covered explicitly in the book. This is
a book about the nuclear and related bits that constitute the key bits that define nuclear
engineering, not the many associated disciplines per se.
CANDU 6 was chosen as the reference design. Specifically it was agreed that Pt. Lepreau (post
refurbishment) would be used as the default reference design. This does not preclude refer-
ence to multi-unit designs, single-unit variants and future designs as appropriate.
It is the root story line that glues a textbook together and guides what material is to be empha-
sized or deemphasized. The steam generator (and the associated heat duty diagram) is the
central node of the station. It is the steam generator that dominates the overall plant dynamics.
The primary side temperature floats on top of the steam drum temperature with just enough ΔT
to transfer the heat from the primary side to the secondary side as discussed in Chapter 6. If
the drum pressure (hence temperature) jumps, the primary side jumps to suit, affecting primary
temps, fuel temps, margins, reactivity feedbacks, etc. But clearly the overall system behavior is
very much dependent on the characteristics and limits of the various sub-systems and compo-
nents. In short, reactor physics – the inevitable focus of typical reactor texts – is important to
be sure and is covered herein. But there is so much more to a nuclear plant than reactor
physics that we would be remiss in telling the CANDU story if we were to overlook the broader
1 CANDU (short for CANada Deuterium Uranium)
vi
picture.
There are a number of existing documents and products that cover significant parts of the
CANDU story. UNENE itself hosts a dozen or so graduate level courses that are significant
resources for preparing a CANDU textbook. See www.unene.ca for details on those courses.
And of course there are many other existing university nuclear and nuclear related courses in
Canada and elsewhere (see for instance http://nuclearcanada.ca). The CANTEACH repository at
http://canteach.candu.org contains a large number of legacy CANDU documents, albeit largely
descriptive. Significant training, design, analysis and operation documents exist within AECL
(and now Candu Energy) , Ontario Power Generation, Bruce Power, the CNSC and other organi-
zations comprising the Canadian nuclear enterprise. The Essential CANDU, with its educational
focus is complementary to all that.
As Editor-in-Chief it has been my pleasure to work with the authors as they laboured well
beyond expectations. I have worked for years on course development with Ben Rouben and
Dorin Nichita on reactor physics, Nik Popov on thermalhydraulics, Victor Snell on safety, and
George Bereznai on plant systems and operations. It is truly satisfying to see the fruits of that
labour appearing in the form of this text. On a more administrative level I have worked with
Robin Chaplin (thermodynamics), Derick Lister and Willy Cook (Chemistry and Corrosion), Jin
Jiang (Electrical) as they developed and delivered courses for UNENE. From its inception in
2000, UNENE has been a world class and world leading nuclear network. The aforementioned
people are part of the UNENE team. Their research and educational contributions are a big part
of that success. Guy Marleau (reactor physics), Al Hepburn (I&C), Ed Waller (Health Physics),
Milan Gacesa and Mukesh Tayal (Fuel topics), and Ms. Diane Damario (Fuel Handling) are all
well known in their areas. It was a pleasure to work with them on this project. Old and new
colleagues and friends - it is an honour to have worked with them all.
Of course it is not just the authors that deserve special mention. As they would readily attest, it
took many creative and hard working professionals to build our Canadian Nuclear Enterprise –
generations of them going all the way back to the early roots of CANDU. They are the past and
present colleagues, professors, managers, researchers, students and others that form our
foundation and upon whose work we are but modest reporters. Two particular subsets of those
are the reviewers of this text and the Working Group for this project. Scholars and authors in
their own right, they have overseen and guided the development of this work. It was a com-
forting ‘defense in depth’. The reviewers are acknowledged at the end of each chapter. We
wish to sincerely thank the following Working Group members for their guidance and technical
oversight:
• George Bereznai (University of Ontario Institute of Technology)

• Basma Shalaby (UNENE)
• John Luxat (McMaster University)
• Mahesh Pandey (University of Waterloo)
• Jatin Nathwani (University of Waterloo)
• Esam Hussein (University of Regina)
I have worked with each of the Working Group members as colleagues on and off over the years
going back to the 70’s. George, Basma, others and I bemoaned the lack of a comprehensive
vii
CANDU textbook from the beginning. George and I even mapped out our vast plans (with half-
vast ideas) and did manage over time to guide a number of courses through to completion, as
did other professors. The interactions and the innumerable chats with these and many other
colleagues through the years collectively set the backdrop for this textbook. But it was Basma,
when stepping in as President of UNENE as I stepped out, who said: “Come on Bill. Let’s go.”
And so, we did. Thanks Basma for that final prod.
This textbook, as large and comprehensive as it is, is far from complete. Each topic is a book,
perhaps several books, on its own. I hope it at least encourages others to build on this modest
start. It is worth noting too that the written word, try as we might, does not capture well the
tacit knowledge, the working knowledge, that the working professionals have. This textbook is
but a humble complement to that. The very process of creating this textbook, however, creates
a learning environment for the authors and we found a considerable sharing of expertise as we
prepared the chapters. This will apply to future authors as well. So, “Come on, let’s go.”
Finally, what can we say of our long suffering families but a truly heartfelt thanks. We shall be
forever in your debt. Thank you all.
Bill Garland
Professor Emeritus
Department of Engineering Physics
McMaster University
viii
Acknowledgments
We wish to sincerely thank the following Steering Committee member organizations for their
financial support and access to technical information, without which this project would not
have been possible:
Ontario Power Generation (OPG)

Bruce Power Inc. (BP)
Atomic Energy of Canada Limited (AECL-CRL)
Korea Hydro & Nuclear Power Company (KHNP)
Societatea Nationala “Nuclearelectrica” – S.A. (SNN)
Third Qinshan Nuclear Power Company (TQNPC)
AMEC-Nuclear Safety Solutions Limited (AMEC-NSS)
Canadian Nuclear Safety Commission (CNSC)
Canadian Nuclear Society (CNS)
Hydro Quebec (HQ)
Kinectrics (Kinectrics)
New Brunswick Power Nuclear (NBPN)
SNC-Lavalin (SNC-Lavalin)
University Network of Excellence in Nuclear Engineering (UNENE)
We wish to acknowledge and thank CANDU Owners Group Inc. (COG), especially Henry Chan
and John Sowagi, for administrative support. Thank you Henry for stepping up to the plate and
securing sponsorship.
ix
x
1
PROLOGUE
CANDU in Context
prepared by
Dr. William J. Garland, Professor Emeritus,
Department of Engineering Physics,
McMaster University, Hamilton, Ontario, Canada
Summary:
Herein, we introduce the CANDU reactor by first looking at the broad social context to see why
we need nuclear power. Then we take a look at the Nuclear Reactor in a nutshell, giving a quick
overview to provide some context for the details found in this textbook.
Table of Contents
1 Energy in Society..................................................................................................................... 2
2 First some basics ..................................................................................................................... 3
2.1 Fission.............................................................................................................................. 3
2.2 The fuel: the source of energy by the fission process ................................................... 4
2.3 The moderator: slowing down those speedy neutrons................................................. 4
2.4 The coolant: to take away the heat generated by fissioning .......................................... 6
2.5 Control: staying within desired and safe limits of power ............................................... 7
2.6 Shielding: providing protection from radiation .............................................................. 8
2.7 The system that pulls it all together ............................................................................... 9
3 Textbook Organization .......................................................................................................... 10
4 Further reading ..................................................................................................................... 11
List of Figures
Figure 1 Fission................................................................................................................................ 3
Figure 2 Nuclear reactor components ............................................................................................ 5
Figure 3 A Conceptual CANDU ........................................................................................................ 6
Figure 4 Basic power reactor schematic ......................................................................................... 7
Figure 5 The overall CANDU nuclear plant...................................................................................... 9
©UNENE, all rights reserved. For educational use only, no assumed liability. CANDU in Context – September 2014
2 The Essential CANDU
1 Energy in Society
Let's step back to take a glimpse of the context into which CANDU must be placed. The simple
fact is that we are here. And there are a lot of us - some 7 billion in the year 2013 and our
population is destined to hit 10 billion in a few generations. On the premise that life is worth
living (and most of us think that way, otherwise our population wouldn't be growing), it follows
that we should make the best of the situation. Quality of life is, thus, a worthy and meaningful
pursuit. To achieve and maintain a reasonable quality of life requires energy. Access to energy is
an enabling force, empowering society and individuals. In short, it is fundamental to our exis-
tence. Energy is not a panacea for the strife of life, to be sure; but without it, there would be no
life at all.
One has only to reflect on early societies to appreciate the central role that access to energy has
played in our development. Food and water are essential commodities for the human race. To
supply these in adequate quantities to society has always been a challenge. Ancient civilizations
have met this challenge with remarkable feats of engineering such as those at Angkor Wat and
Machu Picchu. The Roman aqueducts such as the Pont du Gard also demonstrate significant
ingenuity to provide a single commodity. Most of these were built using human power only,
which limited the rate of implementation and hence development. This changed with the
industrial revolution when first water power and then steam power became a source of energy
to assist in the supply of food and water as well as clothing and transport. This supply of energy
enabled society to implement new ideas and advance technology at a greater rate and to
sustain an increase in population. Thus the human race became dependent upon another
essential commodity, namely energy.
It is well established that for the prosperity of any nation, a plentiful supply of cheap energy is
the key to economic growth. Countries where energy in some form or another is readily availa-
ble have become wealthy. Others have had to use their ingenuity to provide this resource.
Over time natural resources become depleted and new resources have to be found or devel-
oped. Some key resources which have sustained society are wood, coal, oil and now uranium.
Although uranium may not generally be considered an essential resource some countries
currently cannot do without it.
A suitable environment is an important factor for human survival. People are adaptable and
various sectors of humanity have managed to live in all parts of the world from the frigid Arctic
to the burning deserts. Hardships in these regions can be alleviated with a supply of energy
enabling increased development and access to more resources. In essence, an artificial envi-
ronment has been created in which to live. The human race is becoming more and more
dependent upon such an artificial environment or inner environment which includes heating
and air conditioning and promotes productivity. This requires increasing dependence on
energy. On the other hand there is the outer environment into which waste products from the
inner environment and energy production are dumped. No matter how good the inner envi-
ronment the outer environment still needs protection to sustain humanity.
The development of technology has put increasing stress on the outer environment but has also
provided the means to protect it. In the generation of power and the use of energy through
fossil fuels, various solid pollutants such as fly ash and gaseous pollutants such as sulphur
dioxide and nitrogen oxide are produced. The former have been significantly reduced by the
CANDU in Context 3
use of electrostatic precipitators and fabric filters in coal fired plants and the latter by the
implementation of alternative fuels to eliminate smog and the development of on board diag-
nostic systems to reduce automobile emissions. Nevertheless, the outer environment is ad-
versely impacted by the extensive use of fossil fuels. The only practical and economic way that
these effects can be reduced is by burning less hydrocarbon fuel.
This problem of energy use and power production can be overcome by using electrical energy to
replace hydrocarbon fuels and generating this energy by water and wind as well as by the
nuclear fission of uranium. Most sites for hydro power have already been developed, and
future development in this area is limited by possible detrimental local environmental effects.
Wind power is intermittent and variable so that, beyond a certain capacity relative to the grid
capacity, the electrical grid becomes unstable. Moreover, reserve capacity has to be provided
for the loss of wind power. This leaves nuclear power as the only choice for long term base load
energy production with minimal impact on the environment. There are no carbon dioxide
emissions from nuclear power, and a relatively small amount of radioactive waste is produced.
By recycling this waste the very long lived transuranic products can be returned to the reactor
to be destroyed leaving only the shorter life fission products to be stored. A huge advantage of
nuclear energy is the vast amount of energy obtained from a very small amount of uranium.
Nuclear power is the only existing option for large scale power production that transcends the
limitations of nonrenewable alternatives (such as coal, oil and gas) and renewable alternatives
(wind, solar and biomass). To be sure, there are many local, national and international issues
that flavour the ultimate choice of energy source, but nuclear should not be dismissed. The
consequences would be dire.
We conclude, then, that nuclear should be part of the energy mix now and in the future, that is,
we have a functional requirement for nuclear energy.
2 First some basics

2.1 Fission
To make sense of nuclear reactor design in Fission ~200 MeV in kinetic energy
general, and CANDU design in particular,
the reader needs to have some familiarity
with a few key nuclear concepts and 2 -3
phenomena. In a nutshell, slow neutrons fast neutrons
(called thermal neutrons) can initiate a

fission of uranium 235 (U-235), an isotope Fission Products
of uranium that occurs in nature. This is thermal neutron

illustrated in Figure 1. Natural uranium U235 nucleus
that is mined from the ground is 0.7% U-

235 and 99.3% U-238. The results of fission
are fission products that are radioactive, Before After
gamma radiation, fast (or energetic)

neutrons and heat. The fast neutrons have a
low probability of inducing further fissions, Figure 1 Fission
and hence have a low probability of generat-
ing more neutrons and thus sustaining a chain reaction. So the neutrons need to be slowed
down (i.e., thermalized or moderated), which is done by using a moderator such as water. The
heat generated needs to be removed. The process is controlled by controlling the number of
neutrons since the number of fissions per second (and hence the heat produced) is proportional
to the number of neutrons present to induce the fissions.
From this the basic functional requirements of a reactor are directly derived. Needed are:
• a fuel such as U-235
• a moderator to thermalize (i.e., slow down) the fast neutrons
• a coolant to remove the heat
• a control system to control the number of neutrons
• a shielding system to protect equipment and people from radiation
• a system that pulls all this together into a workable device.
In the following, these requirements are discussed in turn to gain some insight on how and why
CANDUs (and other reactor types) are built the way they are.
2.2 The fuel: the source of energy by the fission process
The probability of neutron capture leading to fission is larger for slow neutrons than for fast
neutrons. Hence, most practical reactors are "thermal" reactors, that is, they utilize the higher
thermal cross sections. Possible fuels include some of the various isotopes of uranium (U) and
plutonium (Pu). The only naturally occurring fuel with suitable properties of significant quanti-
ties is U-235, hence most reactors use this fuel.
Naturally occurring uranium is composed of 0.7% U-235. The rest is U-238. This percentage is
too low to sustain a chain reaction when combined with most practical moderators. Hence
either, the probability of fission must be enhanced or the moderator effectiveness must be
enhanced. One group of reactor types (PWR, BWR, AGR, RBMK, HTGR) enrich the fuel (a costly
task) and use a cheap moderator (ordinary water or graphite).
Alternatively, natural uranium (relatively cheap) is used with an excellent but expensive mod-
erator (heavy water). This is the CANDU approach. Which is better? There is no simple answer.
Both work. In engineered systems, there are always tradeoffs and the final design has to be
viewed in the overall context of the end-use environment.
2.3 The moderator: slowing down those speedy neutrons
The best moderator to slow down a speedy neutron is something that is the same size as the
neutron itself. This is true because if a neutron hit a massive target, it would just bounce off in a
different direction but with little loss in energy like a hard ball against a wall. If the neutron hit
an object much smaller that itself, it would just continue on virtually unaffected. But if it hit a
hydrogen atom, which is just a proton and an electron, that is almost exactly the mass of a
neutron, it could lose all its energy in one collision, just like in a game of billiards. However,
hydrogen does absorb neutrons as well which runs counter to the need to preserve these
precious neutrons so that they can cause fission. The deuterium isotope of hydrogen, at twice
the mass of hydrogen, is almost as good a slowing down agent but, since it already has an extra
CANDU in Context 5
neutron in the nucleus, it has a very low absorption probability. So, overall, deuterium is a far
better moderator than hydrogen. By using deuterium in the form of heavy water, natural
uranium can be used as a fuel. If ordinary water is used, the fuel must be enriched in U-235.
Other possible moderators include graphite or beryllium and gases such as carbon dioxide and
helium. A good moderator has a high scattering cross section, a low absorption cross section
and slows down the neutron in the least number of collisions. Figure 2 illustrates the fission
and moderation processes.
Coolant Contol Coolant

Outlet Rod Outlet
Neutron
absorbed
Neutron from
previous generation Neutrons to
Fission next generation
Fission
Fuel Rods Moderator Moderator Fuel Rods

Neutron
leaking
Figure 2 Nuclear reactor components
2.4 The coolant: to take away the heat generated by fissioning
The fissioning process generates energy, predominately in the form of kinetic energy of the
fission products which, after a few collisions with the immediately surrounding material matrix,
manifests itself as heat. A coolant (commonly) is passed over the fuel to remove this heat so
that it can be used productively for energy generation. So far, we have fuel, moderator and
coolant. We can conceptualize our CANDU as in the illustrations in Figure 3, below.
Figure 3 A Conceptual CANDU
CANDU in Context 7
We typically use the 'heat engine' process to turn this heat into a more useable (that is, flexible,
transportable, convenient, etc.) form of energy. As illustrated in Figure 4, this heat is used to boil
water and the resulting steam drives a turbine which drives an electrical generator. Electricity is
a very convenient form of energy - today it is so ubiquitous that it is hard to image life without
it.
Figure 4 Basic power reactor schematic
2.5 Control: staying within desired and safe limits of power
Control of the fissioning process is achieved most easily by simply adding or removing neutron
absorbers. Materials such as cadmium readily absorb neutrons and can be conveniently formed
into solid rods. So by having a number of these control rods partially inserted into the modera-
tor tank (also called the calandria) amongst the fuel and moderator in guide tubes, the neutron
population can be controlled.
Hence the fissioning process and the resultant heat output can be controlled. For safety’s sake,
in CANDUs, these control rods and the associated control system electronics and measurement
devices are built for reliable, fail-safe operation, are highly redundant, and employ additional
safety concepts such as group separation.
2.6 Shielding: providing protection from radiation
Uranium isotopes are not very radioactive by themselves and do not constitute a direct radia-
tion hazard. You can safely hold a fresh CANDU fuel bundle in your hands. Your biggest concern
would be dropping it on your toes since a typical bundle weighs about 20 kg or so. It is the
fissioning process that creates the nasty radioactive fission products. These ARE dangerous and
must be kept isolated from us. Radiation takes on a number of forms. Alpha and beta particles
are energetic charged particles that cannot penetrate solids to any significant degree. So as
long as the radioactive fission products are contained by a fuel sheath or some other pipe or
wall, there is no concern. Neutrons are not charged and can penetrate solid walls. We protect
ourselves from them by thick walls that slow down and absorb the neutrons. Combinations of
hydrogenous materials (like water and hydrocarbons) and absorbing materials (like boron and
cadmium) make good neutron shields. Gamma radiation, essentially very energetic photons
(ordinary light is low energy photons), are best stopped by dense material like lead and con-
crete. So constructing good shielding is not an onerous task, but it is an important one. Like in
the control systems, safety is enhanced by redundancy. In this case, this means layering the
shielding systems, one inside the other like a Russian doll set.
CANDU in Context 9
2.7 The system that pulls it all together
The various requirements related to fuel, moderation, cooling, control and shielding to conceive
a stylized CANDU are pulled together as illustrated in figure 5. At the heart of the plant is the
reactor core containing the fuel and the moderator. Heat generated there is transported away
by the cooling system to the conventional side of the plant (steam generator, turbine and
electrical generator).
Recall the layered, defense-in-depth approach wherein the radioactive fission products are kept
from the environment by multiple protective barriers, culminated by the outer containment
shell.
Figure 5 The overall CANDU nuclear plant
Obviously, a real CANDU is far more complex than the illustrations and designing a nuclear plant
is not a trivial exercise. There are many systems and sub-systems that interact.
3 Textbook Organization
Presenting educational material on a complex physical system such as a nuclear power plant is
problematic in that some knowledge of the physical plant is needed in order to give context to
the physical phenomena studied, yet some knowledge of the physical phenomena is needed to
make sense of the physical plant. So some iteration in presentation is unavoidable and, from a
learning perspective, will prove helpful.
So we begin in the introductory chapters by putting the CANDU reactor in a societal context,
give a very brief look at how the reactor functions, provide some historical context, and de-
scribe how it has evolved. From this the reader will appreciate enough of what a CANDU as a
physical system entails to give context to the more in-depth chapters that follow.
The flow of the Chapters 3 to 11 is dictated by the focus on the Process Systems and the flow of
power from the core to the grid. Subsequently Chapters 12 to 19 explore various key topics that
arise directly as a result of the energy generating process.
The very core of the subject, literally is the fuel. The process (i.e. fissioning) aspects of the fuel
and associated moderator are the subject of Chapters 3, 4 and 5 which cover the basic nuclear
processes and the reactor physics topics of statics and dynamics. This sets the nature of the
beast, as it were.
From both a process perspective and a safety perspective, cooling the reactor core and trans-
porting the heat energy generated by the core is centrally important. Thermalydraulic Design
and Analysis is covered by Chapters 6 and 7.
The overall plant dynamics are not so much determined by the dynamics of the individual
components and systems such as the reactor core or heat transport system since each of these
systems is controlled to keep each system within its own operating envelop. Overall plant
dynamics is more characterized by how these systems interact with each other and is the
subject of the chapters on Plant Systems (8) and Plant Operations (9).
Then a closer look at reactor and process system Instrumentation and Control (10) and Electrical
Systems (11) is given.
Radiation is generated and so must be understood, quantified and handled (Chapter 12 Radia-
tion Protection and Environmental Safety).
Fissioning is inherently an exponential process and entails substantial decay heat even after
shutdown. Hence Safety (13) and Regulations and Licensing (16) are central issues.
The nuclear environment can be a harsh and demanding one. Physical integrity must be main-
tained over the long life of the plant, day in, day out. Hence Materials (14) and Chemistry (14)
are key subjects.
The process characteristics of fuel as they relate to fissioning and heat transfer are treated in
the early chapters. But there is much more to say about fuel, specifically its physics characteris-
tics that enable it to perform so well under conditions that are so demanding. This is the
CANDU in Context 11
subject of Chapter 17. The forward looking topics of Fuel cycles and storage and disposal are
the subjects of Chapters 18 and 19 and are a fitting way to end our introductory textbook on the
Essential CANDU.
4 Further reading
• The Virtual Nuclear Tourist at http://www.virtualnucleartourist.com/ - a very popular

site run by Joe Gonyeau, a dedicated nuclear engineer.
• The CANTEACH website at https://canteach.candu.org/ - be sure to visit the site library
for extensive information on CANDU reactors.
1
CHAPTER 1
Introduction to Nuclear Reactors
prepared by
Dr. Robin Chaplin
Summary:
This chapter provides a top-level introduction to nuclear reactors and surveys the world reactor
situation. The various commercial large power producing reactors are identified and described
against a brief background of nuclear reactor principles and key reactor components. The
progressive expansion of nuclear power production is put into perspective in the global context.
The operation of nuclear plants with respect to changing grid system demand is explained with
some constraints identified. The chapter concludes with a brief review of safety and risk which
are critical aspects in the design and operation of nuclear plants.
Table of Contents
1 Nuclear Reactor Principles ...................................................................................................... 3
1.1 Fission Energy...................................................................................................................... 3
1.2 Nuclear Reactor Principles .................................................................................................. 5
2 Reactor Types.......................................................................................................................... 7
2.1 Prototype Reactors ............................................................................................................. 7
2.2 Commercial Reactors .......................................................................................................... 8
3 World-Wide Commercial Power Reactors ............................................................................ 12
3.1 Reactor Development ....................................................................................................... 12
3.2 Commercial Reactors in Service........................................................................................ 13
3.3 Recent Past and Near Future ............................................................................................ 16
3.4 Status of Large Reactors.................................................................................................... 17
4 Power Production ................................................................................................................. 18
4.1 Energy Transfer ................................................................................................................. 18
4.2 Power Output.................................................................................................................... 19
4.3 Operating Constraints ....................................................................................................... 21
4.4 Fuel Burnup....................................................................................................................... 21
5 Safety and Licensing.............................................................................................................. 22
5.1 Radiation Hazards ............................................................................................................. 22
5.2 Risk Assessment ................................................................................................................ 23
5.3 Licensing Principles ........................................................................................................... 23
6 Problems ............................................................................................................................... 24
7 Bibliography .......................................................................................................................... 24
8 Acknowledgments................................................................................................................. 25
List of Figures
Figure 1 Range of stable and unstable nuclides.............................................................................. 3
Figure 2 Binding energy per nucleon .............................................................................................. 4
©UNENE, all rights reserved. For educational use only, no assumed liability. Introduction to Nuclear Reactors – June 2015
Figure 3 Fission and absorption characteristics of uranium ........................................................... 5

Figure 4 Nuclear reactor components ............................................................................................ 7
Figure 5 Diagrammatic cross section of a typical BWR................................................................... 8
Figure 6 Diagrammatic cross section of a typical PWR................................................................... 9
Figure 7 Diagrammatic cross section of a typical PHWR (CANDU) ............................................... 10
Figure 8 Diagrammatic cross section of a typical AGR.................................................................. 11
Figure 9 Diagrammatic cross section of a typical RBMK (LGR) ..................................................... 11
Figure 10 Simplified flow diagram of a nuclear plant ................................................................... 19
Figure 11 Lagging and leading plant operation............................................................................. 20
List of Tables
Table 1 Nuclear reactors in service and planned .......................................................................... 14
Table 2 Nuclear reactors in service in 2012 .................................................................................. 16
Table 3 Projected reactors in service in 2022 ............................................................................... 17
Table 4 New large reactors (1000 MWe and greater)................................................................... 17
Table 5 Decommissioned large reactors (1000 MWe and greater) .............................................. 18
Introduction to Nuclear Reactors 3
1 Nuclear Reactor Principles

1.1 Fission Energy
All atoms consist of a nucleus of protons and neutrons surrounded by a cloud of electrons.
Generally, for elements of low atomic mass numbers, there are equal numbers of neutrons and
protons in the nucleus because this arrangement represents the most stable configuration. As
the atomic mass number increases, however, the increasing weak repulsive forces of similarly
charged protons in the nucleus are compensated for by an increasing number of neutrons which
contribute additional strong nuclear forces that bind adjacent protons and neutrons together.
Therefore, at high atomic mass numbers, there are approximately one and one-half times as
many neutrons as protons in the nucleus, as shown in Figure 1.
Figure 1 Range of stable and unstable nuclides
Should an element of high atomic mass number split, during a fission process, into two ele-
ments of lower atomic mass number, not as many neutrons will be required to create a stable
configuration, and the surplus neutrons will be released.
The nucleons (neutrons and protons) in the nucleus are bound together by nuclear forces.
When a nucleus is assembled, the binding energy attracting the individual nucleons is released.
Conversely, energy is required to separate individual nucleons from one another or from the
nucleus. This is analogous to the attraction of magnets to one another, but with a different type
of force. This binding energy is a maximum per nucleon for elements near the middle of the
range of atomic mass numbers and is somewhat less for elements with high atomic mass
numbers, as shown in Figure 2. This means that, if a heavy element fissions or splits into two
mid-range elements, the nucleons are bound together with a greater amount of binding energy
per nucleon. The surplus energy is released as the nucleons come together more strongly. This
is analogous to masses with potential energy falling into deeper holes and releasing this energy.
Because of the well-known relationship between mass and energy, the release of binding
energy is accompanied by a very slight decrease in the total mass of the constituents.
Figure 2 Binding energy per nucleon

Elements with high atomic mass numbers are unstable and become more so as the atomic mass
number increases. If a neutron is added to the nucleus of a very heavy element, the additional
binding energy brought into the nucleus may excite it to the extent that it fissions. This occurs
in two isotopes of uranium and two isotopes of plutonium. Of these four fissile materials, only
uranium-235 is naturally occurring. Furthermore natural uranium contains only 0.7 percent of
U-235. Nevertheless, uranium-235 is the primary fuel of all commercial power producing
nuclear reactors.
When a U-235 nucleus fissions, the release of binding energy amounts to about 200 MeV, or 32
x 10-12 J. Although a very small amount, this translates into a heat production rate of 950 MW if
1 kg of U-235 is totally consumed per day. In a typical nuclear power plant operating on a
conventional steam cycle, this could be converted into an electrical power output of about 300
MW.
1.2 Nuclear Reactor Principles

Some heavy elements, such as uranium-235, can be induced to fission by adding a neutron to
their nuclei. When fission occurs, the resultant lighter elements do not require as many neu-
trons in their nuclei to maintain a stable configuration and, on average, between two and three
surplus neutrons are released. These neutrons can cause further fissioning of other U-235
nuclei and so establish a chain reaction. Such a reaction can be allowed to diverge, as in an
atomic bomb, or be controlled, as in a nuclear reactor. Under steady state conditions, just one
neutron on average from each fission should go on to produce another fission event.
When fission occurs, the release of energy drives the lighter elements or fission products and
the surplus neutrons away from one another at high velocity. Most of the energy is thus trans-
formed into kinetic energy carried by the fission products. As heavy strongly charged particles,
they do not travel any significant distance and dissipate their kinetic energy in the fuel by
interaction with other atoms, thus increasing the temperature of the fuel. The high energy fast
neutrons, being uncharged, readily pass through the fuel and other reactor materials.
1000 U-
23
5
Cross section (barns)
100 t
U-235
10 t
U-2
38
f
U-238 f
t
t
1
1 meV 1 eV 1 keV 1 MeV
1/v Resonance Smooth

region region region
Interactions of Importance
s = Scattering
= Radiative capture
a = Absorption
= Fission
f
Figure 3 Fission and absorption characteristics of uranium

There are varying probabilities that they will be absorbed by different nuclei. The probability of
absorption by another U-235 nucleus to cause fission increases if the neutron velocity is re-
duced, and therefore reducing neutron energies is advantageous. Figure 3 shows the variation
of the fission cross section, that is, the probability of a fission reaction, of U-235 versus neutron
energy as a dotted line. The probability increases by a factor of nearly 500 as the neutron
energy is reduced from the fissioning release energy to the natural energy of neutrons in the
local environment, a process known as thermalization. This can be achieved by allowing the
neutrons to experience a series of non-absorbing collisions with light nuclei, which during an
elastic collision receive some of the energy from the neutrons. The resulting low energy slow
neutrons are then absorbed in U-235 nuclei to cause further fissions. Some elements that can
slow down or moderate neutrons effectively without significant absorption are hydrogen,
deuterium, helium, beryllium, and carbon. Hence, light water (H2O), heavy water (D2O), and
graphite (C) all make good moderators in nuclear reactors.
The solid lines in Figure 3 show the total probability of absorption (resulting in both fission and
capture) in U-235 and U-238. In the resonance region (where neutron and nucleus frequencies
coincide) are large spikes with very high probability of capture for U-238, so it is best to avoid
this region during the neutron slowing down process. This can be achieved to some degree by
ensuring that the moderation process is carried out away from the fuel. This leads to a reactor
consisting of a matrix of fuel elements within the moderator with a discrete distance between
fuel elements. Furthermore, to promote heat transfer from the fuel, the fuel elements them-
selves consist of bundles of small fuel rods. This is illustrated diagrammatically in Figure 4.
Most of the heat from fission is generated by dissipation of the kinetic energy of the fission
products. Because this occurs in the fuel near the point of fission, it follows that the fuel
becomes the main source of heat in the reactor. To maintain a thermodynamic cycle to produce
work, this heat must be removed continuously as it is produced. A suitable coolant is therefore
required to flow over the fuel elements and remove the heat. The coolant must not readily
absorb neutrons and must have suitable thermal properties. Coolants such as light water, heavy
water, helium, and carbon dioxide meet these requirements.
Finally, to ensure steady state operation, the number of neutrons allowed to go on to produce
fission must be the same as the number in preceding generations of neutrons. To achieve the
required balance, the reactor as a whole is designed to generate excess neutrons in each
generation and to have a system for absorbing the excess so as to maintain and control the
reactor at a steady load. This also enables the number of neutrons in successive generations to
be increased when more power is required or to be decreased when power must be reduced.
Such control is usually achieved by having movable neutron-absorbing control rods partially
inserted into the reactor. By fully inserting the control rods, the reactor can be shut down.
Therefore, a typical nuclear reactor consists of the following main components as shown in
Figure 4:
 Fuel in which fission occurs and heat is generated
 Moderator to reduce the energy of the neutrons
 Coolant to remove the heat from the fuel elements
 Control rods to maintain the proper neutron balance.
Figure 4 Nuclear reactor components

These components are usually arranged in a two-dimensional matrix so that neutrons generat-
ed by fission in one fuel rod pass through the moderator before entering the next fuel rod. The
spacing of the fuel rods depends upon the moderator properties and the distance required to
reduce the energy of the neutrons. The fuel rods are made small enough to promote heat
removal and are surrounded by coolant. Often the coolant serves as a moderator as well. The
control rods are made to penetrate between the fuel rods to capture excess neutrons effective-
ly. Naturally, some neutrons leak through the boundaries of the system, and many are absorbed
by the reactor materials or in the fuel without causing fission, meaning that the control rods do
not have to absorb a significantly large number of neutrons compared with the number causing
fission. The actual configuration of the matrix and the spacing of the fuel rods depend upon the
fuel and moderator characteristics.
2 Reactor Types
2.1 Prototype Reactors
As nuclear reactors developed around the key elements of fuel, moderator, and coolant, many
different types were proposed and constructed as demonstration models in an endeavour to
prove their technical and commercial viability. Some designs had serious technical problems,
while others had uniquely advanced features. However, due to arbitrary political decisions, one
or two good designs did not go beyond the prototype stage, while others with insurmountable
technical difficulties were abandoned. Nevertheless, some prototypes did operate successfully
for many years, providing valuable technical and operational experience while producing power.
Ultimately, the field narrowed to certain proven designs which were adopted on a commercial
basis. Currently, there are six main types, which are listed below in order of numbers in service.
The first five have proven to be commercially viable, while the sixth can be considered to be still
in the prototype stage, but to hold promise for the future as a reactor which can breed new
fissile fuel:
PWR Pressurized Water Reactor

BWR Boiling Water Reactor
PHWR Pressurized Heavy Water Reactor (CANDU)
GCR Gas Cooled Reactor
LGR Liquid Graphite Reactor
LMFBR Liquid Metal Fast Breeder Reactor.
2.2 Commercial Reactors

The reactors listed above will be described briefly in the following paragraphs for comparison,
starting with the simplest concept.
2.2.1 Boiling water reactor (BWR)
The boiling water reactor consists of a pressure vessel containing the fuel rods in vertical
elements, with each element surrounded by a channel. The channels are flooded with light
water which serves as the moderator and coolant. Pumps circulate water upwards through the
channels, and steam is generated within the channels. The exit steam quality is about 13%, and
the steam is separated from the water in cyclone separators above the reactor core. This
saturated steam is sent directly to the steam turbines.
Figure 5 Diagrammatic cross section of a typical BWR

This simple direct cycle has the disadvantage that steam from the reactor is passed through the
turbine and the condensate is returned through the feedwater heaters, making the whole
steam circuit slightly radioactive during operation. A simplified diagram of a boiling water
reactor is given in Figure 5.
2.2.2 Pressurized water reactor (PWR)

The pressurized water reactor overcomes the problem of a slightly radioactive steam circuit by
having an intermediate heat exchanger to separate the reactor coolant circuit from the turbine
steam circuit. Steam is generated in this steam generator and is sent to the turbine as saturated
steam under conditions similar to those in the boiling water reactor. The reactor coolant circuit
is maintained at high pressure to prevent any boiling in the reactor and operates at a slightly
higher temperature than the BWR to promote heat transfer to the secondary steam circuit.
Because no boiling occurs in the reactor core, it is more compact and does not require channels.
The fuel rods of the individual elements form a continuous vertical matrix in the core. This is
flooded with an upward flow of circulating light water which serves as coolant and moderator.
A simplified diagram of a pressurized water reactor is given in Figure 6.
Figure 6 Diagrammatic cross section of a typical PWR
2.2.3 Pressurized heavy water reactor (PHWR) (CANDU)

The pressurized heavy water reactor is similar to the pressurized water reactor in that it has a
primary coolant circuit and secondary steam circuit with a steam generator producing saturated
steam. The reactor, however, is different in that it has individual pressure tubes passing hori-
zontally through the reactor core. Fuel rods in the form of bundles are contained within these
pressure tubes. The heavy water coolant flows through these tubes and is then circulated
through the steam generator. The pressure tubes are surrounded on the outside by heavy
water at low pressure and low temperature, which serves as the moderator. This is contained in
a large vessel or calandria at ambient pressure. The advantage of this system is that it does not
require a very heavy pressure vessel as in the BWR and PWR. The Canadian version is known as
the CANDU. Other variations such as the steam generating heavy water reactor (SGHWR) had
light water coolant and vertical channels and generated steam in the core. A simplified diagram
of a pressurized heavy water reactor is given in Figure 7.
Figure 7 Diagrammatic cross section of a typical PHWR (CANDU)
2.2.4 Gas cooled reactor (GCR)

The gas cooled reactor, like the pressurized water reactor, has separate reactor coolant and
steam generating circuits with an intermediate steam generator. The coolant, however, is gas,
which enables higher coolant temperatures to be achieved. This enables superheated steam to
be generated and a high efficiency steam circuit similar to that of a fossil fuel fired plant to be
used. Due to the high temperatures and large core volume required with a gas coolant, graphite
is more suitable than water as a moderator. This is installed in the form of blocks with holes to
form channels into which the fuel elements are placed and through which the gas coolant flows.
The early reactors of this type were known as Magnox from the type of fuel used, and later ones
were called advanced gas cooled reactors (AGRs), which run at higher temperatures with a
different type of fuel. By changing the coolant from carbon dioxide to helium, even higher
temperatures could be achieved, hence the designation high temperature reactor (HTR). This
type gave the ultimate advantage of large gas cooled reactors, which have a cycle efficiency as
high as any fossil fuel fired plant.
Later developments in this field, such as the pebble bed modular reactor (PBMR), use a direct
cycle, with the helium coolant passing directly to a helium turbine with the promise of even
better cycle efficiency. A simplified diagram of an advanced gas cooled reactor is given in Figure
8.
Figure 8 Diagrammatic cross section of a typical AGR

2.2.5 Liquid graphite reactor (LGR)
The liquid graphite reactor has a graphite block core similar to that of a gas cooled reactor.
Pressure tubes similar to those of the pressurized heavy water reactor pass through vertical
holes in the graphite, which serves as the moderator. Fuel elements made up of clusters of rods
are located within these vertical tubes, and light water coolant flows upwards over the fuel and
within the tubes. Boiling occurs in the fuel as in the boiling water reactor. The resulting steam
is separated from the water in an external drum and the water recirculated. As with the boiling
water reactor, saturated steam is sent to the turbine. This type of reactor of Russian design is
known as the RBMK. A simplified diagram of a liquid graphite reactor is given in Figure 9.
Figure 9 Diagrammatic cross section of a typical RBMK (LGR)
2.2.6 Liquid metal fast breeder reactor (LMFBR)

The liquid metal fast breeder reactor is different from the others in many ways. The coolant is a
liquid metal, usually sodium, which flows directly over the fuel rods of the vertical fuel elements
in an upward direction. The heat carried in this primary circuit is transferred to a secondary
sodium circuit in a heat exchanger. This secondary circuit then transfers the heat to the steam
circuit in a steam generator. Because sodium remains a liquid at high temperatures without
being pressurized, heavy pressure vessels are not required, and high temperature superheated
steam can be produced to provide good cycle efficiency. There is no moderator to reduce the
energy of the neutrons, hence the term “fast”, and surplus neutrons are used to convert non-
fissile material into fissile fuel, hence the term “breeder”. This type of reactor has not yet
reached the commercially viable stage, and therefore there is no typical design.
3 World-Wide Commercial Power Reactors

3.1 Reactor Development
Following the Second World War, the development of nuclear reactors followed different paths
in different countries depending upon the facilities developed during the war and the perceived
military needs of certain countries following the war. The first nuclear reactors served to
generate plutonium-239, another fissile material formed when uranium-238, the major consti-
tuent of natural uranium, absorbs excess neutrons. Pu-239 could be easily separated from the
original uranium fuel and was needed for atomic bombs. U-235 is also used in nuclear wea-
pons, but is difficult to separate from U-238.
The United States had isotope separation facilities for uranium and therefore was able to pursue
the development of light water reactors requiring the use of uranium fuel slightly enriched in U-
235. The United Kingdom did not have such facilities and therefore was forced to develop
reactors using natural uranium. Canada had developed expertise in heavy water technology as
a result of the war effort and was therefore well positioned to develop heavy water moderated
reactors also using natural uranium.
When using natural uranium as a fuel, the low concentration of U-235 and the absorption of
neutrons in U-238 necessitate use of a moderator with an extremely low absorption cross
section to establish a continuous chain reaction. Only heavy water (deuterium and oxygen) and
graphite (carbon) have the required physical and neutron properties. This led Britain to develop
graphite moderated natural uranium metal fueled reactors with carbon dioxide as a coolant.
Later, when they had uranium enrichment facilities, they were able to switch to uranium dioxide
as the fuel. Canada, which had supplies of heavy water, was able to develop heavy water
moderated natural uranium dioxide fueled reactors. Heavy water is the only moderator that
can be used directly with natural uranium dioxide fuel to sustain the required fission chain
reaction. Although heavy water is an excellent moderator, especially with respect to neutron
absorption, it is very expensive to separate from ordinary water. If enriched uranium is availa-
ble, as was the case in the United States, there is wider scope in the choice of a moderator
because more neutron absorption can be tolerated. Therefore, light water (hydrogen and
oxygen) can be used as a moderator. An advantage of light water as a moderator is that it is
very effective per unit distance in slowing down neutrons, leading to a smaller moderator
volume and a more compact reactor than with the other moderators. However, light water has
the inherent disadvantage of absorbing a greater fraction of neutrons than graphite or heavy
water.
Of the three moderators mentioned above, graphite is the least effective in reducing neutron
energy and requires the largest volume. Graphite moderated reactors are therefore the largest,
leading to high capital costs. Considering capital cost, moderator costs, and enriched fuel
supply and cost, all three of these became economically viable in their respective countries, and
commercial reactors for power plants subsequently evolved.
Over time, enriched uranium became more easily available, and commercial reactors were sold
to other countries which did not have fuel enrichment resources, leading to a free choice of
reactors. It is of interest to review the world use of nuclear reactors for electric power produc-
tion.
3.2 Commercial Reactors in Service

Table 1 shows the number, type, and output of reactors operating and committed or under
construction in different countries. The first column under the two latter headings is for reac-
tors currently in service and the second column for those reactors to be added in the coming
years. This table is representative of current commercial technology because most prototype
and early commercial reactors have served their useful life and been decommissioned.
Table 1 Nuclear reactors in service and planned*
BWR–Boiling (Light) Water Reactor

GCR–Gas Cooled Reactor
LGR–Light Water Graphite Reactor
LMFBR–Liquid Metal Fast Breeder Reactor
PWR–Pressurized (Light) Water Reactor
PHWR–Pressurized Heavy Water Reactor
Number Operational + Existing Capacity +
Country Reactor Type
Planned New Capacity (MWe)
Argentina PHWR 2 1 935 692

Armenia PWR 1 375
Belgium PWR 7 5 885
Brazil PWR 2 1 1 901 1 275
Bulgaria PWR 2 2 1 906 2 000
Canada PHWR 22 15 137
China PWR 12 41 9 748 42 230
PHWR 2 1 300
GCR 1 200
LMFBR 1 20
Czech Republic PWR 6 3 678
Finland PWR 2 1 976 1 600
BWR 2 1 740
France PWR 58 1 63 130 1 600
Germany PWR 7 9 486
BWR 2 2 572
Hungary PWR 4 1 889
India PWR 2 1 834
BWR 2 300
PHWR 18 4 4 091 2 560
LMFBR 1 500
Iran PWR 1 915
Japan PWR 24 19 286
BWR 26 2 24 818 2 756
Kazakhstan LMFBR 1 70
Country Reactor Type Number Operational + Existing Capacity +

Planned New Capacity (MWe)
Mexico BWR 2 1 300
Netherlands PWR 1 487
Pakistan PWR 2 1 600 300
PHWR 1 125
Romania PHWR 2 3 1 300 1 869
Russia PWR 16 11 11 914 9 810
LGR 15 10 219
LMFBR 1 1 560 750
Slovakia PWR 4 2 1 816 810
Slovenia PWR 1 666
South Africa PWR 2 1 800
South Korea PWR 17 7 15 975 8 600
PHWR 4 2 722
Spain PWR 6 6 004
BWR 2 1 510
Sweden PWR 3 2 799
BWR 7 6 504
Switzerland PWR 3 1 700
BWR 2 1 538
Taiwan PWR 2 1 780
BWR 4 2 3 104 2 600
Turkey PWR 4 4 600
Ukraine PWR 15 3 13 107 2 850
United Arab Emirates PWR 4 5 600
United Kingdom PWR 1 1 188
GCR 17 8 732
United States PWR 69 8 68 459 8 990
BWR 35 2 34 935 2 700
* Adapted from Nuclear News, March 2012
It is evident from Table 1 that certain types of reactors became dominant in certain countries,
particularly in those developing their own reactors, for example, PWRs and BWRs in the United
States, GCRs in the United Kingdom, PHWRs in Canada, and LGRs in Russia. However, in later
years, certain types of reactors have been favoured by countries without their own develop-
mental program, leading to the spread of some of these types to other countries. The PWR,
however, has become the most common type and is currently produced by various manufactur-
ers around the world.
3.3 Recent Past and Near Future

Table 2 gives a summary of the situation at the beginning of 2012. It shows that PWRs make up
61% of operational reactors and 67% of total nuclear capacity, due mainly to their larger size
relative to PHWRs and GCRs.
Table 2 Nuclear reactors in service in 2012*
Reactor Type Number in % Total Capacity %

Service (MWe)
PWR 267 61 246 555 67

BWR 84 19 78 321 21
PHWR 51 12 25 610 7
GCR 17 4 8 732 2
LGR 15 4 10 219 3
LMFBR 1 0 560 0
TOTAL 435 100 369 997 100
*Adapted from Nuclear News, March 2012.

A similar review done in 1998 indicated that during the intervening 14 years, the number of
reactors in service had only increased by two, but that total capacity had increased by 11%. This
was due primarily to the decommissioning of smaller older reactors which had reached the end
of their economic life. It does not reflect a lack of new plants coming on-line because the
number of PWRs increased by 17, even though 10 were taken out of service during this period.
Looking forward over 10 years to 2022, Table 3 shows the total number of reactors currently in
service and committed for commissioning or construction. Although the table does not consid-
er possible decommissioning of some reactors, it also does not take account of new orders
which will be completed before 2022. In this context, it should be noted that 80% of new
construction is due to be on-line by 2017, leaving time for new commitments which will likely
overcompensate for possible decommissionings.
Table 3 shows a dramatic increase of 25% and 29% respectively in the number of reactors in
service and available capacity over the following 10 years.
Table 3 Projected reactors in service in 2022*
Reactor Type Number in Change (%) Total Capacity Change (%)

Service (MWe)
PWR 356 +33 339 569 +38

BWR 90 +7 86 377 +10
PHWR 59 +16 30 722 +20
GCR 18 +6 8 932 +2
LGR 15 0 10 219 0
LMFBR 5 +400 2 076 +271
TOTAL 543 +25 477 895 +29
*Adapted from Nuclear News, March 2012
3.4 Status of Large Reactors

Table 4 New large reactors (1000 MWe and greater)*
Country Reactor Type Number New Capacity

Committed (MWe)
Brazil PWR 1 1 275

Bulgaria PWR 2 2 000
China PWR 38 36 400
Finland PWR 1 1 600
France PWR 1 1 600
Russia PWR 6 6 900
South Korea PWR 7 8 600
Taiwan PWR 2 2 600
Turkey PWR 4 4 600
United Arab Emirates PWR 4 5 600
United States PWR 8 8 990
BWR 2 2 700
*Data extracted from Nuclear News, March 2012

Most reactors currently under construction or on order are of large capacity, that is, 1000 MWe
or greater. Considering only reactors of this size, Table 4 shows the type of reactor, total capaci-
ty, and country where these are being built or to be built.
For comparison, the number of large units taken out of service to date is given in Table 5. This
table shows the capacity, date taken out of service, and number of years in service.
Table 5 Decommissioned large reactors (1000 MWe and greater)*
Country Reactor Type Capacity Date Years of

(MWe) Service
France LMFBR 1 200 1998 12

Germany PWR 1 219 2001 14
BWR 1 346 2011 27
PWR 1 345 2011 32
PWR 1 240 2011 34
PWR 1 167 2011 36
United States PWR 1,095 1992 16
PWR 1 040 1998 24
PWR 1 040 1998 25
TOTAL 10 692 220
*Data extracted from Nuclear News, March 2012

The average service life of these excluding the LMFBR is 26 years. Typically with refurbishment
the life expectancy of such reactors would be at least double this.
4 Power Production
4.1 Energy Transfer
Heat, energy, and work all have the same units (joules), but somewhat different meanings, and
power is the rate of doing work (watts). What is important is that not all heat can be converted
into work. Although the first law of thermodynamics states that all work can be dissipated as
heat, the second law states that not all heat can be converted into work. In a typical water
cooled nuclear plant (CANDU, PWR and BWR) approximately 30% of the heat released by the
fuel is ultimately converted into electrical energy. The rest must be discharged as low grade
(low temperature) heat.
During the fission process, heat is generated in the fuel. This heat is removed by the reactor
coolant flowing over the fuel rods and transported in the primary circuit to the steam generator,
where it flows inside the tubes and its heat is transferred to the secondary circuit through the
walls of the tubes in the steam generator. Water in the secondary circuit outside the tubes
absorbs this heat and is converted into steam under saturated conditions. This steam is passed
to the turbine where it expands to low pressure while being directed onto the turbine blades
and in so doing transfers its energy to the turbine rotor. The rotor drives the electrical genera-
tor which produces electric power. The exhaust steam is condensed by cooling water passing
through the condenser tubes and in so doing discharges the bulk of the heat which cannot be
converted into work. The condensate is returned to the steam generator after being preheated
in the feedwater heating system. A simple flow diagram for a nuclear plant is shown in Figure
10. Note that in the BWR and RBMK, steam is generated directly in the reactor, so there is no
separate steam generator.
Figure 10 Simplified flow diagram of a nuclear plant
4.2 Power Output

Consider a very simple system with a nuclear reactor, steam generator, and turbine generator
supplying electric power to an isolated electrical grid, as shown in Figure 11. This power must
be generated at the moment it is required by the consumers connected to the grid. Power
production must follow demand exactly, and any mismatch will cause the grid frequency to fall
or rise as demand increases or decreases. A basic control system works as follows to maintain
appropriate power output from the plant. In the event of an increase in demand the mismatch
will cause the grid frequency to fall. Because the turbine generator is synchronized to the grid,
its speed will drop accordingly. This will be sensed by the turbine governor which will open the
governor valves to admit more steam to increase the power output of the turbine generator.
The additional flow of steam to the turbine will cause a reduction in steam pressure in the
steam generator. This in turn will be sensed by the reactor regulating system which will with-
draw control rods from the reactor core until the increased fission rate generates sufficient
additional heat to restore the steam generator pressure. In the event of a decrease in demand
the reverse occurs. This is known as the reactor following or turbine leading mode of operation
(or normal mode in some plants because it is a natural way of maintaining stable conditions).
Such a system, however, cannot maintain a specified frequency (60 Hz in North America) exactly
without large unstable oscillations, and therefore a certain speed droop is incorporated into the
turbine governor. This enables a progressive increase in governor valve opening (steam flow) as
the turbine speed (grid frequency) falls. A typical droop setting is 4%, which means that, if the
turbine was initially at zero load and full speed, its speed would have to drop to 96% before the
governor valve would be fully open. Such a speed is not acceptable for the turbine due to
possible blade vibration, nor to the grid due to loss in speed of connected motors. Therefore,
the governor is adjusted to bring the speed back to 100% at full load. In the event of a turbine
trip or load reduction to zero under these conditions, the reverse would occur, and the turbine
speed would rise to 104% of full speed.
Figure 11 Lagging and leading plant operation

In the mode of operation just described, the nuclear reactor output follows the electrical grid
demand, and therefore its power level oscillates continuously. This can have adverse effects
depending upon the type and design of the reactor. Excessive oscillations impose temperature
transients on the fuel, which can cause premature failures of the fuel cladding. Large oscilla-
tions near full power can cause power limits to be exceeded, thus tripping the reactor and
losing power production as well as imposing restart transients on it and the turbine. Due to
high capital cost and low fuel cost it is desirable to run nuclear reactors at full power most of the
time. An alternative mode of operation is therefore often used at nuclear plants where the
reactor power output is fixed. This is known as reactor leading or turbine lagging operation. To
maintain stable operation, the reactor power is controlled at a given value by measuring the
neutron flux and adjusting the control rods accordingly. Pressure in the steam generator must
be maintained at the proper value to ensure stable conditions in the reactor coolant circuit.
This is done by opening or closing the turbine governor valves to control steam flow from the
generators. The turbine then delivers power according to the steam flow and the generator
sends this power into the grid system regardless of the grid frequency. The grid frequency must
then be controlled by other turbine generators which feed into the grid system and operate in
the turbine leading mode.
By referring to the figure showing the two modes of operation, it can be seen that steam
generator pressure is a key control parameter in both modes. This highlights the importance of
the steam generator, where a balance of heat input and heat output must be maintained to
maintain pressure. Furthermore, the difference in temperature between the primary coolant
and the secondary working fluid determines the rate of heat transfer. Hence the reactor coo-
lant temperature is determined by saturation temperature and thus by steam generator pres-
sure.
4.3 Operating Constraints

During operation, parts of the reactor, steam system, and turbine are subject to high tempera-
tures. If these parts have thick walls or have a substantial solid mass they will likely suffer
thermal stress during the heating and cooling that arises during startup and shutdown and also
during load transients. If a thick component is heated on one side, that side will tend to ex-
pand. If the side is constrained by the still cold base material so that it cannot expand an
internal stress will be set up. The reverse happens during cooling. Therefore, large rigid com-
ponents which are subject to transient and uneven heating and cooling will suffer low cycle
fatigue damage and may ultimately fail. This effect can be minimized by slow heating and
cooling to reduce temperature differences in single components such as reactor pressure
vessels, steam generators, steam pipes, turbine casings and turbine rotors. This means that all
these components must be preheated slowly before startup and the unit must be loaded slowly.
Similarly, load changes on the unit should be carried out slowly. This requirement imposes
operating restrictions on reactors and turbines. Generally, the larger the unit, the longer the
time to start it up and load it. This makes large units less flexible in operation than smaller
units.
In the reactor, temperature transients in the fuel cause structural changes within the fuel and
stress on the cladding and therefore must be minimized to avoid premature fuel failures. In the
turbine, uneven heating and cooling of the rotor can cause bending, which in turn causes
excessive vibration and in the extreme case contact with the casing and therefore must be
monitored very carefully.
A nuclear reactor is also subject to xenon transients which may inhibit operation for a certain
period. When the load on a reactor is suddenly reduced, xenon, a fission product that absorbs
neutrons very strongly, builds up in the fuel and may force a reactor shutdown. The xenon
eventually decays after about 40 hours, and the reactor can be restarted. During this time,
there is a loss in electrical power production, and the larger the unit, the more serious is this
loss in revenue generating output.
4.4 Fuel Burnup

As the fuel in the reactor is used up, the concentration of uranium-235 decreases. This reduces
the number of fissions occurring with a given number of neutrons. Furthermore, some of the
fission products produced absorb neutrons, thus reducing the number of neutrons available to
produce fission. These changes can be accommodated by withdrawing the control rods from
the reactor and allowing more neutrons to be available in the fuel. After a long period of
operation, however, such changes can no longer be accommodated, and the fuel may have
become depleted in U-235 to the point where a continuous chain reaction can no longer be
sustained. At this point, the reactor has to be refuelled with fresh fuel.
With reactors that are partially refuelled once a year, the control rods do not provide an ade-
quate range of control, and therefore a soluble neutron absorber is added in small quantities to
the moderator. Its concentration is gradually reduced over time to compensate for fuel burnup.
Some reactors are designed for continuous on-load refuelling. This is advantageous because the
effects of fuel burnup and fission product production are negligible with regard to overall
reactor conditions.
5 Safety and Licensing

5.1 Radiation Hazards
Many types of radiation, both natural and human generated, are encountered in everyday life.
Even the human body contains radioactive materials. This radiation can be divided into two
categories. The first is electromagnetic radiation, which is assumed to travel in waves of dis-
crete frequencies, for example, X-rays, γ-rays, and UV rays. The second is particulate radiation
consisting of high-energy particles moving at high velocity, such as -particles, -particles,
neutrons, etc. Charged particles such as -particles and -particles are easily stopped by
ionization of surrounding material, but γ-rays and neutrons can be very penetrating due to their
low interaction with most materials.
Radiation causes damage to biological structures such as cells. However, cells are in a continual
state of dying and being replaced, so such damage is naturally repaired provided that it is not
excessive. Of concern though is possible damage to DNA structures and subsequent genetic
effects. For this reason exposure to radiation is controlled much more rigidly than exposure to
toxic materials and dangerous chemicals. Although γ-rays and neutrons can penetrate the
human body, -radiation and -radiation are generally damaging only if materials that emit
these particles are ingested or come into contact with skin.
In the core of a nuclear reactor, the fission process produces neutrons and highly reactive fission
products which emit -particles and -particles as well as γ-radiation as they decay to stable
isotopes. Neutrons and γ-rays are emitted from the core, but are attenuated to low and safe
levels by adequate shielding around the reactor. The fission products are held within the fuel
and therefore do not pose a hazard unless released to the environment through leakage from
the fuel or rupture of the fuel.
During operation of a nuclear plant, small amounts of leakage inevitably occur because some
fission products are gaseous and attenuation cannot eliminate all radiation. Therefore, the
entire operational area within and surrounding the plant as well as the local environment is
continuously monitored to ensure that radiation is kept well within safe limits and usually kept
much lower.
In the event of a serious accident, for example, fuel overheating and rupture of reactor compo-
nents, the design of all reactors makes provision for barriers to contain these fission products
and prevent their release to the environment. Such barriers may not be perfect, and some
radioactive products may be released. If people in the surrounding area ingest such products,
they will be subject to -radiation or -radiation emitted by these products, along with some γ-
radiation resulting from the decay process. The design of the plant has, as a key objective, the
provision of safety devices and containment barriers to ensure that radiation exposure to the
public is kept well below acceptable limits in the event of an accident.
5.2 Risk Assessment

Humankind is subject to many risks, both natural and human generated. These risks have been
analyzed, and if the severity of an accident is compared with the probability of its occurrence, it
is generally found that the probability of an accident occurring is much higher if the number of
deaths per accident is low (motor vehicle accidents) than if the number of deaths is high (air-
craft accidents). Because the consequences of a serious nuclear accident (for example, the
Chernobyl accident) can be very severe, the probability of its occurrence must be kept very low.
This puts great pressure on the nuclear industry to minimize the risk of serious accidents to
levels far below those of other industries. The result is that the nuclear industry is the most
regulated of all industries. Highly sophisticated design and accident analysis techniques and
requirements are in place to ensure public safety.
5.3 Licensing Principles

All nuclear plants are subject to a rigorous licensing process by an independent regulatory
authority to ensure safety of operating personnel and the general public and to ensure that
there are no detrimental effects to the surrounding environment. All aspects of the plant are
checked and assessed with regard to the effects of possible failures of key components. Initially
this assessment was based on the maximum credible accident due to a single failure. In the
event of the worst possible accident, the likely damage to the plant was assessed and the
containment barriers reviewed to determine the likely amount of radioactive fission product
release to the environment. This release was to be kept below specified values. This concept
did not take account of the possible (though very unlikely) failure of a component deemed not
to fail, nor did it take account of multiple minor failures which could have a cumulative or
cascading effect, causing an otherwise unpredicted failure of a key component.
The modern approach, based on the probability of failure of a key component and the conse-
quences of such a failure, overcomes the shortcomings mentioned above. A component
deemed not to fail can be given a very low but still finite probability of failure. Moreover,
multiple components which are all associated with one type of accident can each be given a
certain probability of failure, and hence the probability of simultaneous failure can be assessed.
Ultimately the entire plant can be assessed to determine the probability of an accident leading
to a certain release of radioactive fission products. Hence, the safety of the surrounding envi-
ronment and population can be ensured to a very high degree, that is, a very low probability of
damage to the environment or excessive radiation exposure to the population.
The transition to a probabilistic approach as opposed to the deterministic method was initially
difficult due to a lack of statistical data and because certain failures could have different root
causes. However, as the industry has matured, more statistical data have become available
from test results and operating experience. Nevertheless, the process remains complex.
Another consideration is the relative importance of component failure and human error. A
quick analysis done many years ago indicated that roughly half of various incidents and acci-
dents in operating nuclear plants around the world were caused by human error and roughly
half by component failure. Many component failures could be ascribed to human error during
component design or manufacture. Rigorous quality control, quality assurance, and quality
management procedures can minimize manufacturing and construction defects. Proper train-
ing, assessment, and licensing of plant operators can minimize human error during operation.
This all contributes to a lower risk of accident and increased safety in nuclear plants.
A further development in the nuclear industry is the concept of passive safety. This means that
in the event of a failure leading to accident conditions, the reactor will naturally revert to a safe
shutdown condition and maintain that condition with minimal operator intervention, even in
the event of the loss of key services. Some newer designs, for example, can maintain reactor
cooling after shutdown by natural coolant circulation and natural heat convection to the envi-
ronment.
6 Problems
1 Explain why the fission of very heavy elements results in release of neutrons as well as
production of excess energy.
2 Explain the purpose of the key components of a nuclear reactor and show how these are
arranged to ensure adequate heat removal and satisfactory control.
3 Describe the key similarities and differences between a CANDU and a PWR as well as be-
tween a CANDU and a BWR.
4 Explain how fission heat is converted into electrical energy and identify all key interfaces and
points of energy conversion in the process.
7 Bibliography
[Cameron1982] I. R. Cameron, Nuclear Fission Reactors, Plenum Press, New York, 1982.
[El-Wakil1982] M. M. El-Wakil, Nuclear Energy Conversion, 3rd ed., American Nuclear Society, La
Grange Park, IL, USA, 1982.
[Foster1983] A. R. Foster and R. L. Wright, Basic Nuclear Engineering, 4th ed., Prentice-Hall,
Englewood Cliffs, NJ. USA, 1983.
[Glasstone1994] S. Glasstone and A. Sesonske, Nuclear Reactor Engineering: Volume 2, Reactor
Systems Engineering, 4th ed., Chapman & Hall, New York, NY, USA, 1994.
[Knief1992] R. A. Knief, Nuclear Engineering, 2nd ed., Hemisphere Publishing Corporation,
Washington, DC, USA.
[Lamarsh2001] J. R. Lamarsh and A. J. Baratta, Introduction to Nuclear Engineering, 3rd ed.,
Prentice-Hall, New Jersey, USA, 2001.
8 Acknowledgments
The following reviewers are gratefully acknowledged for their hard work and excellent com-
ments during the development of this Chapter. Their feedback has much improved it. Of course
the responsibility for any errors or omissions lies entirely with the author.
Bob Tapping
Jeremy Whitlock
Bhaskar Sur
Terry Rogers
George Bereznai
Thanks are also extended to Diana Bouchard for expertly editing and assembling the final copy.
1
Historical Background
prepared by
Dr, Robin Chaplin
Professor of Power Plant Engineering (retired)
University of New Brunswick
Summary:
A review of the historical background for the development of nuclear energy is given to set the
scene for the discussion of CANDU reactors.
Table of Contents
1 Growth of Science and Technology......................................................................................... 2
2 Renowned Scientists ............................................................................................................... 4
3 Significant Achievements........................................................................................................ 6
3.1 Niels Bohr........................................................................................................................ 6
3.2 James Chadwick .............................................................................................................. 6
3.3 Enrico Fermi .................................................................................................................... 6
4 Nuclear Fission ........................................................................................................................ 7
5 Nuclear Energy........................................................................................................................ 7
6 Acknowledgments................................................................................................................... 8
List of Figures
Figure 1 Timeline of significant discoveries .................................................................................... 3
©UNENE, all rights reserved. For educational use only, no assumed liability. Historical Background – September 2014
1 Growth of Science and Technology
The rate of technology increase has accelerated as technology itself has increased, leading to a
sort of exponential growth. Energy growth is a good example of this phenomenon. Water power
developed relatively slowly, steam power somewhat more quickly, electrical power even more
rapidly, and nuclear power very rapidly. In the case of nuclear energy, collaboration between
scientists in different countries initially spurred the discovery of various nuclear particles. Then
the Second World War provided an incentive to harness the fission process, and later the rapid
growth in demand for new energy after the war provided the need for nuclear power. Within
the time span of roughly half a century, an entirely unknown source of energy had become a
major producer of commercial electric power. In addition, an entirely new and sophisticated
technology had been developed.
Before about 1900, most scientific work in this field was related to electricity. Then in the next
two decades following the discovery of radiation and its characteristics, various pieces of the
nuclear puzzle began to come together. Only by 1920 was the basic structure of the atom
understood, and the existence of the neutron was confirmed in 1932. This was a pivotal discov-
ery because the neutron is the key element in establishing a fission chain reaction. Just ten
years later, in 1942, the first self-sustaining chain reaction was established, and by 1956 the
technology had advanced to the point where a nuclear fission reactor could produce electric
power on a commercial scale.
In reviewing this evolution of nuclear energy, it is important to note the recognition given to the
researchers responsible for these discoveries. The Nobel Prize is the most prestigious award
given to such advances in the scientific field. Some twenty Nobel Prizes in nuclear and radiation
physics and nuclear chemistry or closely related fields were awarded over a fifty-year period
from 1901, the year of the first Nobel Prize.
To appreciate these advances in technology properly, it is convenient to consider a timeline of

achievement with key advances shown and to link these with some renowned scientists. A
timeline of these discoveries is shown in Figure 1.
Historical Background 3
Figure 1 Timeline of significant discoveries
©UNENE,
UNENE, all rights reserved. For educational use only, no assumed liability. Historical Background – September 2014
2 Renowned Scientists
Many researchers in different countries around the world have contributed to this new field of
science. Several collaborated with one another or worked or were trained in another’s labora-
tory. Each contributed a piece of the puzzle. Those making a significant advancement in this
field were awarded Nobel Prizes. Some 20 prizes were awarded over a fifty-year period, demon-
strating the importance of this field of science.
The Nobel Prizes awarded in Physics and Chemistry in the nuclear field from 1901 to 1951 with
the recipients and a statement of their contribution, as obtained from the Nobel Foundation
Official Web Site, are listed below.
1901 Physics: Wilhelm Conrad Roentgen

"In recognition of the extraordinary services he has rendered by the discov-
ery of the remarkable rays subsequently named after him"
1903 Physics: Antoine Henri Becquerel

"In recognition of the extraordinary services he has rendered by his discov-
ery of spontaneous radioactivity"
Pierre Curie, Marie Curie (née Sklodowska)
"In recognition of the extraordinary services they have rendered by their
joint researches on the radiation phenomena discovered by Professor Henri
Becquerel"
1906 Physics: Joseph John Thomson

"In recognition of the great merits of his theoretical and experimental
investigations on the conduction of electricity by gases"
1908 Chemistry: Ernest Rutherford

"For his investigations into the disintegration of the elements, and the
chemistry of radioactive substances"
1911 Chemistry: Marie Curie (née Sklodowska)

"In recognition of her services to the advancement of chemistry by the
discovery of the elements radium and polonium, by the isolation of radium
and the study of the nature and compounds of this remarkable element"
1918 Physics: Max Karl Ernst Ludwig Planck

"In recognition of the services he rendered to the advancement of Physics
by his discovery of energy quanta"
1921 Physics: Albert Einstein

"For his services to Theoretical Physics, and especially for his discovery of
the law of the photoelectric effect"
1922 Physics: Niels Henrik David Bohr
"For his services in the investigation of the structure of atoms and of the
radiation emanating from them"
1923 Physics: Robert Andrews Millikan

"For his work on the elementary charge of electricity and on the photoelec-
tric effect"
1927 Physics: Arthur Holly Compton

"For his discovery of the effect named after him"
Charles Thomson Rees Wilson
"For his method of making the paths of electrically charged particles visible
by condensation of vapour"
1929 Physics: Prince Louis-Victor Pierre Raymond de Broglie

"For his discovery of the wave nature of electrons"
1932 Physics: Werner Karl Heisenberg

"For the creation of quantum mechanics, the application of which has, inter
alia, led to the discovery of the allotropic forms of hydrogen"
1933 Physics: Erwin Schrödinger, Paul Adrien Maurice Dirac

"For the discovery of new productive forms of atomic theory"
1935 Physics: James Chadwick

"For the discovery of the neutron"
1938 Physics: Enrico Fermi

"For his demonstrations of the existence of new radioactive elements
produced by neutron irradiation, and for his related discovery of nuclear
reactions brought about by slow neutrons"
1939 Physics: Ernest Orlando Lawrence

"For the invention and development of the cyclotron and for results ob-
tained with it, especially with regard to artificial radioactive elements"
1944 Chemistry: Otto Hahn

"For his discovery of the fission of heavy nuclei"
1945 Physics: Wolfgang Pauli

"For the discovery of the Exclusion Principle, also called the Pauli Principle"
1951 Physics: Sir John Douglas Cockcroft

Ernest Thomas Sinton Walton
"For their pioneer work on the transmutation of atomic nuclei by artificially
accelerated atomic particles'
1951 Chemistry: Edwin Mattison McMillan, Glenn Theodore Seaborg

"For their discoveries in the chemistry of the transuranium elements"
3 Significant Achievements
Three scientists and their respective discoveries tend to stand out in the development of
nuclear physics as applied to energy production. Niels Bohr developed a model to describe the
structure of the atom, James Chadwick established the existence of neutrons, and Enrico Fermi
directed the construction of the first reactors. All these achievements were the result of exten-
sive work in analyzing the results and further developing the ideas of other researchers and
collaborators. Extracts of their biographies from the Nobel Foundation Official Web Site are
given below.
3.1 Niels Bohr
In the autumn of 1911 he made a stay at Cambridge, where he profited by following the ex-
perimental work going on in the Cavendish Laboratory under Sir J.J. Thomson's guidance, at the
same time as he pursued his own theoretical studies. In the spring of 1912 he was at work in
Professor Rutherford's laboratory in Manchester, where just in those years such an intensive
scientific life and activity prevailed as a consequence of that investigator’s fundamental inquir-
ies into the radioactive phenomena. Having there carried out a theoretical piece of work on the
absorption of alpha rays which was published in the Philosophical Magazine, 1913, he passed
on to a study of the structure of atoms on the basis of Rutherford's discovery of the atomic
nucleus. By introducing conceptions borrowed from the Quantum Theory as established by
Planck, which had gradually come to occupy a prominent position in the science of theoretical
physics, he succeeded in working out and presenting a picture of atomic structure that, with
later improvements (mainly as a result of Heisenberg's ideas in 1925), still fitly serves as an
elucidation of the physical and chemical properties of the elements.
3.2 James Chadwick
In 1932, Chadwick made a fundamental discovery in the domain of nuclear science: he proved
the existence of neutrons—elementary particles devoid of any electrical charge. In contrast with
the helium nuclei (alpha rays) which are charged, and therefore repelled by the considerable
electrical forces present in the nuclei of heavy atoms, this new tool in atomic disintegration
need not overcome any electric barrier and is capable of penetrating and splitting the nuclei of
even the heaviest elements. Chadwick in this way prepared the way towards the fission of
uranium 235 and towards the creation of the atomic bomb. For this epoch-making discovery he
was awarded the Hughes Medal of the Royal Society in 1932, and subsequently the Nobel Prize
for physics in 1935.
3.3 Enrico Fermi
In 1934, Fermi evolved the -decay theory, coalescing previous work on radiation theory with
Pauli's idea of the neutrino. Following the discovery by Curie and Joliot of artificial radioactivity
(1934), he demonstrated that nuclear transformation occurs in almost every element subjected
to neutron bombardment. This work resulted in the discovery of slow neutrons that same year,
leading to the discovery of nuclear fission and the production of elements lying beyond what
was until then the Periodic Table.
In 1938, Fermi was without doubt the greatest expert on neutrons, and he continued his work
on this topic on his arrival in the United States, where he was soon appointed Professor of
Physics at Columbia University, N.Y. (1939-1942).
Upon the discovery of fission, by Hahn and Strassmann early in 1939, he immediately saw the
possibility of emission of secondary neutrons and of a chain reaction. He proceeded to work
with tremendous enthusiasm, and directed a classical series of experiments which ultimately
led to the atomic pile and the first controlled nuclear chain reaction. This took place in Chicago
on December 2, 1942, on a squash court situated beneath Chicago's stadium. He subsequently
played an important part in solving the problems connected with the development of the first
atomic bomb. (He was one of the leaders of the team of physicists on the Manhattan Project for
the development of nuclear energy and the atomic bomb).
4 Nuclear Fission
Before 1939, there was no evidence to suggest the practical usefulness of atomic energy. In
1934, Enrico Fermi had reported that, when uranium was subjected to a stream of neutrons,
elements of higher mass number were formed. These transuranic elements attracted the
interest of other researchers, who discovered elements of unexpected mass numbers in the
products. Otto Hahn and Friedrich Strassmann noted in 1938 that the masses of two of these
products added up to the mass of the uranium atom plus a neutron. Subsequently, in early
1939, Lise Meitner and Otto Frisch were able to explain in a published report that "it seems
possible that the uranium nucleus has only small stability of form and may, after neutron
capture, divide itself into two nuclei of roughly equal size". This was called fission, the term
which was commonly used in biology to describe the division of living cells. Otto Frisch soon
after proved, as they had predicted, that the particles released had strong ionizing power. This
was also confirmed by several other researchers. Furthermore, these fission fragments were
found to be ejected at high velocity and to possess radioactive properties. Most of the confir-
matory work of this new revelation was completed within three months of the initial publication
of the theory of nuclear fission. This subsequently was generally referred to as "splitting the
atom".
Lise Meitner and Otto Frisch estimated that the amount of energy release in a single fission
process was in the order of 200 MeV, far greater than any other known nuclear reaction. This
can be shown by comparing the masses of the fission products with the masses of the original
nucleus plus a neutron and converting this difference to energy.
It soon became evident, as had been mentioned by Enrico Fermi, that neutrons should be
emitted during fission due to the general structure of the atoms of uranium and the fission
products. This was subsequently confirmed by different researchers.
5 Nuclear Energy
It had already been recognized that mass could be converted into energy and that so-called sub-
atomic particles existed within atoms. When nuclear fission by neutrons was discovered and
found to produce further neutrons along with fission fragments, the possibility of a branching
chain of fissions became a real prospect. If these occurred in a rapid sequence, given the
amount of energy released in each fission event, the result could be a catastrophic explosion. At
that time, interest became focused on the possibility of a powerful atomic bomb. Subsequently,
from 1940 on, all further work related to the use of nuclear energy to produce an atomic bomb
took place in secrecy. The limits to this possibility became evident with further research that
showed that a nuclear chain reaction could be sustained only in uranium-235 and plutonium-
239. The former could be separated by gaseous diffusion and the latter created by a controlled
neutron chain reaction. Both methods required extensive equipment to produce even small
amounts.
The production of Pu-239 required an operating nuclear reactor with suitable fuel and modera-
tor to establish a chain reaction with U-235 in which neutrons were produced, some of which
would be absorbed in U-238 to create Pu-239. To construct a reactor in which a controlled chain
reaction would occur, Enrico Fermi and Leo Szilard realized that a heterogeneous system of
lumps of uranium embedded in graphite blocks in a lattice formation would be required. The
first experimental lattice was erected in 1941 at Columbia University under the supervision of
Fermi and followed by a larger one shortly thereafter, but impurities in the materials prevented
a chain reaction from being initiated. Towards the end of 1942, sufficient amounts of pure
materials were available, and a sufficiently large pile of graphite blocks and lumps of uranium
was built at the University of Chicago to establish a continuous self-sustaining chain reaction on
December 2, 1942.
The next step was an enormous scale-up to the plutonium production plants at Hanford Works,
where sufficient plutonium was produced to enable the first atomic bomb to be tested on July
16, 1945. Concurrently, a huge gaseous diffusion plant was built at Oak Ridge to separate U-235
from U-238 and so create highly enriched uranium for a second atomic bomb.
6 Acknowledgments
Bob Tapping
Jeremy Whitlock
Bhaskar Sur
Terry Rogers
George Bereznai
1
CHAPTER 2
Genealogy of CANDU Reactors
prepared by
Dr. Robin Chaplin
Summary:
This chapter discusses the historical beginnings and evolution of the CANDU reactor. The
research and prototype reactors leading to the CANDU design are described with some of their
principle technical parameters. Reasons for the choice of key parameters are given. This leads
into a review of the evolution of the design of increasingly larger commercial CANDU reactors.
The CANDU 6 reactor design has been chosen as a reference and a brief description of its main
components with typical design parameters follows with a comparison with the CANDU 9
reactor. The chapter includes a note on the advantages of the CANDU reactor compared with
other water cooled reactors and a general review of reactor safety as applicable to most water
cooled reactors. It concludes with some technical details of the proposed Advanced CANDU
reactor for comparison with existing commercial CANDU reactors.
Table of Contents
1 Canadian Historical Review..................................................................................................... 3
1.1 Canadian Nuclear Beginnings ........................................................................................... 3
2 Canadian Research Reactors................................................................................................... 4
2.1 Small Research Reactors................................................................................................... 4
2.2 The NRX Reactor............................................................................................................... 4
2.3 The NRU Reactor .............................................................................................................. 5
3 Development of the CANDU Reactor...................................................................................... 6
3.1 The NPD Reactor............................................................................................................... 6
3.2 Douglas Point.................................................................................................................... 6
4 Commercial CANDU Reactors ................................................................................................. 7
4.1 Commercial Reactor Development .................................................................................. 7
4.2 CANDU Reactors in Service............................................................................................. 11
4.3 Direct Steam Generation ................................................................................................ 13
5 The Current CANDU Reactor Design ..................................................................................... 14
5.1 Plant Arrangement ......................................................................................................... 14
5.2 Fuel Channel Conditions................................................................................................. 16
5.3 Comparison of CANDU Reactors .................................................................................... 16
5.4 Power Density................................................................................................................. 17
6 CANDU Technical Parameters ............................................................................................... 18
6.1 Reference Plant............................................................................................................... 18
6.2 Reactor Core Arrangement............................................................................................. 19
6.3 Coolant Loop Arrangement ............................................................................................ 26
©UNENE, all rights reserved. For educational use only, no assumed liability. Genealogy of CANDU Reactors – September 2016
6.4 Steam Generators........................................................................................................... 29

6.5 Steam Turbines ............................................................................................................... 30
6.6 Technical Data................................................................................................................. 32
7 Possible CANDU Reactor Development ................................................................................ 32
7.1 CANDU Advantages ........................................................................................................ 32
7.2 Future Prospects............................................................................................................. 33
7.3 Safety Aspects................................................................................................................. 34
7.4 The Advanced CANDU Reactor....................................................................................... 35
8 Problems ............................................................................................................................... 38
9 References and Bibliography................................................................................................. 39
10 Acknowledgements............................................................................................................... 40
List of Figures
Figure 1 Diagrammatic cross section of a typical CANDU (PHWR) ............................................... 14
Figure 2 Moderator and coolant circuits ...................................................................................... 15
Figure 3 Point Lepreau nuclear generating station....................................................................... 19
Figure 4 Reactor vault and assembly ............................................................................................ 20
Figure 5 37-element fuel bundle .................................................................................................. 21
Figure 6 CANDU 6 reactor assembly ............................................................................................. 23
Figure 7 Reactor cross section ...................................................................................................... 24
Figure 8 Reactor longitudinal section ........................................................................................... 25
Figure 9 Reactor plan view............................................................................................................ 26
Figure 10 Fuel channel end fittings on reactor face ..................................................................... 27
Figure 11 Feeder tube assembly on reactor face.......................................................................... 28
Figure 12 Steam generator for CANDU system............................................................................. 29
Figure 13 600 MW steam turbine for nuclear unit ....................................................................... 31
Figure 14 Turbine generator for nuclear unit................................................................................ 31
Figure 15 Typical capacity factors of CANDU and other reactors ................................................. 33
Figure 16 ACR 1000 nuclear systems schematic........................................................................... 36
Figure 17 ACR CANFLEX fuel bundles............................................................................................ 36
Figure 18 Comparison of core sizes .............................................................................................. 38
List of Tables
Table 1 Technical data for NRX........................................................................................................ 5
Table 2 Canadian CANDU reactors: capacity and service date ..................................................... 12
Table 3 Foreign CANDU 6 reactors: capacity and service date ..................................................... 13
Table 4 Comparison of CANDU reactor types ............................................................................... 18
Table 5 CANDU core parameters................................................................................................... 22
Table 6 CANDU coolant system parameters ................................................................................. 28
Table 7 CANDU steam generator parameters............................................................................... 30
Table 8 Steam turbine parameters................................................................................................ 32
Table 9 Advanced CANDU reactor parameter comparison........................................................... 37
Genealogy of CANDU Reactors 3
1 Canadian Historical Review

1.1 Canadian Nuclear Beginnings
During the first part of the 20th Century, significant discoveries were made in the field of nuclear
physics and chemistry, leading to the splitting of the atom and release of energy. Canada
became an important player in the development of nuclear technology during and after World
War II as a result of the Allied war effort to develop the atomic bomb. Although not directly
involved in the development of the bomb, Canada contributed a significant amount of related
nuclear research.
Canadian history in nuclear physics essentially started in 1898 when Ernest Rutherford, a New
Zealand born British scientist, was appointed Professor of Experimental Physics at McGill
University. His early work here was related to radioactivity, and he won a Nobel Prize for this
contribution to scientific knowledge in 1908. One of his students, John Cockcroft, who received
a Nobel Prize jointly with Ernest Walton in 1951, subsequently came to Canada in 1944 to direct
the Montreal Laboratory which had been set up in 1942 as a joint British-Canadian effort. In the
intervening period, much fundamental research had been done, and consideration was then
being given to the design and building of a reactor which would require more extensive labora-
tories to be located at Chalk River.
Concurrent work in the United States had led to the building of the first nuclear reactor, in
which a self-sustaining nuclear reaction took place in 1942. This reactor was built of graphite
blocks in which were distributed lumps of uranium. It was known as a pile due to the method of
construction.
During the war years, emphasis in the United States was directed towards the development of
the atomic bomb. This required highly enriched uranium or plutonium. Whereas the former
required extensive isotope separation facilities, the latter could be created from non-fissile
uranium in a reactor and chemically separated from the spent fuel. There was therefore a
concerted effort in the United States to build suitable reactors. Therefore, after the war, nuclear
research was also directed towards the design of reactors for power production.
After the war, Britain, without enrichment facilities and with no source of heavy water, followed
the route of graphite moderated natural uranium reactors to produce plutonium for nuclear
weapons, while at the same time developing reactors for power production. As a result, Calder
Hall became the first commercial nuclear power plant in the world when the first of four 60
MWe units was connected to the grid on 27 August 1956. It continued to operate until 2003. In
the United States, with uranium enrichment facilities in place, the focus was on the develop-
ment of light water moderated reactors using enriched uranium. Such reactors were highly
suitable for submarine propulsion because the reactors were very compact. Furthermore, the
vessels could remain submerged for extended periods and could operate extensively between
refuellings. The expertise thus developed in steam system components for submarine propul-
sion greatly influenced and facilitated the design of Canadian power reactors. Canada had
supported the war effort by research related to heavy water physics and therefore had the
expertise to direct its efforts into the development of heavy water moderated reactors using
natural uranium as fuel. In summary, each country focused its efforts in different directions
depending upon its relative expertise and resources.
Work in Canada was directed towards use of heavy water as a moderator, and to test the
concept, a Zero-Energy Experimental Pile known as ZEEP was built at Chalk River. It was essen-
tially an aluminum cylinder 2 m in diameter and 2.5 m in height surrounded by blocks of graph-
ite. Uranium metal rods clad in aluminum were hung vertically from the top, and heavy water
could be pumped in slowly at the bottom. ZEEP was completed and went critical on 5 Septem-
ber 1945, with almost exactly the amount of heavy water that had been predicted by theory. It
was the first nuclear reactor to operate outside the United States. This confirmed the design
parameters that would be used for the fuel rod lattice for the first functional reactor, designated
NRX. ZEEP, however, continued to be used in later years for experimental studies of neutron
behaviour.
2 Canadian Research Reactors

2.1 Small Research Reactors
The Zero-Energy Experimental Pile or ZEEP, which had an operating power of just 1 W, enabled
tests to be carried out to determine the characteristics of heavy water and uranium lattices. It
was also used to determine the neutron energy spectrum in various regions of a fuel bundle.
A larger version of ZEEP, known as ZED-2, was used for studies of fuel and irradiation effects as
well as a variety of possible lattices for power reactors. Fuel bundle designs from 7-element to
37-element assemblies were tested to determine their performance.
The Pool Test Reactor, or PTR, was constructed, as its name implies, near the bottom of a deep
pool of light water. Because this light water was both moderator and coolant, slightly enriched
uranium had to be used as fuel. This arrangement with an easily accessible core made it possi-
ble to shuttle samples of fuel and other materials in and out of the reactor core to determine
their effect on reactivity as well as to study short term radiation effects on materials.
2.2 The NRX Reactor

The first Canadian nuclear reactor to be designed to operate with a positive power output due
to its generation of excess heat was the NRX. This reactor started operation in 1947 with a
design thermal power rating of 10 MW, which was increased to 42 MW by 1954. At the time of
its construction, it was the most powerful nuclear research reactor in the world and became the
most intense source of neutrons available anywhere. It also suffered one of the first major
nuclear reactor accidents in the world (power excursion to 80 MW and partial core meltdown),
but was rebuilt and restarted within two years. It continued in service until 1993, which was
quite a remarkable achievement for an experimental reactor.
The NRX reactor consisted of a vertical cylindrical aluminum calandria with a diameter of 8 m
and a height of 3 m. The calandria contained the heavy water moderator. The core consisted of
a hexagonal array of approximately 175 tubes with a diameter of 60 mm. These tubes held fuel
elements, control rods, and experimental devices. Of these, 12 held control rods made of steel
tubes containing boron carbide powder. The fuel elements consisted of uranium metal rods 3.1
m in length and 31 mm in diameter sheathed in aluminum. Each was surrounded by an alumi-
num tube through which flowed cooling water from the Ottawa River. An air flow in the gap
between the calandria tube and the coolant tube was maintained to limit heat flow to the
moderator to maintain it at a low temperature.
The key technical parameters for the NRX reactor are given in Table 1. The modern CANDU
reactor has all these key elements in a similar but horizontal arrangement.
Table 1 Technical data for NRX

Parameter Characteristic
Reactor shape cylindrical

Reactor orientation vertical
Calandria diameter 8m
Calandria height 3m
Fuel channel arrangement vertical hexagonal array
Number of fuel channels ~175
Fuel channel diameter 60 mm
Fuel channel material aluminum
Fuel uranium metal
Fuel rod length 3.1 m
Fuel rod diameter 31 mm
Fuel rod mass 55 kg
Fuel cladding material aluminum
Moderator heavy water
Moderator quantity 14 m3
Coolant light water
Coolant flow 250 kg/s
Annulus gas air
Annulus flow 8 kg/s
Control rods steel tubes containing boron
carbide
2.3 The NRU Reactor

The National Research Universal (NRU) Reactor was a significantly advanced version of the NRX.
Design was started in 1949 after NRX had been put into operation, and it started self-sustained
operation in 1957, some 10 years after NRX. It was designed for a thermal output of 200 MW
using natural uranium, but was converted to 60 MW using highly enriched uranium in 1964 and
converted again to 135 MW using low enriched uranium in 1991. The NRU supported the
Canadian Neutron Beam Centre and served as a test bed for the development of the CANDU
reactor. Like NRX, it was when built the most powerful source of neutrons in the world and
became the world’s leading supplier of radioactive isotopes. It suffered an accident in 1958
when a fuel rod caught fire on removal from the reactor and caused radioactive contamination
within the building and to a lesser degree on the surrounding site. As with NRX, the NRU was
designed to use natural uranium as fuel and heavy water as moderator, but was cooled by heavy
water. With a view to later commercial power applications, it was designed for on-load refuel-
ling.
3 Development of the CANDU Reactor

3.1 The NPD Reactor
The Nuclear Power Demonstration or NPD Reactor was partially modelled on the NRU reactor
and was intended to be the first Canadian nuclear power reactor and a prototype for the
CANDU design. Design work commenced in 1955, with the important components being
designed and fabricated by Canadian General Electric. As a power producer, it had to operate at
high temperatures sufficient to generate steam to drive a power producing turbine. These
requirements presented some unique technical challenges due to the pressures under which
the systems would have to operate. The design electrical output was to be 22 MW. Use of
heavy water as a coolant and on-load refuelling had been proven with the NRU reactor, but the
higher temperatures and pressures had to be accommodated by some significant changes.
Firstly, the fuel was changed from uranium metal to uranium dioxide, which could withstand
higher temperatures and was dimensionally stable over long irradiation periods. Secondly, the
fuel cladding was changed from aluminum to zirconium alloy, which could also withstand higher
temperatures and was corrosion resistant as well as being essentially transparent to neutrons.
Thirdly, the fuel channels rather than the calandria were pressurized, leading to a pressure tube
design rather than a pressure vessel design, and these pressure tubes would be of zirconium
alloy. The pressure tube design led to further considerations to accommodate on-load refuel-
ling. With horizontal pressure tubes rather than vertical tubes as in the NRX and NRU reactors,
the fuel could be made into easily handled bundles which could be pushed in from one end and
recovered from the other during refuelling while the reactor was at full power. Several unique
safety features, such as safety shutdown systems with testability during operation, decay heat
removal by natural circulation after shutdown, and a containment to prevent the release of
radioactive material in the event of an accident, were incorporated. The NPD reactor started
operation in 1962 and was taken out of service in 1987, by which time it had fulfilled its original
purpose as a prototype and would have required re-tubing. It had also been an important
training centre for future CANDU nuclear plant operators.
3.2 Douglas Point

The Douglas Point plant was the first full-scale CANDU nuclear generating station and was built
on the shores of Lake Huron where the Bruce Nuclear Power Complex is now located. The
proposed electrical output was 200 MW, and approval was given in 1959. It was basically a
scale-up of the NPD reactor with similar design and components. It delivered power to the
Ontario Hydro electrical grid system from 1967 until 1984, when replacement of the pressure
tubes became necessary and could not be justified on economic grounds. Valuable experience
was gained on this plant which was subsequently applied to later CANDU nuclear power plants.
4 Commercial CANDU Reactors

4.1 Commercial Reactor Development
The design of the CANDU reactor began with the Nuclear Power Demonstration (NPD) plant.
Certain parameters had to be established within the constraints of existing technology. Basi-
cally, the reactor had to be fuelled with natural uranium, with heavy water as a moderator. This
required a heterogeneous core arrangement with an array of fuel channels through which
coolant flowed. Initial studies suggested a fuel channel cross section of 50 cm2, which trans-
lated into a standard size fuel channel 3.25 in (83 mm) in diameter. To promote good heat
transfer from the fuel, a modified hexagonal array of 19 rods or elements was chosen. The
minimum spacing between rods based on laboratory testing was determined to be 0.050 in
(1.27 mm). Based on irradiation testing, uranium dioxide was shown to have better dimensional
stability and corrosion resistance than uranium metal at high burnup levels. This choice did,
however, reduce the uranium concentration in the fuel and increased the need for good neu-
tron economy. Zirconium had been selected as fuel cladding due to its corrosion resistance and
low neutron absorption, but the quantity of it around the fuel had to be reduced as much as
possible. Although a cladding thickness of 0.020 in (0.51 mm) was recommended by the tube
supplier, further experimental work led to a thickness of 0.015 in (0.38 mm) and later to 0.013
in (0.33 mm). This resulted in collapsible cladding where the tube wall was compressed onto
the fuel once in service.
A major decision was to adopt a pressure tube design rather than a pressure vessel design to
avoid the difficulty of constructing a sufficiently large vessel for a heavy water moderated
reactor. Zirconium alloy pressure tubes had sufficient neutron transparency to make this
possible. A further major decision was to orient the fuel channels horizontally instead of
vertically to facilitate on-load refuelling. The fuel could be made in the form of bundles which
could be pushed through the channels with no need to link them together. These decisions on
reactor configuration made the CANDU unique with regard to other commercial reactor devel-
opments. Although some were designed with pressure tubes or separate fuel bundles, only the
CANDU has horizontal fuel channels.
With a selected fuel channel length of 4 m, the next question was the optimum number of fuel
bundles. A significant advantage of separate fuel bundles was that the entire fuel channel need
not be refuelled at one time. Low burnup fuel could be retained while new fuel was added. An
analysis of diminishing returns indicated that having more than eight bundles would not yield
any advantage, assuming a cosine neutron flux distribution along the fuel channel. This gave a
fuel bundle length of 50 cm in a 4 m long fuel channel. This translated into a standard fuel
bundle length of 19.5 in (495 mm), which was adopted for future CANDU reactors.
A fuelling machine was required at each end of the fuel channel. One supplied fresh fuel from
one end, while the other received spent fuel at the other end. They had to connect to each fuel
channel in turn and remove and re-insert closure plugs at the beginning and end of the refuel-
ling process. The fuel configuration in the form of separate bundles facilitated the design of
these machines because they did not have to handle a very long fuel assembly. In fact, the
design of the fuel string and machines was dictated by the space available in the rock excava-
tion, which had been dimensioned on the assumption that the reactor would have a vertical
orientation. Of necessity, therefore, the fuelling machines were designed each with a rotating
magazine with slots to hold each bundle separately. A significant advantage of this arrangement
was that fuelling could take place from either end of the reactor. Such bi-directional fuelling
made it possible to balance the tendency for flux skewing in a partially refuelled channel by
fuelling in the opposite direction in an adjacent channel. In the NPD design, coolant flowed in
the direction opposite to the fuelling direction, so that the fuel bundles were held in place by
the flow of coolant. This established the characteristic bi-directional coolant flow in CANDU
reactors.
The calandria vessel and calandria tubes were of aluminum. Initially, there was no spacer
between the calandria tubes and the fuel channel pressure tube, but as the design for Douglas
Point evolved, a single spacer in the form of an Inconel wire garter spring was introduced. This
enabled the pressure tubes to withstand the effect of creep sag. The calandria had a light water
neutron reflector surrounding it in a radial direction. To bring about a shutdown, the heavy
water moderator could be rapidly drained into a dump tank, which for a small reactor was an
effective way of ensuring safe shutdown in the event of an accident.
The heat transport system which removed heat from the reactor core and generated steam in a
steam generator required pumps to circulate the coolant. Originally, completely enclosed
canned pumps were proposed, but these were later changed to conventional vertical pumps
with shaft seals. These had the advantage of being able to incorporate a flywheel to provide
increased rotational inertia, which would extend the rundown period after a power failure and
enable extra core cooling during the initial stages of high decay heat. The steam generator
design was based on those being designed for United States nuclear submarines. It consisted
essentially of a U-shaped horizontal heat exchanger in which the heavy water coolant trans-
ferred heat to the light water of the steam system. Because the latter operated at a lower
pressure, some steam was generated around the tubes. Risers and downcomers circulated the
steam and water mixture to a steam drum and returned the water to the heat exchanger. The
long circular horizontal steam drum was similar to that of a conventional fossil fuel fired boiler
and had internal cyclones for effective steam separation. The overall configuration required
inlet and outlet headers at each end of the reactor to accommodate bi-directional coolant flow.
The former were supplied by three 50% capacity pumps, and the latter delivered coolant to a
single steam generator.
The NPD reactor was followed very closely by the Douglas Point reactor, the latter being essen-
tially an enlarged version capable of operating like a commercial power plant and producing
electric power to the grid system on a commercial scale. The site chosen was on Lake Huron,
where there was an adequate supply of cooling water to sustain large scale power production.
The most significant change was an upgrade in power from 20 MW electrical to 200 MW
electrical. This naturally required a larger and more robust calandria, and therefore the struc-
tural material was changed from aluminum alloy to stainless steel and the calandria tubes from
aluminum alloy to zirconium alloy. Because stainless steel is a stronger neutron absorber than
aluminum, the calandria diameter was further enlarged to accommodate an internal heavy
water reflector instead of an external light water reflector. The axial shields were moved
inwards so that the fuel channel end fittings would be outside the shield and therefore more
easily accessible.
The increased size of the Douglas Point reactor required fuel channels with a length of 5 m. This
meant more fuel bundles per channel. The one-eighth length fuel bundle was based on a
cosine flux distribution, but with flux flattening, such short lengths were not required, and one
third length bundles would give virtually optimum fuel burnup. Because the same standard
length fuel bundles would be used, the fuelling machines were therefore redesigned to handle
two bundles in each slot. This enabled the magazine to be smaller in diameter, but able to
accommodate 39 in (991 mm) of fuel with each fuel movement. Other changes in fuelling-
machine design provided positive control from both ends of the fuel channel during refuelling
and enabled refuelling in the same direction as coolant flow. The advantage of this was that
fresh fuel, which would give a higher local power output, would be in contact with cooler water
when only part of the fuel channel had been refuelled.
The larger output required increased steam generating capabilities. Furthermore, the longer
reactor put the inlet and outlet of each channel further apart. Therefore, it was decided to put
steam generators and circulating pumps at each end of the reactor. The same concept of
separate heat exchangers and steam drums was retained, but instead of one horizontal heat
exchanger, several vertical hairpin tube exchangers were used to feed a single steam drum in a
way similar to that in which steam is generated in the water walls of a fossil fuel fired boiler.
The multiple smaller heat exchangers could be replaced more easily than a single large one
should tube problems arise. With steam generators and circulating pumps at each end of the
reactor and with bi-directional flow, the “figure of eight” configuration for the heat transport
system was established for this and all future CANDU reactors. A further feature of Douglas
Point was the introduction of horizontal inlet and outlet headers above the highest fuel chan-
nels, but below the steam generators and circulating pumps. This would enable the system to
be partially drained for maintenance on the circulating pumps and steam generators while still
maintaining fuel cooling to remove decay heat. In normal operation, pressure was maintained
in the heat transport system by a feed-and-bleed system to minimize heavy water inventory.
Pickering was the first large scale multi-unit commercial CANDU plant. This was yet a further
scale-up from the 200 MW Douglas Point plant to a 4 x 500 MW plant. The 500 MW electrical
unit size was consistent with that of the coal fired units on the Ontario Hydro system. Signifi-
cant changes to the reactor included longer and wider fuel channels to accommodate the
increased fuel inventory. The fuel channels were 6 m in length instead of 5 m. Rather than
increasing the number of fuel channels excessively, it was decided to increase the diameter to a
nominal 4 in (102 mm) while retaining the fuel element or rod diameter. This resulted in fuel
bundles with 28 instead of 19 elements. This required a fuel channel 4.07 in (103 mm) in
diameter to maintain the standard minimum element spacing, which was maintained by pads
brazed to the element cladding. This then became the standard for all future CANDU reactors.
The end shields were redesigned, and the annular space between the pressure tube and calan-
dria tube filled with inert gas to minimize the risk of corrosion.
A further change was the adoption of a double “figure of eight” coolant loop with two hydrauli-
cally independent loops. This had the advantage of limiting the consequences of a loss of
coolant accident due to a pressure tube or feeder tube break because the fuel channels in half
the reactor would remain flooded without the need for emergency coolant injection. The
moderator dump system for emergency shutdown was retained, but due to the time needed to
drain the large calandria, it was supplemented with gravity operated shutoff rods. The latter
system was able to handle most reactor trips, and therefore a dump arrest system was added to
prevent dumping, enabling a faster reactor restart for such non-accident related trips.
The larger reactor introduced the possibility of flux oscillation, in which the peak flux wanders
from one region of the reactor to another, induced by xenon transients. To overcome this, the
reactor was divided into 14 zones, each with a neutron flux detector and a chamber in which an
appropriate amount of light water could act as a stronger neutron absorber than the surround-
ing heavy water moderator and thus achieve local flux control. This regional neutron flux
control system became a standard feature of all subsequent CANDU reactors.
An innovative change was the introduction of vertical steam generators which incorporated the
vertical U-tube heat exchange surface and steam separation cyclones into a single vessel with
provision for water recirculation and feedwater preheating. This design became the standard
for CANDU reactors and many pressurized light water reactors, or PWRs.
For a multi-unit station, the concept of a containment surrounding each reactor was maintained
and extended by linking the separate containments to a common vacuum building. In the event
of a major reactor accident and release of steam inside any containment, the vacuum building
would act to suck out and condense the steam, thus establishing a negative pressure in the
containment to prevent egress of radioactive material.
Bruce was the next large multi-unit nuclear generating station to be built. This involved a
change in the heat transport system. In previous designs, the outlet conditions of the fuel
channels varied due to the different power ratings of the channels across the reactor core. To
create more uniform conditions at the channel outlets, the central core region required a
greater degree of subcooling at the inlet. This was accomplished by having separate preheaters
outside some steam generators rather than integral preheaters and passing coolant for the
central region only through these preheaters. This complicated the flow system, but obtained
better matching of coolant outlet conditions. Another change was made to the steam genera-
tors, where the large U-tube heat exchanger as used in Pickering was retained, but without the
integral upper steam separating section. Instead, a single long horizontal steam drum was
attached to the top of all steam generators at each end of the reactor. The larger water-steam
interface reduced level control problems, but introduced thermal stress problems. This design
was used on Bruce A, while Bruce B reverted to the separate integral design as used at
Pickering.
Studies had indicated that a moderator dump system was too slow for a large reactor, and
therefore this was eliminated in favour of faster acting gravity shutoff rods. Because dumped
moderator would not be available for channel cooling in the event of loss of coolant, a new
system was developed for the next generation of reactors, but applied at Bruce. This was the
injection of high pressure water into the heat transport system headers from gas pressurized
storage tanks. For Bruce, the initial supply pressure was 800 lbf/in2 (5.52 MPa), but this was
reduced to 600 lbf/in2 (4.14 MPa) in subsequent reactors.
A further safety issue arose with regard to the shutdown system. Bearing in mind that the
CANDU design is a neatly balanced configuration of fuel and moderator, any disruption of this
configuration will render the reactor subcritical and inoperable. In the event of failure of the
shutdown system to operate in an overpower situation, the reactor would disassemble and shut
itself down. Further consideration of the core disassembly philosophy led to addition of a
second shutdown system which would remove the need to consider core disassembly for
licensing purposes. Because the shutoff rods operated in a vertical plane, a new system operat-
ing entirely independently on a horizontal plane was developed. Solid rods were considered,
but might suffer mechanical interference, so liquid rods were favoured. However, their effect
would be too localized, so liquid dispersion in the moderator was the ultimate choice. By
injecting neutron absorbing gadolinium nitrate solution into the moderator through an array of
nozzles in the reactor core, a very quick shutdown could be ensured. This system was driven by
helium in pressurized storage tanks and became standard on all CANDU reactors. Because
removal of the gadolinium nitrate from the moderator was a lengthy process, this became the
secondary shutdown system, while the shutoff rods remained the primary shutdown system
which operates first in adverse transient conditions requiring a reactor trip.
The next generation of CANDU units included Point Lepreau and Gentilly-II as well as a number
of overseas plants. A design for single unit stations was needed for these to be attractive to
small utilities. The major change was the requirement for a single large containment for the
reactor and heat transport system, including the steam generators, which would condense
released steam and contain radioactive products in the event of a large break and loss of
coolant accident. Other changes followed a natural evolution in reactor design. The fuel
bundles were modified from 28 to 37 element bundles by using smaller diameter elements.
The increased heat transfer surface thus created enabled a power increase from 500 MW to 600
MW electrical, while reducing the number of fuel channels from 390 to 380. This reactor
therefore became known as the CANDU 600. Furthermore, from 12 steam generators and 16
coolant pumps at Pickering, the design had evolved to 4 steam generators and 4 coolant pumps.
Experience in steam generator design had enabled larger components to be built. Comprehen-
sive testing had shown that some boiling could be permitted within the fuel channels, thus
increasing substantially the heat removal capability of the coolant. Darlington followed the
design of the CANDU 600, but with a degree of scaling up. The heat transport system main-
tained the same arrangement of two loops in a figure-of-eight configuration, with each loop
having a steam generator and a circulating pump at each end. For increased output, the reactor
was designed for 480 fuel channels instead of 380. For future reactors of this size, a change in
the figure-of-eight arrangement has been proposed. Instead of the two loops being in separate
halves of the reactor, they would be interleaved alternately between adjacent channels
throughout the reactor. The advantage of this would be to avoid a flux tilt and power pulse in
the event of a loss of coolant accident in one loop while still maintaining the advantages of the
two loop concept.
4.2 CANDU Reactors in Service

Table 2 lists the CANDU reactors in service in Canada with their commercial service start dates.
Of these, Point Lepreau and Gentilly 2 are both CANDU 600 or CANDU 6 reactors, which was the
basic reference design for reactors sold abroad. Table 3 lists CANDU 6 reactors in operation in
foreign countries. In this text, Point Lepreau has been selected as the reference plant because it
is typical of CANDU reactors on a worldwide basis.
Table 2 Canadian CANDU reactors: capacity and service date
Country Station Name Gross (MWe) Net (MWe) Service Date
Canada Pickering 1 542 515 1971

Canada Bruce 1 836 781 1977
Canada Point Lepreau* 680 635 1983
Canada Gentilly 2* 675 635 1983
Canada Darlington 1 935 881 1992
*CANDU-6 Reactor
Table 3 Foreign CANDU 6 reactors: capacity and service date
Country Station Name Gross (MWe) Net (MWe) Service Date
Argentina Embalse 1 648 600 1984

China Quinshan 4 700 640 1984
China Quinshan 5 700 640 2002
Romania Cernavoda 1 706 655 1996
Romania Cernavoda 2 706 655 2007
South Korea Wolsong 1 679 629 1983
4.3 Direct Steam Generation

Some reactors, such as the boiling water reactor or BWR, generate steam within the reactor
core. Due to the cost of heavy water and the formation of tritium, heavy water cannot be used
in the steam cycle because some leakage from this cycle is inevitable. Therefore, light water
must be used as the coolant. The CANDU with its separate moderator and coolant circuits
(unlike the light water PWR and BWR reactors in which the moderator and coolant are the same
and in a single loop) can be adapted for direct steam generation by using heavy water as mod-
erator in the calandria and light water as coolant in the fuel channels. Some boiling already
occurs in CANDU fuel channels, and this can be increased with vertical channels, where the
natural buoyancy effect enhances the forced circulation generated by the circulating pumps.
Steam is separated in horizontal steam drums similar to those in the early CANDU reactors.
Hence, we have the heavy water-light water reactor, or HWLWR.
The most successful of these was the 100 MWe Steam Generating Heavy Water Reactor, or
SGHWR, built in the United Kingdom. This reactor operated as a commercial prototype and
produced reliable power to the national grid system from 1968 to 1990, when it reached the
end of its design life. It had some novel design aspects, including direct injection of emergency
coolant into the centres of the fuel bundles in the event of a loss of coolant accident, and
served as a testing ground for various types of fuel. Because light water, which absorbs more
neutrons than heavy water, was used as coolant, slightly enriched uranium in the form of
uranium dioxide was required. The steam became slightly radioactive due to formation of
nitrogen-16 in the reactor, and therefore certain precautions had to be taken by operational and
maintenance staff when working around the steam turbine and feedwater system. Gentilly-1, a
similar Canadian design, designated the CANDU Boiling Water Reactor or CANDU-BWR and also
known as the CANDU Boiling Light Water Reactor or CANDU BLW, was modelled on the SGHWR,
but fuelled with natural uranium. Built for an output of 250 MWe, first power was produced in
1971, and full power was achieved in 1972. However, it did not operate successfully. The
reactor had a strong positive power coefficient of reactivity when additional voidage due to
steam formation in the coolant channels occurred. Although this could be corrected by ab-
sorber rods in the core, the reactor remained difficult to control, and its operation was sporadic
due to other problems such as condenser corrosion. It was shut down in 1979 and eventually
decommissioned in 1984.
5 The Current CANDU Reactor Design

5.1 Plant Arrangement
The general arrangement of the plant is shown in Figure 1. The active part of the reactor is
cylindrical in shape and set horizontally. The moderator is contained in the calandria, which is
about 6.0 m long and about 7.6 m in diameter for the CANDU 6. Because of the low internal
pressure in the calandria, the thickness of its shell and its tubes needs to be sufficient only to be
structurally sound and to support the weight of the moderator. Only the pressure tubes located
within the calandria tubes are pressurized. These pressure tubes, which contain the fuel and
through which the coolant flows, are arranged horizontally in an axial direction. The coolant is
fed to and from the pressure tubes by feeder tubes, which in turn are connected to headers
situated above the reactor. The headers in turn are connected to steam generators in which
heat from the coolant is used to generate steam. Circulating pumps, usually known as heat
transport pumps, drive the coolant around the primary circuit. A pressurizer maintains pressure
in the primary system to suppress large scale boiling and maintain operating temperatures.
Figure 1 Diagrammatic cross section of a typical CANDU (PHWR)

The moderator absorbs some energy from the neutrons as they are slowed down and also
receives some heat by conduction and radiation through the annulus gas between the pressure
tubes and calandria tubes as well as conduction through the tubes themselves. The low con-
ductivity of the annulus gas provides the greatest resistance to heat transfer. Cooling of the
moderator is therefore required to maintain a temperature of about 70°C, which means that a
circulating and cooling system is required. This is shown diagrammatically in Figure 2.
Figure 2 Moderator and coolant circuits

One characteristic when using heavy water as a moderator is that the neutron slowing down
distance is relatively long. This determines the spacing between the fuel channels because
neutrons need to be able to reach thermal equilibrium with the moderator before entering the
next fuel channel. Too short a distance results in under-moderation and too long a distance in
over-moderation. CANDU reactors are slightly over-moderated, but while this has some disad-
vantages, it does provide additional space between the fuel channels for various control de-
vices. Over-moderation enables additional neutrons to be absorbed in the moderator, and
when voidage occurs in the fuel channels, fewer neutrons are absorbed, thus increasing the
number of neutrons and giving a positive void coefficient of reactivity which must be counter-
acted in other ways. With a separate calandria, the moderator temperature can be kept rela-
tively low compared with the fuel and coolant. This is advantageous in reducing neutron energy
and promoting the fission process due to the increased fission cross section of uranium at lower
temperatures. In all CANDU reactors, the moderator in the calandria is maintained at atmos-
pheric pressure and a temperature of about 70°C.
5.2 Fuel Channel Conditions

In the CANDU reactor, coolant enters the fuel channels of the core in a subcooled state, but not
excessively subcooled. It leaves the most highly rated fuel channels with a vapour content of a
few percent, with boiling having occurred. The lower rated channels have little or no boiling. In
this way, nearly constant reactor coolant outlet temperatures can be maintained even though
some channels produce more heat than others. Furthermore, the outer channels with the
lowest rating can have their flow slightly restricted to enhance their coolant outlet tempera-
tures. Note in this context that for good thermodynamic efficiency, the cycle working fluid
should receive heat at the highest possible temperature. Hence, the reactor primary coolant
should be at the highest possible temperature so that it can supply heat to the secondary
working fluid at a suitably high temperature as well. Limited boiling in some fuel channels of a
CANDU reactor enables the average temperature of the coolant entering the steam generators
to be very near saturated conditions. An alternative view of the above review is that, by allow-
ing some boiling to occur in the fuel channels, the pressure of the primary circuit can be lower
while still maintaining desired steam generator temperatures. This has significant benefits in
the design and manufacture of a rather complex primary system with multiple pressure tubes
and feeder pipes.
Within individual fuel channels where boiling does occur, conditions are not uniform across the
fuel bundles. The circular arrangement of fuel elements in each bundle leads to irregularly
shaped and sized coolant channels. More boiling occurs in narrow channels than in wider ones.
The horizontal arrangement of the fuel bundles also creates non-uniform mixing between the
coolant channels because the lighter vapour tends to migrate upwards, resulting in some
vertical segregation of vapour and liquid across the channel. Although turbulence, particularly
at the abutting ends of fuel bundles, tends to equalize conditions, some fuel pins are subject to
more boiling than others. Boiling does enhance heat transfer, but excessive vapour is detrimen-
tal, and therefore fuel sheath temperatures do vary slightly across any particular fuel bundle.
5.3 Comparison of CANDU Reactors

The CANDU reactor has evolved progressively and has been marketed in several other coun-
tries. Three standardized versions, CANDU 300, CANDU 600, and CANDU 900, with different
capacities were developed. These are now known as CANDU 3, CANDU 6, and CANDU 9 respec-
tively. The original numbers indicate roughly the electrical output in megawatts, but later
developments have enabled much greater outputs to be achieved by the respective models.
Table 4 gives key technical parameters and shows the range in outputs available from these
three models built up with similar basic components.

Although the CANDU 6, for example Point Lepreau, is representative of CANDU reactors in
general and most existing CANDU reactors are of this capacity, the CANDU 3 and CANDU 9, for
example Bruce and Darlington, were marketed as alternate capacities to suit smaller or larger
electrical utility systems respectively. The CANDU 9 is very similar to the CANDU 6, being simply
larger in diameter and with increased heat transport capacity. It therefore has the benefit of
economies of scale. The CANDU 3, however, was a novel development to minimize the adverse
effects of economies of scale, but no examples have been built. It is like a half CANDU 6 with a
single loop instead of a “figure of eight” loop with two steam generators at one end and two
heat transport pumps at the other, thus maintaining essentially the same size components as
the CANDU 6. Coolant flow is in one direction only, and only one refuelling machine is used,
which is located at the coolant outlet end so that coolant flow aids the discharge of fuel bun-
dles. A comparison of these three basic designs is given in Table 4. The technical parameters
given in Table 4 and those given in Section 6 were extracted from the Atomic Energy of Canada
publication “Technical Summary CANDU Nuclear Generating Station” which gives the parame-
ters for the 600 MW and 950 MW units (subsequently the CANDU 6 and CANDU 9). The com-
plete tabulated parameters are given in Appendix C to this chapter. Since publication these
parameters have been revised as new developments have allowed the design to evolve particu-
larly with regard to channel power. The parameters given for CANDU 9 (950 MW unit) are
therefore not representative of the latest plants for example Darlington which has a lesser
number of fuel channels each of which have a higher power rating. However Appendix C is
useful for comparison of the two unit sizes and serves well for educational purposes as the data
is based on the same design philosophy. A later AECL publication in 1997 “CANDU Technical
Summary” gives some technical parameters for the CANDU 6 but not for the CANDU 9.
As the CANDU design evolved with the proposed Bruce B and the later Darlington nuclear plants
these were nominally in the 800 MW range resulting in the CANDU 8 design. The AECL publica-
tion in 1989 “CANDU 8 Technical Outline” lists the net output of Bruce B as 825 MW and of
Darlington as 881 MW. Subsequently the terminology was rationalized as in Appendix B
“CANDU: The Evolution”. This shows the development of the CANDU up to the four-unit CANDU
600 MW class (Pickering) and the subsequent division into the single-unit CANDU 6 or 700 MW
class (Point Lepreau onwards) and the four-unit CANDU 900 MW class (Bruce A & B and Darling-
ton) with the further development of the single-unit CANDU 9 also 900 MW class.
5.4 Power Density

Power density is a measure of how much power is generated per unit volume of core. Gener-
ally, the higher the power density, the more compact is the reactor and the lower the capital
cost. In all reactors, therefore, there is an incentive firstly to maintain a high neutron flux (to
achieve a high heat release rate) right across the reactor core and secondly to maintain a high
heat flux (to achieve a high heat removal rate) everywhere in the reactor.
All CANDU reactors contain roughly the same amount of in-core fuel per channel, but the major
differences are in the fuel bundles themselves and in the number of channels. The total design
power output of a CANDU reactor can be fixed by selecting an appropriate number of fuel
channels for the reactor. In a CANDU 3 reactor there are 232 fuel channels, in the CANDU 6
reactor the number of channels is increased to 380, while in the CANDU 9 reactor there are 600
fuel channels.
A new development is the Advanced CANDU Reactor, known as the ACR, which uses slightly
enriched uranium as fuel and light water as coolant. This has the benefit of significantly reduc-
ing the heavy water inventory. In the proposed ACR, the power density is increased by a smaller
calandria diameter, but with an increased number of elements in the fuel bundle and a smaller
channel pitch, resulting in a more compact reactor.
Table 4 Comparison of CANDU reactor types
Parameter Units CANDU 3 CANDU 6 CANDU 9

Moderator D 2O D 2O D 2O
Coolant D 2O D 2O D 2O
Number of fuel channels 232 380 600
Fuel UO2 UO2 UO2
Number of elements in bundle 37 37 37
Number of bundles in channel 12 12 12
Number of steam generators 2 4 8
Number of heat transport 2 4 4
pumps
Reactor outlet pressure MPa 10.0 10.3 10.3
Reactor outlet temperature °C 310 312 312
Reactor coolant flow rate kg/s 5 300 7 600 13 500
Steam pressure MPa 4.7 4.7 5.1
Steam temperature °C 260 260 265
Steam flow rate kg/s 700 1 050 1 610
Total fission heat MW 1 441 2 156 3 394
Net heat to steam cycle MW 1 390 2 060 3 347
Gross turbine generator output MW 470 676 1 121
Net electrical output MW 450 626 1 031
6 CANDU Technical Parameters

6.1 Reference Plant
Point Lepreau has been chosen as the reference for this text because it is a typical CANDU 6
reactor. Technical parameters for other CANDU reactors are given here for comparison to show
the main differences, but other chapters will generally refer to Point Lepreau characteristics.
Figure 3 shows an external view of Point Lepreau. Appendix A shows a cutaway view of a
CANDU 6 reactor plant.
Figure 3 Point Lepreau nuclear generating station
6.2 Reactor Core Arrangement

The reactor core is cylindrical, but is set horizontally, as shown in Figure 4, with horizontal fuel
channels. The heavy water moderator is contained in a cylindrical calandria with multiple tubes
set in a square array and in an axial direction so that the moderator surrounds all the tubes.
Pressure tubes containing the fuel bundles pass through the calandria tubes. Heavy water
coolant under a pressure of about 10 MPa flows through the pressure tubes to remove heat
from the fuel bundles. Because the coolant is heavy water, there is minimal absorption of
neutrons. Each pressure tube contains twelve fuel bundles, each of which has 37 fuel elements
arranged in a circular pattern, as shown in Figure 5. The fuel elements consist of zirconium-alloy
tubes filled with natural uranium dioxide pellets. A feature of the CANDU is that individual fuel
channels can be refuelled while the reactor is at full power by pushing new fuel elements in at
one end of the pressure tube and removing spent fuel elements at the other end, using special
refuelling machines which can be attached to the ends of any pressure tube.
Figure 4 Reactor vault and assembly
Figure 5 37-element fuel bundle

Control of the reactor is achieved by varying the amount of light water in special liquid level
tubes in separate control zones in the calandria and by manipulation of control rods which are
inserted into special vertical channels within the calandria and pass at right angles between the
calandria tubes. Typical parameters for 600 MWe (electrical) and 950 MWe (electrical) CANDU
reactor cores are given in Table 5.
Table 5 CANDU core parameters
Parameter CANDU 600 CANDU 900
Number of fuel channels in core 380 600

Number of fuel bundles per channel 12 12
Number of fuel elements per bundle 37 37
Number of pellets per rod 30 30
Fuel pellet material UO2 sintered UO2 sintered
Fuel cladding material Zircaloy-4 Zircaloy-4
Fuel channel array square square
Fuel channel lattice pitch 286 mm 286 mm
Fuel element configuration circular circular
Fuel bundle length 495 mm 495 mm
Fuel bundle diameter 102.4 mm 102.4 mm
Fuel element diameter 13.08 mm 13.08 mm
Fuel pellet diameter 12.16 mm 12.16 mm
Mass of uranium dioxide in core 95 Mg 153 Mg
Mass of uranium in core 84 Mg 135 Mg
Fuel type natural U natural U
Average core power density ~ 11 MW/m3 ~ 11 MW/m3
Active core length 5.94 m 5.94 m
Extrapolated core length 6.06 m 6.06 m
Equivalent core diameter 6.27 m 7.90 m
Total core fission power 2 180 MW 3 394 MW
Total core thermal power 2 060 MW 3 237 MW
Maximum channel power 6.5 MW 6.5 MW
Maximum bundle power 0.8 MW 0.8 MW
Active heat transfer area ~ 3 430 m2 ~ 5 420 m2
Average heat flux ~ 600 kW/m3 ~ 600 kW/m3
Maximum heat flux ~ 1 000 kW/m3 ~ 1 000 kW/m3
The reactor core includes monitoring and control devices to measure neutron flux and to
modify the flux profile if required to obtain the best power distribution. Safety devices to bring
about rapid shutdown, such as the shutoff rods and poison injection system, are also built into
the core. The general arrangement of these is shown in Figure 6, and their locations are shown
in the three views given in Figures 7, 8, and 9.
Figure 6 CANDU 6 reactor assembly
Figure 7 Reactor cross section
Figure 8 Reactor longitudinal section
Figure 9 Reactor plan view
6.3 Coolant Loop Arrangement

The pressure tubes contain heavy water coolant at about 10 MPa and form part of the coolant
loop. Flow in the pressure tubes is arranged to be in opposite directions in adjacent fuel chan-
nels. Thus, at each end of the reactor, there are sets of inlets and outlets to and from each fuel
channel, as shown in Figure 10. To carry the coolant to and from each of these inlets and
outlets, small diameter feeder tubes are used, as shown in Figure 11. These are linked to twin
common headers above each end of the reactor core. Coolant from the outlet headers at each
end passes to the steam generators, then through the coolant pumps and back to the inlet
headers at the same end of the reactor core. After a second pass through the fuel channels, the
coolant passes to steam generators and coolant pumps at the other end of the reactor core.
The complete coolant loop thus has a double figure-of-eight configuration. Typically, there are
two steam generators and two coolant pumps at each end of the reactor core, making a total of
four of each. Although each pair of steam generators and its associated coolant loop has a
separate figure-of-eight configuration, the headers are cross-connected so that the whole
system operates at the same pressure. Pressure is maintained by a single pressurizer connected
to one of the coolant loops. The pressurizer is a tall cylindrical vessel containing half water and
half steam and maintained at saturation conditions. By varying the temperature in the vessel
with heaters or water sprays, the pressure in the entire coolant system can be controlled.
Typical parameters for 600 MWe (electrical) and 950 MWe (electrical) CANDU coolant systems
are given in Table 6. Full technical details are given in Appendix C Single Unit Station Data.
Figure 10 Fuel channel end fittings on reactor face
Figure 11 Feeder tube assembly on reactor face
Table 6 CANDU coolant system parameters
Number of pressure tubes 380 600

Pressure tube material Zirconium-niobium alloy Zirconium-niobium alloy
Pressure tube diameter (ID) 103.38 mm 104 mm
Number of primary pumps 4 4
Pump flow rate (each) 2.228 m3/s 3.959 m3/s
Pump total heat (each) 215 m 245 m
Coolant flow rate through core 7 600 kg/s 13 500 kg/s
Coolant inlet temperature 267°C 266°C
Coolant outlet temperature 312°C 312°C
Coolant inlet pressure 11.04 MPa 11.17 MPa
Coolant outlet pressure 10.03 MPa 10.29 MPa
6.4 Steam Generators

The purpose of the steam generators
is to transfer heat from the primary
coolant system to the secondary
steam system and thus supply the
required heat to operate the steam
cycle. Like those of the PWR, the
steam generators are tube or surface
heat exchangers, with high pressure
coolant passing inside the tubes and
lower pressure steam being generated
on the outside of the tubes. The
steam pressure is typically 5 MPa. The
usual configuration is a vertical cylin-
der with inverted U-tubes in the lower
part and steam-water separators in
the upper part, as shown diagram-
matically in Figure 12. Steam gener-
ated within the tube bundle rises and
promotes natural circulation. At the
top of the vessel, the steam-water
mixture passes through cyclone type
primary separators where the steam is
separated from the circulating water.
The water returns down an annulus
between the tube bundle and the
steam generator shell. The separated
steam passes through secondary
separators or steam dryers to remove
moisture and so improve its quality Figure 12 Steam generator for CANDU system
before passing out at the top of the
vessel. Incoming feedwater enters at the bottom of the steam generator. This cooler feedwater
is confined within baffles and made to flow in a criss-cross manner over that part of the tube
bundle near the primary coolant outlet. In this way, the feedwater temperature is raised to
saturation conditions before mixing with the circulating water in the steam generator. This
arrangement improves thermodynamic performance, but limits the range of temperature at
which feedwater can be introduced to the steam generator. A sudden drop in feedwater
temperature during normal operating conditions could cause thermal shock to the vessel and
tubes. Typical parameters for 600 MWe (electrical) and 950 MWe (electrical) CANDU reactor
steam generators are given in Table 7.
Table 7 CANDU steam generator parameters
Number of steam generators 4 8

Heat transfer capacity (total) 2 064 MW 3 258 MW
Primary side water flow (total) 7 600 kg/s 13 500 kg/s
Primary side operating pressure ~ 10 MPa ~ 10 MPa
Primary side inlet temperature 312°C 312°C
Primary side outlet temperature ~ 267°C ~ 266°C
Secondary side steam flow (total) 1 047 kg/s 1 612 kg/s
Secondary side feedwater flow (total) 959 kg/s 1 536 kg/s
Secondary side operating pressure 4.69 MPa 5.07 MPa
Secondary side inlet temperature 187°C 177°C
Secondary side outlet temperature 260°C 265°C
Steam outlet wetness 0.25% 0.25%
Number of tubes (each) 4 663
Heat transfer area per steam generator ~ 3 066 m3
Tube diameter (OD) 22.23 mm 15.9 mm
Tube material Inconel 600 Incoloy 800
6.5 Steam Turbines

Each CANDU unit has a single steam turbine and electrical generator on a single shaft. The
steam turbine consists of a high pressure turbine receiving saturated steam from the steam
generators and three low pressure turbines receiving steam, after moisture separation and
reheating, from the high pressure turbine. Saturated steam from the steam generator is used
for reheating, and the reheated steam becomes superheated at the lower pressure. This is
required to avoid too high a moisture content in the turbine exhaust. Three low pressure
turbines are required due to the large increase in steam specific volume as it expands to con-
denser conditions. Figure 13 shows the 600 MW steam turbine at Point Lepreau viewed from
the steam inlet end, with the steam inlet valves and high pressure turbine in the foreground and
the low pressure turbines further back. Figure 14 shows the same turbine from the electrical
output end, with the exciter and generator in the foreground and the low pressure turbines
behind them. Typical parameters for 600 MW (electrical) and 950 MW (electrical) steam
turbines are given in Table 8.
Figure 13 600 MW steam turbine for nuclear unit
Figure 14 Turbine generator for nuclear unit
Table 8 Steam turbine parameters
Number of HP cylinders 1 double flow 1 double flow

Number of LP cylinders 3 double flow 3 double flow
Main steam flow rate 1 047 kg/s 1 612 kg/s
Live steam flow to valves 957 kg/s 1 533 kg/s
Live steam flow to reheater 90 kg/s 77 kg/s
HP steam inlet pressure 4.55 MPa 4.93 MPa
HP steam inlet temperature 258°C 263°C
HP steam exhaust pressure 0.665 MPa 0.545 MPa
HP steam exhaust quality 88.2% 86.5%
LP steam inlet pressure 0.588 MPa 0.500 MPa
LP steam inlet temperature 242°C 247°C
LP steam exhaust quality 89.5% 91.0%
Steam cycle efficiency 32.8% 34.4%
Electrical generator gross output 678 MW 1 121 MW
6.6 Technical Data

The technical data in this chapter were obtained from various sources, including references in
the bibliography and public relations brochures from Atomic Energy of Canada Limited, and
therefore are not necessarily applicable to a particular nuclear power plant. Some data may be
inconsistent or in conflict due to evolution of the design or to differences between design and
operational data. Furthermore, some missing data were deduced by simple calculation for
illustrative purposes and are indicated as an approximate value, meaning that what is presented
is typical rather than specific.
7 Possible CANDU Reactor Development

7.1 CANDU Advantages
Refuelling can be done while the reactor is on power by pushing new fuel bundles in at one end
of a fuel channel while spent fuel bundles are removed at the other end. This allows the
CANDU reactors to achieve high capacity factors compared with those reactors which have to be
shut down periodically for refuelling. Figure 15 shows a typical annual analysis of capacity
factors for CANDU reactors and other light water reactors in the early years of their life before
major maintenance or refurbishment became necessary. This shows that in that year in the top
25 reactors worldwide, with regard to capacity factor performance, there are proportionately
twice as many CANDU reactors listed as in the general population of water cooled reactors.
One year within the first ten years of operation Point Lepreau was actually ranked first on this
list. This was due to only one short planned annual shutdown for routine maintenance of
systems not able to be accessed during normal operation during the initial years of operation.
Furthermore the relative simplicity of the fuel bundles and the use of natural uranium leads to
low fuel costs for CANDU reactors. With no heavy reactor pressure vessel and the possibility of
modular construction the erection time of a CANDU reactor can be shorter than that of some
other reactor systems. Having been developed to use natural uranium in conveniently small
fuel bundles, the CANDU reactor has the ability of utilizing low enriched fuel from other sources
such as light water reactors and also blended fuel from nuclear weapons. This makes it attrac-
tive as part of a fuel reprocessing and recycling scheme involving different nuclear plants and
facilities.
The refurbishment of CANDU reactors has been proven. Steam generators and reactor pressure
tubes can be replaced. This has the potential of effectively doubling the life of the plant ena-
bling it to operate efficiently for 50 to 60 years. The overall capital investment including the
cost of refurbishment can make the CANDU reactor an attractive investment for bulk base load
power production.
Figure 15 Typical capacity factors of CANDU and other reactors
7.2 Future Prospects

Future prospects for the CANDU reactor are good. About one third of CANDU reactors in
operation are of the CANDU 6 type and all have performed well. The few larger reactors have
also provided good service. A CANDU reactor would be a good choice for any future nuclear
reactors built in Canada. However it would be in competition with light water reactors espe-
cially the PWR where later designs are under construction.
Of particular note is the modular construction proposed for new reactors. The advantage that
this modular construction offers is a decreased construction time and hence lower costs. It also
allows specific parts of the reactor to be more easily replaced during the life of the plant.
New fuel bundles with 43 fuel elements have been developed and tested. The increased
surface area and smaller diameter elements resulting from this change allow for increased heat
transfer and hence more power per bundle. Such design developments are able to reduce the
capital and operating costs of the current CANDU system.
7.3 Safety Aspects

The primary way to avoid reactor accidents is through use of duplicate safety systems, core
cooling systems, and engineered safeguards built into the reactor design. The CANDU reactor
has evolved to be inherently safe, with consideration given to all conceivable accidents. Acci-
dents to early reactors, as mentioned earlier, have highlighted the need for very conservative
and safe design requirements.
In case of an adverse transient which could lead to a potential accident, there are two inde-
pendent shutdown systems. The SDS1 system consists of mechanical shutoff rods which drop
into the core by gravity when a trip signal is received. The SDS2 system consists of an array of
nozzles which inject gadolinium nitrate under gas pressure into the moderator on receipt of a
trip signal.
In the event of loss of coolant from the reactor heat transport system, an emergency core
cooling system injects cooling water into the headers to ensure a water supply to maintain fuel
cooling. This is backed up by alternative systems to maintain cooling for an extended period.
Use of multiple pressure tubes instead of a single large pressure vessel permits thinner walls
and simpler manufacture to the required pressure threshold. Any leaks can be detected by
monitoring moisture content and pressure in the gap between the pressure tube and the
calandria tube, which is done on a continuous basis. When detected, a faulty pressure tube can
be readily replaced.
The heavy water moderator and reflector in the calandria surrounding the pressure tubes are at
a relatively low temperature compared with that of the coolant flowing in the pressure tubes.
Because of this lower temperature, the moderator and reflector act as an energy sink in case of
certain reactor accidents. This heat sink is an important feature in the CANDU design because it
means that if the emergency core cooling system fails, some heat can be transferred from the
fuel to the moderator. The calandria itself is surrounded by light water in the reactor vault, thus
creating an additional short term heat sink for the moderator.
Most important are the engineered safeguards that protect the public from possible release of
fission products in the event of a component failure. As in most nuclear power plants, there are
four barriers that prevent the release of significant quantities of fission products to the envi-
ronment. These barriers are: the fuel itself, the fuel cladding, the primary heat transport system
boundary, and the containment building.
The first barrier is the uranium dioxide fuel, which is chemically inert even in high temperature
water and has a high melting point. Even under high temperature conditions, about 99% of all
radioactivity is trapped in the uranium dioxide matrix. This means that almost all solid fission
products are contained in the fuel under normal non-melting conditions.
The second barrier is the fuel cladding, which is a zirconium alloy sheath. This sheath is de-
signed to withstand the stress associated with fuel expansion and buildup of trapped fission
gases. Zirconium would be subject to damage should dryout and elevated temperatures occur
in the fuel channel. Provided the fuel bundles are kept flooded with coolant during accident
conditions, this barrier will remain intact.
The third barrier is the primary heat transport system boundary. This provides containment for
the coolant, which may contain fission products in the event of fuel cladding leakage. The
primary circuit is a closed loop and does not allow any fission products to go any further unless
it in turn has a leak.
The fourth and final barrier to fission product release is the prestressed low leakage concrete
containment building. The building is maintained under a slightly negative pressure and is lined
with a plastic coating which limits the leakage of fission products in the event of overpressure
due to an accident. In addition, air exchange filters in the ventilation system remove any fission
products in the circulating air which is discharged to the atmosphere.
7.4 The Advanced CANDU Reactor

A possible new development of the CANDU reactor is the Advanced CANDU Reactor (ACR 1000),
with a nominal net electrical output of 1000 MWe. The general arrangement is very much the
same as the basic CANDU reactor, as shown in Figure 16, but some significant technical changes
have been included.
These reactors have been designed so that they can be built within a four-year period and have
an expected plant life of 60 years, with an overall lifetime capacity factor of over 90%. They use
light water instead of heavy water as coolant, thus simplifying several supporting auxiliary
systems. This necessitates use of low enrichment fuel to compensate for increased neutron
absorption in the coolant. The fuel bundles are of the CANFLEX 43 element design, as shown in
Figure 17, which provides increased power output per bundle and hence per channel. This
design has fuel elements of two different sizes. There are 8 central elements 13.5 mm in
diameter and 35 outer elements 11.5 mm in diameter. A quick comparison between the
CANDU 6 and Darlington Generating Station shows the general trend of the conventional
CANDU design towards larger capacities, while a comparison between Darlington and the ACR
1000 shows somewhat of a reversal of this trend with current technical changes. Table 9 shows
some key technical parameters which illustrate this trend.
Figure 16 ACR 1000 nuclear systems schematic
Figure 17 ACR CANFLEX fuel bundles
Table 9 Advanced CANDU reactor parameter comparison
Parameter Units CANDU 6 Darlington ACR 1000
Reactor power output MWt 2 064 2 657 3 187

Moderator D 2O D 2O D 2O
Coolant D 2O D 2O H2O
Fuel Nat UO2 Nat UO2 Enriched UO2
Heavy water inventory in moderator Mg 265 312 250
Heavy water inventory in coolant Mg 192 280 0
Number of fuel channels 380 480 520
Number of bundles per fuel channel 12 13 12
Number of elements per bundle 37 37 43
Fuel burnup MWd/tU 7 500 7 791 20 000
Calandria diameter m 7.6 8.5 7.5
Lattice pitch mm 286 286 240
Pressure tube wall thickness mm 4 4 6.5
Inlet header pressure MPa 11.2 11.3 12.5
Inlet header temperature °C 260 267 275
Outlet header pressure MPa 9.9 9.9 11.1
Outlet header temperature °C 310 310 319
Maximum channel flow kg/s 28 27 28
Number of heat transport pumps 4 4 4
Pump motor rating (each) MWe 6.7 9.6 10.0
Pump rated flow (each) m3/s 2.228 3.240 4.300
Number of steam generators 4 4 4
Tube diameter mm 15.9 15.9 17.5
Steam pressure MPa 4.6 5.0 5.9
Steam temperature °C 260 265 276
Steam quality % 99.75 99.75 99.90
Net power to turbine generator MWt 2 060 2 650 3 180
Steam cycle efficiency 35.3 35.3 ~ 36.6
Gross electrical power output MWe 728 935 1 165
Net electrical power output MWe 666 881 1 085
Turbine inlet steam temperature °C 258 263 273
Final feedwater temperature °C 187 177 217
Condenser vacuum kPa 4.9 4.2 4.9
The table clearly shows the evolution of the ACR 1000. The 43-element fuel bundle provides an
average channel power of 6.13 MW (12 bundles) as opposed to 5.43 MW (12 bundles) and 5.54
MW (13 bundles). This in turn enables fewer channels to be used for an equivalent output,
resulting in a smaller reactor for the same output. Furthermore, the channel pitch has been
reduced by 16% resulting in an even smaller reactor. The overall result is that the ACR 1000
reactor is about the same size as the CANDU 6, as illustrated in Figure 18. The heavy water
inventory has been reduced accordingly. The light water coolant eliminates a pressurized heavy
water circuit altogether, thereby effecting a further and major saving in inventory and heavy
water support systems. The low enriched uranium increases the burnup by a factor of about 2.5
and consequently decreases the amount of spent fuel. An additional reserve water system
provides a passive safety feature, making this a Generation III+ design. Other than these
important changes, there are many small improvements in the general operating conditions of
the heat transport system and steam cycle, giving an improvement in overall efficiency from
35.3% to 36.6%, which ultimately saves on fuel consumption.
Figure 18 Comparison of core sizes
8 Problems
1 Sketch a typical figure-of-eight CANDU heat transport system showing all key components
and describe the functions of these components.
2 Sketch a typical CANDU reactor core and show in the sketch where, with respect to the fuel
channels, the various control devices are installed. Identify the devices and state their purpose.
3 Describe the structure and characteristics of a CANDU fuel channel and the fuel within the
channel. Describe also what happens in the fuel channel as the coolant flows through it.
4 Sketch a CANDU steam generator showing the key components and explain the function of
each component and why they are arranged in this particular configuration.
9 References and Bibliography

[AECL1989] Atomic Energy of Canada Limited, CANDU 8 Technical Outline. September 1989.
[AECL1989] Atomic Energy of Canada Limited, CANDU 3 Technical Outline, Document Number
74-01010-TED-001, Revision 9, 1989.
[AECL1991] Atomic Energy of Canada Limited, Technical Summary CANDU Nuclear Generating
Station. 1991.
[AECL1997] Atomic Energy of Canada Limited, CANDU Technical Summary, 1997.
[AECL1997a] Atomic Energy of Canada Limited, Canada Enters the Nuclear Age, AECL, McGill-
Queen’s University Press, 1997.
[AECL1997b] Atomic Energy of Canada Limited, CANDU 6 Technical Outline, Document Refer-
ence CO/MISC-0131, Revision 1, 1997.
[AECL2007] Atomic Energy of Canada Limited, ACR 1000 Technical Summary. 2007.
[Brooks1993] Brooks, G. L., A Short History of the CANDU Nuclear Power System, Ontario Hydro
Demand/Supply Hearing, Canada, January 1993.
[Brooks2001a] Brooks, G. L., CANDU Origins and Evolution – Part 2 of 5 – Why CANDU, AECL
Retirees Group, Canada, February 2001. [CANTEACH, https://canteach.candu.org]
[Brooks2001b] Brooks, G. L., CANDU Origins and Evolution – Part 3 of 5 – ‘Figure of Eight’ Heat
Transport System Arrangement, AECL Retirees Group, Canada, February 2001.
[CANTEACH, https://canteach.candu.org]
[Brooks2001c] Brooks, G. L., CANDU Origins and Evolution – Part 4 of 5 – Emergency Core
Cooling Systems, AECL Retirees Group, Canada, February 2001. [CANTEACH,
https://canteach.candu.org]
[Brooks2001d] Brooks, G. L., CANDU Origins and Evolution – Part 5 of 5 – Origins and Evolution
of the Second Shutdown System, AECL Retirees Group, Canada, February 2001.
[Chaplin1973] Chaplin, R. A., “The Winfrith Steam Generating Heavy Water Reactor”, South
African Mechanical Engineer, May 1973.
[Chaplin2006a] Chaplin, R. A., “Pressurized Heavy Water Reactors”, Chapter 3.6.2.3 in Nuclear
Energy and Reactors, Encyclopedia of Life Support Systems by UNESCO, EOLSS Publish-
ers, Oxford, UK, 2006.
[Chaplin2006b] Chaplin, R. A., “Heavy Water Light Water Reactors”, Chapter 3.6.2.4 in Nuclear
Energy and Reactors, Encyclopedia of Life Support Systems by UNESCO, EOLSS Publish-
ers, Oxford, UK, 2006.
[Dyke2006] Dyke, J. M. and Garland, W. J., “Evolution of CANDU Steam Generators – A Historical
View”, Canada, May 2006. [CANTEACH, https://canteach.candu.org]
[Foster2001] Foster, J. S. & Brooks, G. L., CANDU Origins and Evolution – Part 1 of 5 – An Over-
view of the Early CANDU Program, AECL Retirees Group, Canada, February 2001.
[Laurence1980] Laurence, G. C., Early Years of Nuclear Research in Canada. AECL, Chalk River
ON, May 1980.
10 Acknowledgements
Acknowledgements are extended to Atomic Energy of Canada Limited and to NB Power for use
of information, diagrams and photographs to support the text of this chapter.
Bob Tapping
Jeremy Whitlock
Bhaskar Sur
Terry Rogers
George Bereznai
1
Genealogy of CANDU Reactors: Appendices

prepared by
Dr. Robin Chaplin
Table of Contents
Appendix A – Cutaway View of CANDU 6 Reactor Plant................................................................. 2

Appendix B – CANDU Evolution ...................................................................................................... 3
Appendix C – Single Unit Station Data (5 pages) ............................................................................ 4
©UNENE, all rights reserved. For educational use only, no assumed liability. Genealogy Appendices - December 2016
Appendix A – Cutaway View of CANDU 6 Reactor Plant
Genealogy of CANDU Reactors: Appendices 3
Appendix B – CANDU Evolution
Appendix C – Single Unit Station Data (5 pages)
1
CHAPTER 3
Nuclear Processes and Neutron Physics
prepared by
Dr. Guy Marleau
École Polytechnique de Montréal
Summary:
In this section, we first describe the nucleus, including its composition and the fundamental
forces that affect its behaviour. Then, after introducing the concepts of radioactivity and
nuclear decay, we discuss the various interactions between radiation and matter that can take
place in and around a nuclear reactor. We finish with a presentation of neutron physics for
fission reactors.
Table of Contents
1 Introduction .............................................................................................................................3
1.1 Overview......................................................................................................................... 3
1.2 Learning Outcomes......................................................................................................... 4
2 Structure of the Nucleus..........................................................................................................4
2.1 Fundamental Interactions and Elementary Particles...................................................... 4
2.2 Protons, Neutrons, and the Nuclear Force ..................................................................... 6
2.3 Nuclear Chart.................................................................................................................. 7
2.4 Nuclear Mass and the Liquid Drop Model.................................................................... 10
2.5 Excitation Energy and Advanced Nuclear Models ........................................................ 12
2.6 Nuclear Fission and Fusion ........................................................................................... 15
3 Nuclear Reactions ..................................................................................................................16
3.1 Radioactivity and Nuclear Decay .................................................................................. 16
3.2 Nuclear Reactions ......................................................................................................... 20
3.3 Interactions of Charged Particles with Matter.............................................................. 22
3.4 Interactions of Photons with Matter ............................................................................ 24
3.5 Interactions of Neutrons with Matter .......................................................................... 30
4 Fission and Nuclear Chain Reaction.......................................................................................34
4.1 Fission Cross Sections ................................................................................................... 35
4.2 Fission Products and the Fission Process...................................................................... 36
4.3 Nuclear Chain Reaction and Nuclear Fission Reactors ................................................. 40
4.4 Reactor Physics and the Transport Equation ................................................................ 42
5 Bibliography ...........................................................................................................................44
6 Summary of Relationship to Other Chapters.........................................................................45
7 Problems ................................................................................................................................46
8 Acknowledgments .................................................................................................................46
©UNENE, all rights reserved. For educational use only, no assumed liability. Nuclear Physics – June 2015
List of Figures
Figure 1 Nuclear chart, where N represents the number of neutrons and Z the number of
protons in a nucleus. The points in red and green represent respectively stable and
unstable nuclei. The nuclear stability curve is also illustrated. ..............................................8
Figure 2 Average binding energy per nucleon (MeV) as a function of the mass number (number
of nucleons in the nucleus)......................................................................................................9
Figure 3 Saxon-Wood potential seen by nucleons inside the uranium-235 nucleus. ...................13
Figure 4 Shell model and magic numbers......................................................................................14
Figure 5 Decay chain for uranium-238 (fission excluded). The type of reaction (β or α-decay)
can be identified by the changes in N and Z between the initial and final nuclides. ............20
Figure 6 Microscopic cross section of the interaction of a high-energy photon with a lead atom.24
Figure 7 Photoelectric effect..........................................................................................................26
Figure 8 Compton effect. ...............................................................................................................28
Figure 9 Photonuclear cross sections for deuterium, beryllium, and lead....................................29
Figure 10 Cross sections for hydrogen (left) and deuterium (right) as functions of energy..........32
Figure 11 Cross sections for 235
92 U (left) and 92 U (right) as functions of energy. ........................33
238
Figure 12 Effect of temperature on U 238 absorption cross section around the 6.64-eV
resonance...............................................................................................................................34
Figure 13 Capture and fission cross sections of 235 92 U (left) and 92 U (right) as functions of
238
energy.....................................................................................................................................36
Figure 14 Relative fission yield for 235 92 U as a function of mass number........................................37
Figure 15 Fission spectrum for 235

92 U ..............................................................................................39
List of Tables
Table 1 Properties of the four fundamental forces .........................................................................5
Table 2 Properties of leptons and quarks ........................................................................................6
Table 3 Main properties of protons and neutrons. .........................................................................6
Table 4 Energy states in the shell model, where g is the degeneracy level of a state...................14
Table 5 Comparison of the energy, U, required for a fission reaction in a compound nucleus
with the energy, Q, available after a neutron collision with different isotopes. ...................35
Table 6 Average numbers of neutrons produced by the fission of common heavy nuclides........37
Table 7 Distribution of kinetic energy among the particles produced by a fission reaction.........38
Table 8 Maximum values of the elastic scattering and absorption cross sections for 11 H , 21 D ,
and 12
6 C in the energy range 0.1 eV < E <0.1 MeV. ...............................................................41
Nuclear Processes and Neutron Physics 3
1 Introduction
Energy production in an operating nuclear reactor is mainly the result of fission reactions
initiated by neutrons. Following these reactions, radiation is produced in the form of alpha
particles, electrons (ߚ-particles), photons, and neutrons, as well as a large number of unstable
nuclei (including actinides and fission products) that may decay, thereby producing additional
radiation and energy. The neutrons generated directly in the fission reaction as well as those
resulting from fission product decay are a key factor in maintaining and controlling the chain
reaction as discussed in Chapters 4 and 5. Radiation also has an impact on the properties of
the various materials in the core as well as on living cells. It is therefore important to under-
stand how radiation interacts with matter if one needs, for example, to evaluate the radio-
toxicity of burned fuel (Chapter 19) or to determine the kind of barrier that must be put in
place to protect the public, the environment, or the personnel in a nuclear power plant
(Chapter 12). Moreover, the energy produced by these decays contributes to heating the fuel,
whether it is still in the reactor or in the spent fuel pools. The aim of this chapter is to provide
the reader with an overview of the nuclear processes that take place in and around a reactor.
1.1 Overview
Several very important processes take place simultaneously in a nuclear reactor, including
neutron slowdown as a result of collisions with nuclei, nuclear fission, and decay of radioactive
nuclei with radiation emissions that subsequently lose energy or are absorbed.
To understand how all these physical processes take place, it is important to possess a general
knowledge of the physics of the nucleus, including its composition and the fundamental forces
that affect its behaviour. Although one rarely uses a description of the nucleus in terms of
elementary particles and their interaction in power-reactor applications, such a description
provides a practical background for a study of nuclear structure. In fact, the nucleus is such a
complex object that it can be described only through simplified models that provide useful
information. For example, the liquid drop model (Section 2.4) gives an empirical description of
how one can extract energy from a nucleus by breaking it into smaller components (fission or
decay) or by combining nuclei (fusion). These models are also used to predict the stability of a
nucleus under various decay processes, the energy associated with its excited states, and the
associated gamma-ray spectrum. In Section 2, an overview of the structure of the nucleus is
provided, starting from a fundamental description and progressing to more practical models.
We will also introduce some language specific to nuclear physics that will be useful throughout
this book.
The majority of the fission reactions taking place in a reactor are initiated following the
absorption of a neutron by a heavy nucleus. Radioprotection also involves the interaction
between radiation, in the form of alpha particles, electrons, photons, and neutrons, and
matter. Two important points must then be addressed before describing the physics of
neutrons in a nuclear reactor: the source of this radiation and its behaviour over time, and the
interaction of the emitted particles with matter. Therefore, Section 3 is divided into three
parts: it starts with a description of radioactivity and the decay of nuclei, continues with a
description of the interaction of ionizing radiation with matter (everything but neutrons),
which is a very important topic for radiation protection studies, and ends with a brief discus-
sion of the interaction of neutrons with matter.
The last section of this chapter is dedicated to the nuclear chain reaction. It starts by giving an
extensive description of the fission reaction and discusses the need to slow down neutrons.
This is followed by a description of the neutron transport equation that can be used to charac-
terize the behaviour of neutrons inside a reactor.
1.2 Learning Outcomes

The goal of this chapter is for the reader to:
 Understand the fundamental forces in nature and the structure of the nucleus;
 Learn how to compute the mass of a nucleus based on the liquid drop model and to
evaluate the energy released by the fission and fusion processes;
 Understand how to evaluate the activity and radio-toxicity of spent fuel;
 Become familiar with nuclear cross-section databases for neutrons and ionization radiation
and the computation of reaction rates between particles and matter;
 Recognize the need to slow down neutrons in thermal reactors and to evaluate the effec-
tiveness of a moderator.
2 Structure of the Nucleus

2.1 Fundamental Interactions and Elementary Particles
Four forces, or interactions, are currently sufficient to understand and explain nature from
particle physics to astrophysics, including chemical and biological processes. These forces are
classified relative to their intensity at the nuclear level (distances of the order of 1 fm) as
follows:
 The strong force: responsible for the cohesion of the proton as well as the nucleus. This
force is always attractive.
 The electromagnetic force: governs the physics of the atom and the associated chemical
and biological processes. This force can be both attractive and repulsive.
 The weak force: responsible for the existence of the neutron as well as many other radio-
active nuclei. This force is always attractive.
 The gravitational force: governs the large-scale structure of the universe. This force
ensures the stability of the solar system as well as the fact that people can keep their feet
on the ground. This force is always attractive.
From the point of view of physics, these forces arise from the specific intrinsic properties of
the particles on which they act. For example, the electromagnetic force arises between two
particles that have electric charges, while the gravitational force appears between two parti-
cles with masses (or more generally, between two particles with energy). For the strong force,
the so-called “colour” of the particle plays a role equivalent to that of electric charge in
electromagnetic interactions, while the weak force is the result of the interaction of two
particles having a “weak” charge. Table 1 provides a brief description of the general proper-
ties of the four fundamental forces, where the ߙ’s are dimensionless coupling constants
related to the intensity of the interaction and the spatial dependence of the force is expressed
in powers of ‫ݎ‬, the distance between the two interacting particles [Halzen1984]. For example,
the electromagnetic force between two particles of charge ‫݁ݖ‬and ܼ݁is given by:
   e 2  zZ zZ
| F  r  |   2  ( e c) 2 , (1)
 4 0  r |r|
where ݁ = 1.6 × 10ିଵଽ C is the charge of an electron, ߝ଴ = 8.854 × 10ିଵଶ F/m is the electric
permittivity of vacuum, ℏ = 1.0546 × 10ିଷସ J × s is the reduced Planck constant, and
ܿ = 2.9979 × 10଼ m/s is the speed of light. For the gravitational force, the coupling constant
is expressed in terms of proton masses. One can immediately see that the range of the
gravitational and electromagnetic forces is substantially larger than that of the weak force due
ଵ
to their ௥మ dependence. Accordingly, the weak force is important only when two particles
with weak charges are in close contact. On the other hand, the strong (colour) force remains
significant irrespective of the distance between particles with colour charge. This observation
led to the concept of colour confinement, meaning that it is impossible to isolate a particle
having a net colour charge different from zero [Halzen1984].
Table 1 Properties of the four fundamental forces
Force Equivalent charge Coupling constant Spatial dependence
Strong Colour charge ߙ௦ ≈ 1 ≈ ‫ݎ‬଴
Electromagnetic Electric charge ߙ௘ ≈ 1/137 ‫ିݎ‬ଶ
Weak Weak charge ߙ௪ ≈ 10ି଺ ‫ିݎ‬ହ to ‫଻ିݎ‬
Gravitational Mass ߙீ ≈ 10ିଷଽ ‫ିݎ‬ଶ
In the standard model of particle physics [Halzen1984, Le Sech2010], these forces are repre-
sented by 12 virtual particles of spin 1 called “gauge bosons”:
 the massless photon that carries the electromagnetic interaction;
 eight massless gluons that carry the strong (colour) force;
 three massive bosons (W± and Z଴) that mediate the weak interaction.
These bosons are exchanged between the 12 “physical fermions” (spin 1/2 particles) that
make up all the matter in the universe:
 six light fermions or leptons that have a weak charge. Three of these have no electric
charge (the neutrinos), while three are charged (electron, muon, and tau);
 six heavy fermions or quarks that have colour, weak, and electric charges. These fermions
cannot exist freely in nature because of colour confinement and must be combined in trip-
lets of quarks or quark-antiquark pairs to form respectively baryons (protons, neutrons)
and mesons (pion, kaon) that have no net colour charge.
Table 2 Properties of leptons and quarks
Leptons Quarks
Particle Mass (GeV/c2) Charge Particle Mass (GeV/c2) Charge
݁ 0.000511 −1 ‫ݑ‬ ≈ 0.003 2/3
ߥ௘ < 2.5 × 10ିଽ 0 ݀ ≈ 0.006 −1/3
ߤ 0.1057 −1 ܿ ≈ 1.2 2/3
ߥఓ < 0.000170 0 ‫ݏ‬ ≈ 0.1 −1/3
߬ 1.7777 −1 ‫ݐ‬ ≈ 173 2/3
ߥఛ < 0.018 0 ܾ ≈ 4.1 −1/3
Table 2 provides a description of the main properties of leptons and quarks with the masses
given in GeV/c2, with 1 GeV/c2≈ 1.782661 × 10ିଶ଻ kg, and the charges in terms of the charge
of one electron [Le Sech2010].
2.2 Protons, Neutrons, and the Nuclear Force

The two lightest baryons that can be created out of a quark triplet are the positively charged
proton, which is composed of two ‫ ݑ‬quarks and one ݀ quark, and the neutral neutron, made
up of two ݀ quarks and one ‫ ݑ‬quark. These are present in the nucleus of all the elements
present naturally in the universe. The main properties of the proton and the neutron are
provided in Table 3 [Basdevant2005]. One can immediately see that the masses of the proton
and neutron are very large compared with those of the ‫ ݑ‬and ݀ quarks. In fact, most of their
mass is a result of the strong colour force (gluon interaction). One can also observe that
protons are stable, while neutrons are unstable and decay relatively rapidly into protons
through the weak interaction. A second observation is that the magnetic moment, which is
created by the movement of charged particles, is negative for the neutron. This confirms, at
least partially, the theory that neutrons are made of quarks because even though the neutron
is neutral, the distribution of charge inside it is not uniform.
Table 3 Main properties of protons and neutrons.
Property Proton Neutron

Mass (GeV/c2) ݉ ௣ = 0.938272 ݉ ௡ = 0.939565
Charge (electron charge) 1 0
Magnetic moment (ߤN ) ߤ௣ = 2.792847 ߤ௡ = −1.913042
Spin 1/2 1/2
Mean-life (s) ߬௣ > 10ଷ଺ ߬௡ = 881.5
Although the proton and neutron have a neutral colour, they interact through the nuclear
force, which is a residual of the strong colour force. However, instead of exchanging gluons
directly, as do the quarks inside the proton or neutron, they interact by exchanging of virtual
pion. As a result, the nuclear force is much weaker than the strong force in the same way that
the Van der Waals force between neutral atoms is weaker than the electromagnetic force.
Moreover, in contrast to the strong colour force, the magnitude of the nuclear force decreases
rapidly with distance, the nuclear interaction being given approximately by the Yukawa poten-
tial:
mp cr
-

e
VYukawa (r) = -a N c , (2)
r
where ݉ గ is the mass of the pion (140 MeV/c2) and ߙே ≈ 14.5 is a constant that controls the
strength of the nuclear interactions between baryons (protons or neutrons). The minus sign
here indicates that the resulting force (‫ = ⃗ܨ‬−∇
ሬܸ⃗) is always attractive. This potential is identi-
cal for proton-proton, proton-neutron, and neutron-neutron interactions.
The strength of this force at short distances is so intense that it can compensate for the
repulsive electromagnetic force between protons, leading to the formation of bonded nuclei
as well as inhibiting neutron decay. For example, the helium nucleus, which is composed of
two protons and two neutrons, is stable. Similarly, the stable deuteron is composed of one
proton and one neutron. The absence in nature of a nucleus composed of two protons is not a
result of the fact that the electromagnetic force is greater than the strong force, but rather
due to the exact structure of the nuclear force. In addition to having a spatial behaviour
provided by the Yukawa potential, this potential is also spin-dependent and repulsive between
particles with anti-parallel spins (a nucleus with two protons having parallel spins is not
permitted because of Pauli’s exclusion principle). This also explains why stable nuclei contain-
ing only neutrons are not seen in nature.
2.3 Nuclear Chart

As indicated in the previous section, not all combinations of neutrons and protons can lead to
bonded states. Figure 1 provides a list of all bonded nuclei [Brookhaven2013, KAERI2013].
The points in red represent stable nuclei, while those in green indicate radioactive nuclei. The
straight black line indicates nuclei having the same number of neutrons and protons. Note
that this chart is quite different from the periodic table, where the classification of the ele-
ments is based on the number of electrons surrounding the nucleus. Because the number of
electrons surrounding the nucleus is identical to ܼ, the number of protons inside the nucleus,
each element corresponds to a combination of the nuclei that are found on a horizontal line of
the nuclear chart. The nomenclature used to classify these nuclei is the following:
 The number of protons in the nucleus is used to identify the element. For example, all
nuclei with ܼ = 8 are known as oxygen (atomic symbol O), independently of the number
of neutrons present in the nucleus.
 For a given element, the specific isotope is identified by ‫ ܼ = ܣ‬+ ܰ , where ‫ ܣ‬is known as
the mass number and ܰ is the number of neutrons in the nucleus. For example, oxygen-
17, denoted by ଵ଻O, corresponds to an oxygen atom where the nucleus has eight protons
and nine neutrons. Instead of the notation ே ା௓ X, with X the element associated with ܼ,
the notation ே ା௓
௓ X is also found in the literature.
 Two other terms are less frequently used: isobars, corresponding to nuclei having the same
mass numbers, but different values for ܼ and ܰ ; and isotones, corresponding to nuclei
having the same number of neutrons, but a different number of protons.
Several conclusions can be reached from Figure 1:
 Isotopes containing a small number of protons ((ܼ ൏ ʹͲ)) will generally be stable only if ܰ
and ܼ are nearly equal. If the difference ȁܰ െ ܼȁis slightly too large,, the nucleus
nucle becomes
unstable. For combination of protons and neutrons with large values of |ܰ െ ܼ|, ܼ no nuclei
can be formed.
 For nuclei with values of ܼ ranging from 20 to 82, the number of excess neutrons
neutron (ܰ െ ܼ)
required to create bonded nuclei increases steadily. One can also observe e that for a spe-
sp
cific value of ܼǡseveral
several stable nuclei can be produced with different numbers of neutrons.
 No stable nucleus exists with ܼ > 83,, although some radioactive nuclides with ܼ ൏ ͻ ͵ can
still be found in nature because they decay at a very slow rrate. No isotope with ܼ above 92
remains in nature.
Approximately 3200 bonded nuclei have been identified
identified, out of which 266 are stable.
stable The
stability of a nucleus is ensured by the presence of a sufficient number of neutrons to com-
co
pensate for the effect of the Coulomb force between protons
protons. Because the nuclear force acts
only at short distances, while the Coulomb force has a longer range, additional neutrons are
required when the radius ܴ of the nucleus increases because of the presence of more nucle-
nucl
ons (protons and neutrons):
R  A1/3R0 , (3)
where ܴ௢ is the average radius of a single nucleon.
Figure 1 Nuclear chart, where N represents the number of neutrons and Z the number of
protons in a nucleus. The points in red and green represent respectively stable and unstable
nuclei. The nuclear stability curve is also illustrated.
The stability of a nucleus
leus does not only depend on the number of nucleons it contains but also
on how these protons and neutrons are paired
paired. For example, out of the 266 stable nuclei:
nuclei
 159 contain an even number of protons and of neutrons,
 53 have an even number of protons
protons, but an odd number of neutrons,
 50 have an odd number of protons and an even number of neutrons, and
 only 4 are made up of an odd number of protons and of neutrons.
In addition, for values of ܼ or ܰ equal to 8, 20, 50, 82, and 126, the number of stable nuclei is
very high. These numbers are known as “magic numbers”
numbers”. They suggest that the neutrons
and protons, just like electrons in atoms, are more tightly bonded when they fill a quantum
energy shell (see Section 2.5).
The mass ݉ ಿ శ ೋ X of a nucleus produced by combining ܼ protons and ܰ neutrons is less than
the sum of the masses of its constituents
constituents:
m N  Z X  Z m p  Nmn  AB / c 2 , (4)
where ‫ ܤ‬൐ Ͳ,, the average binding energy per nucleon, is the result of the negative potential
that binds the nucleons inside the nuclei
nuclei. As can be seen in Figure 2,, where a plot of ‫ ܤ‬as a
function of ‫ ܣ‬is provided for stable nuclei, ‫ ܤ‬first increases rapidly for low mass numbers,
reaching a maximum around ‫ ܣ‬ൌ 60 60, and then decreases slowly with ‫ܣ‬. This is means that, on
the average, nucleons are more tightly bonded inside nuclei having intermediate values of ‫ܣ‬
(ʹͲ ൏ ‫ <ܣ‬120) than for very low or very high values.
Figure 2 Average binding energy per nucleon (MeV) as a function o

off the mass number
(number of nucleons in the nucleus).
Information on the atomic mass of over 3000 isotopes is available in [Livermore2013].
[Livermore2013] Note
that the atomic mass is different from the nuclear mass of an isotope because it includes both
the mass and thehe binding energy of the ܼ electrons gravitating around the nucleus.
nucleus One
simple way to compute the atomic mass is using
M N  Z X  Z M 1 H  Nmn  AB / c 2 , (5)
where the mass of the proton has been replaced by that of the hydrogen atom, which includes
the electron mass as well as its binding energy. The approximation sign comes from the fact
that the binding energy of each of the ܼ electrons around a heavy nucleus is different from
that of a single electron attached to a proton in the hydrogen atom.
Finally, the natural composition of a given element will include stable as well as some radioac-
tive isotopes. Assuming that the relative atomic abundance of isotope ݅of atomic mass ‫ ܯ‬௜ in
the natural element is ߛ௜, then the mass of the natural element is given by:
M X   i M i . (6)
i
In reactor physics, one often works with the mass (or weight) fraction for the abundance of an
isotope, defined as:
 iMi
wi  . (7)
MX
For completeness, one can compute the isotopic concentration ܰX of an element (atoms/cm3)
with atomic mass ‫ ܯ‬X (g/moles) and density ߩ (g/cm3) using:

NX   A, (8)
MX
where ℕA = 6.022 × 10ଶଷ atoms/moles is the Avogadro number. The concentration of
isotope ݅is then given by:

Ni   i A   iNX . (9)
MX
More information related to the isotopic contents of natural elements can be found in
[KAERI2013, WebElements2013].
2.4 Nuclear Mass and the Liquid Drop Model

As discussed in Section 2.3, the mass of the nucleus is smaller than its classical mass (the mass
of its constituents) because the strong force linking the nucleons reduces the global energy of
the nuclear system, the binding energy being related to the missing mass Δ݉ X according to
ΔmX c 2  AB. (10)
This missing mass is very difficult to compute based on theoretical models because the form of
the nuclear potential is known only approximately and because quantum models involving
more than two particles in interaction cannot be solved exactly. For evaluation of the missing
mass, one generally relies on a semi-empirical model known as the liquid drop model. The
main assumption used in this model is that the nuclear force is identical for protons and
neutrons.
The liquid drop model first assumes that each of the ‫ ܣ‬nucleons in the nucleus sees the same
nuclear potential [Basdevant2005]. Moreover, this potential is independent of the number of
nucleons in the nucleus. This second assumption can be justified by the fact that the range of
the nuclear force is small, and therefore each nucleon feels the effect of only the limited
number of nucleons that are close by. Accordingly, the contribution to the missing mass from
the ‫ ܣ‬nucleons is given by:
ΔmV c2  aV A, (11)
where the constant ܽ௏ ≈15.753 MeV has been determined experimentally. This contribution
is known as the volume effect because the volume of a nucleus can be assumed to be propor-
tional to the number of nucleons it contains. This term would be sufficient if the nucleus had
an infinite size. However, this is not the case, and the protons and neutrons that are located
near the boundary of the nucleus will interact with a smaller number of nucleons. If the
surface of the nucleus is given by 4ߨܴଶ ∝ ‫ܣ‬ଶ/ଷ, one can assume that the ‫ܣ‬ଶ/ଷ nucleons near
the outer surface of the nucleus see a reduced uniform potential. This leads to a negative
surface contribution to the mass defect of the form:
ΔmS c2  aS A2/3 , (12)
where ܽௌ ≈17.804 MeV.
Until now, only the nuclear force has been taken into account. However, the protons present
inside the nucleus repel each other through the Coulomb force, thereby decreasing the
binding energy. Because the potential associated with this force is long-range (ܸ(ܴ) ∝ 1/ܴ ∝
1/‫ܣ‬ଵ/ଷ), each of the ܼ protons will interact with the remaining ܼ − 1 protons, leading to a
contribution of the form:
Z ( Z  1) Z2
ΔmC c 2  aC   a C , (13)
A1/3 A1/3
where only the dominant term in ܼ has been preserved, and ܽ஼ ≈0.7103 MeV.
It can be observed in the nuclear chart that stable nuclei with a low mass number generally
contain an equal number of protons and neutrons. As the number of excess neutrons or
protons in a nucleus increases, it becomes more and more unstable until no bonded state can
be produced. This means that nuclear binding should decrease as a function of |ܰ − ܼ| =
|‫ ܣ‬− 2ܼ|, the form of the contribution being
( A  2Z )2
Δm Ac 2  a A , (14)
A
where ܽ஺ ≈23.69 MeV is known as the asymmetry term. The nuclear chart also shows that
few stable nuclei with an odd number of neutrons and protons (odd-odd nuclei) can be found
in nature, while the number of isotopes with an even number of protons and neutrons (even-
even nuclei) dominates. This means that the binding energy should be large for even-even
and small for odd-odd nuclei. Assuming that the contribution to the missing mass presented
previously remains valid only for odd-even or even-odd nuclei (‫ ܣ‬odd), an additional empirical
correction of the form
(1   1 )
A
ΔmP c  ( 1) a P
2 Z
(15)
2 A3/4
is required. Here, ܽ௉ ≈33.6 MeV is known as the pairing term.
The final formula for the missing mass is therefore
(1   1 )
A
2
Z2 ( A  2 Z )2
ΔmX c  aV A  aS A  aC
2 3
1
 aA  ( 1) Z aP , (16)
3
A 2 A3/4
A
which is known as the Bethe-Weizsäcker or semi-empirical mass formula (SEMF). Even if this
mechanistic model is relatively simple, it provides a very good approximation to the mass of
the nuclei found in nature. It can also be used to determine the value of ‫ ܣ‬that minimizes the
mass of a nucleus for a specific value of ܼ, thereby maximizing the probability that it is stable.
Neglecting the pairing term, one then obtains
4 Aa A  A  mn  m p  c 2
Z A  2
. (17)
8a A  2aC A 3
This equation, which is also illustrated in Figure 1 (the nuclear stability curve), closely follows
the nuclear stability profile.
The main weakness of the liquid drop model is its lack of predictive power. For example, it
neither explains the presence of “magic numbers”, nor the gamma-ray absorption spectrum,
nor the decay characteristics of different nuclei. These explanations can be obtained only by
looking at the interactions between protons and neutrons using quantum theory, as explained
in the next section.
2.5 Excitation Energy and Advanced Nuclear Models

All chemists and physicists are familiar with atomic “magic numbers”, even though the explicit
term is rarely used. For example, the number of electrons in noble gases could be considered
“magic” because these gases are odourless, colourless, and have very low chemical reactivity.
For such atoms, all the states associated with principal quantum energy levels 1 to ݊ are
occupied by electrons (closed or filled shells). In addition, the absorption (emission) spectrum
of light by different atoms is the result of electrons being excited (de-excited) from one elec-
tronic shell to another. One can therefore expect nuclear magic numbers to have a similar
nature, namely, that only specific energy levels are permitted for the protons and neutrons
when they are bonded inside a nucleus (which would explain also the photon emission and
absorption spectra). These energy levels are not uniformly distributed, but occur in bands
called shells. The fact that all the quantum states in a shell (a neutron or proton shell) are
occupied by a nucleon increases the stability of the nucleus compared to nuclides with par-
tially filled shells.
To obtain the energy levels permitted for protons and neutrons inside the nucleus, one needs
to solve the quantum-mechanics Schrödinger equation for ‫ ܣ‬strongly coupled nucleons
[Griffiths2005]. However, when trying to solve this equation, two problems arise:
 the Schrödinger equation has no analytic solution for problems where three or more
particles are coupled;
 the interaction potential between nucleons is not known exactly.
As a result, this many-body problem is generally simplified using the following assumptions
that are inherent to the shell model [Basdevant2005]:
 each nucleon in the nucleus is assumed to move independently without being affected by
the displacement of other nucleons;
 a nucleon moves in a potential ܸ(‫ )ݎ‬that is relatively uniform inside the nucleus and
increases sharply near its outer
er surface, namely when ‫ ݎ‬ൎ ܴ (mean field approximation).
Note that in this model, there are no one
one-to-one
one interactions between the nucleons.
nucleons This
means that instead of solving the Schrödinger equation for ‫ ܣ‬strongly coupled nucleons, the
problem is reduced
ced to the solution of ‫ ܣ‬independent Schrödinger equations for a single
nucleon with an averaged potential
potential. The conventional nuclear potential used in the shell
model is the Saxon-Wood
Wood potential (see Figure 3), which has the form [Basdevant2005]:
[Basdevant2005]
 1 
V  r   V0  r  R/ R , (18)
1 e 
30 MeV is the potential depth and ܴ ‫ܣ ן‬ଵȀଷ is the radius of the nucleus in fm.
where ܸ଴ ≈30
Figure 3 Saxon-Wood
Wood potential seen by nucleons inside the uranium
uranium-235
235 nucleus.
The energy levels are then very similar to those associated with the harmonic potential,
namely
 2V0  3
EN   N  , (19)
R mn  2
where ݉ ௡ is the mass of a nucleon (proton or neutron), ܰ ൌ ሺʹ݊ ൅ ݈െ ʹሻൌ 0,1,2, 0 ⋯ with
݊ ൌ ͳǡʹǡ‫ ڮ‬the principal quantum numbernumber, and l  0,1, 2, the angular momentum quantum
number that takes into account spin
spin-orbit coupling. In general, one assumes that the energy is
independent of the magnetic quantum number ݉ . The spin ݆of of each of the nucleons, which
is important when determining the probability of interaction between a nucleus (spin ‫ )ܬ‬and
other particles, is a combination of its internal spin and its angular momentum:
 1
j   l  . (20)
 2
The energy states of the shelll model are presented in Table 4, whereas the association be-
b
tween the magic numbers and these states is illustrated in Figure 4.
The model is never used to evaluate the mass of nuclei (the SEMF model is used for this
purpose), ), but it is useful for classifyi
classifying
ng the energy levels of excited nuclei (denoted as
( ேା௓ ∗
௓ X) ) and for determining the magic numbers.
Table 4 Energy states in the shell model, where g is the degeneracy level of a state.
ࡺ 0 1 2 2 3 3 4 4 4
࢔ 1 1 1 2 1 2 1 2 3
࢒ 0 1 2 0 3 1 4 2 0
ࢍ 2 6 10 2 14 6 16 10 2
orbital 1s 1p 1d 2s 1f 2p 1g 2d 3s
Figure 4 Shell model and magic numbers.

The second model that is often used is the collective model of the nucleus
nucleus. Here, instead of
looking at the motion of the individual nucleons, one considers the overall rotation and
vibration movement of the nucleus
nucleus. The justification for this model is that most nuclei take
the shape of an ellipsoid (only nuclei that contain magi
magicc numbers of protons and neutrons are
spherical). Distorted nuclei can then acquire rotational and vibrational motions,
motions to which
quantized energy levels can be associated
associated. In classical mechanics, the energy of a rotating solid
having a moment of inertia I and an angular momentum ‫ ܬ‬is given by [Basdevant2005,
Wong2004]:
J2
E , (21)
2I
In quantum mechanics, the angular momentum is quantized, with the energy levels being
given by
EJ =  2
(
J J +1 ), (22)
2I
with ‫ =ܬ‬0,1,2, ⋯. For nucleus having an ellipsoid shape, only even values of ‫ܬ‬are allowed,
and the energies satisfy the following relation:
3 1 1
E2  E4  E6  E8 , (23)
10 17 12
with
15 2
E2 = , (24)
2MR 2
where ‫ ܯ‬and ܴ are respectively the mass and radius of the nucleus. These energy levels are
generally much smaller than those associated with the shell model.
2.6 Nuclear Fission and Fusion

Let us look back at Figure 2, where the average binding energy per nucleon is provided as a
function of mass number. As mentioned previously, this curve indicates that the total binding
of one heavy nuclide containing ‫ > ܣ‬180 nucleons, which is given by
E A  AB  A , (25)
is larger than that of two identical nuclides containing ‫ܣ‬/2 protons and neutrons:
 A
2 E A/2  AB    E A. (26)
2
In physical terms, this means that two lighter nuclides generally represent a lower energy state
than a heavier nucleus. Accordingly, a spontaneous reaction where the heavy nuclide, ‫ ܪ‬,
breaks down into two smaller components (‫ܮ‬ଵ and ‫ܮ‬ଶ) is permitted from an energy-
conservation point of view, with the energy released by the reaction given by ‫ܧ‬௅భ + ‫ܧ‬௅మ − ‫ܧ‬ு .
This spontaneous reaction, called fission, has been observed experimentally for long-lived
actinides such as ଶଷଶTh, ଶଷହU, and ଶଷ଼U, even though this is not their preferential decay
mode (seven in 109 decays of ଶଷହU follow this path). For example, the spontaneous fission
reaction
235
92
U Þ 2 10 n + 133
50
Sn + 100
42
Mo, (27)
which releases 116 MeV, has been observed.
A second observation is that, even if a spontaneous fission reaction is very improbable, the
fission process can be initiated externally by an absorption collision between a projectile and a
heavy nuclide. For example, the neutron-induced reaction,
1
0 n  235
92U  3 0n  50 Sn  42 Mo,
1 133 100
(28)
is possible even with very low-energy neutrons and produces the same amount of energy as
the spontaneous fission reaction above. This reaction is facilitated by the fact that the main
interaction between the neutron and a nucleus is the attractive nuclear force. The fact that
several neutrons are generally produced following such reactions finally leads to the concept
of a controlled chain reaction, if sufficient neutrons are produced following a fission reaction
to initiate another fission reaction. The conditions required to achieve and maintain such a
chain reaction are discussed in Section 4.
Figure 2 also indicates that the nucleons inside very light nuclides are loosely bonded com-
pared to intermediate-mass nuclei. Combining two such nuclides could therefore result in a
nucleus that is more strongly bonded, again releasing energy. This process is called nuclear
fusion. Examples of such exothermal fusion reactions are
2
1 H  13 H  10n  24 He, (29)
2
1 H  32 He 11 H  24 He, (30)
which release respectively 17.6 and 18.3 MeV in the form of kinetic energy for the final fusion
products. Producing such reactions is much more difficult than neutron-induced fission
because the two initial nuclides have positive charges and must overcome the Coulomb force
to come into close contact. This can be achieved by providing sufficient kinetic energy (plasma
temperature) for a sufficient long time (confinement) to ensure that the energy released by
fusion is greater than the energy required to reach these conditions (the break-even point).
The two main technologies that are currently able to attain the conditions necessary for
controlled nuclear fusion are based respectively on confinement by magnetic fields (Tokamak)
and inertial confinement (fusion by laser).
3 Nuclear Reactions
From the nuclear chart (Figure 1), it is clear that only 266 nuclides are stable. This means that
most of the bonded nuclei that can be created will decay. This raises two questions:
 How do they decay?
 How are they created?
Another observation is that if one succeeds in creating a nucleus (stable or not), it may not
necessarily be in its fundamental (ground) energy state (see Section 2.5). These are the main
questions to answer in the first part of this section. We start by discussing the decay process
before continuing with a description of nuclear reactions. The second topic involves the
interaction of particles with matter, with most of these particles being produced during the
decay process.
3.1 Radioactivity and Nuclear Decay

Nuclear decay, the emission of radiation from an unstable nucleus or radioisotope, is a sto-
chastic process controlled by the laws of statistics and physics. Therefore, it cannot take place
unless several fundamental quantities are conserved [Basdevant2005, Smith2000]:
 total energy (including mass, which is a form of energy according to special relativity);
 linear momentum;
 angular momentum;
 electric charge;
 leptonic charge; and
 baryonic charge.
The latter two are required because of the intrinsic properties of quarks and leptons. For
common nuclear reactions, these last two conservation relations can be translated to:
 conservation of electrons and neutrinos where the electron and the neutrino have a
leptonic charge of +1 and the positron (anti-electron) and anti-neutrino have a leptonic
charge of -1; and
 conservation of the total number of protons plus neutrons (mass number).
A decay process transforms, through emission of particles, a parent nucleus into a more stable
daughter nucleus (a nucleus with a lower mass where the nucleons are more strongly bonded
or equivalently have a lower mass). In fact, if one considers a nucleus at rest that decays
according to the following reaction:
N
A
Z X  WB Y  VCii b, (31)
i 1
then total energy conservation implies that
   
N
m A X c 2  m B Y c 2  TB Y   mCi bc 2  TCi b , (32)
Z W W Vi Vi
i 1
where ܶ is the kinetic energy of the different decay products. This reaction can take place
spontaneously only if
N N
TB Y  TCi b  m A X c 2  m B Y c 2  mCi bc 2  0, (33)
W Vi Z W Vi
i 1 i 1
N
Z  W  Vi , (34)
i 1
N
A  B  Ci ,
i 1
(35)
that is, when the energy balance is favourable (the last two equations are for mass number
and charge conservation). For a nucleus in an excited state, the daughter nucleus will often be
the same as the parent nucleus, but at a different energy level (ground or lower-lying excited
state), the secondary particle being a photon (a ߛ-ray because the photon is produced follow-
ing a nuclear reaction).
Nuclear decay is characterized by a constant ߣ defined according to [Smith2000, Nikjoo2012]:
ΔN (t )
N (t )
  lim , (36)
Δt 0 Δt
where ܰ(‫ )ݐ‬is the number of nuclei present at time ‫ݐ‬, Δܰ(‫ )ݐ‬is the change (negative because
of the decay process) in the number of nuclei after time Δ‫ݐ‬, and ߣ is known as the decay
constant. This definition leads to the Bateman equation [Bell1982]:
dN (t )
  N  t   S (t ), (37)
dt
where ܵ(‫ )ݐ‬represents the net rate of production of nuclei (from the decay of other nuclei or
from nuclear reactions). One can also define
At    N t  , (38)
the activity of a radionuclide that represents the number of decays per second taking place at
time ‫ݐ‬. This activity is generally stated in Becquerels (Bq), where 1 Bq corresponds to one
decay per second. For the case where ܵ(‫ = )ݐ‬0, the solution to the Bateman equation has the
form
N (t0  Δt )  N  t0  e  Δt , (39)
where ܰ (‫ݐ‬଴) is the number of nuclei at time ‫ݐ‬଴.

Two other concepts related to decay constants are also commonly used. The mean life, ߬,
defined as
1
 , (40)

represents the average time required for the final number of nuclei to reach a value of
ܰ (‫ݐ‬଴)/݁:
N  t0 
N (t0   )  . (41)
e
Similarly, the half-life, ‫ݐ‬ଵ/ଶ,
ln( 2)
t1/2  ln  2   , (42)

represents the average time required for the initial number of nuclei to decrease to a value of
ܰ (‫ݐ‬଴)/2.
For some radionuclides, several decay reactions (decay channels) may be observed, each being
characterized by a specific relative production yield ܻ௝ such that
J
Y
j 1
j  1. (43)
The decay constant associated with each channel is given by ߣ௜ = ߣܻ௜. For example, potas-
ସ଴
sium-40 ( ଵଽ K), which has a mean life of 1.805 × 10ଽ years, decays 89.28% of the time into
calcium-40 (ܻ రబ
మబCa
= 0.8928) and 10.72% of the time into argon-40 (ܻ రబ
భఴAr
= 0.1072).
Note that the decay constant, ߣ, associated with a reaction can be evaluated using Fermi’s
second golden rule [Schiff1968]:
2
  
2
 f |H ' |i   ( f , Ξ) dΞ, (44)
f 
where |݅⟩ is the wave function associated with the initial state, ⟨݂|is the wave function associ-
ated with a final state, and ‫ ܪ‬ᇱ is the interaction potential associated with the reaction. The
integral is over phase space ݀Ξ, with the term ߩ(݂, Ξ) being the density of state for the final
wave function.
The most common decay reactions are the following:
ߛ-ray emission for an excited nucleus:
( ௓஺ X)∗ ⇒ ௓஺ X + ߛ. (45)
Internal conversion for an excited nucleus:

( ௓஺ X)∗ ⇒ [ ௓஺ X]ା + ݁ି . (46)
This process corresponds to the direct ejection of an orbital electron leaving the nu-
cleus in an ionized state ([ ௭஺ X]ା ).
ߚି decay (weak interaction) for a nucleus containing too many neutrons:

஺ ஺ ା
௓ X ⇒ [ ௓ାଵY] + ݁ି + ߥҧ . (47)
Neutron emission (nuclear interaction) for a nucleus having a very large neutron ex-
cess:
஺ ஺ିଵ ଵ
௓X ⇒ ௓ X + ଴݊. (48)
ߚା decay (weak interaction) for a nucleus containing too many protons:

஺ ஺ ି
௓ X ⇒ [ ௓ିଵY] + ݁ା + ߥ. (49)
Orbital electron capture (weak interaction) for a nucleus containing too many protons:
஺ ஺
௓ X ⇒ ௓ିଵY + ߥ. (50)
Proton emission (nuclear interaction) for a nucleus having a very large proton
excess:
஺ ஺ିଵ ି ଵ ஺ିଵ ଵ
௓ X ⇒ [ ௓ିଵ Y] + ଵ‫ = ݌‬௓ିଵ Y + Hଵ. (51)
This process corresponds to the ejection of an atom of hydrogen.
ߙ-emission for heavy nuclei (nuclear interaction):

஺ ஺ିସ ଶି
௓ X ⇒ [ ௓ିଶ Y] + ସଶߙ = ௓ିଶ
஺ିସ
Y + ସଶHe. (52)
This process corresponds to the ejection of an atom of helium.
Spontaneous fission for heavy nuclei (nuclear interaction):

஺ ଵ ஻ ஺ି஻ି஼
௓ X ⇒ ‫ ܥ‬଴݊ + ௒ F + ௓ି௒ G. (53)
This process corresponds to the fragmentation of a heavy nucleus into two or more
smaller nuclei with the emission of several neutrons.
For heavy or very unstable nuclei (see, for example, Figure 5 for uranium-238), several of these
reactions must often take place before a stable isotope is reached.
The Janis software from OECD-NEA can be used to retrieve and process the decay chain for all
isotopes [OECD-NEA2013].
Figure 5 Decay chain for uranium-238 (fission excluded). The type of reaction (β or α-decay)
can be identified by the changes in N and Z between the initial and final nuclides.
3.2 Nuclear Reactions

As we saw in the previous section, several types of particles are produced through decay of
radioactive nuclides. These particles have kinetic energy, and as they travel through matter,
they interact with orbital electrons and nuclei. A nuclear reaction takes place if these particles
produce after a collision one or more nuclei that are different from the original nucleus (an
excited state of the original nucleus or a new nucleus). Otherwise, the collision gives rise to a
simple scattering reaction where kinetic energy is conserved.
For the case where charged particles are traveling through a material, scattering collisions with
the orbital electrons or the nucleus are mainly observed (also called potential scattering).
More rarely, the charged particles are captured by the nucleus to form a compound nucleus (a
very short-lived, semi-bonded nucleus) that can decay through various channels.
Similarly, the electric and magnetic fields associated with the photons interact mostly with the
nucleus or the orbital electrons. They can also initiate nuclear reactions after being absorbed
in the nucleus. Finally, neutrons rarely interact with electrons because they are neutral
particles and the electrons are not affected by the nuclear force (electromagnetic interaction
with the quarks inside the nucleon, magnetic interactions with unpaired electrons, and weak
interactions are still present). However, they can be scattered (nuclear potential scattering) or
absorbed (nuclear reaction) by the nucleus. Again, this absorption can lead to the formation
of a very unstable compound nucleus that will decay rapidly.
The main nuclear reactions of photons, electrons, neutrons, and charged heavy particles with
nuclei are often the inverses of the decay reactions presented earlier:
  ray absorption:
  ZA X   ZA X  .
*
(54)
ߛ-ray scattering (elastic and inelastic):

  ZA X   ZA X    *.
*
(55)
electron capture:

e  ZA X   Z 1AY     γ. (56)
positron capture:

e  ZA X   Z 1A Y      (57)
proton capture:

p11  ZA X   ZA11Y    . (58)
ߙ capture:
2
 24  ZA X   ZA42 Y    . (59)
neutron capture:
Z X .
A1
0n  Z X  (60)
1 A
neutron scattering (elastic and inelastic):

n  ZA X   ZA X   10n   .
*
1
0 (61)
photon-induced fission:
  ZA X  C 10n  YB F  ABZCY G   . (62)
neutron-induced fission:
Z Y G   .
A1 B C
0n  Z X  C 0n  Y F  (63)
1 A 1 B
These reactions must obey the same conservation laws as those described earlier for the
decay reaction. For most reactions (except elastic scattering), the final nucleus will be in a
highly excited state and will decay rapidly by ߛ-ray emission (prompt photons).
The probability that any given reaction will take place between a nucleus and a particle is
represented by the microscopic cross section ߪ. Much like the decay constant, the cross
sections have a quantum mechanical interpretation and can be expressed in terms of the wave
function |݅⟩ associated with the initial state (the product of the wave functions of the projec-
tile and the nucleus), the wave function ⟨݂|associated with the final state, and ‫ ܪ‬ᇱ, the interac-
tion potential. The differential cross section is then given by Fermi’s first golden rule, which
can be written as:
d 2V 2
  f |H ' |i   ( f ), (64)
dΞ vi
where ܸ is the volume of the system, ‫ݒ‬௜ is the velocity of the projectile (assuming that the
initial nucleus is at rest), and ߩ(݂) is again the density of states in the phase space Ξ for final
particles. The cross section, which is then the integration over the final phase space of the
differential cross section, has units of surface. The most common unit used for the micro-
scopic cross section is the barn (b), with 1 b = 10ିଶସ cmଶ = 10ିଶ଼ mଶ.
These cross sections are very difficult to evaluate explicitly because of the complexity of the
wave functions (the wave function for a nucleus is the product of the wave functions of all the
nucleons in the nucleus bonded by the nuclear potential) and the interaction potential. These
are generally evaluated experimentally, and the results are stored in evaluated nuclear data
files. A software package such as Janis can be used to retrieve and analyze these cross sec-
tions [OECD-NEA2013]. For computational reactor physics, the NJOY program is generally
used to process the cross-section databases for neutron-induced reactions [MacFarlane2010].
3.3 Interactions of Charged Particles with Matter

As mentioned earlier, charged particles traveling through matter interact mainly with atomic
electrons because they are much more numerous than nuclei (ܼ electrons for a nucleus with
atomic number ܼ) and also because the wave functions associated with these electrons have a
very large spatial extension (the size of an atom) compared to that associated with the nu-
cleus.
Heavy particles, such as protons, ߙ-particles, and ions, are scattered by the Coulomb potential
produced by the electron, with the cross section given by the Mott scattering formula
[Leroy2009, Bjorken1964]. As a result, a heavy particle loses a very small quantity of energy
and momentum to secondary electrons and travels mainly in a straight line as it slows down.
The secondary electrons, on the other hand, can gain enough energy to be knocked loose from
the atom, leaving it in an ionized state.
For electrons and positrons traveling through matter, the problem is somewhat more complex.
The cross section for electron-electron collisions is given by the Möeller scattering formula,
while Bhabha scattering is used for positron-electron collisions [Leroy2009, Bjorken1964].
Positrons lose, on average, half their energy in collision with atomic electrons, and their
trajectory is somewhat erratic. Positron-electron annihilation reactions can also take place,
producing two ߛ-rays. Finally, the electrons and positrons that are accelerated in the strong
electromagnetic field of a heavy nucleus lose some of their energy by ߛ-ray emission due to
synchrotron radiation (also known as Bremsstrahlung) [Leroy2009, Bjorken1964].
One can approximate the energy loss of a particle of charge ‫ ݖ‬and mass ݉ inside a material of
density ߩ, atomic number ܼ, and atomic mass ‫ ܯ‬using the continuous slowing-down approxi-
mation (CSDA), where one assumes that the particle is interacting constantly with an electron
gas of density ݊ given by

nZ  A. (65)
M
The energy loss per unit distance traveled by the particle inside the material (−݀‫ܧ‬/݀‫)ݔ‬,
known as the stopping power, is then given by the Bethe-Block formula [Smith2000,
Leroy2009]:
 z e c  ( L v  F v ),
2
 dE 
   4 n     (66)
 dx collision mv 2
where ‫ ݒ‬is the speed of the particle and ‫ )ݒ(ܮ‬the stopping number, which can be written as
 2mv 2 
L  ln  2 
 (v / c ) 2 , (67)
 I (1  ( v / c ) ) 
where ‫ܫ‬is the mean excitation potential for the atomic electrons in this material. The correc-
tion term, ‫)ݒ(ܨ‬, takes into account the fact that the atomic electrons are bonded as well as
other quantum effects. For electrons, one must also consider the energy lost by synchrotron
radiation, which is given approximately by [Leroy2009]:
 dE   e3  2
   Zn , (68)
 dx  radiation me
where ݉ ௘ is the mass of the electron. Here, it is assumed that the kinetic energy of the
electrons ܶ௘ ≪ ݉ ௘ܿଶ.
For heavy particles (protons, ߙ-particles), nuclear slowing-down, (−݀‫ܧ‬/݀‫)ݔ‬nuclear, becomes
important when they reach a low energy.
The CSDA range ܴ of a particle is defined as the distance (path of flight) that it must travel to
lose all its kinetic energy. It is computed as
T
1
R dE , (69)

0 
dE 
 
 dx total
where ܶ is the initial kinetic energy of the particle and
 dE   dE   dE   dE 
          . (70)
 dx total  dx collision  dx  radiation  dx nuclear
These CSDA expressions are helpful to understand the general behaviour of particles slowing
down in a material. However, for practical applications, tabulated values for −݀‫ܧ‬/݀‫ ݔ‬and ܴ
are more useful. This type of information for electrons, protons, and ߙ-particles can be found
on-line for various elements and different material compositions on the National Institute of
Standard and Technology Web sites [Berger2013a]. The databases available are:
 ESTAR, for electrons;
 PSTAR, for protons;
 ASTAR, for ߙ-particles.
For heavy particles, this site also provides information on nuclear stopping power. Also note
that the stopping power and range are defined in a somewhat different way from the notation
above. The following expressions are used instead:
1
dE / dx (MeV  cm2 / g )   dE / dx (MeV / cm) (71)

and
R(g / cm2 )   R(cm) (72)
because these are the conventional units used in radiation shielding studies
studies. Note that the
number of electron/ion pairs produced per unit distance in a material is given approximately
by
dn pairs 1  dE 
   , (73)
dx W  dx total
where ܹ is the average energy required to ionize an atom in this material.
3.4 Interactions of Photons with Matter

The behaviour of photons as they travel in a material is somewhat diffe different
rent from that of
charged particles, which lose their kinetic energy mainly by continuously interacting with
electrons and nuclei, the Coulomb
oulomb field being a long
long-range force. The photon is the quantum
particle that carries the energy ‫ ߥ݄ = ܧ‬and momentum ‫݌‬Ԧൌ ݊ ݄⃗Ȁߣ of an electromagnetic field
ሬ
of frequency ߥ and wavelength ߣ ൌ ܿȀߥ travelling in direction ݊ ሬ⃗. Depending on its wave-
wav
length, the photon behaves as a wave (diffraction and interference for low
low-energy
energy photons) or
as a particle (photoelectric effectt for high
high-energy
energy photons) when it interacts with matter.
matter
Here, we will consider only the interactions with matter of the relatively high
high-energy
energy photons
(‫ ܧ‬൐ ͳ)) found in nuclear reactors
reactors. In this case, the photon behaves as a particle that is
first absorbed
sorbed by the atomic electrons or the nucleus
nucleus. Following this absorption, secondary
particles can be produced, including photons (scattering reactions), electrons and positrons
(pair creation), and neutrons (photonuclear reaction).
Figure 6 Microscopic cross section of the interaction of a high

high-energy
energy photon with a lead
atom.
The five main reactions taking place between high-energy photons and matter that are of
interest in nuclear reactor-related studies are:
 Photoelectric effect;
 Rayleigh scattering;
 Compton scattering;
 Pair production; and
 Photonuclear reactions.
With each such reaction is associated a macroscopic cross section Σ௠ ,௫(‫ )ܧ‬representing the
probability per unit distance travelled by a photon of energy ‫ ܧ‬that an interaction of type ‫ݔ‬
has taken place in material ݉ . The macroscopic cross section takes into account the probabili-
ties of interaction both between single photons and between particles and the fact that the
material contains ܰ௠ ,௜ particles of type ݅per unit volume:
 m , x ( E )   N m ,i i , x ( E ), (74)
where ߪ௜,௫(‫ )ܧ‬is the microscopic cross section of the interaction of a photon of energy ‫ ܧ‬with
a particle of type ݅through reaction ‫( ݔ‬see Section 3.2). The total microscopic photon cross
section can be expressed as:
 ( E )   Photoelectric ( E )   Rayleigh ( E )   Compton ( E )   Pair production ( E )   Photonuclear ( E ). (75)
Examples of microscopic cross sections for the interaction of photons with a lead atom are
presented in Figure 6.
Let us now describe in detail the five reactions contributing to ߪ(‫ )ܧ‬and discuss their respec-
tive dependences on the energy of the incident photon.
3.4.1 Photoelectric effect

This effect is illustrated in Figure 7. A photon with an energy ‫ = ܧ‬ℎߥ that is greater than the
binding energy ܹ of atomic electrons around the nucleus is absorbed. Following this absorp-
tion, the electron becomes free, leaving behind an ion. The energy balance for this reaction is
Te  TN  E  W , (76)
where ܶ௘ is the kinetic energy of the final electron and ܶே that of the recoil nucleus (generally
very small compared with ܶ௘).
Figure 7 Photoelectric effect.

From quantum mechanics, the binding energy of electrons around a nucleus of charge ܼ݁ can
be approximated using [Griffiths2005]
[Griffiths2005]:
13.61 Z 2
Wn  eV , (77)
n2
where ݊ ൌ ͳ for K-shell
shell electrons, 2 for LL-shell electrons, and so on. Therefore,, as the energy
of the photon decreases, the electrons in the more tightly bonded shells (K, LL, and M shells)
are successively excluded, producing a large drop in the photoelectric cross section when the
energy falls below ܹ ௡ (see Figure 6). The evaluation of this cross section is therefore very
difficult, and only approximate relations are available
available. For a K-shell
hell electron, one can use
[Leroy2009]:
3
s Photoelectricc,K ( E ) »
(
32 2p Z 5a e6 me c 2 ) ( c) ,
2
2
(78)
7
( )
3 E 2
which is valid, within 2%, for photon with energies ranging from 100 keV to 200 MeV.
MeV The
total photoelectric cross section (the
the sum over all energy shells) is given approximately by
[Leroy2009]:

 Photoelectric  E    Photoelectric , K  E   1  0.01481  ln  Z    0.000788  ln  Z   .
2 3
 (79)
Clearly, this cross section decreases rapidly with photon energy

energy. It is dominant only for pho-
ph
tons of energies less than a few tenths of an MeV.
Several processes can take place after this reaction
reaction. If the electron ejected from the atom was
initially in the valence band, the ionized atom is in its ground state
state, and no further radiation
emission is observed. When electrons coming from more tightly bonded energy levels level are
ejected, the ionized atom is left in an excited state
state. This excited ion
on can then decay to its
ground state by two different means
means. First, an electron in a much higher energy state (i.e.,
(i.e. a
valence electron) can emit a photon and fill the vacant energy levellevel. The electron can also
come from a slightly higher energy level, leaving the ion in a less excited state that will further
decay before reaching its ground state
state. When the photons emitted in these successive decays
are absorbed by less-bonded electrons, so-called soft Auger electrons are ejected from the
atom, producing an Auger electron cascade.
3.4.2 Rayleigh scattering

This elastic scattering reaction takes place when a photon, after being absorbed by a bound
atomic electron, is re-emitted with the same energy, but in a different direction, by the elec-
tron, returning the atom to its original state [Leroy2009, Jauch1976]. The cross section for this
reaction is very difficult to evaluate because it depends strongly on the wave function of the
bonded electrons surrounding the nucleus. It is generally written as an integral of the form
  c 
2
 Rayleigh  E     e 2   1  cos      F  E,  , Z  sin    d  ,

2
2
(80)
 mec  0
where ‫ܧ(ܨ‬, ߠ, ܼ) is the atomic form factor that represents the effect of the ܼ electrons sur-
rounding the nucleus on the scattering of a photon at an angle ߠ with respect to the initial
photon direction. Several theoretical expressions for ‫ܧ(ܨ‬, ߠ, ܼ) are available in the literature
for different values of ܼ and various energy ranges. For example, when the photon energy is
greater than 100 keV, the main contributors to Rayleigh scattering are the electrons in the K-
shell, and the relativistic Bethe-Levinger form factor is used [Hubbell1975]:
sin(2g atan(Q))
(
F E,q ,Z = ) , (81)
( )
g
g Q 1+ Q 2
with
2E sin(q / 2)
(
Q E,q ,Z = ) a e Zme c 2
, (82)
g ( Z ) = 1+ (a e Z ) .
2
(83)
Because Rayleigh scattering is never the dominant photon reaction (see Figure 6) for the
photon energy range considered in nuclear reactors, its contribution to the total cross section
is often neglected.
3.4.3 Compton scattering

The Compton effect is the incoherent inelastic scattering of a photon by an electron that is
weakly bonded to a nucleus (see Figure 8). Here, a photon with momentum ݇ ሬ⃗ (ห݇
ሬ
⃗ ห= ℎߥ/ܿ =
‫ܧ‬/ܿ) is scattered at an angle ߠ with respect to its original direction, with the final momentum
of the photon being ݇ ሬ⃗ ′. The electron is then ejected with a kinetic energy ܶ௘, obtained by
conservation of momentum and energy:
 (1  cos  ) 
Te  E 2  , (84)
 me c  E (1  cos  ) 
2
where the binding energy of the electron is assumed to be small compared to the photon
energy, and the recoil energy carried by the nucleus is neglected. For a nearly free electron,
the cross section for this reaction is given by the Klein-Nishina formula, which can be written
as
   c    2 1     1 4 
2
1
 Compton  E    e 2  ln 1  2  1     2
, (85)
  mec     2
 2  2 1  2   
with   E / (mec 2 ) This cross section decreases with energy
energy, but much more slowly than
tha for
the photoelectric effect. It dominates the photoelectric reaction for energies above approxi-
approx
mately 0.5 MeV.
Figure 8 Compton effect.
3.4.4 Pair production

A photon with energy ‫ ܧ‬ൌ ݄ߥ ൐ 2 2݉ ௘ܿଶ traveling through the Coulomb field of a nucleus of
charge ܼ (or an electron) can be converted into an electron
electron-positron pair. After the reaction,
the nucleus carries a small part of the kinetic energy of the photon, with the bulk of the
energy being transferred
ansferred in the form of mass and kinetic energy to the pairpair.. For photon
energies in the range found in nuclear reactors, the cross section for this reaction is given
approximately by [Leroy2009]:
  c   28
2
218 129 
 Pair production  E   e Z  e 2   ln  2  
2
  , (86)
 e 
m c  9 17 20 
with   E / (mec 2 ). This cross section vanishes for ‫ ܧ‬൏ ͳǤͲʹ MeV and is larger than the
Compton cross section above a few MeV.
3.4.5 Photonuclear reactions

Photonuclear reactions are the result of photons being absorbed by a nucleus to form an
excited nucleus, which gets rid of its energy through different decay processes, namely the
emission of a secondary photon, the ejection of one or more nucleons (proton
(protonss and neutrons),
the ejection of ߙ-particles, or through a fission reaction. Photonuclear reactions occur only
above a threshold, which lies between 7 and 9 MeV for heavy nuclides, but is considerably
lower for very light isotopes (1.666 MeV and 2.226 Me MeVV respectively for beryllium and deute-
deut
rium). The absorption cross sections ions for these reactions (see Figure 9)) are generally very
difficult to evaluate,, and only approximate expressions are available.
Figure 9 Photonuclear cross sections for deuterium, beryllium, and lead.

For deuterium, which is important to CANDU reactor
reactors because they use large amounts of
heavy water, the expression [Leroy2009]
( )
3
E - 2.226
s 2 D,Phhotonuclear (E) » C 2 D 3
(87)
E
is often used, where ‫ܧ‬ǡthe
the photon energy
energy, is given in MeV, ‫ ܥ‬మD varies between 0.061 and
0.0624, and ߪD (‫ )ܧ‬is in barns. The end result of such a photonuclear reaction is a hydrogen
atom and a fast neutron that can initiate a fission reaction (see Section 2.6).
Tabulated values for Rayleigh scattering (coher
(coherent
ent scattering), photoelectric effect, Compton
scattering (incoherent scattering), and pair production cross sections with different atoms or
materials and for energies ranging from 1 keV to 100 GeV are available on
on-line
line on the Web site
of the National Institute of Standards and Technology (NIST) [Berger2013b]. For photonuclear
reactions, the energy-dependent
dependent cross sections for various isotopes can be obtained using the
OECD-NEA Janis software [OECD-NEA2013]
NEA2013].
For each reaction presented
esented above, the photons first transfer all their energy to an electron or
a nucleus following a collision. As a result, starting with ݊ሺ‫ܧ‬ǡ‫ݎ‬ ሬ
ሬ⃗) photons of energy ‫ ܧ‬and
Ԧǡȳ
direction ሬ
ሬ⃗ incident at point ‫ݎ‬
Ω on a material of uniform macroscopic cross section
Ԧon on ߑሺ‫ܧ‬ሻ, the
number of initial photons ݊ሺ‫ܧ‬ǡ‫ݎ‬ Ԧ+ ሬ
ሬ
⃗ ሬሬ
⃗
+ ‫ݏ‬ȳ , Ω) still present after a distance ‫ ݏ‬has been crossed is
given by [Basdevant2005]:
      E s
  
n E , r  sΩ,Ω  n E , r , Ω e   .  (88)
The number of photons therefore decreases exponentially with distance. This behaviour is
similar to what is observed when light is attenuated by passing through a material with a
linear attenuation coefficient ߤ(‫)ܧ(ߑ = )ܧ‬. Secondary photons produced with energy ‫ ܧ‬and
direction ሬ
Ωሬ⃗ following the absorption of photons of energy ‫ܧ‬′ and direction ሬ Ωሬ⃗′ will also con-
tribute to the total population of photons at point ‫ݎ‬ Ԧ+ ‫ݏ‬Ω ሬሬ⃗. The equation that describes the
photon population in a material, assuming a constant neutron incident population, is the static
photon transport equation, which takes the form [Pomraning1991]:
          
    
Ω   n E , Ω, r    E  n E , Ω, r  dE ' dΩ'  s E '  E , Ω '  Ω n E ' , Ω ' , r ,  
(89) 
where Σ௦൫‫ܧ‬ᇱ → ‫ܧ‬, Ω ሬሬ⃗ᇱ → Ωሬ
ሬ⃗൯ is the differential scattering cross section representing the prob-
ability that the absorption of a photon of energy ‫ܧ‬ᇱ and direction ሬ ሬ⃗ᇱ produces a photon of
Ω
energy ‫ ܧ‬and direction ሬ ሬ⃗.
Ω
3.5 Interactions of Neutrons with Matter

Neutron interaction with matter is the result of the nuclear force, although weak and electro-
magnetic interactions, the latter being the result of anomalous magnetic moment, are also
possible. As a result, only neutron-nucleus interactions are generally considered, including
 elastic potential scattering with cross section σĞůĂƐƟĐ;
 inelastic scattering with cross section σŝŶĞůĂƐƟĐ;
 neutron absorption with cross section σĂďƐŽƌƉƟŽŶ ; and
 spallation reactions, σƐƉĂůůĂƟŽŶ .
In the first three reactions, the neutron interacts with the nucleus as a whole. In spallation
reactions, the neutrons interact directly with the nucleons in the nucleus. The latter reaction
can take place only when the energy of the incident neutron ‫ܧ‬௡,௜ is sufficiently high (threshold
reactions with ‫ܧ‬௡,௜ > 0.1 MeV), and therefore σspallation is either neglected, because it is too
small, or simply included in the absorption cross section. The probability that the neutron
interacts with a nucleus is then given by the total cross section, σ, where
σ( En ,i )  σabsorption ( En,i )  σ scattering ( En,i ) , (90)
σ scattering  En,i   σelastic  En,i   σinelastic  En,i  . (91)
Elastic potential scattering is an interaction in which kinetic energy is conserved and the
neutron does not penetrate the nucleus. As a result, the neutron is slowed down, transferring
part of its energy to the nucleus. For a collision between a neutron with an initial kinetic
energy ‫ܧ‬௡,௜ and a nucleus of mass ‫ ܯ‬X at rest, the final neutron kinetic energy ‫ܧ‬௡,௙ (momentum
and energy conservation) will be given by [Bell1982]:
A  1  2 A cos  ,
En ,i
En , f  2
(92)
1  A
2
where ‫ ܯ = ܣ‬X /݉ ௡ and ߴ, the scattering angle for the final neutron in the of mass system, is
related to the scattering angle in the laboratory system ߠ by
A cos  1
cos   . (93)
 A2  1  2 A cos 
On average, the neutron will lose an energy Δ‫ܧ‬:
E 
1    E , (94)
n ,i
2
 A  1 ,
2
α (95)
 A  1
2
after each collision. Because Δ‫ ܧ‬decreases as ‫ ܣ‬increases, neutron slowing-down (also called
moderation) is more efficient when scattering collisions with light rather than heavy nuclei are
involved.
For low-energy neutrons, the potential scattering cross section in the centre of mass system is
nearly constant and has the form [Bell1982]:
s = 4p R 2 , (96)
where ܴ is the nuclear radius for this material. At higher energy, one can use Fermi’s golden
rules to obtain an approximation for ߪ(‫ )ܧ‬of the form
 2 sin 2 d
s (E) = 4p , (97)
2mn E
where ߜ represents the phase shift of the neutron wave function in isotropic scattering.
Inelastic scattering and absorption reactions generally proceed by creation of a compound
nucleus:
n  ZA X   AZ1 X   n'   ZA X  ,
* *
(98)
n  ZA X   AZ1 X   other decay channels.

*
(99)
The creation of the compound nucleus is a resonant reaction because the resulting weakly
bonded nucleus can exist only if it corresponds to an excitation state of ஺ାଵ ௓X. The lifetime of
ିଵ଺
this excited nucleus is generally very short (߬ ≈ 10 s), and with each decay channel ݇,
including inelastic scattering, is associated a resonance width given by [Bell1982]:

Γk  , (100)
k
with ߬௞ being the mean life for decay through this channel. The total resonance width for the
creation of the excited nucleus Γ is then given by
Γ  Γ k . (101)
For light nuclei, Γinelastic ≫ Γabsorption , and the inelastic scattering process dominates. For
heavy nuclei, the reverse is true (absorption dominates). The resonance widths for inelastic
scattering also increase with energy, the explicit dependence being
inelastic  E   n  E   0 E , (102)
where Γ଴ is a constant that depends on the nucleus considered.

The probability (cross section) for the creation of a compound nucleus with resonant energy
‫ܧ‬଴ is given by
 
 n  E   
  2
 A1 X  E    2l  1 , (103)
Z
2mE   
2
  E  E0     
2
 2 
where ݈ is the orbital-angular-momentum
momentum quantum numb
number. For a nucleus that decays
through channel ݇, the cross section is
k
 k  E    A1 X  E  . (104)
Z

Note that at low energy (‫ܧ ا ܧ‬଴), Γ௡ = Γ଴√‫ܧ‬,
 2 0 k  0
 k  E    2l  1  , (105)
2m E v
and the cross section is a function of ͳȀ‫ݒ‬, with ‫ ݒ‬being the incoming neutron velocity.
Figure 10 Cross sections for hydrogen (l

(left) and deuterium (right) as functions of energy.
Examples of cross sections for light (hydrogen and deuterium) and heavy nuclei ( ଶଷହ ଽଶU and
ଶଷ଼
ଽଶU) are presented in Figure 10 and Figure 11. For or light nuclides, the scattering process
clearly dominates. One can also observe that the total cross section of hydrogen is about ten
times larger than that of deuterium (better scattering nucleus for neutron moderation).
moderation) With
absorption
bsorption representing 10% of the total cross section of hydrogen, one out of ten collisions
will lead to neutron capture. For deuterium, the ratio of neutron absorptions to collisions is
closer to 1%, making this isotope a better neutron moderator overall. For heavy nuclides, the
absorption reaction is the most prevalent and exhibits a very large number of resonances.
resonances At
very high energy, these resonances do not disappear, but are so closely packed that they
produce a nearly continuous cross section.
Figure 11 Cross sections for 235

92 U (left) and 238
92 U (right) as functions of energy.
The derivations above are valid only if one assumes that the nucleus is at rest
rest. This is clearly
not the case for material having temperatures ܶ ൐ Ͳ, because atoms acquire an average
kinetic energy that is proportional to the temperature of their medium. The resulting effect
on the cross sections can be taken into account for a material at temperature ܶ using
[Bell1982]:
   
 v X  k  vn  v X  p  E X  dE X ,
1
 k  E, T   v n (106)
2mE
where ‫ݒ‬Ԧ௡ and ‫ݒ‬ԦX are the velocity of the neutron and the nucleus respectively. Here, ‫ܧ(݌‬X )݀‫ܧ‬X
is the Maxwell-Boltzmann
Boltzmann distribution
distribution, which represents the probability of finding
ing a nucleus
with energy ‫ ܧ‬in the range ‫ܧ‬X ൏ ‫ ܧ‬൏ ‫ܧ‬X ൅ ݀‫ܧ‬X due to thermal motion:
2
p  E X  dE X  E X e  E X /( k BT )dE X , (107)
  k BT 
3/2
where ݇࡮ is the Boltzmann constant

constant. This thermal motion of nuclei has no impact on cross
sections exhibiting ͳȀ‫ ݒ‬behaviour (i.e.
(i.e., absorption cross sections of hydrogen at low energy).
energy)
At low neutron energies,ies, the scattering cross section for light nuclides also exhibits ͳȀ‫ݒ‬
behaviour and is not affected by this transformation, although the energy distribution of the
final neutrons resulting from these collisions will be such that they end up in thermal equilib-
eq
rium with the nuclides in the medium [Bell1982].
The thermal motion of atoms also changes the form of resonant cross sections by reducing the
height of the resonance while inc
increasing its width (see Figure 12). This widening of the
resonance is known as the Doppler effect.
Figure 12 Effect of temperature on U 238 absorption cross section around the 6.64-eV reso-
nance.
The neutron-nucleus interaction cross sections are very complex and in most cases are evalu-
ated using a combination of theory and experiment. They are generally tabulated in evaluated
nuclear data files that can be read by nuclear calculation software (NJOY, for example)
[MacFarlane2010]. They can also be extracted on-line from nuclear databases or off-line using
the Janis software [KAERI2013, Livermore2013, OECD-NEA2013].
Once the energy-dependent cross sections are known, one can easily evaluate the number of
reactions of type ‫ ݔ‬per second (the reaction rate ܴ௠ ,௫(‫ ))ݒ‬taking place inside a region of
volume V containing ܰ௠ nuclei per cm3 of a material ݉ (cross section σ௠ ,௫(‫ ))ݒ‬with a popula-
tion of ݊(‫ )ݒ‬neutrons per cm3 having a velocity ‫ = ݒ‬ඥ2‫ܧ‬௡ /݉ :
N m m , x  v  n  v 
Rm, x  v   V  V m, x  v    v  , (108)
v
where ߶(‫)ݒ(݊ = )ݒ‬/‫ ݒ‬is the neutron flux and Σ௠ ,௫(‫ܰ = )ݒ‬௠ σ௠ ,௫(‫ )ݒ‬the macroscopic cross
section for a reaction of type ‫ ݔ‬in this material.
4 Fission and Nuclear Chain Reaction

The concept of controlled energy production in a nuclear reactor is based on three fundamen-
tal ideas that were discussed in Sections 2.6, 3.1, and 3.2, namely:
 energy is released if a heavy nucleus is fragmented into lighter nuclei, the fragmentation
process being called nuclear fission;
 the spontaneous nuclear fission process, which has a very low yield for several heavy
nuclei, generates neutrons in addition to the light nuclei as end products;
 the nuclear fission process can be initiated by neutrons and photons.
Accordingly, a neutron of kinetic energy ܶ௡ travelling through a medium containing heavy
nuclei at rest could initiate a fission reaction:
A B C 1
1
o n  ZA X  C o1n  YB F  Z YG. (109)
The kinetic energy ܶோ released by this reaction is then
TR  Tn  (1  C )mnc 2  m A X c 2  m B F c 2  m AB C 1G c2 , (110)
Z Y Z Y
with part of the energy carried by the neutrons and the lighter nuclei. These lighter nuclei
rapidly lose their energy in matter (see Section 3.3), thereby producing heat that is removed
by the coolant and transformed to electricity.
Not all the neutrons produced after this reaction initiate a fission reaction. A number of them
are absorbed by material without producing a fission reaction. Others simply leave the system
(with or without collisions). Provided that on average a single neutron per fission reaction can
initiate a new fission, a controlled exothermal chain reaction can be established. Establishing
such a chain reaction looks relatively simple; however, this is far from the case, as shown in
the following subsections.
4.1 Fission Cross Sections

The probability of fission is given by the value of the fission cross sections associated with a
nucleus and depends on the energy of the neutron that initiates the reaction. All isotopes
with atomic number ܼ > 90 are fissionable (have a non-zero probability of undergoing fission
following the capture of a neutron). Table 5 compares the average energy, ܷ, required to
initiate a fission reaction in different nuclei with the energy, ܳ, available in the compound
nucleus after a neutron with a vanishing kinetic energy has been captured. For nuclei contain-
ing an even number of protons and an odd number of neutrons, ܳ > ܷ. Such isotopes are
called fissile isotopes because the internal energy of the compound nucleus resulting from
neutron capture is sufficient to surmount the potential barrier for fission even when the
incident neutrons are very slow (ܶ௡ ≈ 0). For isotopes with ܳ < ܷ, the fission barrier can be
overcome only using neutrons with relatively high kinetic energy. Fission for these nuclei is a
threshold reaction.
Table 5 Comparison of the energy, U, required for a fission reaction in a compound nucleus
with the energy, Q, available after a neutron collision with different isotopes.
Isotope ࡽ(MeV) ࢁ(MeV)
ଶଷଶ
ଽ଴Th 5.1 5.9
ଶଷଷ
ଽଶU 6.6 5.5
ଶଷହ 6.4 5.8

ଽଶU
ଶଷ଼
ଽଶU 4.9 5.9
ଶଷଽ
ଽସPu 6.4 5.5
As shown in Figure 13, the energy dependence of the fission cross sections for ଶଷହ ଶଷ଼
ଽଶU and ଽଶU
reflects the behaviour described above. The probability that a neutron of energy ‫ܧ‬௡ < 1.0
MeV will collide with ଶଷ଼ ଶଷହ
ଽଶU and result in a fission reaction is very low. For ଽଶU, there is no
lower limit on the neutron energy, and the fission probability increases as the energy of the
neutron decreases. The neutron capture cross section (neutron absorption followed by
photon emission) for ଶଷ଼ ଶଷହ
ଽଶU dominates at low energy, while for ଽଶU, it remains relatively small
compared to the fission cross section (approximately 10%) over the full energy range. As a
result, most neutron absorption in ଶଷହ
ଽଶU will lead to fission followed by production of new
neutrons (neutrons are regenerated). For ଶଷ଼
ଽଶU, a collision with a neutron having energy below
the fission threshold will lead to a net neutron loss. As a result, the natural uranium fuels used
in CANDU reactors (0.72 atomic% of ଶଷହ ଽଶU) will be much less reactive (prone to fission) than
the enriched fuels (enrichment factors ߳ ranging from 2.5 to 5 atomic% of ଶଷହ ଽଶU) used in light
water reactors (LWR).
Figure 13 Capture and fission cross sections of 235

92 U (left) and
238
92 U (right) as functions of
energy.
4.2 Fission Products and the Fission Process

The fission process generally produces two light nuclei, called fission products, although three
or more nuclei can also be generated. The relative rate of production of a given nuclide ‫ܫ‬,
called the fission yield, is denoted by ܻூ and satisfies
Y
I
I  2, (111)
assuming that only two nuclei are produced by each fission reaction. As one can see in Figure
14, where the fission yields of all the nuclei resulting from the fission of ଶଷହ
ଽଶU are plotted as a
function of their mass number, the fission process favours the production of relatively light
nuclides (90 < A<105) combined with heavier isotopes (130 < A<145).
Figure 14 Relative fission yield for 235

92 U as a function of mass number.
number
The average number ߭ of neutrons emitted by the fission process (see Table 6)) is relatively
small compared to the neutron excess ܰ െ ܼ ൐ ͶͲ required for heavy isotopes to exist. exist The
value of ߭ for a given reaction depends on the nature of the fissile nuclides, the fission prod-
pro
ucts released, and the energy of the neutrons
neutrons. Most fission products are therefore very
processes ߚି decay
neutron-rich and unstable. They will reach stability through two main processes:
and neutron emission. All the transformations that take place within 10ିଵସs of the fission
reaction are generally included in the so so-called
called prompt contribution (prompt neutrons,
prompt electrons, and prompt photons)
photons). Particles produced on a longer time scale (from
milliseconds to minutes) due to decay of isotopes with longlonger mean lifetimes
lifetime are called
delayed.
Table 6 Average numbers of neutrons produced by the fissio
fission
n of common heavy nuclides.
Isotope ࣏(ࡱ࢔ <1 eV) ࣏(ࡱ࢔ >1 MeV)

ଶଷଶ
ଽ଴Th 0 2.2
ଶଷଷ
ଽଶU 2.50 2.6
ଶଷହ 2.43 2.6
ଽଶU
ଶଷ଼
ଽଶU 0 2.6
ଶଷଽ
ଽସPu 2.89 3.1
The fission reaction (see Section 2.6) also releases a large amount of kinetic energy (on the
order of 200 MeV) that is distributed between fission products, neutrons, ߛ-rays, rays, etc.
etc (see
Table 7). Clearly,, the fission products carry most of the energy, and because of their very short
range in matter, this energy heats the nuclear fuel electron (ߚି
fuel. Similarly, the charged electrons
particles) also lose most of their energy in the fuel
fuel. The highly penetrating ߛ-rays
rays will end up
depositing their energy in the reactor structures, while the weakly interacting neutrinos will
generally be lost to the environment
environment. Finally, for LWR and CANDU reactors, only a very small
sm
part of the neutron energy will be transferred to the fuel in the form of heat, with most of it
being lost to the environment (mostly through slowing-down in the moderator).
Table 7 Distribution of kinetic energy among the particles produced by a fission reaction.
Particles Kinetic energy

(MeV)
Fission products 168
Prompt ߛ-rays 7
Delayed ߛ-rays 7
Electrons 8
Neutrinos 12
Neutrons 7
Total 207
Note that the data presented in Table 7 represent average values, meaning that the neutrons
produced in the fission process have an energy distribution. The probability that a neutron of
energy ‫ܧ‬௡ is produced in the fission process, which is known as the fission spectrum, ߯(‫ܧ‬௡ ), is
often approximated by the Watt relationship [Watt1952]:
  En   Ce  aEn sinh  
bEn , (112)
where ܽ and ܾ are constants associated with the isotope undergoing fission and depend
weakly on the incident neutron energy. The normalization factor ‫ ܥ‬is defined so that

  E  dE
0
n n 1 (113)
and has the value

2a 2/3e -b/4a
C= . (114)
pb
This fission spectrum for ଶଷହ
ଽଶU is illustrated in Figure 15. Clearly, most of the neutrons are
released with energies greater than 1 MeV and can initiate fission reactions with fissionable
material. However, the fission cross section for this energy range is relatively low (see Figure
13), and fissions at such energies are highly improbable. Nevertheless, a controlled chain
reaction that relies mainly on fast fission can be achieved in so-called “fast neutron reactors”,
where ଶଷଽଽସPu is often used because of its large fission cross section for high-energy neutrons.
The alternative to fast neutron reactors is moderated reactors, in which the fast neutrons are
slowed down to energies where the fission cross section is significantly larger (below 1 eV).
This is the concept used in “thermal neutron reactors”, which are discussed in the next sec-
tion.
Now, let us look at the properties of fission products. These nuclides have intermediate
masses and vanishing fission cross sections. However, they can have very large neutron-
capture cross sections and thereby can have a considerable negative impact on the neutron
population in the fuel where they are produced. For this reason, they are often called neutron
poisons. Examples of poisons that have a large impact on nuclear-reactor operation are ଵଷହହସXe
଺ ଵସଽ ସ
(ߪ௔ = 3.1 × 10 b) and ଺ଶSm (ߪ௔ = 6.1 × 10 b). A second observation is that most of these
nuclides are unstable and decay naturally through one of the reactions described in Section
3.1. These reactions produce both particles, including delayed neutrons, and energy that is
often deposited in the fuel. These delayed neutrons play an important role in the control of
nuclear reactors, as will be explained in Chapter 5. Moreover, energy will still be produced in
the fuel for a long time after the controlled fission reaction has been stopped, due to the
decay of fission products. This is the main source of the heat produced in neutron-irradiated
nuclear fuels (some energy is also produced through photonuclear fission). This residual heat
is removed from shut-down reactors and irradiated fuel pools by continuously circulating the
coolant.
Figure 15 Fission spectrum for 235

92 U.
Finally, we must address the problem of the changes in the composition of a material exposed
to a neutron flux. In Section 3.1, we introduced the Bateman equation, which takes the form
[Smith2000, Nikjoo2012, Magill2005]:
ௗே ಺(௧)
= −ߣூܰூ(‫ )ݐ‬− ‫ܮ‬ூ(‫ )ݐ‬+ ܲூ(‫)ݐ‬, (115)
ௗ௧
where ܰூ(‫ )ݐ‬is the concentration of isotope ‫ܫ‬in the material and the net source ܵ(‫ )ݐ‬is now
divided into a production ܲூ(‫ )ݐ‬and a loss ‫ܮ‬ூ(‫ )ݐ‬term. The loss term, from sources other than
decay, is the result of a neutron being absorbed (captured) by isotope ‫ܫ‬:

LI  t   N I  t  σ I ,absorption ( E )  E  dE , (116)
0
where ߶(‫ )ܧ‬is the neutron flux. The production term has two contributions: decay from other
nuclei, and transformation of a different nucleus into ‫ܫ‬following a neutron-induced nuclear
reaction:

PI  t   YJd, I J N J  t   YJx, I N J  t  σ J , x  E    E  dE , (117)
J J x 0
ௗ ௫
where ܻ௃,ூ is the yield for the production of isotope ‫ܫ‬from the decay of isotope ‫ܬ‬, while ܻ௃,ூ is
the isotope ‫ ܫ‬production yield from a reaction of type ‫( ݔ‬fission, capture, etc.) between a
neutron and isotope ‫ܬ‬.
4.3 Nuclear Chain Reaction and Nuclear Fission Reactors

The previous sections have introduced all the concepts required to design a conceptual
nuclear fission reactor:
 The nuclear fuel contains heavy isotopes that are broken into lighter nuclides (fission
products) following a neutron-induced fission reaction. Such a reaction also generates en-
ergy and secondary neutrons that can be used to maintain a chain reaction.
 Most of the energy released as a result of the fission process is carried away by the fission
products and deposited in the fuel because of the short range of these nuclei.
 The secondary neutrons produced in the fission process are fast neutrons.
 These neutrons can lose their energy through scattering collisions with the various materi-
als present in our conceptual reactor. Neutron slowing-down by collisions with light iso-
topes is more efficient than by collision with heavy isotopes (see Section 3.5).
 This energy deposition increases the temperature of the fuel, which must then be cooled
to make use of the heat and avoid damage to the reactor. For a power reactor, this heat is
collected in a coolant (generally gas or liquids circulating around the fuel) and used to gen-
erate electricity.
Now let us look at the fuel and coolant materials that are the most readily available.
Three heavy fissionable isotopes are present in nature: ଶଷଶ ଶଷ଼ ଶଷହ
ଽ଴Th, ଽଶU, and ଽଶU. Of these, only
ଶଷହ 3
ଽଶU is fissile by neutrons of all energies. Because the density of thorium is 11,000 kg/m and
it has a microscopic fission cross section of 0.1 barn for 1 MeV neutrons, the mean free path of
a neutron for fission in this material is 3.5 m. The mean free path for neutron capture in
thorium is very similar in magnitude. Therefore, nearly half the 2.2 neutrons emitted by
fission in thorium will be lost to unproductive capture. For ଶଷ଼ ଽଶU, which represents 99.27% of
all the atoms found in natural uranium (density of 18,000 kg/m3), the odds are slightly better,
with mean free paths for fission and capture respectively of 50 and 300 cm and fewer neu-
trons emitted per fission are lost to capture.
The main advantage of ଶଷହ ଽଶU, even though it represents only 0.72% of natural uranium compo-
sition, is that its fission cross section is quite high (1000 barns at 0.01 eV), giving a mean free
path for fission of 2 cm. The capture cross section of ଶଷହ ଽଶU at this energy is also six times
smaller than the fission cross section. It is possible to decrease the mean free path for fission
with ଶଷହଽଶU by increasing its relative concentration in uranium-based fuel (fuel enrichment).
The ratio of capture to fission in uranium-enriched fuels remains small even if the fuel contains
a large amount of ଶଷ଼ ଽଶU (capture cross section less than 1 barn).
The heat produced in the fuel can then easily be extracted by conventional means (by circulat-
ing a fluid such as water). The main problem is to slow down the secondary fission neutrons
to low energies (< 1 eV) while ensuring that few are lost by capture inside other materials or
by leaving the reactor. A neutron slowing-down material, the moderator, must therefore be
introduced into the reactor to serve this purpose.
Table 8 Maximum values of the elastic scattering and absorption cross sections for 11 H , 21 D ,
and 12
6 C in the energy range 0.1 eV < E <0.1 MeV.
Isotope ߪƐĐĂƩ ĞƌŝŶŐ (b) ߪĂďƐŽƌƉƟŽŶ (b)

ଵ
ଵH 18 0.17
ଶ
ଵD 3.4 0.00025
ଵଶ
଺C 4.6 0.0016
For this moderator, two questions are of prime importance: what should be its composition,
and where should it be located? The first question is easily answered by combining the kinet-
ics of the neutron-nucleus collision and a survey of the scattering and absorption cross sec-
tions of the most promising candidates. In Section 3.5, we concluded that the slowing-down
efficiency of a nucleus increases as its mass number decreases, which points to the use of light
nuclides as moderators. The most promising candidates are the isotopes of hydrogen and
carbon that are readily available in water and graphite. Table 8 provides the maximum values
of the elastic scattering and absorption cross sections for ଵଵH, ଶଵD, and ଵଶ଺C in the energy range
0.1 eV < ‫ < ܧ‬0.1 MeV. As one can see, ଵଵH is the best moderator if one does not take into
account its absorption cross section. However, up to one in ten collisions of a neutron with ଵଵH
could lead to absorption (for 1eV neutrons). On the other hand, even if the number of colli-
sions required to slow down neutrons using ଶଵD and ଵଶ଺C is significantly higher (the distance the
neutron will travel between collisions is longer), less than 0.007% of the collisions with ଶଵD and
0.03% of the collisions with ଵଶ଺C (1 eV neutron) lead to capture. Accordingly, if one wants to
build a core that is compact, a ଵଵH-based moderator is a requirement (light water, for example
in LWR); otherwise the best candidate is a moderator that contains large concentrations of ଶଵD
(heavy water in CANDU).
Now let us try to answer the question of the spatial distribution of this moderator. Two very
different options can be considered: a homogeneous mixture of fuel and moderator, and small
isolated fuel elements placed strategically in the moderator. Again, the answer to this ques-
tion can be inferred from Figure 13. The neutrons in a homogeneous mixture of fuel and
moderator materials have a large chance of colliding with fuel isotopes as they are slowing
down. This means that neutrons with energies in the “resonance” range (4 eV < ‫ < ܧ‬10 keV)
will be in direct contact with fuel nuclides. Because their capture cross sections at resonance
energies increase, often by several orders of magnitude with respect to the background value,
neutron capture will have a detrimental effect on the number of neutrons that can reach
energies below 1 eV. On the other hand, by separating different fuel elements by a sufficient
amount of moderating materials to ensure that a maximum number of neutrons escape the
resonance energy range, one decreases neutron loss by capture and increases the probability
of fission.
These are the general physics considerations that led to current reactor designs, including the
pressurized water (PWR) and CANDU reactors. In the next section, we conclude this chapter
by introducing the neutron transport equation, which provides a means to evaluate numeri-
cally the neutron population in the reactor and to determine whether a given reactor concept
can lead to a sustained chain reaction. This equation also provides the neutron flux that is
required to follow the evolution of the fuel and the fission products in the reactor using the
Bateman equation.
4.4 Reactor Physics and the Transport Equation

When designing a nuclear power reactor, physicists have two main objectives:
 controlling the nuclear chain reaction;
 extracting the energy produced inside the fuel to avoid damage to the core and to use this
energy to produce electricity.
Because most of the energy produced in a nuclear reactor is the result of the fission process,
the power produced in the core can be expressed by:

  
P  t   d r dE  J  f , J (r , t , E )   (r , t , E , Ω)d 2 Ω,
3
(118)
V 0 J 4
where ܸ is the reactor volume, ߶(‫ݎ‬ Ԧ, ‫ݐ‬, ‫ܧ‬, Ω ሬ

ሬ⃗) is the flux of neutrons with energy ‫ ܧ‬and direction
ሬ
ሬ⃗ at a point ‫ݎ‬
Ω Ԧin space and a time ‫ݐ‬, and Σ௙,௃(‫ݎ‬ Ԧ, ‫ݐ‬, ‫ )ܧ‬is the macroscopic fission cross section
for isotope ‫ܬ‬given by
 
 f ,J  r , t, E   N J  r , t   f ,J  E  , (119)
where ܰ௃(‫ݎ‬ Ԧ, ‫ )ݐ‬is the time-dependent concentration of ‫ܬ‬, the solution of the flux-dependent
Bateman equations (see Section 4.2). Finally, ߢ௃ represents the average energy produced by
the fission of isotope ‫ܬ‬, which we assume to be energy-independent.
Now let us consider the flux, which is a measure of the neutron population in the reactor
(Section 3.5). For a given time interval ݀‫ݐ‬, the change of the neutron population
݀݊(‫ݎ‬Ԧ, ‫ݐ‬, ‫ܧ‬, ሬ
ሬ⃗) = ݀߶(‫ݎ‬
Ω Ԧ, ‫ݐ‬, ‫ܧ‬, ሬ
ሬ⃗)/‫ ݒ‬due to loss of neutrons by collisions is [Bell1982, Duder-
Ω
stadt1979, Hébert2009, Lewis1993]:
    
  
LC r , t, E ,Ω dt    r , t , E   r , t , E ,Ω dt ,  (120)
where Σ is the total cross section and ‫ܮ‬஼ (‫ݎ‬ ሬ

ሬ⃗) the rate of collision per unit volume of
Ԧ, ‫ݐ‬, ‫ܧ‬, Ω
neutrons with energy ‫ ܧ‬and direction ሬ ሬ⃗ at point ‫ݎ‬
Ω Ԧand time ‫ݐ‬. This term includes both ab-
sorbed and scattered neutrons. Some neutrons will be lost to migration when the net flow of
neutrons reaching a region in space of volume ݀ଷ‫ ݎ‬is different from that leaving this same
region. This rate of neutron loss due to migration is given by
     
( )
LM r,t, E, Ω dt = -Ω × Ñf r ,t, E, Ω dt, ( ) (121)
meaning that neutrons, which can be considered to form a gas, tend to move towards regions
ሬ
ሬ⃗
of low neutron population (the kinetic theory of gas). Neutron with energy ‫ ܧ‬and direction Ω
can also be produced in this region of space due to scattering of neutrons with different
directions and energies:
  
    
SS     
r , t , E , Ω dt  dE'  d 2 Ω'  s r , t , E '  E , Ω '  Ω  r , t , E ' , Ω' dt ,  (122)
0 4
where Σ௦(‫ݎ‬ ሬ
ሬ⃗′ → Ω
Ԧ, ‫ݐ‬, ‫ܧ‬ᇱ → ‫ܧ‬, Ω ሬ
ሬ⃗) is the probability (also known as the double differential
scattering cross section) that a neutron of energy ‫ܧ‬ᇱ moving in direction ሬ ሬ⃗′ is scattered in
Ω
direction ሬ
ሬ⃗ with energy ‫ܧ‬. For the case where fission takes place, the neutron production rate
Ω
is
  
  
 
S F r , t , E , Ω dt   j ( E ) dE'  j  f ,J ( r , t , E ' )  d 2 Ω ' ( r , t , E', Ω')dt , (123)
j 0 4
where ߯௝(‫ )ܧ‬is the fission spectrum and ߥ௝ the average number of neutrons produced by
fission of isotope ‫ܬ‬. Neutron productions from other sources ܵா ൫‫ݎ‬ Ԧ, ‫ݐ‬, ‫ܧ‬, ሬ
ሬ⃗൯can also be added
Ω
to the balance, for example, sources for other types of nuclear reactions (photonuclear fission
and photoneutron production in ଶଵD) or flux-independent contributions from the decay of
radioactive isotopes.
The global neutron population will then satisfy the following balance equation:
 
(
1 df r,t, E, Ω  )         
v dt
( ) ( ) ( )
= LC r ,t, E, Ω + LM r,t, E, Ω + SS r,t, E, Ω + S F r,t, E, Ω + S E r,t, E, Ω , (124) ( ) ( )
which is known as the Boltzmann transport equation. If one assumes that the cross sections
are nearly constant over time (the change in the composition of the materials in the reactor
varies substantially only over long time periods) and that the core is at static equilibrium (the
flux remains constant), one obtains the following time-independent neutron transport equa-
tion [Lewis1993]:
      
 
Ω   r , E , Ω    r , E   r , E , Ω   

 ' '   ' ' 1

 '  ' '
0 4
dE'
d Ω 
2
s
'
r , E E , Ω  Ω  r , E , Ω   j 
k j
  E   
dE '
 
j f ,J r , E   d Ω
2 '
 r , E ,Ω ,  
0 4
(125)
where ݇ is known as the multiplication factor. It is inserted in the transport equation to
modify the fission rate in such a way as to reach this static equilibrium. When ݇ = 1, equilib-
rium is achieved, and the reactor is critical. For ݇ < 1, the reactor is sub-critical, and neutron
generation is insufficient to compensate for losses, whereas for ݇ > 1, the reactor is super-
critical, and neutron production exceeds losses.
For reactor calculations, where fission reactions take place only inside the core, boundary
conditions of the form
   
f rB , E, Ω = 0 for Ω × N < 0 ( ) (126)
are generally imposed on the transport equation. Here, the point ‫ݎ‬ Ԧ஻ is on the outer boundary
ሬ⃗ being a unit vector normal to this boundary and directed outwards. This
of the reactor, with ܰ
condition simply means that no neutrons will enter the reactor from the outside.
The static transport equation is an eigenvalue equation, with ݇ being the inverse of the
eigenvalue. Therefore, the neutron flux is the eigenvector for this equation and is only defined
to within a normalization constant. Therefore, if ߶൫‫ݎ‬Ԧ, ‫ܧ‬, ሬ

ሬ⃗൯ is a solution, so is ܽ߶൫‫ݎ‬
Ω Ԧ, ‫ܧ‬, ሬ
ሬ⃗൯,
Ω
with ܽ an arbitrary constant. For reactor calculations, the normalization constant is directly
related to the power produced in the core.
Solving the transport problem is not trivial, even for very simple cases (static, one velocity
neutrons in one dimension). In Chapters 4 (static) and 5 (time-dependent), approximation to
and simplification of the transport equation commonly used in reactor simulations will be
discussed. One should then be able to extract, from the solutions of the resulting simplified
equations, information useful for the design, operation, and control of nuclear reactors.
5 Bibliography
[Basdevant2005] J.-L. Basdevant, R. James, S. Michel, Fundamentals in Nuclear Physics.
Springer, New York, 2005.
[Bell1982] G. I. Bell, S. Glasstone, Nuclear Reactor Theory. Robert E. Krieger, Malabar, 1982.
[Berger2013a] M. J. Berger, J. S. Coursey, M. A. Zucker, J. Chang, J., ESTAR, PSTAR, and ASTAR:
Computer Programs for Calculating Stopping-Power and Range Tables for Electrons,
Protons, and Helium Ions, http://physics.nist.gov/Star, visited April 2013.
[Berger2013b] M. J. Berger, J. H. Hubbel, S. M. Seltzer, J. Chang, J. S. Coursey, R. Sukumar, D. S.
Zucker, K. Olsen, XCOM: Photon Cross Section Database,
http://www.nist.gov/pml/data/xcom, visited April 2013.
[Bjorken1964] J. D. Bjorken, S. D. Drell, Relativistic Quantum Mechanics, McGraw-Hill, New
York, 1964.
[Brookhaven2013] National Nuclear Data Center at Brookhaven National Laboratory:
http://www.nndc.bnl.gov/chart/, visited April 2013.
[Griffiths2005] D. J. Griffiths, Introduction to Quantum Mechanics, Pearson Canada, Toronto,
2005.
[Duderstadt1979] J. J. Duderstadt, W. R. Martin, Transport Theory. Wiley, New York, 1979.
[Halzen1984] F. Halzen, A. Martin, Quarks and Leptons: An Introductory Course in Modern
Particle Physics. Wiley, New York, 1984.
[Hébert2009] A. Hébert, Applied Reactor Physics. Presses Internationales Polytechnique,
Montréal, QC, 2009.
[Hubbell1975] J. H. Hubbell, W. J. Veigele, E. A. Briggs, R. T. Brown, D. T. Cromer, and R. J.
Howerton, “Atomic form factors, incoherent scattering functions, and photon scatter-
ing cross sections”, J. Phys. Chem. Ref. Data, 4, 471-538, 1975.
[Jauch1976] J. M. Jauch, F. Rohrlich, The Theory of Photons and Electrons. Springer-Verlag,
New York, 1976.
[KAERI2013] Nuclear Data Center at Korea Atomic Energy Research Institute:
http://atom.kaeri.re.kr/ton/, visited April 2013.
[Leroy2009] C. Leroy, P.-G. Rancoita, Principle of Radiation Interaction in Matter and Detection,
2nd edition. World Scientific, Singapore, 2009.
[Le Sech2010] C. Le Sech, C. Ngô, Physique nucléaire : Des quarks aux applications. Dunod,
Paris, 2010.
[Lewis1993] E. E. Lewis and W. F. Miller, Jr., Computational Methods of Neutron Transport, 2nd
edition. Wiley, New York, 1993.
[Livermore2013] Isotope Evaluation at Lawrence Livermore National Laboratory:
http://ie.lbl.gov/toi2003/MassSearch.asp, visited April 2013.
[MacFarlane2010] R. E. MacFarlane, A. C. Kahler, “Methods for Processing ENDF/B-VII with
NJOY”, Nuclear Data Sheets, 111, 2739-2890, 2010.
[Magill2005] J. Magill and J. Galy, Radioactivity, Radionuclides, Radiation. Springer: Heidelberg,
2005.
[Nikjoo2012] H. Nikjoo, S. Uehara, D. Emfietzoglou, Interaction of Radiation with Matter. CRC
Press, Boca Raton, FL, 2012.
[OECD-NEA2013] OECD-NEA Janis 3.0, Java-Based Nuclear Data Display Program.
https://www.oecd-nea.org/janis, visited April 2013.
[Pomraning1991] G. C. Pomraning, Linear Kinetic Theory and Particle Transport in Stochastic
Mixtures. World Scientific, London, 1991.
[Schiff1968] L. I. Schiff, Quantum Mechanics. McGraw-Hill, New York, 1968.
[Smith2000] F. A. Smith, A Primer of Applied Radiation Physics. World Scientific, Singapore,
2000.
[Watt1952] B. E. Watt, “Energy spectrum of neutrons from thermal fission of ଶଷହU”, Phys. Rev.
87, 1037-1041, 1952.
[WebElements2013] WebElements: The Periodic Table on the Web.
http://www.webelements.com/isotopes.html, visited April 2013.
[Wong2004] S. S. M. Wong, Introductory Nuclear Physics. Wiley-VCH, Weinheim, 2004.
6 Summary of Relationship to Other Chapters

This chapter explains the structure of the nucleus and introduces the various interactions of
charged particles, photons, and neutrons with matter. It also defines the basic quantities and
concepts needed to study the fission chain reaction and to approach the analysis of reactors.
The Bateman equations are presented and used in other chapters to evaluate the rate of
evolution of fuel composition. The Boltzmann transport equation, which provides a way to
evaluate the neutron population in a nuclear reactor, is also introduced. This is the fundamen-
tal equation from which the neutron diffusion equation used in Chapters 4 and 5 is derived.
Chapter 4 starts with the static transport equation to derive the time-independent neutron-
diffusion equation for the analysis of steady-state (time-independent) neutron distributions
either in the presence of external sources or in fission reactors. It also discusses the evolution
of the fuel composition in an operating reactor based on the coupled solution of the transport
and Bateman equations using a quasi-static approximation.
Chapter 5 covers the time-dependent neutron-diffusion equations and studies the phenomena
of fast-neutron kinetics, fission-product poisoning, reactivity coefficients, etc., in reactors.
Chapter 12 deals with radioprotection and uses the concept defined in the present chapter for
the interactions of charged particles, photons, and neutrons with matter with the goal of
reducing the damage to living tissues.
Chapter 19 examines the behaviour of fuel after irradiation in a nuclear reactor, particularly its
composition and the residual energy that it produces.
7 Problems
1. Evaluate the intensity of the electromagnetic force acting between two protons in
an atom of helium.
2. Determine the atomic mass of natural chlorine if it is composed of 75.77 %atomic
ଷହ ଷ଻
ଵ଻Cl and 24.23 %atomic ଵ଻Cl.
3. What is the atomic concentration of ଶଷହ ଽଶU in a sample of uranium oxide (UOଶ) of
ଷ
density ߩ=10.4 g /cm enriched at a level of 3.5 %weight?
4. Using the semi-empirical mass formula, evaluate the atomic mass of ହ଺Fe and
compare with the exact value.
5. What is the ground state energy level of a neutron in the Saxon-Wood potential for
ଶଷ଼
ଽଶU?
6. Complete the following nuclear reactions and determine the net energy in MeV re-
leased by these reactions:
a) ଶଵ‫ ܪ‬+ ଷଵ‫ = ܪ‬஺௓ܺ + ଵ଴݊
ଽ଴
b) ଵ଴݊+ ଶସଵ ஺ ଵ
ଽ଺‫ = ݉ܥ‬௓ܺ + ଷ଺‫ ݎܭ‬+ 3 ଴݊
7. What is the activity of one gram of potassium-40?
8. Assuming that the activity of a sample of water due to tritium is 100 Bq, determine
the mass of tritium in this sample. What will be the activity of this sample after
three years?
9. Using the information available on the ESTAR web site, plot the CSDA range in cm
for an electron in dry air at sea level as a function of its energy for 100 keV <E < 10
MeV.
10. Draw the deuterium cross section for photonuclear reactions as a function of en-
ergy in the range 2.23 MeV < E < 10MeV (here assume ‫ ܥ‬మ஽ = 0.062). Compare
with the deuterium cross section for neutron radiative capture at 300 K in the
same energy range.
11. How many collisions are required on average to slow down a neutron to half its
speed (1/4 of its energy) through elastic collisions with deuterium nuclei?
12. Evaluate the potential scattering cross section for oxygen and hydrogen at low en-
ergy.
13. A CANDU bundle contains 19.5 kg of natural uranium (enrichment of 0.711
%weight). Compute the fission rate in a bundle due to ଶଷହ ଽଶU if it is exposed to a flux
of 10ଵସ neutrons/cm2/s of energy ‫=ܧ‬0.01 eV.
14. Show that the mean free path of 1 MeV neutrons in thorium is 3.5 m.
8 Acknowledgments
Reviewers Marv Gold and Esam Hussein are gratefully acknowledged for their hard work and
excellent comments during the development of this Chapter. Their feedback has much im-
proved it. Of course the responsibility for any errors or omissions lies entirely with the author.
Thanks are also extended to Diana Bouchard for expertly editing and assembling the final
copy.
1
CHAPTER 4
Reactor Statics
Prepared by
Dr. Benjamin Rouben, 12 & 1 Consulting, Adjunct Professor, McMaster University & University
of Ontario Institute of Technology (UOIT)
and
Dr. Eleodor Nichita, Associate Professor, UOIT
Summary:
This chapter is devoted to the calculation of the neutron flux in a nuclear reactor under special
steady-state conditions in which all parameters, including neutron flux, are constant in time.
The main calculation method explored in this chapter is the neutron-diffusion equation. Analyti-
cal solutions are derived for simple neutron-diffusion problems in one neutron energy group in
systems of simple geometry. Two-group diffusion theory and the approximate representation of
the diffusion equation using finite differences applied to a discrete spatial mesh are introduced.
The rudimentary reactor-physics design of CANDU reactors is presented. The two-step approach
to neutronics calculations is presented: multi-group lattice transport calculations, followed by
full-core, few-group diffusion calculations. Finally, the chapter covers fuel-property evolution
with fuel burnup and specific features of CANDU neutronics resulting from on-line refuelling.
Table of Contents
1 Introduction ............................................................................................................................ 3
1.1 Overview ............................................................................................................................. 3
1.2 Learning Outcomes ............................................................................................................. 3
2 Neutron Diffusion Theory ....................................................................................................... 4
2.1 Time-Independent Neutron Transport ............................................................................... 4
2.2 Fick’s Law and Time-Independent Neutron Diffusion......................................................... 7
2.3 Diffusion Boundary Condition with Vacuum at a Plane Boundary ..................................... 9
2.4 Energy Discretization: The Multi-Group Diffusion Equation............................................. 12
3 One-Group Diffusion in a Uniform Non-Multiplying Medium.............................................. 12
3.1 Plane Source...................................................................................................................... 13
3.2 Point Source ...................................................................................................................... 14
3.3 Flux Curvature in Source-Sink Problems and Neutron In-Leakage ................................... 16
4 One-Group Diffusion in a Uniform Multiplying Medium with No External Source .............. 16
4.1 Uniform Infinite Reactor in One Energy Group................................................................. 17
4.2 Uniform Finite Reactors in One Energy Group.................................................................. 18
4.3 Uniform Finite Reactors in Various Geometries ............................................................... 21
4.4 Flux Curvature and Neutron Out-Leakage ........................................................................ 30
5 Reactors in Two Neutron-Energy Groups with No External Source...................................... 31
5.1 The Neutron Cycle and the Four-Factor Formula ............................................................. 31
5.2 The Two-Energy-Group Model.......................................................................................... 34
5.3 Neutron Diffusion Equation in Two Energy Groups .......................................................... 34
5.4 Uniform Infinite Medium in Two Energy Groups.............................................................. 34
5.5 Uniform Finite Reactors in Two Energy Groups ................................................................ 36
©UNENE, all rights reserved. For educational use only, no assumed liability. Reactor Statics – September 2014
6 Solution for Neutron Flux in a Non-Uniform Reactor ........................................................... 38

6.1 Finite-Difference Form of the Neutron-Diffusion Equation.............................................. 38
6.2 Iterative Solution of the Neutron-Diffusion Equation....................................................... 42
7 Energy Dependence of Neutron Flux.................................................................................... 43
7.1 The Fission-Neutron Energy Range................................................................................... 43
7.2 The Thermal Energy Range ............................................................................................... 44
7.3 The Slowing-Down Energy Range ..................................................................................... 44
7.4 Neutron Flux over the Full Energy Range ......................................................................... 46
8 Basic Design of Standard CANDU Reactors........................................................................... 47
9 CANDU Reactor Physics Computational Scheme.................................................................. 50
9.1 Lattice Calculation............................................................................................................. 51
9.2 Reactivity-Device Calculation............................................................................................ 52
9.3 Full-Core Calculation ......................................................................................................... 52
10 Evolution of Lattice Properties.............................................................................................. 53
11 Summary of Relationship to Other Chapters........................................................................ 56
12 Problems ............................................................................................................................... 58
13 Appendix A: Reactor Statics .................................................................................................. 62
List of Figures
Figure 1 Negative flux curvature in homogeneous reactor .......................................................... 19
Figure 2 Infinite-slab reactor......................................................................................................... 22
Figure 3 Flux with physical and unphysical values of buckling ..................................................... 23
Figure 4 Infinite-cylinder reactor .................................................................................................. 24
Figure 5 Ordinary Bessel functions of first and second kind ........................................................ 25
Figure 6 Rectangular-parallelepiped reactor ................................................................................ 26
Figure 7 Finite-cylinder reactor..................................................................................................... 28
Figure 8 Energies of fission neutrons............................................................................................ 31
Figure 9 Sketch of cross section versus neutron energy............................................................... 32
Figure 10 Neutron cycle in thermal reactor.................................................................................. 33
Figure 11 Face view of very simple reactor model ....................................................................... 38
Figure 12 Flux in a cell and in immediate neighbours .................................................................. 39
Figure 13 Linear treatment of flux in central cell and one neighbour.......................................... 40
Figure 14 Flux in cell at reactor boundary .................................................................................... 42
Figure 15 Energy distribution of fission neutrons......................................................................... 44
Figure 16 Variation of flux with energy if absorption is smooth .................................................. 46
Figure 17 Variation of flux with energy in the presence of resonances ....................................... 46
Figure 18 Sketch of flux variation over full energy range ............................................................. 47
Figure 19 CANDU basic lattice cell ................................................................................................ 51
Figure 20 Supercell for calculating device incremental cross sections......................................... 52
Figure 21 Simple model with superimposed reactivity device..................................................... 53
Figure 22 Evolution of fuel isotopic densities............................................................................... 54
Figure 23 Infinite-lattice reactivity vs. irradiation......................................................................... 55
Figure 24 Finite-reactor reactivity vs. irradiation.......................................................................... 55
Figure 25 Averaging reactivity with daily refuelling...................................................................... 56
Reactor Statics 3
1 Introduction
1.1 Overview
This chapter is devoted to the calculation of the neutron flux1 in a nuclear reactor under special
steady-state conditions in which all parameters, including neutron flux, are constant in time.
The position-, energy-, and angle-dependent steady-state neutron-transport equation is derived
by writing the detailed neutron-balance equation. This is done in Chapter 3 of the book.
However, it is also done here for completeness. The steady-state diffusion equation is subse-
quently derived using a linear approximation of the angular dependence of the neutron flux.
Multi-group neutron-energy discretization is also introduced.
Analytical solutions are derived for simple neutron diffusion problems. First, the one-group
diffusion equation is solved for a uniform non-multiplying medium in simple geometries.
Subsequently, the one-group diffusion equation is solved for a uniform multiplying medium (a
homogeneous “nuclear reactor”) in simple geometries. The importance of neutron leakage and
the concepts of criticality and the neutron cycle are introduced.
Of course, real reactors are almost never homogeneous, and rarely can neutron energies be
accurately represented by a single energy group. However, two energy groups are often suffi-
cient to represent neutron diffusion in a thermal reactor because most of the fissions are
induced by thermal neutrons. This chapter therefore proceeds to introduce two-group diffusion
theory and the approximate representation of the diffusion equation using finite differences
applied to a discrete spatial mesh. The latter enables the treatment of non-uniform reactor
cores.
Following this treatment of the basic theory of neutron transport and diffusion, the rudimen-
tary reactor-physics design of CANDU reactors is presented, and the associated neutron energy
spectrum is discussed. Subsequently, more general core modelling concepts are presented,
including the two-step approach to neutronics calculations: multi-group lattice transport
calculations, followed by full-core, few-group diffusion calculations. The final section in the
chapter concentrates on fuel-property evolution with fuel burnup and specific features of
CANDU neutronics resulting from on-line refuelling.

The goal of this chapter is for the student to understand:
 The neutron transport resulting from the detailed neutron balance;
 The neutron-diffusion equation resulting from the linear approximation of the angu-
lar dependence of the neutron flux;
 Analytical solutions of the diffusion equation in simple geometries for both multiply-
ing and non-multiplying media;
1 In this document, the term “neutron flux” is used to denote the quantity  = n, where n is neutron density and
 is neutron speed. The units of flux are 1cm-2s-1 or 1m-2s-1. Some newer texts use the term flux density or
fluence rate. All three terms refer to the same physical quantity.
 The concepts of neutron cycle, criticality, and buckling;

 Energy discretization of the diffusion equation using neutron energy groups and spa-
tial discretization using finite differences;
 The two steps of neutronics calculations: lattice-cell transport calculations and full-
core diffusion calculations; and
 Basic design elements of CANDU reactors and their implications for the physical at-
tributes of such reactors.
2 Neutron Diffusion Theory

2.1 Time-Independent Neutron Transport
The time-independent neutron-transport equation is an equation in six variables: three vari-

ables r for the position of the neutron in the reactor, two variables ̂ for the direction of
neutron motion, and one variable E for neutron energy (we could also use neutron speed v
instead of E, because there is a one-to-one relationship between v and E).
Because we are dealing with time independence (i.e., reactor statics), the neutron-transport
equation expresses the fact that the number of neutrons at any position, moving in any direc-
tion, and with any energy, is unchanging over time. In other words, if we consider a differential
  ˆ
volume in the six-dimensional variable space, i.e., dr d̂dE about r ,  
, E , the rate of neutrons
“entering” the differential volume must be equal to the rate of neutrons “exiting” the differen-
tial volume.
To write the equation, we must therefore consider all phenomena by which neutrons enter or
exit the differential volume or are created or destroyed within the volume. The phenomena
which we must consider are:
 Neutron production by nuclear fission;
 Neutron scattering, which can change the neutron’s energy and/or its direction of
motion;
 Neutron absorption;
 Neutron spatial leakage into or out of the differential volume.
[Note: Another process for neutron production is the (n, 2n) nuclear reaction. However, this
reaction produces many fewer neutrons than fission and is neglected here.]
The rates at which neutrons enter or exit the differential volume by virtue of these various
phenomena, expressed per unit differential volume per second, are given one by one in the
following equations.
The rate of birth of neutrons by fission is given by:
 E   ˆ
 
4 E ' ˆ '

 f  r , E '   r ,  
ˆ ' dE ' ,
', E ' d  (1)
where
 
   f r , E ' is the neutron-yield cross section at position r for fissions induced by
neutrons of energy E’;
Reactor Statics 5
 ˆ
 
 r, 
' , E ' is the angular flux of neutrons at position, of energy E’, and moving in di-

rection ' ; and
 (E) is the fraction of all fission neutrons born with energy E. Note: In this chapter,
for simplicity, a single fission spectrum (E) is used for fissions induced in all nu-
clides, whereas in reality the fission spectrum should be taken as different for differ-
ent nuclides. This fully correct treatment is what is done in Chapter 3.
Note: In the above equation, delayed neutrons are not referred to separately. The neutron
production rate includes both prompt and delayed neutrons. This is acceptable because it can
be shown that in (true) steady state, accounting separately for delayed-neutron production
reduces in any case to the above expression.
The rate of production of neutrons from an “external” source (not related to fissions in the
reactor fuel) is given by:

S r , E , (2)
 
where S r , E  is the external source strength for neutrons of energy E at position r .
Note: We will assume that the source is isotropic.
The rate of neutron gain from neutrons entering the differential volume by scattering from
other neutron directions of motion or other neutron energies is given by:
 

   s r , E '  E , ˆ '  ˆ  r , ˆ ', E ' d ˆ ' dE ',

  (3) 
E ' '

 ˆ ' 
where  s r , E '  E ,  ˆ  is the cross section for scattering neutrons from energy E’ to
energy E and from direction ' to direction ̂ .
The rate of neutron loss from absorption and from neutrons exiting the differential volume by
scattering is given by:
  ˆ

t  r , E  r ,  ,E ,  (4)
 
where  t r , E  , the total cross section at position r for neutrons of energy E,
 

  a  r , E     s r , E  E ', 
ˆ 
ˆ ' d
ˆ ' dE ', 

where  a r , E  is the absorption cross section.
The net rate of neutron spatial leakage (i.e., diffusion) out of the differential volume is given by:
  
 ˆ ,E ,
  J r,  (5)

ˆ , E is the angular current at position r of neutrons moving in direction ̂ with
where J r ,  
energy E.

ˆ , E   r , 
Note: Because J r ,    
ˆ ,E  
ˆ , the neutron leakage [Eq. (5)] can also be written as:

ˆ    r , 
  ˆ ,E .  (6)
Taking into account all the above rates, the neutron balance can then be expressed in the time-
independent neutron-transport equation as follows:

   E   ˆ ˆ ' dE '  r, E '  E,  ˆ  r, 
1
4
S r, E  
4 E ' ˆ ' 
 f  r , E '   r ,  
', E ' d    s  ˆ '  
ˆ ', E ' d
ˆ ' dE ' 
ˆ
E ' ' (7)
  ˆ   
 
 t  r , E   r , , E   J r ,   
ˆ , E  0.
Note again that for simplicity a single fission spectrum (E) has been used here. See Chapter 3
for the fully correct treatment. The second term of the equation can be simplified because the
fission cross section does not depend on ̂' , so that the integral over ̂' in the second term
reduces to the angle-integrated flux, and Eq. (7) becomes:
  E    ˆ  r, 
1
4
S r, E 
4 E '  ˆ

 f  r , E '    r , E '  dE '   s r , E '  E, 
ˆ '   
ˆ ', E ' d  
ˆ ' dE '
E ' '
  ˆ   
  
 t  r , E   r , , E   J r , , E  0, ˆ  (7)’
  ˆ

where  r , E '    r ,  
ˆ ' is the total (angle-integrated) flux of neutrons with energy E
' , E ' d
ˆ'
 
at position r .
The neutron-transport equation (7) or (7)’ is an exact statement of the general steady-state
neutron-balance problem. However, it can immediately be seen that it is very complex: in
addition to its dependence on six independent variables, it is an integrodifferential equation.
Because of its complexity, this equation cannot be solved analytically except for problems in the
very simplest geometries, and real problems require numerical solution by computer.
Note that when there is no external source S, the equation appears to be a linear homogeneous
equation, which does not generally have a solution (except a trivial zero solution for the flux) for
arbitrary values of the nuclear properties. In this case, a solution can be found by modifying the
nuclear properties. Mathematically, this can be done by modifying the yield cross section f by
dividing it by a quantity, keff, which can be selected to ensure a non-trivial solution. This quan-
tity keff is called the multiplication constant.
Two general categories of codes exist to solve the neutron-transport equation: deterministic
codes (which solve the equation directly by numerical means) and Monte Carlo codes (where
stochastic methods are used to model a very large number—typically millions or even hundreds
of millions—of neutron births and their travel and event histories, from which the multiplication
constant and flux and power distributions can be evaluated using appropriate statistics of these
histories). Although the application of either type of transport computer code to full-core
reactor models requires very significant computer resources and execution time to achieve a
high degree of accuracy, both methods, especially Monte Carlo codes, have seen much greater
application in whole-reactor analysis in the last decade or so. While core-wide pin-power
reconstruction and time-dependent kinetics calculations are still beyond reach, static eigenvalue
calculations and global flux and power distributions can now be carried out routinely using
Monte Carlo codes. However, detailed discussion of either type of transport code will not be
covered in the present work.
The traditional way of attacking neutronics problems in reactors has been to solve the transport
equation numerically in relatively small regions (such as a basic lattice cell or a small collection
Reactor Statics 7
of cells) to compute region-averaged properties, and then to use these to solve the full-core
reactor problem with a simplified version of the transport equation, the neutron-diffusion
equation. This computational scheme is discussed in greater detail in Section 9.
2.2 Fick’s Law and Time-Independent Neutron Diffusion

To derive a simplified version of the neutron-transport equation, we note that, because the
fission and total cross sections do not depend on the neutron direction of motion ̂ (i.e.,
“nuclei do not care from which direction neutrons interact with them”), and also because the
most important quantity in reactor physics is the fission rate (which determines power produc-
tion), it may be very effective to try to obtain an equation which is independent of ̂ and which

involves only the angle-integrated flux  r, E  [also called the integral flux].
To achieve this, we can attempt to remove the neutron direction of motion by simply integrating
Eq. (7) over ̂ . Let us see what this gives, term by term. The first two terms in (7) do not
contain ̂ , and because the integral of ̂ over a sphere is 4, their integrals, which we can call
T1 and T2, can be written directly as:

T1  S  r , E  (8)
 
T2    E    f  r , E '    r , E '  dE '. (9)
E'
Similarly, integrating the fourth term gives a result that depends on the integral flux only:
  ˆ ˆ    r, E    r, E  .
T4   t  r , E   r , 
ˆ

, E d t  (10)

Integrating the third term gives a more complex relation:

 ˆ  r, 
 ˆ '
T3      s r , E '  E , 
ˆ E' 
ˆ'
ˆ ', E ' d  
ˆ ' dE ' d 
ˆ.  (11)

We can, however, simplify T3 by noting that, because no absolute direction in space is “special”,
the scattering cross section s cannot depend on the absolute directions ̂ and ̂ ' , but only on
the scattering angle between the directions of the incoming and scattered neutrons, or even
more specifically on the cosine  of that angle, i.e., on   
ˆ '
ˆ . Using this fact, we can then
simplify the integral in Eq. (11) by integrating over ̂ first (actually over , because s does not
depend on the “azimuthal” angle of scattering, and integrating over that azimuthal angle simply
gives 2). The result is then a product of two integrals:
 ˆ ' 2  r, E '  E ,   d dE '
 ˆ ' d
T3     r , E '  E , 
ˆ'
 s 
E'  
  (12)
    r , E '   s  r , E '  E  dE ',
E'
where we have defined

 
 s  r , E '  E   2   s  r , E '  E ,   d  . (13)


Equation (12) is a useful result because it depends on the integral flux  r , E' only.
For the fifth term, we get:
  
 ˆ   
ˆ d
T5     J r , E ,   ˆ   r , E ,  ˆ 
ˆ d 
 (14)
ˆ
  ˆ 
  
   J r, E ,
where
   

J r , E    J r , E, 
ˆ d
ˆ  (15)

is the current of neutrons of energy E at position r .
Unfortunately, this result is not at all of the form we would like, i.e., it is not at all expressible in
terms of the integral flux, because it is clear that integrating a function of the vector quantity

 
J r , E , ̂ over ̂ will obviously give in general a result totally unrelated to the integral flux
 r, E  !
In summary, integrating Eq. (7) over ̂ gives:
    
S  r , E     E    f  r , E '    r , E '  dE '   s  r , E '  E    r , E '  dE '
E' E'
    
 t  r , E    r , E     J  r , E   0 , (16)
but the last term on the left-hand side (the leakage term) has thwarted our efforts to achieve an
equation in the integral flux only.
However, one approximation which is often used in diffusion problems (diffusion of one mate-
rial through another) can help us here. This approximation, called Fick’s Law, applies in low-
absorption media if the angular flux varies at most linearly with angle and if the neutron source
is isotropic [already assumed in writing Eq. (7)]. Under these conditions it states that the
  
current J r , E  is in the direction in which the integral flux  r, E  decreases most rapidly, i.e.,

it is proportional to the negative gradient of  r, E  :
   
J  r , E    D  r , E    r , E  , (17)

where the quantity Dr , E  is called the diffusion coefficient and can be written as 1/(3tr),
with tr being the neutron transport cross section.
The reader is referred to Appendix A (Reactor Statics) of this chapter for the derivation of Fick’s
Law. Here, we will continue to derive the final form of the neutron-diffusion equation.
Equation (17), when substituted into Eq. (16), finally gives an equation in the integral flux only,
the time-independent neutron-diffusion equation:
    
S  r , E     E    f  r , E '    r , E '  dE '   s  r , E '  E    r , E '  dE ' 
. (18)

E' E'
   
 t  r , E    r , E     D  r , E    r , E   0
Reactor Statics 9
The time-independent neutron-diffusion equation is much simpler than the neutron-transport

equation and can be used to solve for the flux in specially prepared full-core reactor models.
Note that the out-leakage term (the last term on the left-hand side) has a positive sign, even
though this term nominally represents a loss of neutrons; the + sign arises as a result of the –
sign in relationship (17) between the current vector and the flux gradient.
Before the diffusion equation can be used, it is important to understand that Fick’s Law is only
an approximation; it is valid only in low-neutron-absorption media, when the integral flux does
not vary too quickly and when angular flux varies weakly with angle (at most linearly). There-
fore, Fick’s Law is not an especially good approximation in the vicinity of strong absorbers,
where the spatial flux variation is very large, or, for the same reason, near interfaces between
regions with large variations in nuclear properties or near external surfaces.
2.3 Diffusion Boundary Condition with Vacuum at a Plane Boundary

To solve the diffusion equation, which is a second-order partial differential equation, through-
out the reactor volume, we need to define boundary conditions at the surface of the reactor.
The vacuum boundary conditions in transport theory are quite clear: the angular current (or
angular flux) at the boundary must be zero for any direction pointing to the inside of the reactor
(assuming that the reactor has no re-entrant surface):
 ˆ

 r, ,E  0  (19)
 ˆ  eˆ  0, where ê is the unit outgoing normal
for r at the boundary and any ̂ such that   

at r .
This condition cannot be applied in diffusion theory, which depends on the integral flux only,
because we have lost all the directional information of the angular flux. In diffusion theory, we
need boundary conditions, at most, on the integral flux and its gradient. The conditions can be
generalized to demand that the total rate of incoming neutrons be zero. We will now proceed
to derive the boundary conditions used in diffusion theory.
We first derive a general relationship between the angular flux and the current using the
approximation that the angular flux is linear in angle. Linearity in angle means that the angular
flux can be written as a linear function of the x-, y-, and z-components of the angle ̂ :
 
ˆ ab  b  b ,
  x x y y z z
(20)
where a, bx, by, bz are constants which can be determined in terms of the integral flux and
current.
Let us first consider the integral flux     ˆ dˆ .

From Eq. (20),
ˆ  b  d
  a d x x
ˆ  b  d
y y
ˆ  b  d
z z
ˆ. (21)
   
The first integral in Eq. (21) has value 4, whereas the others have value 0 (it is clear that in
integrating a single component of the angle over all solid angles, the + and – components cancel
out). This means that  = 4a, which implies that

a . (22)
4
Now let us consider the current:

J     ˆ  ˆ d ˆ .

(23)
From Eq. (20),


ˆ d
J  a  ˆ b  
x x
ˆ d
ˆ b  
y y
ˆ d
ˆ b  
z z
ˆ d
ˆ. (24)
   
It is clear that the first integral in Eq. (24) is zero, for the same reason that the + and – compo-
nents cancel out. Consider the other integrals in Eq. (24), for instance,   x  ˆ d
ˆ . This is a

vector integral, and only the x-component will survive because   x  y or  z d 
ˆ  0 , for the

same reason, cancellation.

ˆ , which can easily be shown to be equal to
On the other hand, the x-component gives 

 2x d 
ˆ  2 1  2 d  2  3 1  4 ].
4/3 [by symmetry, it must be equal to 
 2z d 1 3 1 3
Therefore, we can conclude that
 4
J
3

bxiˆ  by ˆj  bz kˆ ,  (25)
i.e.,
3
bx  Jx (26)
4
(and similarly for y and z).
Now, substituting Eqs. (22) and (26) into Eq. (20),

ˆ    3 J   J   J      3 J 
 
  ˆ. (27)
4 4 4 4
x x y y z z
Let us now apply Eq. (27) at a plane boundary with vacuum. Assume that the boundary plane is
perpendicular to the z-axis and the polar axis is the outward normal to the boundary, i.e., the
unit vector k̂ in the positive z-direction. The total rate of incoming neutrons is:

  
J   
ˆ 
ˆ  kd ˆ, (28)
where the integral over the “half-space” - means integrating over the entire azimuthal angle 
(i.e., over 2), but only over half the polar angle , i.e., over  from /2 to  (or over   cos 
from -1 to 0).
Using Eq. (27), the total rate of incoming neutrons is:
  3 
   
ˆ 
  ˆ 
ˆ  kd ˆ   
 ˆ ˆ
  4  4 J ˆ  ˆ ˆ  kd . (29)
ˆ  k̂     , then, in the integral of the second term, only the part involving z will
Because  z
Reactor Statics 11
survive (for the same reason stated before), and:


 3   J  kˆ
Total rate of incoming neutrons =
0
* 2 *  d  ˆ 0
* 2 * J  k *   d   
2
.
4 1 4 1 4 2
Applying Fick’s Law and setting the total rate of incoming neutrons to 0 gives:
 D ˆ D d 
    k  0, i.e.,   0 (30)
4 2 2 dz 4
1 d 1
  ,
 dz boundary 2D
which can be written as
1
 ,
d extr
which leads to
2
d extr  2 D  , (31)
3 tr
2
d extr  2 D  , (32)
3 tr
where tr is the transport cross section.
Actually, a more advanced analysis gives a slightly higher, more correct value for dextr:
2.1312 0.7104
d extr  2.1312 D   . (33)
3 tr  tr
Equations (30)-(33) result in a zero net incoming current at the physical boundary.
The geometric interpretation of Eq. (30) is that the relative neutron flux near the plane bound-
ary has a slope of -1/dextr, i.e., the flux would extrapolate linearly to 0 at a distance dextr beyond
the boundary. This is often stated as “the flux goes to 0 at an extrapolation distance dextr
beyond the boundary”. Such an interpretation is not literally correct: the flux cannot go to zero
in a vacuum, because there are no absorbers to remove the neutrons; the flux only appears to
be heading to the zero value at the extrapolation point. In fact, instead of using the “slope”
version (30), the boundary condition is often applied by “extending” the reactor model to the
extrapolation point and demanding a zero flux at that point.
Now, typical values for the diffusion coefficients D in CANDU reactors are  1 cm, and therefore
the extrapolation distance dextr is of the order of 2–3 cm. For systems of large dimension, e.g.,
large power reactors which are several metres in size, the extrapolation distance is sometimes
neglected if a slightly approximate answer is sufficient.
[Note: Equation (33) applies in principle to plane boundaries only. Slightly different formulas for
the extrapolation distance would apply to curved boundaries; however, the difference is small
unless the radius of curvature of the boundary is of the same order of magnitude as dextr.]
2.4 Energy Discretization: The Multi-Group Diffusion Equation

Neutron energy E is of course a continuous variable. However, the diffusion equation (18) with
continuous E is not usually solved as is. Instead, the equation is rewritten in “multi-group” form
by subdividing the energy range into a number G of intervals, called energy “groups” and
labelled with g from 1 to G (g =1 corresponding to the interval with the highest energies and g =

G to the interval with the lowest energies). In each group g, the continuous flux  r, E  is

replaced by an average group flux,  g r  .
The nuclear cross sections in each group are assumed to have been appropriately averaged over
the corresponding energy intervals. In multi-group notation, the cross sections and variables
are written as follows:
 
S  r , E   Sg  r 
  E   g
 
 f  r , E     f , g  r 
 
 s  r , E '  E    s , g ' g  r 
 
 t  r , E   t ,g  r 
 
D  r , E   Dg  r  .
The multi-group diffusion equation then takes the form:

 G
 
Sg  r    g  f ,g '  r  g '  r  
g ' 1
G
      

g ' 1
s , g ' g  r  g '  r   t ,g  r  g  r     Dg  r  g  r   0, g  1,..., G.
3 One-Group Diffusion in a Uniform Non-Multiplying Medium

We will start the analysis of systems with simple systems and increase complexity gradually.
Assume that all neutrons are lumped into a single energy group, i.e., G = 1. There is then no
need for the subscript g. With a single energy group, the notion of scattering across groups has
no meaning, and the “incoming scattering” term in s cancels the “outgoing scattering” term
inherent in the total cross section t (another way of saying this is that the “out-scattering” and
“in-scattering” terms cancel one another if there is only one group). This means that we can
drop the second term in the diffusion equation and write a instead of t in the third term.
In this section, we will deal with neutron diffusion in non-multiplying media, i.e., in media
where the fission cross section is zero and the neutron flux is driven by an external neutron
source. This type of problem is sometimes called a “source-sink” problem. We will assume that
the medium is uniform outside the source, and also that it is infinite in size. Uniform properties
also mean that the diffusion coefficient can be taken outside the divergence operator. With
these assumptions, the diffusion equation becomes:
    
S  r    a  r   D    r   0,
   (34)
or S  r   a  r   D2  r   0.
Reactor Statics 13
3.1 Plane Source

We take the source as an infinite plane source, which we can place, without loss of generality, in
the y-z plane. In this case, the flux is a function of x only, (x), and the Laplacian can be written
as:
d2
2  .
dx 2
The diffusion equation outside the source, i.e., for any x different from 0, becomes:
d 2  x 
D  a  x   0. (35)
dx 2
We can simplify this equation by dividing by D. If we define
D
L2  , (36)
a
called the diffusion area (and L called the diffusion length), the equation becomes:
d 2  x  1
 2   x   0. (37)
dx 2 L
For x > 0, Eq. (37) has mathematical solutions exp(x/L) and exp(-x/L), which give a general
solution:
  x   Ae x / L  Ce  x / L .
However, the term ex/L goes to  as x   and therefore cannot be part of a physically accept-
able solution for x > 0. The solution for x > 0 must then be  x   Ce  x / L .
By left-right symmetry, we can see that the full solution for any x must be
  x   Ce  x / L , (38)
with C being a constant which we can determine from the boundary condition at x = 0.
If S is the source strength per unit area of the plane, then the number of neutrons crossing
outwards per unit area in the positive x-direction must tend to S /2 as x  0. Therefore,
S
lim  J  x   
x 0 2
 d  x   S
 lim   D  2
x 0
 dx 
 CD  x / L  S
 lim  e 
x 0
 L  2
CD S SL
  , i.e., C  .
L 2 2D
The neutron flux outside the source is then finally:
SL  x / L
  x  e . (39)
2D
It is interesting to try to interpret the “physical” meaning of the diffusion area L2. Let us calcu-
late the mean square distance that a neutron travels in the x-direction from the source (at x =
0) to its absorption point. We can do this by averaging x2 with the absorption rate a as a
weighting function:

 x  dx
2
a
x2  0

  dx a
0 .
With the form (39) for the flux, we can evaluate the integrals and show that
 x 2  2L2 ,
i.e., we can interpret L2 as one-half the square of the average distance (in one dimension)
between the neutron’s birth point and its absorption point.
3.2 Point Source

Let us now take the source as a single point source, assumed isotropic. Without loss of general-

ity, we can place the source at the origin of co-ordinates, r  0.
To solve for the flux in this geometry, we need to write the Laplacian  2 in spherical co-
ordinates. In view of the spherical symmetry of the problem, there is no dependence on angle
(whether polar or azimuthal), the flux is a function of radial distance r only, i.e., (r), and the
Laplacian is then:
1 d  2 d  d2 2 d
  2 r
2
  . (40)
r dr  dr  dr 2 r dr
The diffusion equation outside the source, i.e., for all points except the origin, is then:
D 2  r    a  r   0,
 d 2 2 d  (41)
i.e., D  2    a  r   0,
 dr r dr 
D 2  r    a  r   0,
 d 2 2 d  (42)
i.e., D  2    a  r   0,
 dr r dr 
d 2  r  2 d  r  1
  2   r   0. (43)
dr 2 r dr L
 r 
To solve this equation, we can write  in the form  r   . Then, in terms of , Eq. (43)
r
becomes:
Reactor Statics 15
d  r 
dr    r  ,
r r2
d 2  r  d  r  d  r 
d 2  r  2 2  r 
2
 dr  dr2  dr2  .
dr r r r r3
.
Substituting these forms into Eq. (43) results in the simpler form:
d 2  r  1
 2   r   0. (44)
dr 2 L
Equation (44) has mathematical solutions exp(r/L) and exp(-r/L), which give a general solution:
  r   Ae r / L  C e  r / L
er/ L e r /L
  r  A C .
r r
er / L
Now, the term   as r   and therefore cannot be part of a physically acceptable
r
solution. The solution must then be
e r / L
 r  C , (45)
r
with C being a constant which remains to be determined. To determine C, we can use the
continuity condition at the origin.
If S is the source strength, then the number of neutrons crossing the surface of a small sphere
outwards must tend to S as the sphere’s radius tends to 0, and therefore:
lim 4 r 2 J  r    S
r 0
  d  
 lim  4 r 2   D  S
r 0
  dr  
   r  r/ L  r/ L  
   e e  
 lim  4 r 2  CD  L 2    S
r 0 r
    
     
S
 4 CD  S , i.e., C  .
4 D
The neutron flux outside the source is then, finally:
S e r / L
 r  . (46)
4 D r
Again, it is interesting to interpret the physical meaning of the diffusion area L2. Let us calculate
the mean square distance that a neutron travels outwards from the source (at r = 0) to its
absorption point. We can do this by averaging r2 with the absorption rate a as a weighting
function:

 r   4 r dr
2 2
__ a
r 
2 0

.
   4 r dr
2
a
0
With the form (46) for the flux, we can evaluate the integrals and show that
 r 2   6L2 , (47)
i.e., we can interpret L2 as one-sixth of the square of the average distance outwards between
the neutron’s birth point and its absorption point.
3.3 Flux Curvature in Source-Sink Problems and Neutron In-Leakage

Let us consider the “out-leakage” term (call it Leakout) in the above examples.
     
Leakout  leakage out per differential physical volume dr    J    D r   r  . Applying
this to the flux from a plane source (Eq. 39) gives, for x > 0:
 SL  x / L   S   x/ L
Leakout   D 2  e    e  0.
 2D   2L 
We can also show that Leakout < 0 for x < 0. If we apply the formula to the flux from a point
source (Eq. 46):
 S er/L   d 2 2 d   S e r/L 
Leakout   D 2    D  2 
r dr   4 D r 
, which we can show
 4 D r   dr
S e r/ L
  0.
4 L2 r
We see that in all cases, Leakout < 0, i.e., the leakage is inwards (in-leakage) at any point outside
the source. This can actually be seen in the general case from
  
S  r    a  r   D2  r   0,
 
stated earlier as Eq. (34), which for any point outside the source gives  D 2 r    a  r   0
. In other words, we can say that any point outside an external source in a non-multiplying
medium sees a positive flux curvature and neutron in-leakage, i.e., any differential volume is a
net sink or absorber of neutrons:
Positive flux curvature  neutron in-leakage.
4 One-Group Diffusion in a Uniform Multiplying Medium with No

External Source
We will now switch to the analysis of reactor systems with no external source, meaning that all
neutrons are produced by neutron-induced nuclear fission. In this section, as in the previous
one, we will start by assuming that all neutrons are lumped into a single energy group, i.e., G =
Reactor Statics 17
1. There is then, again, no need for the subscript g. Also, as explained in Section 3, we can drop
the term in s and replace t by a.
The neutron-diffusion equation in this case is:
  
 f   r   a  r   D 2  r   0. (48)
A very important point to note is that without the external source, this is a linear homogeneous
equation in the flux (if the properties are truly constant and independent of the flux). This
means that if we find one solution of the equation, then any multiple is also a solution. There-
fore, the absolute value of the flux cannot possibly be deduced from the diffusion equation
(incidentally, not from the transport equation either). This is totally different from problems
with external sources, which drive the absolute value of the flux.
4.1 Uniform Infinite Reactor in One Energy Group

Let us first assume a uniform reactor, infinite in size. This assumption makes all points in space
“equivalent”, which means that the neutron flux  would also have to be constant throughout
space. Moreover, there can be no leakage from one point to another; therefore, the leakage
term can be dropped from the equation.
The equation in the infinite uniform multiplying medium is then:
 f   a  0. (49)
This equation is interesting. The only solution is a trivial solution, i.e., a null flux,  = 0, unless
 f  a . (50)
In other words, unless the composition of the medium is exactly balanced so that Eq. (50) is
satisfied, the uniform infinite reactor cannot really operate in steady state (except in a trivial
zero-flux situation). We can call Eq. (50) the criticality condition for a uniform infinite reactor.
What happens if the criticality condition in Eq. (50) is not satisfied? Then there is no non-trivial
solution, but is this all that we can say? Actually, we can ensure that there is always a non-trivial
solution if we modify Eq. (49) by “tuning” the neutron-yield cross section by dividing it by a new
“modifying factor” which we call k∞, as in:
 f
   a  0. (51)
k
A non-trivial solution of Eq. (51) can always be guaranteed by selecting the value of k∞ as:
 f
k  . (52)
a
What this means is that if the composition of the uniform infinite reactor is modified from the
original composition by dividing the neutron-yield cross section by k∞ defined as in Eq. (52), the
modified uniform infinite reactor can then be operated in steady state with a non-zero flux; this
modified reactor is now critical. What is the value of the flux? We can see from Eq. (51) that
with the modified neutron-yield cross section, the flux can have any value; in other words, the
critical uniform infinite reactor can operate at any flux (and therefore power) value! This is a
direct consequence of the (apparent) homogeneity of the equation (i.e., the equation is of the
form F = 0, with F an arbitrary operator independent of the flux).

Comparing Eqs. (50) and (51), the criticality condition for a uniform infinite reactor can be seen
to be:
 f
k  1 (53)
a
for criticality. The deviation of k∞ from 1 tells us how far the reactor with the original composi-
tion ( f ,  a ) is from criticality. k∞ is called the infinite reactor multiplication constant and can
have the following values:
k∞ < 1: the infinite reactor is said to be subcritical;
k∞ = 1: the infinite reactor is critical;
k∞ > 1: the infinite reactor is said to be supercritical.
The physical interpretation of k∞ can be stated as follows:
 f  f  neutron production rate
k    . (54)
a  a neutron loss rate
A quantity related to k∞, but “centred” at 0 instead of at 1, is the “reactivity”, ∞, defined as:
1
  1  . (55)
k
Therefore:
k∞ < 1   < 0: the infinite reactor is subcritical
k∞ = 1   = 0: the infinite reactor is critical
k∞ > 1   > 0: the infinite reactor is supercritical.
Real-life reactors are designed to be operated at critical or very close to critical. Therefore, in
most situations, we would expect k∞ to be very close to 1 and ∞ to be very close to 0. As a
result, while ∞ (and k∞) are absolute, dimensionless numbers, a new fractional “unit” of 1
milli-k (or mk) is defined for reactivity, where:
1 mk  0.001. (56)
For example, ∞ = + 2 mk (or – 1 mk) means that ∞ = +0.002 (or – 0.001). Another reactivity
unit, “pcm”  0.01 mk, is commonly used in Europe.
4.2 Uniform Finite Reactors in One Energy Group

We will now analyze uniform reactors of finite size (at least in some dimensions). The nuclear
properties are assumed uniform throughout the reactor, but not all points in the reactor are
equivalent spatially (some are further from or closer to the reactor boundary), so that the flux is

a function of space, r  , and also the leakage term must remain in the diffusion equation:
  
 f   r   a  r   D 2  r   0,
presented earlier as Eq. (48). This can be rewritten as:
  
 D2  r    f   r   a  r  . (57)
Reactor Statics 19

The left-hand side of Eq. (57) is the leakage out of the “point” r . If we integrate Eq. (57) over
the volume V of the reactor:
   
  D2  r  dr   f   a     r  dr . (58)
V V
The left-hand side of Eq. (58) is the total leakage out of the reactor. This of course must be
positive: neutrons can only go out of the reactor, not into it, because there are no sources of
neutrons outside which can “feed” neutrons into the reactor. Therefore, if Eq. (58) is true as is,
the right-hand side must be positive, and because the integral of the flux must also be positive,
so must the quantity ( f   a ) . We can therefore write Eq. (57) as:
    a 
2  r   f  r 
D (59)
 
i.e.,  2  r   Bg2  r  ,
where we have defined

 f   a
B g2  . (60)
D
The quantity Bg2 , as it appears in Eq. (59), is called the geometrical buckling of the reactor. Its
physical meaning can be understood by rewriting Eq. (60) as:

2  r 
Bg  
2
 . (61)
 r 
In other words, the geometrical buckling is the negative relative curvature of the neutron flux.
Because we have established that B g2 is definitely positive, this means that the flux curvature is
negative (see Figure 1), and in fact, in view of Eq. (59), the relative curvature is uniform (and
negative) in a homogeneous finite reactor. Note that this statement is strictly true for homo-
geneous reactors only. In real reactors, the relative flux curvature (and the buckling) can vary
and even change sign.
Figure 1 Negative flux curvature in homogeneous reactor

Because the neutron flux curvature/geometrical buckling must “bend” the flux to bring it to 0
almost at the reactor boundary (actually, at an extrapolation distance dextr beyond the physical
boundary), it is clear that the flux curvature will be large for reactors of small dimensions and
small if the reactor dimensions are large. Therefore, B g2 in Eq. (59) is purely a geometrical
quantity.
Returning now to Eq. (57) and using the definition of geometrical buckling, the diffusion equa-
tion for homogeneous reactors can be rewritten as:
  
 f   r   a  r   DBg2  r   0. (62)

This equation has the same characteristics as Eq. (49). The only solution is a null flux,  r   0 ,
unless
 f  a  DBg2 . (63)
In other words, unless the composition of the reactor is exactly balanced so that Eq. (63) is
satisfied, the uniform reactor cannot really operate in steady state (except in a trivial zero-flux
situation). We can call Eq. (63) the criticality condition for a uniform finite reactor.
What happens if the criticality condition Eq. (63) is not satisfied? Then there is no non-trivial
solution. However, just as we did for the infinite medium, we can ensure that there is always a
non-trivial solution if we modify Eq. (62) by tuning the neutron-yield cross section by dividing it
by a similar parameter, called keff, as in:
 f   
  r   a  r   DBg2  r   0. (64)
keff
A non-trivial solution of Eq. (64) can always be guaranteed by selecting the value of keff as:
 f
keff  , (65)
a  DBg2
which can also be written as:
 f
 a
keff
Bg 
2
. (66)
D
This means that if the composition of the uniform reactor is modified from the original compo-
sition by dividing the neutron-yield cross section by keff defined as in Eq. (65), the modified
uniform reactor can then operate in steady state (i.e., as a critical reactor) with a non-zero flux.
Again, as in the infinite medium, the flux in the modified critical reactor can have any value, that
is, the critical uniform reactor can operate at any flux (and therefore power) value!
In summary, the criticality condition for a uniform finite reactor is:
 f
keff  1 (67)
a  DBg2
or
Reactor Statics 21
 f
1
 f   a a k 1
Bg 
2
  2 . (68)
D D L
a
Equation (67) is clearly a generalization of the criticality condition Eq. (53) for the infinite
medium, where the buckling was 0, i.e., B g2  0 (flat flux). The deviation of keff from 1 tells us
how far the reactor with the original composition ( f ,  a , D ) is from criticality.
Equation (68) is intriguing. The left-hand side, B g2 , is a geometrical quantity, as already noted.
The right-hand side, on the other hand, is a function of the nuclear properties only and is not a
geometrical quantity. It is nonetheless called a “buckling”, the material buckling:
k  1
Bm2  . (69)
L2
In view of this, the criticality condition for a uniform reactor in one neutron-energy group can
be expressed as:
Geometrical Buckling  Material Buckling, Bg2  Bm2 . (70)
The physical interpretation of keff can be obtained from:


 f  f   r  neutron production rate
keff      . (71)
 a  DBg  a  r   DBg   r  neutron loss rate (by absorption and leakage)
2 2
This ratio of production to loss is the same at any point in a homogeneous reactor and is of
course then also the same as the ratio of the reactor-integrated production and loss.
Incidentally, Eq. (64), a linear equation with a boundary condition (zero flux at the extrapolation
distance beyond the boundary), is mathematically an eigenvector problem, which can best be
seen by rewriting the equation in the form:
 a  DBg2   r    f   r  , (72)
where
1
 (73)
keff
is the eigenvalue of the problem.

Eigenvalues of eigenvector problems can take on only distinct, non-continuous values. This tells
us that keff cannot have just any value, but it can have only a certain number of distinct values.
Later, we will see that the “physical” keff is the largest of these distinct values.
4.3 Uniform Finite Reactors in Various Geometries

Equations (67) and (68) relate the reactor multiplication constant to the reactor properties and
to geometrical buckling, but we do not yet have a value for the buckling or for the distribution
of the neutron flux in the reactor. In this section, we will solve for the flux distribution and for
the value of the reactor multiplication constant keff for reactors of various geometries (and we
can therefore also derive the value of the buckling). We will do this by showing how to solve
the diffusion equation in the following form [rewritten from Eq. (59)]:
 
2  r   Bg2  r   0. (74)
4.3.1 Infinite slab reactor

Let us consider a reactor in the shape of a slab of physical width a in the x-direction and infinite
in the y- and z-directions. The reactor is centred at x = 0; see Figure 2. This is a one-dimensional
d2
problem, meaning that  is a function of x only. The Laplacian reduces to , and the diffu-
dx 2
sion equation (74) is now:
d 2  x 
2
 Bg2  x   0. (75)
dx
This equation has mathematical solutions (x)  sin(Bgx) and (x)  cos(Bgx). However, by left-
right symmetry, sin(Bgx) is not a viable solution (because the flux would be negative in half the
space). Therefore, the only physical solution must have the form (x) = Acos(Bgx), where A is a
constant.
Figure 2 Infinite-slab reactor

a 
However, we must also satisfy the boundary condition for the problem, which is   ex   0 ,
 2 
where aex is the “extrapolated width” and x = aex /2 is the point at which the flux appears to go
 a 
to zero. Therefore, we must have A cos Bg ex   0 , which means that the values of Bg are
 2 
n
limited to B g  , where n = any odd integer.
a ex
While any odd integer value of n gives a mathematical solution of Eq. (75), only n = 1 is a physi-
cally acceptable solution, because higher values of n would give cosine functions which would
Reactor Statics 23
become negative for some values of x before returning to 0 at aex (see Figure 3). Therefore, the
final solution for the flux distribution in an infinite slab reactor is
 x 
  x   A cos   , (76)
 aex 
and the buckling is
2
 
B   .
2
g (77)
 aex 
Figure 3 Flux with physical and unphysical values of buckling

Note: The value of buckling which gives the only physical solution is the smallest of the mathe-
matically possible values, and the corresponding value of keff for the physical solution is the
largest of the mathematically possible values (i.e., the physical eigenvalue is the smallest of the
possible values).
Equation (76) gives the flux distribution in the reactor. However, the amplitude A, which gives
the absolute value of the flux, cannot be obtained from the diffusion equation, as noted at the
beginning of Section 4. To find A, an extraneous condition has to be imposed, for instance the
actual value of the flux at a point, or the total power of the reactor, or the power in some region
of the reactor.
For example, if we assume a value P for the power per unit area of the y-z plane and note that Ef
is the energy released per fission (typically 200 MeV  3.2*10-11 joules), we have:
a /2 a /2
 x 
P  Ef f    x  dx  E f  f  A1 cos  dx
 a /2  a /2  ex 
a
a /2
a  x  a  a 
i.e., P  ex E f  f A sin    2 A ex E f  f sin  
  aex   a /2   2aex 
P
A ,
 a 
2aex E f  f sin  
 2aex 
and the absolute flux in the slab is
P  x 
  x  cos  . (78)
 a   aex 
2aex E f  f sin  
 2aex 
If the extrapolation distance is ignored, i.e., aex = a, this reduces to:
P x 
  x  cos   . (79)
2aE f  f  a 
4.3.2 Infinite cylindrical reactor

We now consider a reactor in the shape of a uniform cylinder of physical radius R and of infinite
length in the axial (z) direction, as shown in Figure 4. The problem is independent of the
azimuthal angle, and the flux is a function of the r variable only.
Figure 4 Infinite-cylinder reactor

The Laplacian can now be written as
1 d  d 
 r ,
r dr  dr 
and the diffusion equation (74) becomes:
Reactor Statics 25
1 d  d  r  
  Bg   r   0
2
r
r dr  dr 
(80)
d   r  1 d  r 
2
i.e., 2
  B g2  r   0.
dr r dr
This differential equation is actually well known to mathematicians: it is called Bessel’s equation
of order 0, and its mathematical solutions are the ordinary Bessel functions of the first and
second kind, J (Br) and Y (Br) respectively; see Figure 5.
0 0
Figure 5 Ordinary Bessel functions of first and second kind

From Figure 5, we can see that Y (Br) goes to - as r → 0 and is therefore not acceptable as a
0
physical solution for the neutron flux. The only solution is then:
  r   AJ 0  Bg r  . (81)
The flux must go to 0 at the extrapolated radial boundary Rex, i.e., we must have:
J 0  Bg Rex   0. (82)
Figure 5 shows that J0(r) has several zeroes, called ri: the first is at r1  2.405, and the second at
r2  5.6. However, because the neutron flux cannot have regions of negative values, the only
physically acceptable value for Bg is
2.405
Bg  . (83)
Rex
As a result, the flux in the infinite cylinder is given by:
 2.405r 
  r   AJ 0  , (84)
 Rex 
and the buckling for the infinite cylinder is
2
 2.405 
B 
2
g  . (85)
 Rex 
If we use the reactor power P per unit axial height to determine A and neglect the extrapolation
distance, the following value is obtained:
0.738P
A , (86)
E f  f R2
and the absolute flux is

0.738P  2.405r 
 r  J0  . (87)
E f  f R2  R 
4.3.3 Parallelepiped reactor

We now look at a fully finite reactor geometry (i.e., the reactor is finite in every dimension, and
is uniform): the parallelepiped of sides a, b, c (see Figure 6). In this case, we use the Cartesian

co-ordinate system, and the flux is written as  r     x, y , z  . The diffusion equation becomes:
d 2  x, y , z  d 2  x, y , z  d 2  x, y , z 
2
 2
 2
 Bg2  x, y , z   0. (88)
dx dy dz
Figure 6 Rectangular-parallelepiped reactor

To solve this equation, we try a separable form for the neutron flux, i.e.,
  x, y , z   f  x  g  y  h  z  , (89)
where f, g, and h are functions to be determined. Substituting form (89) into Eq. (88), we get:
d 2 f  x d 2g  y d 2h  z 
2
g ( y ) h ( z )  f ( x ) 2
h ( z )  f ( x ) g ( y ) 2
 Bg2 f  x  g ( y )h( z )  0. (90)
dx dy dz
Dividing the equation by f(x)g(y)h(z):
1 d f  x 1 d g  y 1 d h z
2 2 2
    Bg2 . (91)
f  x  dx 2
g  y  dy 2
h  z  dz 2
The left-hand side of Eq. (91) is a sum of three terms which are functions only of x, y, and z
Reactor Statics 27
respectively. The right-hand side is a constant. The only way in which this can happen is if each
of the three terms on the left-hand side is a constant on its own, i.e., if we can write:
1 d f  x 1 d g  y 1 d h z
2 2 2
  Bx ,
2
  By ,
2
  Bz2 , (92)
f  x  dx 2
g  y  dy 2
h  z  dz 2
where B x2 , B y2 , B z2 are constants and
Bx2  By2  Bz2  Bg2 . (93)
Each of the equations in (92), e.g.,

1 d f  x d 2 f x
2
  Bx 
2
 Bx2 f  x   0, (94)
f  x  dx 2
dx 2
is exactly the same equation as for the slab reactor, with the same solution
f  x   cos  Bx x  , (95)
where

Bx  . (96)
aex
The full solution for the neutron flux in the parallelepiped reactor is therefore:
 x   y  z 
  x, y, z   A cos   cos   cos   . (97)
 aex   bex   cex 
The quantities
2 2 2
     
B    , By2    , Bz2    ,
2
x (98)
 aex   bex   cex 
are called the “partial bucklings” in the three directions, and the total buckling is
2 2 2
     
B       .
2
g (99)
 aex   bex   cex 
If we normalize the flux to the total fission power P of the reactor and neglect the extrapolation
distance, we can show that the normalization constant, A, is given by
 3P
A . (100)
8abcE f  f
4.3.4 Finite-cylinder reactor

In the case of a finite-cylinder uniform reactor of radius R and height H (see Figure 7), we use

the cylindrical co-ordinate system to write the flux as  r    r , z  . The Laplacian can be
written as:
d2 1 d d2
2    , (101)
dr 2 r dr dz 2
and the diffusion equation becomes
d 2  r, z  1 d  r, z  d 2  r, z 
2
  2
 Bg2  r, z   0. (102)
dr r dr dz
Figure 7 Finite-cylinder reactor

To solve this equation, we again try a separable form for the neutron flux, i.e.,
  r, z   f  r  g  z  , (103)
where f and g are functions to be determined. Substituting Eq. (103) into Eq. (102), we get:
d 2 f r 1 df  r  d 2g  z
2
g  z    f  r  2
 Bg2 f  r  g  z   0. (104)
dr r dr dz
Let us divide this equation by f(r)g(z):
1 d f r 1 1 df  r  1 d g z
2 2
    Bg2 . (105)
f  r  dr 2
f  r  r dr g  z  dz 2
The left-hand side of Eq. (105) is the sum of a function of r and a function of z. The right-hand
side is a constant. The only way in which this can happen is if the parts in r and z are each
individually equal to a constant, i.e., if we can write:
1 d f r 1 1 df  r  1 d g z
2 2
   B 2
,   Bz2 , (106)
f  r  dr f  r  r dr g  z  dz
2 r 2
where Br2 ,B g2 are constants and
Br2  B g2  Bg2 . (107)
The equations in (106) have been seen before; they are the same equations as for the infinite
cylinder and the slab reactor respectively and therefore have the same solutions:
f  r   J 0  Br r  , g  z   cos  Bz z  , (108)
Reactor Statics 29
where
2.405 
Br  and Bz  , (109)
Rex H ex
where Rex and Hex are the extrapolated radius and the extrapolated axial dimension of the
reactor. The full solution for the neutron flux in the finite-cylinder reactor is therefore:
 2.405r   z 
  r, z   AJ 0   cos  , (110)
 Rex   H ex 
2 2
 2.405    
B 
2
r  , Bz  
2
 (111)
 Rex   H ex 
are called the radial and axial bucklings respectively, and the total buckling is
2 2
 2.405    
B 2
g    . (112)
 Rex   H ex 
If we normalize the flux to the total fission power P of the reactor and neglect the extrapolation
distance, we can show that the normalization constant is:
3.63P
A . (113)
 R 2 HE f  f
4.3.5 Spherical reactor

The last geometry we will look at is a spherical uniform reactor of radius R. We use the spheri-
cal co-ordinate system, and because there is spherical symmetry, the flux can be written as

 r    r  . The Laplacian can be written as:
1 d  2 d 
2  r , (114)
r 2 dr  dr 
and the diffusion equation becomes
1 d  2 d  r  
  Bg  r   0. (115)
2
2 r
r dr  dr 
Let us try to represent (r) in the form
 r
 r  , (116)
r
where the function   r  is to be determined. Equation (115) then reduces to:
1  d  2  1 1 d  r    2  r
 r   2   r       Bg 0
r2  dr   r r dr   r
1 d  d  r   2  r
i.e.,     r   r    Bg 0
r2  dr  dr   r
1  d  r  d  r  d 2  r   2  r
i.e,     r   Bg  0,
r2  dr dr dr 2
 r
d 2  r 
which finally reduces to 2
 Bg2  r   0. (117)
dr
The general solution of this equation is:
  r   A sin  Bg r   C cos  Bg r  ,
which gives
sin  Bg r  cos  Bg r 
 r  A C . (118)
r r
However, the cosine term goes to  as r → 0, which is physically not acceptable, whereas the
sine term is acceptable because it has a finite limit as r → 0 (as can be veriﬁed using L’Hôpital’s
rule). Therefore, the final solution for the spherical reactor is
sin  Bg r 
 r  A ,
r
where, for the same reason as in the other geometries (to guarantee no negative flux), Bg must
take the lowest allowable value, i.e.,

Bg  , (119)
Rex
so that finally
 r 
sin  
 r  A  Rex 
. (120)
R
If we normalize the flux by imposing a value P for the total fission power of the reactor and
neglect the extrapolation distance, we find that:
P  4 AR 2 E f  f
P (121)
 A .
4R E f  f
2
4.4 Flux Curvature and Neutron Out-Leakage

Let us consider the “out-leakage” term (call it Leakout) in the above examples: Leakout  outleak-
Reactor Statics 31
     
age per differential physical volume dr    J    D  r    r  . Applying this to the flux in a
uniform slab reactor (Eq. 76) gives, for x > 0:
   x 
2
   x 
Leakout   D  A cos 
2
    DA   cos    0.
  aex    aex   aex 0 
We can similarly show for all other uniform reactors that Leakout > 0. This can actually be seen
  
in the general case from  f   r   a  r   D2  r   0 (presented earlier as Eq. (48)), which
  
for any point in the uniform reactor gives  D 2 r    f   a  r   DB 2 r   0 . This
means that we can say that any point in a uniform reactor sees neutron out-leakage (i.e., any
differential volume is a source of neutrons).
Note again that the above finding applies to uniform reactors, not to all reactors in general.
Real reactors may have regions which are net sinks of neutrons, where the flux curvature is
positive and where there is a net in-leakage of neutrons. The general rule that we can infer
from the above and also from what we learned in source-sink problems is that:
Positive flux curvature  Net neutron in-leakage, while
Negative flux curvature  Net neutron out-leakage.
5 Reactors in Two Neutron-Energy Groups with No External Source

5.1 The Neutron Cycle and the Four-Factor Formula
Neutrons born in fission have high energy (see the sketch of their energy distribution in Figure
8).
Figure 8 Energies of fission neutrons
It can be seen that the peak in their energy distribution is ~0.73 MeV. A 1-MeV neutron has a
speed of ~13,800 km/s: fission neutrons are fast neutrons. There are essentially no fission
neutrons born with thermal energies ~1 eV; the reference thermal energy is 0.025 eV, speed
2,200 m/s, T = 293.6 K.
Figure 9 Sketch of cross section versus neutron energy

However, the probability of neutrons inducing fission in fissile nuclei (such as 235U) is orders of
magnitude larger for thermal neutrons (with energies of a small fraction of 1 eV) than for fast
neutrons (see Figure 9). This is why thermal reactors (such as CANDU reactors) use a neutron
moderator (such as heavy water in CANDU) to slow neutrons down. This leads to the neutron
cycle in thermal reactors: neutrons are born “fast” (i.e., with relatively high energy, typically in
the range of MeV) and are “encouraged” to slow down to the thermal range, where they induce
more fissions. Of course, fast neutrons do induce a small number of fissions; in fact, only fast
neutrons can fission non-fissile nuclides (e.g., 238U), because there is usually a minimum energy
threshold for such fission to occur. Moreover, while following the cycle, neutrons can also
escape (leak out of the reactor) or be captured in non-productive absorptions at any energy, in
particular in the “resonance range” (between ~1 eV and ~105 eV), where many strong reso-
nance-absorption peaks (mostly leading to non-productive captures) exist. To reduce resonance
capture, most power reactors are designed with the fuel lumped into (mostly) cylindrical
elements in fuel-rod assemblies (inside fuel channels in the case of CANDU) surrounded by a
moderator. Note also that some up-scattering (increase in neutron energy on collision with a
nucleus) can occur in the neutron cycle; this is much less than the rate of down-scattering
(moderation) of neutrons, and the up-scattering cross section is sometimes neglected. The
neutron cycle in a thermal reactor and with no up-scattering (i.e., scattering from a low-energy
group to a high-energy group neglected) is illustrated in Figure 10.
Reactor Statics 33
Figure 10 Neutron cycle in thermal reactor

We can derive an equation for the infinite-lattice multiplication constant k∞ by “following” the
various phases of the neutron cycle in the “neutron-generation” view, as follows:
 Imagine that N neutrons are born in thermal fissions in one generation (each ther-
mal fission produces about 2.5 neutrons on average).
 There will also be some neutrons born in fast fissions in the same generation. Define
 as the ratio of the total number of fission neutrons born in both fast and thermal
fissions to the number born in thermal fissions.  is called the fast-fission ratio ( >
1). Therefore the total number of fast neutrons is now N.
 These neutrons start to slow down, but some are captured in non-productive reso-
nances. If we define p as the resonance-escape probability (p < 1), the number of
neutrons which survive to thermal energies is now Np.
 These thermal neutrons can be absorbed in fuel or in non-fuel components of the
medium. If we define f as the ratio of thermal neutrons absorbed in fuel nuclides to
the total number of thermal neutrons absorbed in this same generation, then the
number of neutrons absorbed in fuel in the current generation is Npf; f is called the
fuel utilization (f  1).
 Some of the neutrons absorbed in fuel may induce more fissions, and some may not.
Define  as the average number of fission neutrons released per thermal absorption
event in the fuel. Then the total number of fission neutrons born (in this new gen-
eration!) from these thermal fissions is now Npf.  is called the reproduction fac-
tor ( > 1).
 We have now gone around the cycle one full time, and we can compare the number
of neutrons in the same phase in successive generations: the ratio in successive gen-
erations is k∞, given by:
N  pf 
k    pf . (122)
N
5.2 The Two-Energy-Group Model

The neutron energy range in the neutron cycle is very wide, eight or nine orders of magnitude
(from ~10 MeV to ~0.01 eV). From this it is obvious that the one-neutron-energy-group diffu-
sion treatment, while very instructive, cannot be very accurate. Because a very large majority
of fissions are induced by thermal neutrons and the fission rate is the most important rate to
calculate correctly (because it is essentially the heat-production rate), a two-energy-group
treatment is often sufficiently accurate in diffusion calculations. The nuclear properties in two
groups must of course be obtained by proper averaging of the detailed many-group properties
determined by a transport-theory calculation. Next, we will analyze the two-energy-group
model.
5.3 Neutron Diffusion Equation in Two Energy Groups

With two energy groups, we will need:
 
 Two fluxes 1  r  and 2  r  , with index 1 for the fast (sometimes called the slowing-
down) group and index 2 for the thermal group;
 
 Two absorption cross sections,  a1 r  and  a 2 r  ;
 
 Two neutron-yield cross sections,  f 1 r  and  f 2 r  , acting on the fast and thermal
fluxes respectively;
 
 Down-scattering and up-scattering cross sections, 12 r  and  21 r  respectively;
 
 Two diffusion coefficients, D1 r  and D2 r  .
The diffusion equation for two energy groups is actually two equations, one for each group.
There must be neutron balance in each group. As previously indicated, all neutrons are born in
the fast group:
     
 f 1  r  1  r    f 2  r  2  r   21  r  2  r 
      
 a1  r  1  r   1 2  r  1  r     D1  r   1  r   0,
(123)
        
12  r  1  r    a 2  r  2  r   21  r  2  r     D2  r  2  r   0. (124)
5.4 Uniform Infinite Medium in Two Energy Groups

As we did in one energy group, let us first assume a uniform reactor, infinite in size. Again, all
points in space are “equivalent”, which means that the neutron fluxes are constant through
space and there is no leakage from one point to another; therefore the leakage terms can be
dropped from the equations.
The diffusion equations (123) and (124) in the infinite uniform multiplying medium then be-
come:
 f 11   f 22   212  a11  121  0
121   a 22   212  0,
which can be written as a linear homogeneous system of two equations:
Reactor Statics 35
f1  a1  12  1   f 2  21  2  0 (125)
121   a 2  21  2  0. (126)
Now, this system can have a non-trivial flux solution if and only if the determinant of the flux
coefficients is zero, i.e.:
  f 1   a1  12    a 2   21    f 2  2 1  12  0. (127)
Satisfying the criticality condition, Eq. (127), requires a very fine balance between the nuclear
properties. What if that balance is not achieved and Eq. (127) is not satisfied? Then there is no
non-trivial solution, i.e., we do not have a real operating infinite-medium reactor. However, we
can ensure that there is always a non-trivial solution if we “tune” the neutron-yield cross
sections by dividing them by a parameter, k∞ (to be determined). The diffusion equations then
become:
  f 1    
  a1  12  1   f 2  21  2  0 (128)
 k   k 
121    a 2  21  2  0, (129)
and the criticality condition is:

  f 1   
  a1  12    a 2  21    f 2  21  12  0. (130)
 k   k 
From Eq. (130), we can determine the value that k∞ must have for criticality:
 f 1  a 2  21    f 212
k  . (131)
a1a 2  a121  a 212
As before, we can define the reactivity as
1
  1  .
k
The difference of k∞ from unity (or of ∞ from 0) tells us how far from critical the original
uniform medium is.
When the fast-fission and up-scattering cross sections are neglected, the form obtained for k∞ is
simpler and instructive:
 f 2 12
k, no up  scattering & no fast  . (132)
 a 2 a1  12
fission
The second factor in Eq. (132),

12
,
 a1  12
gives the probability that fast neutrons will down-scatter relative to their probability of being
absorbed or down-scattered; i.e., it is the probability of fast neutrons surviving to thermal
energies, which is simply the resonance-escape probability p in Eq. (122). The first factor in Eq.
(132) is the number of fission neutrons produced per thermal absorption, i.e., it is the repro-
duction factor  .
Another important quantity to determine in the two-group model, which has no meaning in the
one-group treatment, is the ratio of group fluxes. In the uniform infinite medium, this can be
obtained most simply from Eq. (129):
2 1 2
 . (133)
1  a 2   21
Note that the criticality condition (130) ensures that the same value would be obtained from Eq.
(128).
5.5 Uniform Finite Reactors in Two Energy Groups

We now analyze uniform finite reactors. The diffusion equations are now:
     
 f 11  r    f 22  r   212  r   a11  r   121  r   D121  r   0 (134)
   
121  r    a 22  r   212  r   D2 22  r   0. (135)
We can try to find a solution for the flux which is separable in space and energy, i.e., where the
two-row flux vector

  1  r  
 r     
  r  
 2 
can be written as a group-dependent amplitude times a group-independent flux shape. If we
can solve the equation with such a solution, then it is a good solution:
 
 1  r    A1  r  
       . (136)

 2  r  A
  2   r  
Substituting Eq. (136) into Eq. (135), we get:
 
 12 A1  a 2 A2  21 A2   r   D2 A22  r   0, (137)

and if we divide by  r  :

 2  r  1 2 A1   a 2 A2   21 A2
  . (138)
 r  D2 A2
The right-hand side of Eq. (138) is a single number, independent of space, which we can write as
–B2. [Note: We would have reached a similar conclusion if we had substituted Eq. (136) into Eq.
(134).] Equation (138) is analogous to Eq. (61), i.e., B2 is a group-independent geometric
buckling that is applicable to both the fast and the thermal groups:
 
2g  r    B 2g  r  , g  1, 2. (139)
Because this is exactly the same equation as we found with one energy group, the flux distribu-
tion in two groups is exactly the same as in one energy group: i.e., a product of cosines in a
Reactor Statics 37
parallelepiped reactor, an axial cosine times a radial Bessel function in a cylindrical reactor, and
a sine-over-r function in a spherical reactor. The only new quantity is the ratio of the group
fluxes!
Let us now return to the criticality conditions for a finite reactor. Substituting Eq. (139) into Eqs.
(134) and (135) gives a linear homogeneous system, and therefore we can divide the yield cross
sections by keff as usual to ensure a non-trivial solution:
  f 1      
  a1  12  D1B 2  1  r    f 2  21  2  r  0 (140)
  k 
 keff   eff 
 
121  r   ( a 2  21  D2 B 2 )2  r   0. (141)
The criticality condition for finite reactors in two energy groups is found by equating the deter-
minant of this system to zero, which yields:
 f 1   a 2   21  D2 B 2    f 212
k eff  . (142)
 a1  D1 B 2   a 2   21  D2 B 2   12   a 2  D2 B 2 
If we consider the simpler case obtained by neglecting up-scattering and fast fission, as we did
following Eq. (131), we get:
 f 212
keff , no up  scattering & no fast fission 
 a1  D1B 2   a 2  D2 B 2   12  a 2  D2 B 2 
(143)
 f 212
 .
 a1  12  D1 B   a 2  D2 B 
2 2
The ratio of Eq. (143) to Eq. (132) gives the effect of leakage:
 f 212
keff   a1  12  D1B  a 2  D2 B 
2 2
a1  12 a 2
  . (144)
k   
f 2 12  a1  12  D1B a 2  D2 B 2
2
 a 2  a1  12 
The first factor is the ratio of (fast absorption + down-scattering) to (fast absorption + down-
scattering + fast leakage). Therefore, the first factor represents the fast non-leakage probability,
which we can call P1NL. In the same way, the second factor represents the thermal non-leakage
probability, P2NL. Therefore, Eq. (144) is equivalent to
keff  k P1NL P2 NL , (145)
and the total non-leakage probability is

PNL  P1NL P2 NL . (146)
If we use the four-factor formula for k∞, then Eq. (146) becomes a six-factor formula for keff:
keff   pf P1NL P2 NL . (147)
6 Solution for Neutron Flux in a Non-Uniform Reactor

For uniform (homogeneous) reactors of various geometries, we were able to solve the diffusion
equation analytically and find closed forms for the neutron-flux distribution. However, real
reactors are not homogeneous. The steady-state neutron-diffusion equation must then be
solved numerically. This section shows one method for doing this in the two-neutron-energy
group case.
We start with the two-group neutron-diffusion-equation system, Eqs. (123) and (124), which we
rewrite with the required keff added as a divisor to the fission cross sections (6.1):
       
( f 1  r  1  r    f 2  r  2  r ) / keff  21  r  2  r   a1  r  1  r  
    
1 2  r  1  r     D1  r   1  r   0 (148)
   
12  r  1  r    a 2  r  2  r  
    
 21  r  2  r     D2  r   2  r   0. (149)
All cross sections in a lattice cell are homogenized (averaged spatially within the cell) and
condensed to two groups using a multi-group transport code for each lattice cell, but in a real
reactor, they must also incorporate the effects of reactivity devices superimposed upon the
basic lattice.
6.1 Finite-Difference Form of the Neutron-Diffusion Equation

The simplest method for solving the neutron-diffusion equation in a non-homogeneous reactor
starts by developing a finite-difference form of the equations.
Figure 11 Face view of very simple reactor model

Consider a two-dimensional view of a reactor model, where each spatial dimension has been
subdivided into intervals (see Figure 11). The model is then a collection of homogeneous cells
(which may generally all have different nuclear properties). These cells can be basic lattice cells
or subdivisions of these cells. The finite-difference method then consists of solving for the flux
distribution at a single point in each cell, the centre of the cell in the mesh-centred formulation,
which we will use here.
Reactor Statics 39
Let us look at one of these cells (which we label with a superscript C for Central) and its six
nearest neighbours in the three directions, labelled with a superscript n = 1 to 6 (1 to 4 in x-y
geometry). We get the finite-difference form of the diffusion equation if we integrate Eqs. (148)
and (149) over the volume of cell C:
 
Cf11  r   Cf 22  r     C   
C 
   C

21 2  r   C
a1  C

12 1 r     D1 
1  r  dr  0 (150)
 keff 
    
  1C 21  r     Ca 2   C2 1  2  r     D2C  2  r   dr  0. (151)
C  

where we have dropped the dependence on r for the properties because the cells are homo-
geneous.
There are two types of integrals in Eqs. (150) and (151). The first type does not involve the

divergence operator "". In these integrals, the (homogeneous) cross sections can be taken
out of the integral sign, for example:
 
  Ca 22  r    Ca 2  2  r  dr.
C C
For the integral on the right-hand side, the approximation is made that the integral is equal to
the value of the flux at the centre of the cell multiplied by the volume V C  x C y C z C of the
cell ( x C , y C , z C being the dimensions of the cell in x, y, and z). Then:

  Ca 22  r    Ca 22CV C . (152)
C
The second type of integral has the divergence operator in the integrand. This type of integral
simply represents the leakage out of cell C to its neighbours. By Gauss’s theorem, the volume
integral over the divergence is equal to a surface integral, for example,
    
   D1C  1  r  dr   D1C  1  r   nds
ˆ   J 1  nds
ˆ , (153)
C S S

where S is the surface of the cell, n̂ is the outer normal to the surface, and by Fick’s Law, J 1 is
the net outward current at a point on the surface of the cell.
Figure 12 Flux in a cell and in immediate neighbours

Figure 13 Linear treatment of flux in central cell and one neighbour

Because the cell has six faces, the surface integral is evaluated over the six interfaces between
cell C and its six neighbours, n = 1-6. If we take as an example the face towards the cell with the
lowest x-value (n = 1) for cell C, we can evaluate the surface integral of the current in the fast
group over that face (which we call face C1) as follows. First, denote the group-1 flux at the
centre of the face by 1C1 (see Figure 12). Then let us assume that the flux is linear between the
centre of the cell and the centre of the face (see Figure 13). Then the group-1 current from cell
C to its neighbouring cell 1, at the centre of the face, is:
d 1C  C  1C1 
C  1
J 1,C 1   D C
1   D1 . (154)
dx 1 C
x
2
Similarly, the group-1 current from cell 1 to cell C is:
1C1  11 
J 1,1C  D 1
1 . (155)
1 1
x
2
However, the current must be continuous at the face, i.e.,
J1,1C  J1,C1. (156)
Then by equating Eqs. (154) and (155) and solving for the interface flux, we find that:
D1C x11C  D11xC11
1C1  . (157)
D1C x1  D11x C
We can then substitute this into Eq. (155) and calculate the current at the centre of the face:
1C  1C1  2 D1C D11 1C  11 
J 1,C 1   D C
 C 1 . (158)
  
1
1 C D x D 1
x C
x 1 1
2
Reactor Statics 41
If we assume that the average current over the face is equal to the current at the centre of the
face, the total current over face C1 is:
 2 D1C D11y C z C C
C1 1
J  ˆ
nds  
D1C x1  D11xC
1  11   A1C1 1C  11  , (159)
where A1C1 is a coupling coefficient:
2 D1C D11y C z C
A  C 1
C1
. (160)
D1 x  D11x C
1
A very similar process can of course be performed for all six faces between cell C and its
neighbours. The total outward current in Eq. (153) is then:

ˆ   A1Cn 1C  1n  .
6
S
J 1  nds
n 1
(161)
Using all the above results, the finite-difference neutron-diffusion Eqs. (150) and (151) then
become:
Cf 11C   f 22C C
V  C212CV C   Ca1  1C2  1CV C   A1Cn 1C  1n   0
6
(162)
keff n 1
1C21CV C   Ca 2  C21  2CV C   A2Cn 2C  2n   0.

6
(163)
n 1
These equations couple cell C to its closest neighbours. These and similar equations for all the
other cells in the model make up a coupled system of linear homogeneous equations for the
fluxes at the centres of the cells. If there are N cells in the model, the system is composed of 2N
equations. To solve this coupled system, we need also the boundary conditions at the model
edges, which we look at now.
The edge cells have neighbour cells only towards the “interior” of the model. In directions
outward from the model, the diffusion boundary condition with vacuum is that the flux goes to
zero at the extrapolation distance beyond the boundary.
However, we can express the boundary condition in the same form as Eqs. (162) and (163) by
creating a “dummy” neighbour cell of width 2 dextr, where dextr is the extrapolation length =
2.1312 DC [see Eq. (33)] and forcing the flux to be zero at the extrapolation distance (see Figure
14).
Figure 14 Flux in cell at reactor boundary
6.2 Iterative Solution of the Neutron-Diffusion Equation

The finite-difference neutron-diffusion-equation system described above is a coupled set of 2N
homogeneous equations, where N is the number of real cells in the model. However, actually
we seem to have (2N + 1) unknowns: two values of flux for each of the N cells, plus the value of
keff. In fact, however, there are only 2N unknowns because the overall flux normalization is
arbitrary in the homogeneous diffusion equation and only relative fluxes can be determined.
The absolute flux normalization factor must be determined using an extraneous condition such
as the total reactor fission power.
In real-life reactors, the number of cells N is very large (typically tens of thousands). Conse-
quently, the system of equations can be solved only by iterative techniques. The first step is
that a flux guess must be selected; this can be as simple as a “flat” flux distribution, i.e., the
same value in all cells. A first guess must also be made for keff; for instance, we can use the
guess keff = 1.
Once a flux guess has been established, the flux iterations can begin. At each iteration, the
computer program progresses from cell to cell in the x-direction, the y-direction, and finally the
z-direction, with each cell considered as the central cell C. The two equations (162) and (163)
are solved simultaneously to obtain the fluxes in cell C, 1C and 2C , using the latest values of the
fluxes in the neighbouring cells. As the sweep progresses through the cells, all the fluxes are
updated and used as the latest flux values when the cells are called upon as neighbours. During
this sweep, the value of keff is kept constant. However, when a full sweep over all reactor cells
(a “full iteration”) is completed, the value of keff is updated. It can be recalculated from the
definition of the reactor multiplication constant:
Reactor Statics 43
Neutron Production
keff 
Neutron Absorption  Neutron Leakage
    
r [ f 1  r  1  r    f 2  r  2  r ]dr
           ,

  a1 1
  r    r    a2  r   2  r     J 1  r     J 2  r  
 dr
r
where the latest fluxes from the latest iteration are used and the integrals are evaluated as
sums over the cells.
One iteration is not sufficient to obtain a self-consistent solution of the entire system of finite-
difference equations. We must repeat the iterations until the fluxes converge, i.e., until the
relative difference in flux in each cell from one iteration to the next is very small, smaller than
an accepted tolerance, typically ~10-5. A convergence criterion is also needed for keff; typically, a
difference of 0.001 or 0.01 mk from one iteration to the next is used. Once convergence has
been reached, we have the sought-after solution for the flux shape, i.e., we have the unnormal-
ized (relative) flux distribution. To find the absolute values for the flux, we can normalize the
flux distribution to the total reactor fission power, for example.
7 Energy Dependence of Neutron Flux

In Section 4, we studied the spatial dependence of the neutron flux in one energy group in
uniform reactors. In Sections 5 (uniform reactors) and 6 (non-uniform reactors), we studied
how to determine the neutron flux both in space and in two energy groups. But what can be
said about the variation of the neutron flux with energy when treated as a continuous variable
(not just in two groups)? We examine this in the present section.
When we consider the range of neutron energies from ~1 (or a few) MeV down to a fraction of
1 eV, we can subdivide the flux energy spectrum into three broad regions (note that the region
boundaries cannot be considered “sharp”):
 Fission-neutron energies (>~0.5 MeV)
 Slowing-down (or moderation) range (~0.5 MeV - ~0.625 eV)
 Thermal range (< ~0.625 eV)
We will look at the variation of neutron flux with energy in these three regions in turn.
7.1 The Fission-Neutron Energy Range

This range consists of energies above ~0.5 MeV, up to several MeV, with a maximum flux
magnitude around 0.7 MeV. The energy distribution (“spectrum”) of neutrons in this range is
determined by experimental measurement. It is found to be well approximated by:
  E   0.453e 1.036 E sinh 2.29 E , (164)
where E is in MeV (see Figure 15). [Note: In Chapter 3, the Watt spectrum is used.]
Note that this is a distribution of the number of neutrons, but the flux can be obtained simply
by multiplying by the neutron speed  (not to be confused with , the fission neutron yield).
Figure 15 Energy distribution of fission neutrons
7.2 The Thermal Energy Range

Next, let us look at the thermal energy range, with neutron energies less than ~0.625 eV. Here,
the neutrons are in thermal-equilibrium balance with the ambient medium at temperature T.
Analogously to the atoms in an ideal gas at thermal equilibrium, the neutron population has a
Maxwellian (or approximately Maxwellian) distribution. In terms of flux:
E
E 
 E  e kT
. (165)
 kT 
2
Room temperature is by convention taken as T = 293.6 K = 20.4oC, which gives kT = 0.0253 eV.
The energy value E = kT is the most probable neutron energy in a Maxwellian flux distribution,
and the corresponding “thermal neutron” speed is:
v  kT  0.0253 eV   2200 m / s. (166)
7.3 The Slowing-Down Energy Range

The slowing-down energy range is intermediate between the fission-energy range and the
thermal-energy range. This is the most complex range because of its very broad size and the
highly complex resonance scheme presented to neutrons by heavy nuclides, e.g., 238U.
Therefore, especially when considering neutron absorption in the resonance range in fuel,
neutron slowing-down can be very difficult to calculate and is now mostly done by numerical
computation using complex lattice codes. However, under certain approximations, it is possible
to derive an analytic form for the energy distribution of neutrons as they slow down in the
moderator through collisions with light nuclei.
7.3.1 Slowing down in a hydrogen moderator

The simplest case is slowing down in hydrogen. Using analysis of the kinematics of neutron
Reactor Statics 45
collisions with hydrogen nuclei, the slowing-down flux (E) in hydrogen can be derived. It can
be shown that below the lower boundary Es of the fission-neutron energy range and neglecting
neutron absorption relative to scattering (a fairly good approximation), the slowing-down flux is
inversely proportional to energy E:
S
 E  , (167)
E s
where S is the fission source and s is the scattering cross section (assumed to be independent
of energy).
This provides an important, simple, basic formula for the slowing-down spectrum, even if it is
somewhat of an approximation. Another way of interpreting this relationship is that the prod-
uct E(E) is nearly constant with energy below Es.
7.3.2 Slowing down in a non-hydrogenous moderator

The analysis of the slowing-down spectrum in a non-hydrogenous moderator is more complex,
but under similar approximations, the same form of the slowing-down flux can be found:
S
 E  , (168)
E s
This has the same form as in hydrogen, with an additional factor of  in the denominator.  is
the average lethargy gain per collision, where the lethargy u is defined as:
E 
u  ln  s  . (169)
E
Refer to Chapter 3 for greater detail.
7.3.3 Relaxing the no-absorption approximation

If the absorption cross section is no longer neglected, but is assumed to vary smoothly with
energy (i.e., in the absence of resonances), the 1/E flux spectrum can be modified to:
S  Es   E '  dE ' 
 E  exp    a  . (170)
Et  E     E '  E '
 E t 
The exponential factor in Eq. (170) represents the probability that the neutron survives slowing
down to energy E; i.e., it is the resonance-escape probability to energy E, which we can denote
as p(E).
From the general form of the slowing-down flux, the following simplified statements about the
product E(E) can be deduced:
 If absorption is neglected, and under the assumption that the scattering cross sec-
tion does not depend on energy, E(E) is constant (flat) with E.
 If absorption is included, and assuming a smooth variation of the absorption cross
section with E, then E(E) will decrease smoothly for decreasing E.
These statements are shown in graphical form in Figure 16.
Figure 16 Variation of flux with energy if absorption is smooth
7.3.4 Slowing down in the presence of resonances

Let us now consider absorption resonances. At a resonance energy, the neutron flux decreases
significantly because of the very high absorption cross section at that energy.
As a result of the dip in flux, the absorption rate in the resonance (which is proportional to the
flux) is reduced relative to the absorption rate with the “background” flux value at energies
above and below the resonance, i.e., the product a is smaller than if the flux were unaffected.
This is called “resonance self-shielding”.
On account of resonances, then, the slowing-down flux will be distorted relative to the smooth
curve in the previous subsection, as shown in Figure 17.
Figure 17 Variation of flux with energy in the presence of resonances
7.4 Neutron Flux over the Full Energy Range

With the results in the previous subsections, we are able to “piece together” the neutron flux
over the energy range from fission energies to the thermal range, using:
 the fission spectrum at energies above about 500 keV;
Reactor Statics 47
 the slowing-down spectrum to about 0.625 eV. [Note that in the thermal energy
range, neutrons can gain as well as lose energy in collisions. To be consistent with
the approximation of no up-scattering in the derivation of the slowing-down spec-
trum, the “boundary” between thermal and epithermal energies should be selected
sufficiently high to ensure negligible up-scattering from the thermal region to the
epithermal region. This is one reason for the typical choice of 0.625 eV, which is
about 25 times the most probable energy of 0.025 eV at room temperature, as the
lower energy boundary of the epithermal range.]
 the Maxwellian spectrum at thermal energies. [Note that it is not a perfect Maxwel-
lian, being distorted somewhat by neutron absorption].
The piecing together of the neutron-flux portions in the various energy regions is shown in the
sketch in Figure 18.
Figure 18 Sketch of flux variation over full energy range
8 Basic Design of Standard CANDU Reactors

In this section, we look at the basic features of the standard CANDU reactor design. The CANDU
design is based on the following design principles and the rationale for each:
 Use natural uranium (NU) as fuel. The reason for this choice is that Canada has substantial,
high-grade, domestic uranium mineral resources and, for a reactor cooled and moderated
by heavy water (D2O), it is not necessary to use enriched uranium fuel. Developing a ura-
nium-enrichment industry is extremely expensive, primarily because of the sophisticated
technology involved as well as the costs associated with the stringent materials safeguards,
physical security, and criticality safety requirements associated with enriched nuclear mate-
rials. Moreover, the production cost of uranium-isotope enrichment is also high using cur-
rent technology, in part due to the energy requirements. Consequently, natural-uranium
fuel is much cheaper and simpler to fabricate than enriched-uranium fuel.

 Use heavy water (D2O) as neutron moderator and reactor coolant. The fission chain
reaction cannot be self-sustaining with light water (H2O) as moderator and NU as fuel. The
reason is that normal hydrogen (H) has a fairly large absorption cross section for thermal
neutrons. Hence, light water robs neutrons from the chain reaction, and the fuel must be
enriched to restore a self-sustaining balance between neutron production and capture. On
the other hand, deuterium (D), or heavy hydrogen, has a very small neutron-absorption
cross section and enables the use of NU fuel: heavy water provides high neutron economy.
Although D2O is very expensive to produce due to the low abundance of D relative to H in
nature, the large mass difference between D and H makes separation relatively easy to
achieve using chemical technology processes, and, unlike enriched fuel, the D2O retains its
favourable nuclear properties for the full life of the reactor.
 Use a pressure-tube rather than a pressure-vessel design. The coolant takes heat away
from the fuel (to create steam and produce electricity in a turbine generator) and becomes
very hot. If we want to keep the coolant liquid (or to limit boiling to a small amount), we
must pressurize it (to about 100 atmospheres, ~10 MPa). Therefore, a pressure boundary
must be provided. In light-water reactors (LWRs) a large, thick, steel pressure vessel is used
that surrounds the entire reactor core. The first prototype CANDU reactor [Nuclear Power
Demonstration (NPD), the first reactor to produce electricity in Canada (about 20 MWe)],
was initially conceived with a pressure-vessel design. However, using D2O as the moderator
results in a larger vessel than for an LWR, because the smaller scattering cross section of
D2O (compared to that of H2O) requires a larger moderator volume. Concerns were ex-
pressed about the huge size of the pressure vessels which would be required for larger
CANDU reactors, because such vessels would be difficult to fabricate and very expensive.
The design of NPD, and that of all later CANDU reactors, was therefore changed to a pres-
sure-tube design, in which the pressure boundary occurs in a small, relatively thin, fuel
channel that surrounds each fuel assembly module and its associated coolant. A pressure-
tube design was made possible only by emerging research on zirconium, which showed that
zirconium had a very small neutron-absorption cross section and would be a very good can-
didate metal for the pressure tubes. Using steel, with its much larger absorption cross sec-
tion, would not permit a working pressure-tube design for a reactor using natural-uranium
fuel. The pressure-tube design has the additional safety feature that the break of one pres-
sure tube is more manageable than the break of a large pressure vessel.
Some additional CANDU design features and some consequences of the design decisions are
explored here:
 The natural-uranium lattice has low excess reactivity because the fuel is not en-
riched. This is a safety advantage because it limits the amount of reactivity increase
available in an accident. It is also another rationale for adopting a pressure-tube de-
sign which allows on-power refuelling of the reactor to sustain the core reactivity
without shutting down the reactor. Otherwise, batch refuelling of a CANDU reactor
in a pressure vessel would be required every few months, which would lower its ca-
pacity factor on account of the lost operating time during refuelling and during the
associated, time-consuming, shutdown/cool-down and start-up/heat-up transients.
 CANDU NU fuel achieves a discharge burnup in the ~7,500–~9,500 MW.d/Mg(U)
range, depending on the specific reactor design. This is much lower than the dis-
Reactor Statics 49
charge burnup in LWRs (~30,000–50,000 MW.d/Mg(U)). However, considering fuel

enrichment for LWRs (about 3.5%–5% 235U) and the depleted-uranium tailings which
are discarded, CANDU reactors are more efficient than LWRs and extract ~25% more
energy per Mg of mined natural uranium.
 The pressure-tube design means that the main reactor vessel (the calandria) is not
pressurized. The moderator is insulated from the hot coolant by placing a calandria
tube concentric with each pressure tube, with an insulating gas in the intermediate
(annulus) space. The moderator can be maintained relatively cool at ~70°C by circu-
lating it in its own heat-exchanger circuit, which means that more of the neutrons
will have lower thermal energies than in an LWR and consequently will more readily
induce fission in fissile nuclides, e.g., 235U and 239Pu.
 CANDU reactivity devices are located interstitially between pressure tubes and are
therefore in a benign (low-temperature and unpressurized) environment, which is a
safety advantage. CANDU reactivity devices are not subject to being ejected from
the reactor by coolant pressure.
 Heavy water is used as coolant instead of light water in all operating CANDU reactors
for additional neutron economy. Light water was used as coolant in the Gentilly-1
prototype, and organic coolant was used in the Whiteshell Reactor (WR-1) engineer-
ing test unit, both of which are no longer in operation. Light water is also used as
coolant in the proposed Advanced CANDU Reactor (ACR) design, to reduce capital
costs associated with heavy water, and also in the Supercritical-Water-Cooled Reac-
tor (SCWR) design.
 All operating CANDU reactors have horizontal pressure tubes (the Gentilly-1 proto-
type and the WR-1 had vertical pressure tubes). This orientation promotes symme-
try because coolant can be circulated in opposite directions in alternate tubes (i.e.,
using bi-directional coolant flow), making average neutronics conditions essentially
identical at the two ends of the reactor (unlike the situation in LWRs). With vertical
pressure tubes, the gradient in coolant temperature and density as the coolant picks
up heat and moves upward makes the neutronic and thermal-hydraulic conditions
asymmetric.
 Horizontal pressure tubes also facilitate bi-directional refuelling (i.e., fuelling in op-
posite directions in adjacent channels), which further promotes symmetry. When
the fuel is manufactured as short (~50-cm-long) fuel bundles which can easily be
pushed through the pressure tube from either end, the associated fuelling machines
will also be short, less cumbersome, and require less radiation shielding. With verti-
cal tubes, the entire fuel string (consisting of a vertical stack of perhaps 12 or 13
bundles) would have to be removed from the top of the reactor, so that a very long
fuel assembly would have to be used (together with a long and heavy transfer flask),
or short fuel bundles would have to be connected to a common stringer.
 Whereas the loss of moderator shuts any (thermal) reactor down, the loss of coolant
in a CANDU reactor does not have the same effect because the coolant is separated
from the moderator. Instead, the loss of coolant in operating CANDU reactors gen-
erates a positive reactivity insertion (called the coolant void reactivity, or CVR); this is
due to a number of spectral effects which change mostly the fast-fission factor and
the resonance-escape probability. As a consequence, a loss-of-coolant accident in
CANDU results in a positive power pulse, which must be quickly turned around by
shutdown-system action to avoid overheating the fuel. In fact, the postulated large-
loss-of-coolant accident (LLOCA) scenario is the reason for the adoption of redun-
dancy in CANDU shutdown capability design. CANDU reactors have two fast shut-
down systems, which are physically and logically independent of one another and
each fully capable of shutting the reactor down from any credible configuration.
Shutdown system 1 consists of cadmium shut-off rods which fall under gravity (ini-
tially spring-assisted) into the reactor from above. Shutdown system 2 consists of
the injection of a solution of neutron-absorbing gadolinium (a high neutron ab-
sorber) under high pressure through nozzles directly into the moderator.
 The lattice pitch (distance between the centres of neighbouring tubes) in all operat-
ing CANDU reactors is 28.575 cm. This is not the optimum value in the sense of
maximizing the lattice reactivity (therefore minimizing the refuelling rate and maxi-
mizing the average fuel discharge burnup), which is closer to 34 cm or so. However,
the larger volume of D2O moderator would result in a higher capital cost; the shorter
pitch was selected to minimize the levellized unit-energy cost. A shorter lattice pitch
of 24 cm was selected for the ACR (to reduce moderator cost, among other reasons),
but the reduced moderation would not allow the chain reaction to be self-sustaining
with natural-uranium fuel, and the ACR would need to use enriched fuel.
 The “workhorse” control devices in all operating CANDU reactors are “liquid” zone
controllers. These control 14 compartments in which amounts of light water (used
for its much higher neutron-absorption cross section than heavy water) can be var-
ied uniformly across all compartments to control reactivity or to shape the power
distribution differentially.
9 CANDU Reactor Physics Computational Scheme

Because of the strong heterogeneity of reactor lattices (see the example of the CANDU basic
lattice cell in Figure 19) and because nuclear fuel is a strong neutron absorber, the diffusion
equation cannot be used in reactors based on the detailed lattice geometry. For cases where it
is not desired to apply neutron transport in full-core reactor models, a two-stage or three-stage
process has been developed to enable calculation of the neutron flux and the power distribu-
tion using the diffusion equation. This process is illustrated here for CANDU reactors.
Reactor Statics 51
Figure 19 CANDU basic lattice cell

The computational scheme for CANDU neutronics consists of three stages. Separate computer
programs have been developed to perform the calculations corresponding to each stage. The
three stages are:
 lattice-cell calculation
 reactivity-device (supercell) calculation
 finite-core calculation.
9.1 Lattice Calculation

The first stage involves solving for the flux distribution in the basic lattice cell, which consists of
the fuel, coolant, pressure and calandria tubes, and moderator, but no reactivity devices (see
Figure 19). This calculation is performed using a deterministic multi-group neutron transport
code such as WIMS-AECL. Then the transport code averages the nuclear properties (for absorp-
tion, moderation, fission, etc.) over the cell according to the calculated reaction rates in each
sub-region of the lattice cells to determine the effective neutronic properties of the whole
lattice cell (i.e., few (typically 2)-group macroscopic cross sections and diffusion coefficients).
This averaging (homogenization) of the properties of the basic lattice cell dilutes the strong
absorption of the fuel with the much weaker absorption in the moderator. With these ho-
mogenized properties, Fick’s Law becomes a good approximation over most of the reactor
model, and the diffusion equation can be used to calculate the full-core neutron-flux and power
distributions. The lattice properties govern the neutron-multiplying behaviour of the reactor
lattice, and therefore they must be obtained for all ages of the reactor fuel, i.e., the transport
code must perform a “depletion” calculation to evolve the lattice-cell neutronic properties to
reflect the changes in the nuclide composition of the fuel with irradiation/burnup. The lattice
properties must also reflect any modified configurations of the lattice. For example, if the
coolant is assumed “voided” in a hypothetical loss-of-coolant accident, or the temperature of
the material in a sub-region is assumed to change, then the lattice properties must change
accordingly.
9.2 Reactivity-Device Calculation

This type of calculation determines the effect of a reactivity device on the nuclear properties in
its vicinity. This effect is expressed in the form of “incremental” cross sections which are added
to the (unperturbed) homogenized properties of neighbouring lattice cells in a modelled vol-
ume around the device, called a “supercell”; see Figure 20.
For instance, shutoff-rod incremental cross sections dictate the local efficacy of the shutoff rods
in absorbing neutrons and shutting down the fission chain reaction. The incremental cross
sections of a device are obtained by calculating the differences in the supercell homogenized
properties between the case when the device is inserted into the supercell and the case when it
is withdrawn from it. The incremental cross sections are added to the lattice-cell cross sections
in the modelled volume around the device’s position in the reactor core, when that particular
device is inserted.
Figure 20 Supercell for calculating device incremental cross sections

Because the supercell also contains highly absorbing material, transport theory must again be
used. CANDU reactivity devices are perpendicular to the fuel channels. Therefore, realistic
supercell models are three-dimensional, and as a result supercell calculations are best done
using the DRAGON transport code, which can perform calculations in three dimensions.
9.3 Full-Core Calculation

In the final stage, a full-core reactor model is assembled. The full-core model must incorporate,
for each cell in the reactor, the homogenized unperturbed-lattice-cell cross sections which apply
there, together with the incremental cross sections of nearby reactivity devices in their appro-
priate locations. The model therefore consists of homogenized basic-lattice cells on which are
superimposed homogenized subcells representing reactivity devices. One of many such “device
model areas” is shown in Figure 21; in the white subcells, the nuclear cross sections are ob-
tained by summing the basic-cell properties and the appropriate device incremental properties.
Reactor Statics 53
A full, 3-D reactor model constructed in this way is used to calculate the three-dimensional flux
and power distributions in the core. Because homogenized properties are used, few-group
diffusion theory can be applied to the full-core reactor model.
Figure 21 Simple model with superimposed reactivity device
10 Evolution of Lattice Properties

When nuclear fuel is “burned” in a reactor, changes occur in the material composition of the
fuel. These changes are generally called “fuel (isotopic) depletion”. In this section, we look at
how the properties of the CANDU lattice evolve with fuel irradiation (or burnup).
The following changes occur cumulatively in CANDU nuclear fuel with time:
 The 235U depletes (i.e., its concentration, which starts at 0.72 atom% for fresh natural
uranium, decreases).
 Fission products accumulate; most of these are radioactive, and many have a signifi-
cant neutron-absorption cross section.
239
 Pu is produced by neutron absorption in 238U and two subsequent beta decays:
238
U  n  239 U*  239 Np  
239
Np 239 Pu  .
235
 Pu participates strongly in the fission chain reaction (because it is fissile, like 235U),
but while it continues to be created at about the same rate from 238U, its net rate of
increase slows.
 Further neutron absorptions lead from 239Pu to 240Pu (non-fissile), and then to 241Pu
(fissile).
 Other higher actinides are also formed (e.g., curium, americium).
 The total fissile fraction in the fuel (235U + 239Pu + 241Pu) decreases monotonically.
The evolution of the concentration of these three nuclides is shown in Figure 22.
A typical graph of reactivity ∞ of the infinite CANDU lattice versus fuel irradiation, obtained
from a cell calculation using a transport code, is shown in Figure 23. The following points are of
note:
 The fresh-fuel infinite lattice (where the fuel has not yet received any irradiation) has
a high reactivity (~78 mk when the 135Xe and other saturating fission products have
built up). To achieve a steady state in the infinite lattice with fresh fuel, a corre-
sponding amount of negative reactivity must be added to the lattice [e.g., by dissolv-
ing a neutron poison (i.e., a material with a large neutron absorption cross section)
in the moderator] to suppress the initial supercriticality.
 The reactivity starts to decrease immediately on account of 235U depletion.
 It then starts to increase for a while, on account of production of 239Pu, which is
slightly more effective than 235U. Note the slight delay due to the 239Np ~2-day half-
life.
 However the rate of increase of reactivity slows (because the net rate of plutonium
production decreases), and the reactivity proceeds through a maximum, called the
plutonium peak, with increasing burnup (note that this is not a peak in 239Pu concen-
tration, but in lattice-cell reactivity!).
 Following the plutonium peak, the reactivity decreases monotonically on account of
the continuing depletion of 235U and the continuing accumulation of fission products.
 The infinite lattice reaches zero reactivity at an irradiation corresponding to a burnup
of ~6,700 MW.d/Mg(U).
 A homogeneous infinite lattice with fuel beyond that burnup would be subcritical.
Figure 22 Evolution of fuel isotopic densities
Reactor Statics 55
Figure 23 Infinite-lattice reactivity vs. irradiation

However, remember that the infinite lattice does not experience reactivity losses due to neu-
tron leakage; moreover, it does not account for neutron absorption in all the reactivity devices
within the core. Consequently, a homogeneous reactor with all fuel at the same irradiation
would reach zero reactivity at a much lower burnup than the infinite lattice.
In the CANDU-6 reactor, leakage is about -30 mk, and the average in-core reactivity-device load
is ~-18 mk. Therefore, an estimate of the reactivity eff of a finite homogeneous CANDU-6
reactor can be obtained by subtracting 48 mk from the reactivity of the infinite lattice; see
Figure 24.
Figure 24 Finite-reactor reactivity vs. irradiation

In a homogeneous reactor, a significant amount of additional negative reactivity (~30 mk) is
required to suppress the initial supercriticality. A homogeneous CANDU-6 would reach zero
reactivity at an irradiation corresponding to a burnup of ~4,000 MW.d/Mg(U). A homogeneous
CANDU-6 with fuel beyond that burnup would be subcritical [note that the argument is hypo-
thetical in any case because a homogeneous finite reactor would not remain homogeneous with
burnup]. Therefore, if the CANDU-6 were to be batch-refuelled, the discharge burnup would be
only ~4,000 MW.d/Mg(U).
However, CANDU reactors are refuelled on-power, and therefore there is always (except near
start of life) a mixture of fresh fuel and fuel with high irradiation. The fuel with high irradiation
has negative “local reactivity”, but this is compensated for by the positive local reactivity of low-
irradiation fuel. The proper mixture of fuel in this inhomogeneous reactor (obtained by the
proper rate of daily refuelling) maintains the reactor critical day-to-day. Physically, the older
fuel does the job of reducing the high reactivity of the young fuel, a job which moderator poison
does in the batch-refuelled reactor. The difference is that whereas the poison is just a “para-
sitic” absorber, the older fuel does provide fissions and therefore additional energy.
The mixture of new and old fuel makes it possible to drive the discharge burnup to a much
higher value than that deduced from the “homogeneous” reactivity curve. We can guess (or
calculate) approximately how far we can drive the exit burnup by determining what value gives
equal “positive” and “negative” areas “under” the reactivity curve [this tells us where the
average eff would be 0]; see Figure 25.
From the figure, we can see that positive and negative areas are equal when the average exit (or
discharge) burnup to which we can take the fuel with daily refuelling is ~7,500 MW.d/Mg(U).
This is almost twice the discharge-to-burnup value attainable in the “batch-refuelled” reactor
and represents quite a benefit provided by on-power refuelling!
Figure 25 Averaging reactivity with daily refuelling

 The Neutron Physics chapter explains the interactions that neutrons have with matter and
defines all the basic quantities needed in the study of the fission chain reaction and the
analysis of reactors.
 The present chapter covers mostly the time-independent neutron-diffusion equation and
the analysis of steady-state (time-independent) neutron distributions, either in the presence
of external sources or in fission reactors. It has not covered time-dependent phenomena,
which are left for the following chapter.
 The Reactor Kinetics chapter covers the time-dependent neutron-diffusion equations and
Reactor Statics 57
studies the phenomena of fast-neutron kinetics, fission-product poisoning, reactivity coeffi-

cients, and others in reactors.
12 Problems
Problem 1
There is an infinite homogeneous non-multiplying medium with diffusion length 8 cm and
diffusion coefficient 2 cm. There are four isotropic point sources of neutrons on the x-y plane:
 Source S1 = 108 s-1 at (x, y) = (10 cm, 0)
 Source S2 = 108 s-1 at (x, y) = (-10 cm, 0)
 Source S3 = 1010 s-1 at (x, y) = (0, 20 cm)
 Source S4 = 1010 s-1 at (x, y) = (0, -20 cm)
(a) Find the total flux at (x, y) = (0, 0)

(b) Find the magnitude and direction of the total current at (x, y) = (0 , 0)
(c) Find the value and the direction of the total current at (x, y) = (0, -10 cm).
Problem 2
An isotropic point source of strength S n.s-1 is located at the origin of axes in a homogeneous
non-multiplying material. The material is characterized by an absorption cross section a and a
diffusion constant D.
(a) Imagine a sphere of arbitrary radius R centred at the origin of axes. Calculate the integrated
absorption rate of neutrons (per s) within the sphere.
(b) What is happening to the remaining neutrons (the difference between the number emitted
and the number absorbed per s)? Prove this by calculation.
Problem 3
Suppose that you have an infinite plane source of neutrons in the y-z plane (i.e., at x = 0), which
emits a total of N neutrons per cm2 per s (half in the positive and half in the negative x-
direction).
Of course, we can think of the plane source as an infinite number of point sources, one at each
point of the plane. Let us take the emission from each point source as isotropic. If the point
source actually has differential area dA (which can be as small as we want), then the point-
source strength will be NdA neutrons per s.
Calculate the flux at any point x in space by integrating the flux from all point sources (by
symmetry, this will of course be the same for any values of y and z). Show that you get the
formula for the flux from the infinite plane source.
Problem 4
This is an exercise on the quantitative aspects of the neutron cycle.
Refer to the figure below which pertains to a critical reactor. Refer also to the notes in the
figure. You are asked to calculate how many thermal neutrons escape from the reactor per unit
time.
Remember that the two main things that can happen when a neutron is absorbed in the fuel
Reactor Statics 59
are capture (when the neutron is absorbed and a gamma ray is emitted) and fission (when the
neutron is absorbed and fission is induced). Therefore, the ratio of capture to fission, /f, is
an important parameter. To solve the problem, use the data given and find the right sequence
(up and/or down) for filling numbers into the boxes.
Note that numbers need not be exact integers. In each box, retain non-integer numbers to 3
decimal places.
1000 fast  1000 fast neutrons are born from ther-

neutrons mal fissions
 A thermal fission releases 2.38 fast neu-
Fast n from
trons
Fast Fissions
 For the fuel,    /f = 1.059
 30 fast neutrons are produced by fast
Fast Leakage fission
 32 thermal neutrons are absorbed else-
where than in the fuel
Epithermal and  9 fast neutrons escape from the reactor
Fast Absorption  95 non-thermal neutrons are absorbed
Thermal
Leakage
Non-Fuel Thermal
Absorption
Thermal
Absorptions
in Fuel
Thermal
Captures in
Fuel
Thermal
Fissions
Problem 5
A homogeneous, bare cylindrical reactor with extrapolated axial length 5.8 m is critical. It is
operated at a fission power of 900 MW. In one energy group, the reactor material is known to
have  = 2.38, f = 0.0042 cm-1, and D = 1.14 cm a = 0.0099 cm-1.
The leakage is 9.6 mk. [Note: neglect the extrapolation distance.]
(a) Calculate the reactor buckling and the material’s absorption cross section.
(b) Determine the reactor’s extrapolated diameter.
(c) What is the average flux in the reactor?
(d) What is the ratio of the flux on the cylindrical reactor axis at 50 cm from the reactor face to
the maximum flux in the reactor?
Problem 6
Design the proportions of a cylindrical reactor which minimize leakage [neglect the extrapola-
tion distance].
Problem 7
A research reactor is in the shape of a parallelepiped with a square base of side 5.2 m and a
height of 6.8 m. The reactor is filled uniformly with a fuel of one-group properties f = 0.0072
cm-1 (and  = 2.45) and a = 0.0070 cm-1. The reactor operates steadily at a fission power of 15
MW. The average value of energy per fission Ef = 200 MeV, and 1 eV =1.6*10-19 J. [Neglect the
extrapolation distance.]
(a) What is the value of the diffusion coefficient?
(b) What is the average value of the neutron flux?
(c) What is the maximum value of the neutron flux?
(d) At what rate is the fuel consumed in the entire reactor (in nuclides.s-1) and at the centre of
the reactor (in nuclides.cm-3.s-1)?
Problem 8
A critical homogeneous reactor in the shape of a cube loses 4% of produced neutrons through
leakage.
a) Calculate k for an infinite reactor made of the same material.
b) The initial reactor is re-shaped into a sphere. Calculate the new keff .
NOTE: Use one-group diffusion theory and ignore the extrapolation length.
Problem 9
A reactor is made in the shape of a cone, with the base radius equal to the height and both
equal to 3 m. The reactor is homogeneous, with the following properties:
f = 0.002 cm-1
a = 0.0018 cm-1
Calculate the number of neutrons leaking out of the reactor per second knowing that the
reactor is critical and that the average flux in the reactor is 1013 n/cm2/s.
Problem 10
Imagine we have nuclear material with the following properties in two energy groups:
a1 = 0.0011 cm-1, 12 = 0.0068 cm-1, a2 = 0.0043 cm-1, f2 = 0.00528 cm-1
D1 = 1.07 cm, D2 = 0.92 cm.
(a) Calculate the reactivity of an infinite lattice made of this material.

Reactor Statics 61
(b) In this infinite lattice, what fraction of neutrons is captured in resonances?

(c) We are asked to make a critical homogeneous reactor with this material, cylindrical in
shape, with a diameter 10% greater than the axial dimension [neglect the extrapolation dis-
tance]. What will be this reactor’s dimensions?
(d) Calculate the fast and thermal non-leakage probabilities for this reactor. Also, what is the
total leakage in mk?
(e) What is the ratio of the group-1 flux to the group-2 flux in the reactor?
13 Appendix A: Reactor Statics

In this Appendix, we derive Fick’s Law.
We copy here the angle-integrated version of the time-independent neutron-transport equa-
tion:
    
S  r , E     E    f  r , E '    r , E '  dE '   s  r , E '  E    r , E '  dE ' 
E' E'
    
t  r , E   r , E     J  r , E   0
(presented earlier as Eq. (16)). Our overall objective is to rewrite the last term (the leakage

term) in terms of the total flux  r, E  only.
 
We first try to obtain an equation for the total current J r , E  by multiplying the transport
equation (7), which depends on angle, by ̂ and integrating over it, yielding Eq. (A.1):
1 
S r, E    ˆ   E
ˆ d   ˆ
 f  r , E '   r ,   
ˆ ' dE ' ˆ d
ˆ 
4 ˆ 4   
ˆ'
', E ' d  
ˆ
 E'  
 
   
    s r , E '  E , ˆ '  ˆ  r , ˆ ', E ' ˆ dE ' d ˆ ' d ˆ 
ˆ 
ˆ ' E'
(A.1)

  ˆ  

t  r , E    r ,   ˆ    J r, 
ˆ d ˆ ,E 
ˆ d 
ˆ 0 
ˆ

,E   ˆ

In the first two terms, which we can call T1 and T2, we used the assumption of isotropy in the
ˆ d
external and fission sources, so that these terms are proportional to the integral   ˆ . This is
ˆ

a vector quantity; for instance, the x component is  ˆ dˆ .

ˆ

x It is clear that this integral (and
similarly for the other components) must have a zero value because there are as many positive
ˆ ' s as there are negative. Therefore, the first two terms drop out of the equation:
x
T1  T2  0.
(A.2)
Let us jump to the fourth term, which by definition is:
 ˆ  
ˆ  J r, 

ˆ d  ˆ  J  r , E .
ˆ , E d  
  r , 
ˆ
, E   ˆ
 
Therefore, the fourth term, which we call T4, becomes:

  
T4   t  r , E  J  r , E 
(A.3)
Let us look at the third term, T3:
 ˆ  r , 
 ˆ '
T3      s r , E '  E , 
ˆ'
ˆ
ˆ 
ˆ d 
ˆ ', E ' d 
ˆ ' dE '  
E' 
 
      r , E '  E , ˆ '  ˆ 
ˆ'
E'  ˆ
s
ˆ ˆ '  ˆ  r , 
ˆ 'd ˆ ', E '  d 
ˆ ' dE '.
ˆ '
where we have introduced into the integral the unity factor  ˆ '  1 . Also, recall that s is not
Reactor Statics 63
ˆ 
a function of absolute angles, but only of the cosine  ˆ '.
ˆ '   r, E '  E , 

 ˆ 
ˆ' 
ˆ  ˆ  r, 
ˆ ' d  
ˆ ', E ' d 
ˆ ' dE '. 
T3  E ' ˆ ˆ
 s
'  
The integral over ̂ can be written as
 
 sa  r , E '  E   2   s  r , E '  E ,    d  ,
1
1 (A.4)
and
  ˆ  
 T3    sa  r , E '  E    r ,   ˆ ' d  ' dE '    r , E '  E  J  r, E '  dE '.

', E '   sa
E' ˆ'
 E (A.5)
Note that if the scattering cross section were isotropic, s would not depend on , and there-
fore sa would be 0. Although this is not a good approximation, one that is more reasonable is

to neglect the energy change in anisotropic scattering, i.e., to assume that  sa  r , E '  E  is a

delta function    E' - E  sa  r , E  , which leads to
  
T3   sa  r , E  J  r , E  ,
(A.6)
where
  
1
 sa  r , E   2   s  r , E ,   d   s  r , E 
1 (A.7)
and  is the average cosine of the scattering angle.
Moving on to the fifth term, T5:
   

ˆ ,E 
T5      J r ,  ˆ      r, 
ˆ d  ˆ , E 
ˆ ˆ d
ˆ.  

ˆ
 ˆ
 (A.8)
To calculate the integral, we adopt again the assumption of weak angular dependence of the
angular flux, which leads to the following approximate expression [Eq. (27)] for the angular flux
in terms of the total flux and the total current expression; see derivation in Section 2.3:

 3J  ˆ
ˆ
  
4
 
4
.
(27)
Substituting this expression into Eq. (A.8), we get
    r, E  3  
T5        ˆ 
J  ˆ ˆ d
ˆ.
  
ˆ
  4 4  (A.9)
Now it can be shown (a reasonable result) that the integral of the product of two components
of ̂ , e.g., ij, is equal to 4/3 if i = j, but 0 if i  j. Similarly, the integral of the product of
three components is 0.
Using these two results, Eq. (A.9) becomes
1 
T5     r , E  .
3 (A.10)
Incorporating all these results for the various terms [Eqs. (A.2), (A.3), (A.6), (A.10)], Eq. (A.1)
becomes:
      1 
0  0   sa  r , E  J  r , E   t  r , E  J  r , E     r   0,
3 (A.11)
which yields the final approximation for the total current:
  1 
J r, E      r , E  ,
3  t  r , E    sa  r , E  
(A.12)
which is Fick’s Law, giving for the diffusion coefficient,
 1
D r, E    .
3  t  r , E    sa  r , E  
(A.13)

The quantity in parentheses in Eq. (A.13) is defined as the transport cross section, tr r  .
14 Acknowledgements
the responsibility for any errors or omissions lies entirely with the authors.
Marv Gold
Ken Kozier
Guy Marleau
Bruce Wilkin
1
CHAPTER 5
Reactor Dynamics
prepared by
Eleodor Nichita, UOIT
and
Benjamin Rouben, 12 & 1 Consulting, Adjunct Professor, McMaster & UOIT
Summary:
This chapter addresses the time-dependent behaviour of nuclear reactors. This chapter is
concerned with short- and medium-time phenomena. Long-time phenomena are studied in the
context of fuel and fuel cycles and are presented in Chapters 6 and 7. The chapter starts with an
introduction to delayed neutrons because they play an important role in reactor dynamics.
Subsequent sections present the time-dependent neutron-balance equation, starting with
“point” kinetics and progressing to detailed space-energy-time methods. Effects of Xe and Sm
“poisoning” are studied in Section 7, and feedback effects are presented in Section 8. Section 9
is identifies and presents the specific features of CANDU reactors as they relate to kinetics and
dynamics.
Table of Contents
1 Introduction ............................................................................................................................ 3
1.1 Overview ............................................................................................................................. 3
2 Delayed Neutrons ................................................................................................................... 4
2.1 Production of Prompt and Delayed Neutrons: Precursors and Emitters............................ 4
2.2 Prompt, Delayed, and Total Neutron Yields........................................................................ 5
2.3 Delayed-Neutron Groups .................................................................................................... 5
3 Simple Point-Kinetics Equation (Homogeneous Reactor)....................................................... 6
3.1 Neutron-Balance Equation without Delayed Neutrons ...................................................... 6
3.2 Average Neutron-Generation Time, Lifetime, and Reactivity............................................. 8
3.3 Point-Kinetics Equation without Delayed Neutrons ........................................................... 9
3.4 Neutron-Balance Equation with Delayed Neutrons.......................................................... 11
4 Solutions of the Point-Kinetics Equations............................................................................. 13
4.1 Stationary Solution: Source Multiplication Formula......................................................... 14
4.2 Kinetics with One Group of Delayed Neutrons................................................................. 15
4.3 Kinetics with Multiple Groups of Delayed Neutrons ........................................................ 17
4.4 Inhour Equation, Asymptotic Behaviour, and Reactor Period .......................................... 18
4.5 Approximate Solution of the Point-Kinetics Equations: The Prompt Jump Approximation21
5 Space-Time Kinetics using Flux Factorization ....................................................................... 24
5.1 Time-, Energy-, and Space-Dependent Multigroup Diffusion Equation ........................... 24
5.2 Flux Factorization .............................................................................................................. 25
5.3 Effective Generation Time, Effective Delayed-Neutron Fraction, and Dynamic Reactivity27
©UNENE, all rights reserved. For educational use only, no assumed liability. Reactor Dynamics – December 2016
5.4 Improved Quasistatic Model............................................................................................. 28

5.5 Quasistatic Approximation................................................................................................ 28
5.6 Adiabatic Approximation .................................................................................................. 28
5.7 Point-Kinetics Approximation (Rigorous Derivation) ........................................................ 29
6 Perturbation Theory.............................................................................................................. 29
6.1 Essential Results from Perturbation Theory ..................................................................... 29
6.2 Device Reactivity Worth.................................................................................................... 31
7 Fission-Product Poisoning ..................................................................................................... 32
7.1 Effects of Poisons on Reactivity ........................................................................................ 32
7.2 Xenon Effects..................................................................................................................... 33
8 Reactivity Coefficients and Feedback ................................................................................... 37
9 CANDU-Specific Features ...................................................................................................... 38
9.1 Photo-Neutrons: Additional Delayed-Neutron Groups .................................................... 39
9.2 Values of Kinetics Parameters in CANDU Reactors: Comparison with LWR and Fast
Reactors .................................................................................................................................... 39
9.3 CANDU Reactivity Effects .................................................................................................. 39
9.4 CANDU Reactivity Devices ................................................................................................ 42
11 Problems ............................................................................................................................... 43
12 References and Further Reading........................................................................................... 46
List of Figures
Figure 1 Graphical representation of the Inhour equation........................................................... 19
Figure 2 135Xe production and destruction mechanisms .............................................................. 33
Figure 3 Simplified 135Xe production and destruction mechanisms ............................................. 33
Figure 4 135Xe reactivity worth after shutdown ............................................................................ 36
Figure 5 CANDU fuel-temperature effect ..................................................................................... 40
Figure 6 CANDU coolant-temperature effect................................................................................ 40
Figure 7 CANDU moderator-temperature effect .......................................................................... 41
Figure 8 CANDU coolant-density effect ........................................................................................ 41
List of Tables
Table 1 Delayed-neutron data for thermal fission in 235U ([Rose1991])......................................... 6
Table 2 CANDU reactivity device worth ........................................................................................ 42
Reactor Dynamics 3
1 Introduction
1.1 Overview
The previous chapter was devoted to predicting the neutron flux in a nuclear reactor under
special steady-state conditions in which all parameters, including the neutron flux, are constant
over time. During steady-state operation, the rate of neutron production must equal the rate of
neutron loss. To ensure this equality, the effective multiplication factor, keff, was introduced as a
divisor of the neutron production rate. This chapter addresses the time-dependent behaviour
of nuclear reactors. In the general time-dependent case, the neutron production rate is not
necessarily equal to the neutron loss rate, and consequently an overall increase or decrease in
the neutron population will occur over time.
The study of the time-dependence of the neutron flux for postulated changes in the macro-
scopic cross sections is usually referred to as reactor kinetics, or reactor kinetics without feed-
back. If the macroscopic cross sections are allowed to depend in turn on the neutron flux level,
the resulting analysis is called reactor dynamics or reactor kinetics with feedback.
Time-dependent phenomena are also classified by the time scale over which they occur:
 Short-time phenomena are phenomena in which significant changes in reactor prop-
erties occur over times shorter than a few seconds. Most accidents fall into this
category.
 Medium-time phenomena are phenomena in which significant changes in reactor
properties occur over the course of several hours to a few days. Xe poisoning is an
example of a medium-time phenomenon.
 Long-time phenomena are phenomena in which significant changes in reactor prop-
erties occur over months or even years. An example of a long-time phenomenon is
the change in fuel composition as a result of burn-up.
This chapter is concerned with short- and medium-time phenomena. Long-time phenomena
are studied in the context of fuel and fuel cycles and are presented in Chapters 6 and 7. The
chapter starts with an introduction to delayed neutrons because they play an important role in
reactor dynamics. Subsequent sections present the time-dependent neutron-balance equation,
starting with “point” kinetics and progressing to detailed space-energy-time methods. Effects
of Xe and Sm “poisoning” are studied in Section 7, and feedback effects are presented in
Section 8. Section 9 identifies and presents the specific features of CANDU reactors as they
relate to kinetics and dynamics.

 The production of prompt and delayed neutrons through fission.
 The simple derivation of the point-kinetics equations.
 The significance of kinetics parameters such as generation time, lifetime, reactivity,
and effective delayed-neutron fraction.
 Features of point-kinetics equations and how they relate to reactor behaviour (e.g.,
reactor period).
 Approximations involved in different kinetics models based on flux factorization.
 First-order perturbation theory.

 Fission-product poisoning.
 Reactivity coefficients and feedback.
 CANDU-specific features (long generation time, photo-neutrons, CANDU start-up,
etc.).
 Numerical methods for reactor kinetics.
2 Delayed Neutrons
2.1 Production of Prompt and Delayed Neutrons: Precursors and Emitters
AX
Binary fission of a target nucleus ZX X occurs through the formation of a compound nucleus
AX 1
ZX X which subsequently decays very rapidly (promptly) into two (hence the name “binary”)
fission products Am and Bm, accompanied by the emission of (prompt) gamma photons and
(prompt) neutrons:
AX 1
1
0 n  ZAXX X  ZX X, (1)
(2)
The exact species of fission products Am and Bm, as well as the exact number of prompt neu-
trons emitted,  pm , and the number and energy of emitted gamma photons depend on the
mode m according to which the compound nucleus decays. Several hundred decay modes are
possible, each characterized by its probability of occurrence pm. On average, p prompt
neutrons are emitted per fission. The average number of prompt neutrons can be expressed as:
 p   pm pm . (3)
m
Obviously, although the number of prompt neutrons emitted in each decay mode,  pm , is a
positive integer (1, 2, 3…), the average number of neutrons emitted per fission,  p , is a frac-
tional number.  pm as well as  p depend on the target nucleus species and on the energy of
the incident neutron.
The initial fission products Am and Bm can be stable or can further decay in several possible
modes, as shown below for Am (a similar scheme exists for Bm):
 

 Am1   (mode 1)
 
   Am 2   0
1 HE (mode 2)
 
Am    A 'm 3   0
1 LE (mode 3)
  (fast)

 Am 3  n (delayed neutron)
.
Fission products Am that decay according to mode 3, by emitting a low-energy beta particle, are
Reactor Dynamics 5
called precursors, and the intermediate nuclides A'm3 are called emitters. Emitters are daugh-
ters of precursors that are born in a highly excited state. Because their excitation energy is
higher than the separation energy for one neutron, emitters can de-excite by promptly emitting
a neutron. The delay in the appearance of the neutron is not caused by its emission, but rather
by the delay in the beta decay of the precursor. If a high-energy beta particle rather than a low-
energy one is emitted, the excitation energy of the daughter nuclide is not high enough for it to
emit a neutron, and hence decay mode 2 does not result in emission of delayed neutrons.
2.2 Prompt, Delayed, and Total Neutron Yields

Although one cannot predict in advance which fission products will act as precursors, one can
predict how many precursors on average will be produced per fission. This number is also equal
to the number of delayed neutrons ultimately emitted and is called the delayed-neutron yield,
 d . For incident neutron energies below 4 MeV, the delayed-neutron yield is essentially inde-
pendent of the incident-neutron energy. If delayed neutrons are to be represented explicitly,
the fission reaction can be written generically as:
n  X  A  B   p n p   d nd     . (4)
The total neutron yield is defined as the sum of the prompt and delayed neutron yields:
   d  p . (5)
The delayed-neutron fraction is defined as the ratio between the delayed-neutron yield and the
total neutron yield:
d
 . (6)

For neutron energies typical of those found in a nuclear reactor, most of the energy dependence
of the delayed-neutron fraction is due to the energy dependence of the prompt-neutron yield
and not to that of the delayed-neutron yield. This is the case because the latter is essentially
independent of energy for incident neutrons with energies below 4 MeV.
2.3 Delayed-Neutron Groups

Precursors can be grouped according to their half-lives. Such groups are called precursor groups
or delayed-neutron groups. It is customary to use six delayed-neutron groups, but fewer or
more groups can be used. For analysis of timeframes of the order of 5 seconds, six delayed-
neutron groups generally provide sufficient accuracy; for longer timeframes, a greater number
of groups might be needed. Partial delayed-neutron yields  dk are defined for each precursor
group k. The partial delayed-neutron group yield represents the average number of precursors
belonging to group k that are produced per fission. Correspondingly, partial delayed-neutron
fractions can be defined as:
 dk
k  . (7)

Obviously,
kmax

k 1
k . (8)
235
Values of delayed-group constants for U are shown in Table 1, which uses data from
[Rose1991].
Table 1 Delayed-neutron data for thermal fission in 235U ([Rose1991])
Group Decay Constant, k (s-1) Delayed Yield,  dk (n/fiss.) Delayed Fraction, k
1 0.01334 0.000585 0.000240

2 0.03274 0.003018 0.001238
3 0.1208 0.002881 0.001182
4 0.3028 0.006459 0.002651
5 0.8495 0.002648 0.001087
6 2.853 0.001109 0.000455
Total - 0.016700 0.006854
3 Simple Point-Kinetics Equation (Homogeneous Reactor)

This section presents the derivation of the point-kinetics equations starting from the time-
dependent one-energy-group diffusion equation for the simple case of a homogeneous reactor
and assuming all fission neutrons to be prompt.
3.1 Neutron-Balance Equation without Delayed Neutrons

The time-dependent one-energy-group diffusion equation for a homogeneous reactor without
delayed neutrons can be written as:

n(r , t )   
  f (r , t )  D2(r , t )  a (r , t )
t , (9)
where n represents the neutron density,  f is the macroscopic production cross section, a is
the macroscopic neutron-absorption cross section,  is the neutron flux, and D is the diffusion
coefficient. Equation (9) expresses the fact that the rate of change in neutron density at any

given point is the difference between the fission source, expressed by the term  f ( r , t ) , and

the two sinks: the absorption rate, expressed by the term a (r , t ) , and the leakage rate,

expressed by the term  D2(r , t ) . If the source is exactly equal to the sum of the sinks, the
reactor is critical, the time dependence is eliminated, and the static balance equation results:
  
0   f  s (r )  D2s (r )  a s (r )
, (10)
which is more customarily written as:
  
D2s (r )  a s (r )   f s (r )
. (11)
Reactor Dynamics 7
To maintain the static form of the diffusion equation even when the fission source does not
exactly equal the sum of the sinks, the practice in reactor statics is to divide the fission source
artificially by the effective multiplication constant, keff, which results in the static balance
equation for a non-critical reactor:
  1 
 D 2 s (r )   a  s (r )   f  s ( r )
keff
. (12)
Using the expression for the geometric buckling:
 f
 a
k
Bg2 
eff
D , (13)
Eq. (12) becomes:
 
2s (r )  Bg2s (r )  0
. (14)
Note that the geometrical buckling is determined solely by the reactor shape and size and is
independent of the production or absorption macroscopic cross sections. It follows that
changes in the macroscopic cross sections do not influence buckling; they influence only the
effective multiplication constant, which can be calculated as:
 f
keff 
 a  DBg2
. (15)
Because the value of geometrical buckling is independent of the macroscopic cross section, the
shape of the static flux is independent of whether or not the reactor is critical.
To progress to the derivation of the point-kinetics equation, the assumption is made that the
shape of the time-dependent flux does not change with time and is equal to the shape of the
static flux. In mathematical form:
 
(r , t )  T (t )s (r )
, (16)
where T(t) is a function depending only on time.

It follows that the time-dependent flux  ( r , t ) satisfies Eq. (14), and hence:
 
2(r , t )  Bg2(r , t )
. (17)
Substituting this expression of the leakage term into the time-dependent neutron-balance
equation (9) the following is obtained:

n(r , t )   
  f (r , t )  DBg2(r , t )  a (r , t )
t . (18)
The one-group flux is the product of the neutron density and the average neutron speed, with
the latter assumed to be independent of time:
 
 ( r , t )  n ( r , t )v . (19)
The neutron-balance equation can consequently be written as:


n( r , t )   
  f vn(r , t )  DBg2 vn(r , t )   a vn(r , t )
t . (20)
Integrating the local balance equation over the entire reactor volume V, the integral-balance
equation is obtained:
d    

dt V
n(r , t )dV   f v  n(r , t )dV  DBg2 v  n(r , t )dV  a v  n(r , t )dV
V V V . (21)
The volume integral of the neutron density is the total neutron population nˆ (t ) , which can also
be expressed as the product of the average neutron density n (t ) and the reactor volume:

 n(r , t )dV  nˆ(t )  n (t )V
V . (22)
The volume-integrated flux ˆ (t ) can be defined in a similar fashion and can also be expressed
as the product of the average flux (t ) and the reactor volume:

 (r , t )dV  ˆ (t )  (t )V
V . (23)
It should be easy to see that the volume-integrated flux and the total neutron population satisfy
a similar relationship to that satisfied by the neutron density and the neutron flux:
ˆ (t )   ( r , t ) dV  n( r, t )vdV  nˆ (t )v
  
V V . (24)
With the notations just introduced, the balance equation for the total neutron population can
be written as:
dnˆ (t )
  f vnˆ (t )  DBg2 vnˆ (t )  a vnˆ (t )
dt . (25)
Equation (25) is a first-order linear differential equation, and its solution gives a full description
of the time dependence of the neutron population and implicitly of the neutron flux in a homo-
geneous reactor without delayed neutrons. However, to highlight certain important quantities
which describe the dynamic reactor behaviour, it is customary to process its right-hand side
(RHS) as follows:
dnˆ (t )
 
  f  DBg2  a vnˆ (t )
dt . (26)
3.2 Average Neutron-Generation Time, Lifetime, and Reactivity

In this sub-section, several quantities related to neutron generation and activity are defined.
Reactivity
Reactor Dynamics 9
Reactivity is a measure of the relative imbalance between productions and losses. It is defined
as the ratio of the difference between the production rate and the loss rate to the production
rate.
ˆ  
production rate - loss rate  f  ˆ  DB 2 
ˆ
 
a g

production rate ˆ
 f 
 f   a  DBg2  a  DBg2 loss rate 1
 1  1  1
 f  f production rate keff
. (27)
Average neutron-generation time
The average neutron-generation time is the ratio between the total neutron population and the
neutron production rate.
neutron population nˆ nˆ 1
   
production rate ˆ  f nˆv  f v
 f 
. (28)
The average generation time can be interpreted as the time it would take to attain the current
neutron population at the current neutron-generation rate. It can also be interpreted as the
average age of neutrons in the reactor.
Average neutron lifetime
The average neutron lifetime is the ratio between the total neutron population and the neutron
loss rate:
neutron population nˆ nˆ 1
   
loss rate 2 ˆ
 a  DBg  
 a  DBg nˆ v
2

 a  DBg2 v    . (29)
The average neutron lifetime can be interpreted as the time it would take to lose all neutrons in
the reactor at the current loss rate. It can also be interpreted as the average life expectancy of
neutrons in the reactor.
The ratio of the average neutron-generation time and the average neutron lifetime equals the
effective multiplication constant:
1


  a  DBg2 v  
 f
 keff
 1  a  DBg2
 f v
. (30)
It follows that, for a critical reactor, the neutron-generation time and the neutron lifetime are
equal. It also follows that, for a supercritical reactor, the lifetime is longer than the generation
time and that, for a sub-critical reactor, the lifetime is shorter than the generation time.
3.3 Point-Kinetics Equation without Delayed Neutrons

With the newly introduced quantities, the RHS of the neutron-balance equation (26) can be
written as either:
  DBg2  a   v nˆ(t )   nˆ(t )

 f  DB  a
2
g  vnˆ(t ) 
f
 f
  f

(31)
or
  DBg2  a  keff  1

 f  DB  a 2
g  vnˆ(t ) 
f
DBg2  a
 DB
2
g 
 a vnˆ(t ) 

nˆ(t )
. (32)
The neutron-balance equation can therefore be written either as:
dnˆ (t ) 
 nˆ (t )
dt  (33)
or as:
dnˆ (t ) keff  1
 nˆ(t )
dt  . (34)
Equation (33), as well as Eq. (34), is referred to as the point-kinetics equation without delayed
neutrons. The name point kinetics is used because, in this simplified formalism, the shape of
the neutron flux and the neutron density distribution are ignored. The reactor is therefore
reduced to a point, in the same way that an object is reduced to a point mass in simple kinemat-
ics.
Both forms of the point-kinetics equation are valid. However, because most transients are
induced by changes in the absorption cross section rather than in the fission cross section, the
form expressed by Eq. (33) has the mild advantage that the generation time remains constant
during a transient (whereas the lifetime does not). Consequently, this text will express the
neutron-balance equation using the generation time. However, the reader should be advised
that other texts use the lifetime. Results obtained in the two formalisms can be shown to be
equivalent.
If the reactivity and generation time remain constant during a transient, the obvious solution to
the point-kinetics equation (33) is:

t
nˆ (t )  nˆ (0)e .

(35)
If the reactivity and generation time are not constant over time, that is, if the balance equation
is written as:
dnˆ (t )  (t )
 nˆ(t )
dt (t ) , (36)
the solution becomes slightly more involved and usually proceeds either by using the Laplace
transform or by time discretization.
Before advancing to accounting for delayed neutrons, one last remark will be made regarding
the relationship between the neutron population and reactor power. Because the reactor
power is the product of total fission rate and energy liberated per fission, it can be expressed as:
Reactor Dynamics 11
ˆ (t )  E  nˆ (t )v
P (t )  E fiss  f  fiss f
. (37)
It can therefore be seen that the power has the same time dependence as the total neutron
population.
3.4 Neutron-Balance Equation with Delayed Neutrons

In the previous section, it was assumed that all neutrons resulting from fission were prompt.
This section takes a closer look and accounts for the fact that some of the neutrons are in fact
delayed neutrons resulting from emitter decay.
3.4.1 Case of one delayed-neutron group

As explained in Section 2.1, delayed neutrons are emitted by emitters, which are daughter
nuclides of precursors coming out of fission. Because the neutron-emission process occurs
promptly after the creation of an emitter, the rate of delayed-neutron emission equals the rate
of emitter creation and equals the rate of precursor decay. It was explained in Section 2.3 that
precursors can be grouped by their half-life (or decay constant) into several (most commonly
six) groups. However, as a first approximation, it can be assumed that all precursors can be
lumped into a single group with an average decay constant  . If the total concentration of

precursors is denoted by C ( r , t ) , the total number of precursors in the core, Cˆ (t ) , is simply the
volume integral of the precursor concentration and equals the product of the average precursor
concentration C (t ) and the reactor volume:

 C (r , t )dV  Cˆ (t )  C (t )V
V . (38)

It follows that the delayed-neutron production rate Sd (r , t ) , which equals the precursor decay
rate, is:
 
Sd (r , t )  C (r , t )
. (39)
The corresponding volume-integrated quantities satisfy a similar relationship:
Sˆd (t )  Cˆ (t ) . (40)
The core-integrated neutron-balance equation now must account explicitly for both the prompt
ˆ (t ) , and the delayed-neutron source:
neutron source,  p  f 
dnˆ(t )
  p  f vnˆ (t )  Sˆd (t )  DBg2 vnˆ (t )   a vnˆ (t ) 
dt
 p  f vnˆ (t )  Cˆ (t )  DBg2 vnˆ (t )  a vnˆ(t )
. (41)
Of course, to be able to evaluate the delayed-neutron source, a balance equation for the
precursors must be written as well. Precursors are produced from fission and are lost as a result
of decay. It follows that the precursor-balance equation can be written as:
dCˆ (t ) ˆ (t )   Cˆ (t ) 
 d f 
dt
 d  f vnˆ (t )   Cˆ (t )
. (42)
The system of equations (41) and (42) completely describes the time dependence of the neu-
tron and precursor populations. Just as in the case without delayed neutrons, they will be
processed to highlight neutron-generation time and reactivity. Because reactivity is based on
the total neutron yield rather than the prompt-neutron yield, the prompt-neutron source is
expressed as the difference between the total neutron source and the delayed-neutron source:
dnˆ (t )
  f vnˆ(t )  d  f vnˆ (t )  Cˆ (t )  DBg2 vnˆ (t )  a vnˆ(t ) 
dt
 
 f  DBg2  a vnˆ(t )  d  f vnˆ (t )  Cˆ (t )
. (43)
The RHS is subsequently processed in a similar way to the no-delayed-neutron case:
  DB    vnˆ(t)   vnˆ(t )  Cˆ (t ) 

f
2
g a d f
   DB      
2
    v  nˆ(t )  Cˆ (t ) 

f g a d f
  
f

 f f


nˆ(t )  Cˆ (t )
 . (44)
The neutron-balance equation can hence be written as:
dnˆ (t )   
 nˆ (t )  Cˆ (t )
dt  . (45)
The RHS of the precursor-balance equation can be similarly processed:
  
 d  f vnˆ (t )   Cˆ (t )  d f  f vnˆ (t )   Cˆ (t )  nˆ (t )   Cˆ (t )
 f 
, (46)
leading to the following form of the precursor-balance equation:
dCˆ (t ) 
 nˆ(t )  Cˆ (t )
dt  . (47)
Combining Eqs. (45) and (47), the system of point-kinetics equations for the case of one de-
layed-neutron group is obtained:
dnˆ (t )   
 nˆ (t )  Cˆ (t )
dt 
dCˆ (t ) 
 nˆ (t )  Cˆ (t )
dt  . (48)
Reactor Dynamics 13
3.4.2 Case of several delayed-neutron groups

If the assumption that all precursors can be lumped into one single group is dropped and
several precursor groups are considered, each with its own decay constant k , then the de-
layed-neutron source is the sum of the delayed-neutron sources in all groups:
kmax
Sˆd (t )   k Cˆ k (t )
k 1 , (49)
where Cˆ k (t ) represent the total population of precursors in group k.

The neutron-balance equation then becomes:
dnˆ (t )
  p  f vnˆ (t )  Sˆd (t )  DBg2 vnˆ (t )  a vnˆ (t ) 
dt
kmax
 p  f vnˆ (t )   k Cˆ k (t )  DBg2 vnˆ (t )   a vnˆ(t )
k 1 . (50)
Processing similar to the one-delayed-group case yields:
dnˆ (t )    kmax
 nˆ (t )   k Cˆ k (t )
dt  k 1 . (51)
Obviously, kmax precursor-balance equations must now be written, one for each delayed group
k:
dCˆ k (t )
  dk  f vnˆ (t )  k Cˆ k (t ) (k  1...kmax )
dt . (52)
Processing the RHS of Eq. (52) as in the one-delayed-group case yields:
  
 dk  f vnˆ (t )  k Cˆ k (t )  dk f  f vnˆ (t )  k Cˆ k (t )  k nˆ (t )  k Cˆ k (t )
 f 
. (53)
Finally, a system of kmax+1 differential equations is obtained:
 nˆ (t )   k Cˆ k (t )
dt  k 1
dCˆ k (t )  k
 nˆ (t )  k Cˆ k (t ) (k  1...kmax )
dt  , (54)
representing the point-kinetics equations for the case with multiple delayed-neutron groups.
4 Solutions of the Point-Kinetics Equations

Following the derivation of the point-kinetics equations in the previous section, this section
deals with solving the point-kinetics equations for several particular cases. The first case
involves a steady-state (no time dependence) sub-critical nuclear reactor with an external
neutron source constant over time. An external neutron source is a source which is independ-
ent of the neutron flux. The second case involves a single delayed-neutron group, for which an
analytical solution can be easily found if the reactivity and generation time are constant. Finally,
the general outline of the solution method for the case with several delayed-neutron groups is
presented.
4.1 Stationary Solution: Source Multiplication Formula

Possibly the simplest application of the point-kinetics equations involves a steady-state sub-
critical rector with an external neutron source, that is, a source that is independent of the
neutron flux. The strength of the external source is assumed constant over time. If the total
strength of the source is Ŝ (n/s), the neutron-balance equation needs to be modified to include
this additional source of neutrons. The precursor-balance equations remain unchanged by the
presence of the external neutron source. Because a steady-state solution is sought, the time
derivatives on the left-hand side (LHS) of the point-kinetics equations vanish. The steady-state
point-kinetics equations in the presence of an external source can therefore be written as:
  kmax
0 nˆ   k Cˆ k  Sˆ
 k 1

0  k nˆ  k Cˆ k (k  1...kmax )
 . (55)
Equation (55) is a system of linear algebraic equations where the unknowns are the neutron and
precursor populations. This can be easily seen by rearranging as follows:
  kmax
nˆ   k Cˆ k  Sˆ
 k 1
k
nˆ  k Cˆ k  0 (k  1...kmax )
 . (56)
The system can be easily solved by substitution, by formally solving for the precursor popula-
tions in the precursor-balance equations:

Cˆ k  k nˆ (k  1...kmax )
k  (57)
and substituting the resulting expression into the neutron-balance equation to obtain:
  kmax

nˆ   k nˆ  Sˆ
 k 1  . (58)
Noting that the sum of the partial delayed-neutron fractions equals the total delayed-neutron
fraction, as expressed by Eq. (58), the neutron-balance equation can be processed to yield:
  1  kmax     
nˆ  nˆ     k  nˆ  nˆ  nˆ  nˆ  nˆ  Sˆ
    k 1     
. (59)
The neutron population is hence equal to:

n̂   Sˆ
. (60)
Reactor Dynamics 15
Note that the reactivity is negative, and therefore the neutron population is positive. Equation
(60) is called the source multiplication formula. It shows that the neutron population can be
obtained by multiplying the external source strength by the inverse of the reactivity, hence the
name. The closer the reactor is to criticality, the larger the source multiplication and hence the
neutron population. Substituting Eq. (60) into Eq. (57), the individual precursor concentrations
become:
Sˆ  k
Cˆ k  ( k  1...kmax )
  k . (61)
The source multiplication formula finds applications in describing the approach to critical during
reactor start-up and in measuring reactivity-device worth.
4.2 Kinetics with One Group of Delayed Neutrons

Another instance in which a simple analytical solution to the point-kinetics equations can be
developed is the case of a single delayed-neutron group. This sub-section develops and ana-
lyzes the properties of such a solution. The starting point is the system of differential equations
representing the neutron-balance and precursor-balance equations:
dnˆ (t )   
 nˆ (t )  Cˆ (t )
dt 
dCˆ (t ) 
 nˆ (t )  Cˆ (t )
dt  . (62)
For the case where all kinetics parameters are constant over time, this is a system of linear
differential equations with constant coefficients, which can be rewritten in matrix form as:
  
   nˆ (t ) 
d  nˆ (t )   
   
dt Cˆ (t )      Cˆ (t ) 
   
 . (63)
According to the general theory of systems of ordinary differential equations, the first step in
solving Eq. (63) is to find two fundamental solutions of the type:
 n  t
C  e
  . (64)
The general solution can subsequently be expressed as a linear combination of the two funda-
mental solutions:
 nˆ (t )   n0  0t  n1  1t
ˆ   a0   e  a1 e
 C (t )   0
C  1
C
. (65)
Coefficients a0 and a1 are found by applying the initial conditions.
To find the fundamental solutions, expression (64) is substituted into Eq. (63) to obtain:
  
 n
d   n  t     
 e     e t
dt  C    
   
C
   , (66)
and subsequently:
  
  n
n  
   et       e t
C       
C
   . (67)
Dividing both sides by the exponential term and rearranging the terms, the following homoge-
neous linear system is obtained, called the characteristic system:
  
     n  0
      
 
       
C 0
   . (68)
This represents an eigenvalue-eigenvector problem, for which a solution is presented below.
First, the system is rearranged so that the unknowns are each isolated on one side, and the
system is rewritten as a regular system of two equations:
   
    n   C
  

n      C
 . (69)
Dividing the two equations side by side, an equation for the eigenvalues k is obtained:
   
  
    
    
 . (70)
This is a quadratic equation in  , as can easily be seen after rearranging it to:
    
2     0
   
. (71)
The two solutions to this quadratic equation are simply:
Reactor Dynamics 17
2
        
       4
      
0,1 
2 . (72)
Once the eigenvalues are known, either the first or the second of equations (69) can be used to
find the relationship between n and C . In doing so, care must be taken that the right eigen-
value (correct subscript) is used for the right n - C combination. Note that only the ratio of n
and C can be determined. It follows that either n or C can have an arbitrary value, which is
usually chosen to be unity. For example, if the second equation (69) is used, and if n0 and n1 are
chosen to be unity, the two fundamental solutions are:
 1 
   e0t
 
 
    0 

 1 
   e1t
 
 
   1
 . (73)
The general solution is then:
 1   1 
 nˆ (t )    e0t  a   e1t
 ˆ   a0    1  
C (t )       0        1  
. (74)
4.3 Kinetics with Multiple Groups of Delayed Neutrons

Having solved the kinetics equations for one delayed-neutron group, it is now time to focus on
the solution of the general system, with several delayed-neutron groups. The starting point is
the general set of point-kinetics equations:
 nˆ (t )   k Cˆ k (t )
dt  k 1
ˆ
dCk (t )  k
 nˆ (t )  k Cˆ k (t ) (k  1,..., kmax )
dt  . (75)
As long as the coefficients are constant, this is simply a system of first-order linear differential
equations, whose general solution is a linear combination of exponential fundamental solutions
of the type:
 n 
C 
 1  et
  
 
Ckmax  . (76)
There are kmax+1 such solutions, and the general solution can be expressed as:
 nˆ(t)   nl 
 ˆ  k  l 
 C1 (t)  max  C1  lt
    al    e
  l 0  l 
Cˆkmax (t) Ckmax 
. (77)
4.4 Inhour Equation, Asymptotic Behaviour, and Reactor Period

By substituting the general form of the fundamental solution, Eq. (76), into the point-kinetics
equations and following steps similar to those in the one-delayed-group case, the following
characteristic system can be obtained:
   kmax
n  n  kCk
 k 1

Ck  n  kCk (k  1,..., kmax )
 . (78)
The components Ck can be expressed using the precursor equations in (78) as:
k
Ck  n (k  1,..., kmax )
   k 
. (79)
Substituting this into the neutron-balance equation in (78) yields:
  kmax
k
n  n  k n
 k 1     k  . (80)
Note that the component n can be simplified out of the above and that by rearranging terms,
the following expression for reactivity is obtained:
kmax
k
      k
k 1   k  . (81)
Equation (81) is known as the Inhour equation. Its kmax+1 solutions determine the exponents of
the kmax+1 fundamental solutions. To understand the nature of those solutions, it is useful to
attempt a graphical solution of the Inhour equation by plotting its RHS as a function of  and
observing its intersection points with a horizontal line at y   . Such a plot is shown in Fig. 1
for the case of six delayed-neutron groups.
Reactor Dynamics 19

 6  5  4  3  2  1
Figure 1 Graphical representation of the Inhour equation

For large (positive or negative) values of  , the asymptotic behaviour of the RHS can be ob-
tained as:
kmax
k
      k    
k 1  k . (82)
The resulting oblique asymptote is represented by the blue line in Fig. 1, and the RHS plot
(shown in bright green) approaches it at both  and  . Whenever  equals minus the
decay constant for one of the precursor groups, the RHS becomes infinite, and its plot has a
vertical asymptote, shown as a (red) dashed line, at that value. For   0 , the RHS vanishes, as
can be seen from Eq. (81), and hence the plot passes through the origin of the coordinate
system. Three horizontal lines, corresponding to three reactivity values, are shown in violet.
The two thin lines correspond to positive values, and the thick line corresponds to the negative
value.
Figure 1 shows that the solutions to the Inhour equation are distributed as follows:
 kmax - 1 solutions are located in the kmax - 1 intervals separating the kmax decay con-
stants taken with negative signs, such that k  k 1  k 1 . All these solutions are
negative.
 The largest solution, in an algebraic sense, is located to the right of 1 and is either
negative or positive, depending on the sign of the reactivity. It will be referred to as
max or 0 .
 The smallest solution, in an algebraic sense, lies to the left of   k and will be re- max
ferred to as min or  k . It is (obviously) negative as well. Note that because the

max
generation time is usually less than 1 ms, and often less than 0.1 ms, the slope of the
oblique asymptote is very small. Consequently, min is very far to the left of   k , max
and hence kmax  kmax  kmax 1 . The importance of this fact will become clearer
later, when the prompt-jump approximation will be discussed.
Overall, the solutions are ordered as follows:
min  kmax  kmax 1  .....  0  max

. (83)
It is worth separating out the largest exponent in the general solution described by Eq. (77) by
writing:
 nˆ(t)   n0   nl 
 ˆ   0  
 C1 (t)   C1  0t kmax  C1l  lt
    a0    e  al    e
   0 
l 1
 l 
Cˆkmax (t) Ckmax  Ckmax 
. (84)
Furthermore, it is worth factoring out the first exponential term:
 nˆ(t)    n0   nl  
 ˆ    0  k  l  
 C1 (t)  0t   C1  max  C1  l 0 t 
    e  a0     al    e 
    0  l 1  l  
Cˆkmax (t)  Ck  C  
  max   max 
k . (85)
Note that because 0 is the largest solution, all exponents l  0  are negative. It follows
l 0 t
that for large values of t, all exponentials of the type e nearly vanish, and hence the
solution can be approximated by a single exponential term:
 nˆ(t)   n0 
 ˆ   0
 C1 (t)   C1  e0t
    a0
  
   0 
Cˆkmax (t) Ckmax  . (86)
This expression describes the asymptotic transient behaviour.
The inverse of 0  max is called the asymptotic period:
1
T
max . (87)
With this new notation, the asymptotic behaviour can be written as:
 nˆ(t)   n0 
 ˆ   0 t
 C1 (t)   C1  eT
   0    a
   0 
Cˆkmax (t) Ckmax  . (88)
Before ending this sub-section, a few more comments are warranted. In particular, it is worth
considering the solution to the point-kinetics equation (PKE) for three separate cases: negative
Reactor Dynamics 21
reactivity, zero reactivity, and positive reactivity.
Negative reactivity
In the case of negative reactivity, all exponents in the general solution are negative. It follows
that over time, both the neutron population and the precursor concentrations will drop to zero.
Of course, after a long time, the asymptotic behaviour applies, which has a negative exponent.
Zero reactivity
In the case of zero reactivity, max vanishes, and all other l are negative. The general solution
can be written as:
 nˆ(t)   n0   nl 
 ˆ   0  kmax  l 
 C1 (t)   C1   a  C1  elt
   
    a0 l
  
   0 
l 1
 l 
Cˆkmax (t)  kmax 
C Ckmax  . (89)
After a sufficiently long time, all the exponential terms die out, and the neutron and precursor
populations stabilize at a constant value. Note that these populations do not need to remain
constant from the beginning of the transient, but only to stabilize at a constant value.
Positive reactivity
In the case of positive reactivity, max is positive, and all other l are negative. Hence, after
sufficient time has elapsed, all but the first exponential term vanish, and the asymptotic behav-
iour is described by a single exponential which increases indefinitely.
4.5 Approximate Solution of the Point-Kinetics Equations: The Prompt Jump

Approximation
It was mentioned in the preceding sub-section that the smallest (in an algebraic sense) solution
of the Inhour equation is much smaller than the remaining kmax solutions. This important
property will make it possible to introduce the prompt jump approximation, which is the topic
of this sub-section.
By inspecting the Inhour plot in Figure 1, and keeping in mind the expression of the oblique
asymptote given by Eq. (82), it is easy to notice that the oblique asymptote intersects the x-axis
at:
 
as 
 . (90)
It is also easy to see that:
kmax 1  as
. (91)
Assuming a reactivity smaller than approximately half the delayed-neutron fraction (equal to
0.0065 according to Table 1), and assuming a generation time of approximately 0.1 ms, the
resulting value of as is approximately -32.5 s-1, which is much smaller than even the largest
decay constant in Table 1 taken with a negative sign. That value is only -3 s-1. This shows that
the following inequality holds true:
min  kmax  as  kmax 1  .....  0  max
. (92)
The general solution of the point-kinetics equations expressed by Eq. (77) can be processed to
separate out the term corresponding to kmax 1 :
 nˆ(t)    n0   nkmax 1   nl  
 ˆ    0  kmax 1  kmax 2  l  
 C1 (t)  kmax1t   1   kmax kmax1   a
C C
 1  C  
al  1  e l kmax 1 
   
  t   t
   e  akmax    e kmax 1
  
    0   kmax 1 
l 0
 l  
Cˆkmax (t)  C  C  C  
  max 
k  max 
k  max 
k  . (93)

According to Eq. (92), for l  kmax  2 , all exponents of the type l  kmax 1 t are positive. The 
 
only negative exponent is kmax  kmax 1 t , which is also much larger in absolute value than all
other exponents. It follows that after a very short time, t, the first term of the RHS of Eq. (93)
becomes negligible, and the solution of the point-kinetics equations can then be approximated
by:
 nˆ(t)    nkmax 1   nl    nl 
 ˆ    kmax 1  kmax 2  l    l 
 C1 (t)  kmax 1t   C1   C1  l kmax1 t  kmax 1  C1  el t
  e  akmax 1      al    e    al   
    kmax 1 
l 0
 l   l 0  l 
Cˆkmax (t)  
 Ckmax  Ckmax   Ckmax  . (94)
Concentrating on the neutron population, its expression is:
kmax 1
nˆ(t )   a n e l
l lt
l 0 . (95)
Substituting this into the neutron-balance equation of the point-kinetics system, the following is
obtained:
kmax 1 kmax 1
   l l t kmax ˆ
 l al nl elt   
al n e  k Ck
l 1 l 1 k 1 . (96)
Noting that the following inequality holds true:
 
l  l  0,...kmax 1
 , (97)
the LHS of Eq. (96) can be approximated to vanish, and hence the equation can be approxi-
mated by:
Reactor Dynamics 23
kmax 1
   l lt kmax ˆ
0  
al n e  k Ck
l 1 k 1 , (98)
which is equivalent to:
  kmax
0 nˆ (t )   k Cˆ k (t )
 k 1 . (99)
By adding the precursor-balance equations, the following approximate point-kinetics equations
are obtained:
  kmax
0 n(t )   k Cˆ k (t )
ˆ
 k 1
dCˆ k (t )  k
 nˆ (t )  k Cˆ k (t ) (k  1,..., kmax )
dt  . (100)
This system of kmax differential equations and one algebraic equation is known as the prompt
jump approximation of the point-kinetics equations. The name comes from the fact that
whenever a step reactivity change occurs, the prompt jump approximation results in a step
change, a prompt jump, in the neutron population. To demonstrate this behaviour, let the
reactivity change from 1 to 2 at time t0. The neutron-balance equation before and after t0
can be written as:
1   kmax
nˆ (t )   k Cˆ k (t ) (t  t0 )
 k 1
2   kmax
nˆ(t )   k Cˆ k (t ) (t  t0 )
 k 1 . (101)
The limit of the neutron population as t approaches t0 from the left, symbolically denoted as
nˆ (t0 ) , is found from the first equation (101) to be equal to:
 kmax
nˆ (t0 ) 
1  
  Cˆ (t )
k k 0
k 1
. (102)
Similarly, the limit of the neutron population as t approaches t0 from the right, symbolically
denoted as nˆ (t0 ) , is found from the second equation (101) to be equal to:
 kmax
nˆ (t0 ) 
2  
  Cˆ (t )
k k 0
k 1
. (103)
Taking the ratio of the preceding two equations side by side, the following is obtained:
nˆ (t0 ) 1  

nˆ (t0 ) 2   . (104)
There is therefore a jump nˆ (t0 ) equal to:
1     2 
nˆ (t0 )  nˆ (t0 )  nˆ (t0 )  nˆ (t0 )  nˆ (t0 )  1 nˆ (t0 )
2   2   . (105)
Of course, the actual neutron population does not display such a jump; it is continuous at t0.
Nonetheless, a very short time  t after t0 (at t0  t ), the approximate and exact neutron
populations become almost equal. Note also that Eq. (105) is valid only if both reactivities 1
and 2 are less than the effective delayed-neutron fraction  .
5 Space-Time Kinetics using Flux Factorization

In the previous sections, the time-dependent behaviour of a reactor was studied using the
simple point-kinetics model, which disregards changes in the spatial distribution of the neutron
density. This section will improve on that model by presenting the general outline of space-
time kinetics using flux factorization. The approach follows roughly that used in [Rozon1998]
and [Ott1985]. A complete and thorough treatment of the topic of space-time kinetics is
beyond the scope of this text. This section should therefore be regarded merely as a roadmap.
The interested reader is encouraged to study the more detailed treatments in [Rozon1998],
[Ott1985], and [Stacey1970].
5.1 Time-, Energy-, and Space-Dependent Multigroup Diffusion Equation

The space-time description of reactor kinetics starts with the time-, space-, and energy-
dependent diffusion equation. An equivalent treatment starting from the transport equation is
also possible, but using the transport equation instead of the diffusion equation does not
introduce fundamentally different issues, and the mathematical treatment is somewhat more
cumbersome. The time-, space- and energy-dependent neutron diffusion equation in the
multigroup approximation can be written as follows:
1  
 g (r , t ) 
vg t
     
   Dg (r , t ) g (r , t )    rg (r , t ) g (r , t )    sg ' g (r , t ) g ' (r , t ) 
g ' g
    kmax
 
 pg (r , t ) p (r , t ) fg ' (r , t ) g ' (r , t )    dgk (r , t )k Ck (r , t )
g' k 1
. (106)
The accompanying precursor-balance equations are written as:
     
ck (r , t )   pk (r , t ) fg ' (r , t ) g ' (r , t )  k Ck (r , t )
t g'
. (107)
Equations (106) and (107) represent the space-time kinetics equations in their diffusion ap-
proximation. Their solution is the topic of this section.
It is advantageous for the development of the space-time kinetics formalism to introduce a set
of multidimensional vectors and operators, as follows:
Flux vector
Reactor Dynamics 25
 
Φ  r , t    g  r , t  
(108)
Precursor vector
 
ξ k (r , t )    dgk Ck (r , t ) 
(109)
Loss operator
      
M  r , t      Dg (r , t ) g (r , t )   rg (r , t ) g (r , t )    sg ' g (r , t ) g ' (r , t )
g ' g
(110)
Prompt production operator
    
Fp (r , t )   pg ( r , t ) p ( r , t ) fg ' ( r , t ) g ' ( r , t )
g'
(111)
Precursor production operator for precursor group k
  
Fdk (r , t )  dgk (r , t )k Ck (r , t )
(112)
Inverse-speed operator
1
v 1   g'g
vg
, (113)
where  g ' g is the Kronecker delta symbol.
Using these definitions, the time-dependent multigroup diffusion equation can be written in
compact form as:
 1      k max

v Φr , t   M (r , t )Φr , t   Fp (r , t )Φr , t    k ξ k (r , t )
t k 1 . (114)
The precursor-balance equations can be written as:
    
ξ k (r , t )  Fdk (r , t )Φ  r , t   k ξ k (r , t ) (k  1,..., kmax )
t . (115)
 
As a last definition, for two arbitrary vectors Φ ( r , t ) and Ψ ( r , t ) , the inner product is defined
as:
 
Φ, Ψ    g (r , t )g (r , t )dV
g Vcore
. (116)
5.2 Flux Factorization

Expressing a function as a product of several (simpler) functions is known as factorization. It is a
well-known fact from partial differential equations that trying to express the solution as a
product of single-variable functions often simplifies the mathematical treatment. It is therefore
reasonable to attempt a similar approach for the space-time kinetics problem. A first step in
this approach is to factorize the time-, energy-, and space-dependent solution into a function
dependent only on time and a vector dependent on energy, space, and time. The function
dependent only on time is called an amplitude function, and the vector dependent on space,
energy, and time is called a shape function. The sought-for flux can therefore be expressed as:
 
Φr , t   p(t )Ψ(r , t ) . (117)
Such a factorization is always possible, regardless of the definition of the function p(t). In this
case, the function p(t) is defined as follows:
 
p(t )  w (r ), v 1Φ(r , t )
, (118)

where w ( r ) is an arbitrary weight vector dependent only on energy and position:
 
w  r    wg  r  
. (119)
According to its definition, p(t) can be interpreted as a generalized neutron population. Indeed,
if the weight function were chosen to be unity, p(t) would be exactly equal to the neutron
population.

From the definition of the flux factorization, it follows that the shape vector Ψ ( r , t ) satisfies the
following normalization condition:
 
w (r ), v 1Ψ(r , t )  1
. (120)
Substituting the factorized form of the flux into the space-, energy-, and time-dependent
diffusion equation, the following equations (representing respectively the neutron and precur-
sor balance) result:
dp (t ) 1     
v Ψ  r , t   p (t )  v 1Ψ  r , t     p (t )M (r , t ) Ψ  r , t  
dt t
  kmax

p (t )Fp (r , t ) Ψ  r , t    k ξ k (r , t )
k 1 . (121)
    
ξ k ( r , t )  p (t )Fdk ( r , t ) Ψ  r , t   k ξ k ( r , t ) ( k  1,..., k max )
t . (122)
The precursor-balance equation can be solved formally to give:
   
t
ξ k ( r , t )  ξ k ( r , 0)e  k t   e  k (t t ') p (t ') Fdk ( r , t ') Ψ ( r , t ') dt '
0 . (123)

By taking the inner product with the weight vector w ( r ) on both sides of the neutron-balance
equation and the precursor-balance equation, the following is obtained:
   
w r ; v 1Ψr , t   p(t ) w r ; v 1Ψr , t  
dp(t ) d
dt dt
  
 p(t ) w r ; M (r , t )Ψ r , t  
   k max
 
p(t ) w r ; Fp (r , t )Ψr , t    k w r ; ξ k (r , t )
k 1 . (124)
Reactor Dynamics 27
     
w ( r ); ξ k (r , t )  p (t ) w (r ); Fdk ( r , t ) Ψ  r , t  
t
 
k w ( r ); ξ k ( r , t ) ( k  1,..., k max )
. (125)
Equations (124) and (125) can be processed into more elegant forms akin to the point-kinetics
equations. To do this, some quantities must be defined first which will prove to be generaliza-
tions of the same quantities defined for the point-kinetics equations.
5.3 Effective Generation Time, Effective Delayed-Neutron Fraction, and Dy-

namic Reactivity
The following quantities and symbols are introduced:
Total production operator
  
F (r , t )  Fp (r , t )  Fd ( r , t )
. (126)
Dynamic reactivity
     
w ( r ), F ( r , t )Ψ ( r , t )  w (r ), M ( r , t ) Ψ ( r , t )
 (t )    
w ( r ), F ( r , t )Ψ (r , t )
. (127)
Effective generation time
 
w (r ), v 1Ψ(r , t )
(t )    
w (r ), F (r , t )Ψ(r , t )
. (128)
Effective delayed-neutron fraction for delayed group k
  
w ( r ), Fdk ( r , t ) Ψ ( r , t )
 k (t )    
w ( r ), F ( r , t ) Ψ ( r , t )
. (129)
Total effective delayed-neutron fraction
kmax
 (t )    k (t )
k 1 . (130)
Group k (generalized) precursor population
 
Cˆ k (t )  w (r ), ξ k ( r , t )
. (131)
With the newly introduced quantities, Eqs. (124) and (125) can be rewritten in the familiar form
of the point-kinetics equations:
 (t )   (t ) kmax ˆ
p (t )  p(t )   k Ck (t )
(t ) k 1
 ˆ  (t )
Ck (t )  p (t ) k  k Cˆ k (t ) ( k  1,..., kmax )
t (t ) . (132)
Of course, to be able to define quantities such as the dynamic reactivity, the shape vector

Ψ ( r , t ) must be known or approximated at each time t. Different shape representations

Ψ ( r , t ) lead to different space-time kinetics models. All flux-factorization models alternate

between calculating the shape vector Ψ ( r , t ) and solving the point-kinetics-like equations (132)
for the amplitude function and the precursor populations. The detailed energy- and space-
dependent flux shape at each time t can subsequently be reconstructed by multiplying the
amplitude function by the shape vector.
5.4 Improved Quasistatic Model


The improved quasistatic (IQS) model uses an exact shape Ψ ( r , t ) . By substituting the formal
solution to the precursor equations (123) into the general neutron-balance equation (121), the
following equation for the shape vector is obtained:
dp (t ) 1     
v Ψ  r , t   p (t )  v 1Ψ  r , t     p (t )M (r , t ) Ψ  r , t  
dt t
 
p (t )Fp (r , t )Ψ  r , t  
   
kmax t
 k  ξ k (0)e   e k (t t ') p (t ')Fdk (r , t ') Ψ(r , t ')dt ' 

 k t
k 1  0  . (133)
The IQS model alternates between solving the point-kinetics-like equations (132) and the shape
equation (133). The corresponding IQS numerical method uses two sizes of time interval.
Because the amplitude function varies much more rapidly with time than the shape vector, the
time interval used to solve the point-kinetics-like equations is much smaller than that used to
solve for the shape vector. Note that, other than the time discretization, the IQS model and

method include no approximation. The actual shape of the weight vector w ( r ) is irrelevant.
5.5 Quasistatic Approximation

The quasistatic approximation is derived by neglecting the time derivative of the shape vector in
the shape-vector equation (133). The resulting equation, which is solved at each time step, is:
dp(t ) 1   
v Ψ  r , t    p(t )M (r , t )Ψ  r , t  
dt
     
kmax t
p(t )Fp (r , t )Ψ  r , t    k  ξ k (0)e  k t   e  k (t t ') p (t ')Fdk (r , t ')Ψ (r , t ')dt ' 
k 1  0 . (134)
The resulting shape is used to calculate the point-kinetics parameters, which are then used in
the point-kinetics-like equations (132). As in the case of the IQS model, Equation (134) is solved
in conjunction with the point-kinetics-like equations (132). Aside from the slightly modified
shape equation, the quasistatic model differs from the IQS model in the values of its point-
kinetics parameters, which are now calculated using an approximate shape vector.
5.6 Adiabatic Approximation

The adiabatic approximation completely does away with any time derivative in the shape
equation and instead solves the static eigenvalue problem at each time t:
Reactor Dynamics 29
  1  
M ( r , t ) Ψ  r , t   F (r , t ) Ψ(r , t )
k . (135)
the point-kinetics-like equations (132). As in the case of the IQS and quasistatic models, Equa-
tion (135) is solved in conjunction with the point-kinetics-like equations (132).
5.7 Point-Kinetics Approximation (Rigorous Derivation)

In the case of the point-kinetics model, the shape vector is determined only once at the begin-
ning of the transient (t=0) by solving the static eigenvalue problem:
  1  
M(r , 0)Ψ  r   F ( r , 0)Ψ(r )
k . (136)
the point-kinetics-like equations (132). Because the shape vector is not updated, only the
point-kinetics-like equations (132) are solved at each time t. In fact, they are now the true
point-kinetics equations because the shape vector remains constant over time. This discussion
has shown that the point-kinetics equations can also be derived for an inhomogeneous reactor,
as long as the flux is factorized into a shape vector depending only on energy and position and
an amplitude function depending only on time.
6 Perturbation Theory
It should be obvious by now that different approximations of the shape vector lead to different
values for the kinetics parameters. It is therefore of interest to determine whether certain
choices of the weight vector might maintain the accuracy of the kinetics parameters even when
approximate shape vectors are used. In particular, it would be interesting to obtain accurate
values of the dynamic reactivity, which is the determining parameter for any transient. The
issue of determining the weight function that leads to the smallest errors in reactivity when
small errors exist in the shape vector is addressed by perturbation theory. This section will
present without proof some important results of perturbation theory. The interested reader is
encouraged to consult [Rozon1998], [Ott1985], and [Stacey1970] for detailed proofs and
additional results.
6.1 Essential Results from Perturbation Theory

Perturbation theory analyzes the effect on reactivity of small changes in reactor cross sections
with respect to an initial critical state, called the reference state. These changes are called
perturbations, and the resulting state is called the perturbed state. Perturbation theory also
analyzes the effect of calculating the reactivity using approximate rather than exact flux shapes.
First-order perturbation theory states that the weight vector that achieves the best first-order
approximation of the reactivity (e.g., for the point-kinetics equations) when using an approxi-
mate (rather than an exact) flux shape is the adjoint function, which is defined as the solution to
the adjoint static eigenvalue problem for the initial critical state at t=0:
   
M* (r ,0)Ψ*  r ,0  F* (r ,0)Ψ*  r ,0 
. (137)
The adjoint problem differs from the usual direct problem in that all operators are replaced by
their adjoint counterparts. The adjoint A* of an operator A is the operator which, for any
 
arbitrary vectors Φ ( r , t ) and Ψ ( r , t ) , satisfies:
Φ, AΨ  A*Φ, Ψ
. (138)
The reactivity at time t can therefore be expressed as:
     
Ψ* (r , 0), F(r , t )Ψ(r , t )  Ψ* (r , 0), M(r , t )Ψ(r , t )
 (t )    
Ψ* (r , 0), F(r , t )Ψ(r , t )
. (139)
The remaining point-kinetics parameters can be expressed similarly using the initial adjoint as
the weight function.
An additional result from perturbation theory states that when the adjoint function is used as
the weight function, the reactivity resulting from small perturbations applied to an initially
critical reactor can be calculated as:
     
Ψ* (r ,0), F(r , t )Ψ(r ,0)  Ψ* (r ,0), M(r , t )Ψ(r ,0)
 (t )    
Ψ* (r ,0), F(r ,0)Ψ(r ,0)
, (140)
where the  symbols represent perturbations (changes) in the respective operators with respect
to the initial critical state. Equation (140) offers a simpler way of calculating the reactivity than
Eq. (139) because it does not require recalculation of the shape vector at each time t. Note
that, within first-order of approximation, the calculated reactivity is also equal to the static
reactivity at time t, defined as:
1
 (t )  1 
keff (t )
. (141)
In fact, perturbation theory can also be used to calculate the (static) reactivity when the initial
unperturbed state is not critical. In that case, the change in reactivity is calculated as:
1
Ψ*0 , 0
 FΨ 0  Ψ*0 ,  MΨ 0
1 1 keff
  0
 
keff keff Ψ*0 , FΨ 0
. (142)
In Eq. (142), the “0” subscript or superscript denotes the unperturbed state. Finally, for one-
energy-group representations, the direct flux and the adjoint function are equal. It follows that
in a one-group representation, the reactivity at time t can be expressed as:
     
Ψ ( r , 0),  F ( r , t ) Ψ ( r , 0)  Ψ ( r , 0),  M ( r , t ) Ψ ( r , 0)
 (t )     
Ψ ( r , 0), F ( r , 0) Ψ ( r , 0)

  2 ( r , 0)  f   a  dV
Vcore

  2 ( r , 0) f dV
Vcore
. (143)
Reactor Dynamics 31
More generally, the static reactivity change between any two states, which is the equivalent of
Eq. (142), can be expressed using one-group diffusion theory as:
  1 

 02 (r )  0  f   a  dV
 keff 
1
  0 
1
 core
V  
keff keff 2 
  0 (r ) f 0 dV
Vcore
. (144)
6.2 Device Reactivity Worth

Reactivity devices are devices, usually rods, made of material with high neutron-absorption
cross section. By inserting or removing a device, the reactivity of the reactor can be changed,
and hence the power can be decreased or increased. The reactivity worth of a device is defined
as the difference between the reactivity of the core with the device inserted and the reactivity
of the same core with the device removed. Looking at this situation through a perturbation-
theory lens, the reactor without the reactivity device can be regarded as the unperturbed
system, and the reactor with the reactivity device can be regarded as the perturbed system.
Perturbation theory offers interesting insights into reactivity worth. Considering a device that is
inserted into a critical reactor and which, after insertion, occupies volume Vd in the reactor,
according to the perturbation formula for reactivity, the reactivity worth of the device is:
 1 
  keff
 2
0  0   fd    f 0     ad   a 0   dV
1 1 Vd  
d  0  d 
keff keff
 0 f 0 dV
2
Vcore
. (145)
Note that the integral in the numerator is over the device volume only and that the integral in
the denominator does not change as the device moves, thus simplifying the calculations.
Moreover, if two devices are introduced, their combined reactivity worth is:
 1 
2
 
 0  keff0 fd1 f 0
          ad 1   
a 0  dV
d 1 d 2 
Vd 1   
   f 0 dV
2
0
Vcore
 1 
V  keff
 2
0  0   fd 2    f 0     ad 2   a 0   dV
 
d2

 0 f 0 dV
2
Vcore
 d 1   d 2 . (146)
The interpretation of this equation is that as long as devices are not too close together and do
not have too large reactivity worths (so that the assumptions of perturbation theory remain
valid), their reactivity worths are additive.
7 Fission-Product Poisoning
Poisons are nuclides with large absorption cross sections for thermal neutrons. Some poisons
are introduced intentionally to control the reactor, such as B or Gd. Some poisons are produced
as fission products during normal reactor operation. Xe and Sm are the most important of
these.
7.1 Effects of Poisons on Reactivity

The effect of poisons on a reactor will be studied for a simple model of a homogeneous reactor
modelled using one-group diffusion theory. For such a reactor, in a one-energy-group formal-
ism:
 f
keff0 
 a 0  DBg2
. (147)
Uniform concentration
If a poison such as Xe with microscopic cross section  ax is added with a uniform concentration
(number density) X, the macroscopic absorption cross section increases by:
 aX  X aX . (148)
The total macroscopic absorption cross section is now:
 a   a 0   aX , (149)
and the new effective multiplication constant is:
 f  f
keff  
a  DBg a 0  aX  DBg2
2
. (150)
Addition of the poison induces a change in reactivity:
 1   1  1 1
     0   1     1    0  
 keff   keff  keff keff
0
   
 a 0  DB 2  a 0   aX  DB 2
 
 f  f
 aX X  aX
 
 f  f
. (151)
To calculate the reactivity inserted by the poison, the concentration of poison nuclei, X, must
first be determined.
Non-uniform concentration
In the case of non-uniform poison concentration, the perturbation formula for reactivity can be
Reactor Dynamics 33
used:
     
  (r ) (r )dV   (r )   (r ) X (r )
2 2 2
a aX (r )dV aX dV
1 1
     V  V
 V
k0 k      
  (r ) f (r )dV   (r ) f (r )dV   (r )
2 2 2
f (r )dV
V V V . (152)
It can easily be seen that if the distribution is uniform, the previous formula is recovered:

 aX   2 (r )dV
1 1  aX
    V

k0 k   f
 f   2 (r )dV
V . (153)
In the next sub-section, specific aspects of fission-product poisoning will be illustrated for the
case of Xe.
7.2 Xenon Effects

135
7.2.1 Xe production and destruction
The mechanisms of 135Xe production and destruction are illustrated in Fig. 2.
absorption of a neutron

    
 ,T1/2 1sec  ,T1/2 11sec  ,T1/2  6.7 h  ,T1/2 9.2 h  ,T1/2  2.3 x10 y 6
135
Sb  135
Te   135 I   135 Xe   135Cs   135 Ba (stable )
   
fiss fiss fiss fiss
Figure 2 135Xe production and destruction mechanisms

Because 135Sb decays very rapidly into 135Te, which in turn decays very rapidly into 135I, as an
approximation, 135I can be considered to be produced directly from fission. Because 135Cs has a
very long half-life, as an approximation, it can be considered stable. As a consequence of these
approximations, a simplified 135Xe production and destruction scheme can be used, as illus-
trated in Fig. 3.
absorption of a neutron

 
 ,T1/2  6.7 h  ,T1/2 9.2 h
135
I   135 Xe   135Cs ( stable)
 
fiss fiss
Figure 3 Simplified 135Xe production and destruction mechanisms
7.2.2 Determining the Xe concentration

To find the numerical density of Xe nuclei, the balance equation for iodine nuclei is first written
as:
dI
  I  f   I I
dt , (154)
where  is called the fission product yield and equals the average number of I nuclides created
per fission. The balance equation for Xe nuclei can be subsequently written as:
dX
 I I   X  f    X X   aX  X
dt . (155)
Steady-state conditions
Equilibrium conditions are attained after the reactor operates for a very long time (  ) at a
steady-state flux level ss . Under equilibrium conditions, the concentration of I nuclei is easily
found to be:
 I  f  ss
I 
I . (156)
Similarly, the Xe concentration can be determined as:
I I    X  f  ss ( I   X ) f  ss
X  
X   aX  ss X   aX  ss . (157)
Note that both I and Xe concentrations depend on flux level. However, whereas the I concen-
tration increases indefinitely with flux level, the Xe concentration levels off, and it can, at most,
become equal to:
( I   X ) f
X max 
 aX . (158)
The Xe macroscopic absorption cross section is:
( I   X ) f  ss aX
 aX  X  aX 
X   aX  ss . (159)
Using the notation:
X
X   0.770 1013 cm2 sec1
 aX , (160)
the Xe macroscopic absorption cross section can be rewritten as:
( I   X ) f  ss
 aX 
 X   ss . (161)
Reactor Dynamics 35
If Xe is assumed to be uniformly distributed, then its resulting reactivity worth is:
 aX 1 ( I   X ) f  ss    X   ss
 Xe     I
 f  f  X   ss   X   ss
. (162)
For high reactor fluxes, in the case where  ss   X ,  X can be neglected in the denomina-
tor, and the reactivity becomes independent of the flux level and equal to its maximum value of:
I  X
 Xe  
 . (163)
This is to be expected given that the Xe concentration has been found to “saturate” with in-
creased flux. The reactivity expressed in Eq. (163) is nothing but the corresponding reactivity
for the maximum Xe concentration shown in Eq. (158).
If, on the contrary, the flux is very low, in the case where  ss   X , then  SS can be neglected
in the denominator, and the Xe equivalent reactivity increases linearly with flux level:
 I   X   ss
 Xe  
 X . (164)
Xe load after shutdown: reactor dead time
If, the reactor is shut down (   0 ), I and Xe production from fission ceases, as well as Xe
destruction through neutron absorption. The concentration of I begins to decrease exponen-
tially due to decay. If the I concentration at the time of shutdown is I0, the I concentration as a
function of time can be expressed simply as:
I (t )  I 0 e  I t
. (165)
Substituting this expression into the Xe balance equation and setting the flux to zero leads to
the following expression:
dX
 I I 0 e  I t   X X
dt . (166)
Denoting the Xe concentration at the time of shutdown by X0, the solution can be written as:
I I 0
X (t )  X 0e X t  (e  X t  e  I t )
I  X . (167)
If the reactor is shut down after operating for a long time at steady state, the resulting Xe
concentration is:
( 1   X ) f  ss  I  f  ss  X t  I t
X (t )  e  X t  (e e )
X   aX  ss 1  X . (168)
The equivalent reactivity for uniformly distributed Xe (and assuming that the reactor was shut
down after operating for a long time at steady state) is:
1  (   X ) ss  X t   
   I e  I ss (e  xt  e  I t ) 
   X   ss I   X , (169)
where:
I
I   1.055  1013 cm 2 sec 1
 aX . (170)
The Xe concentration, and consequently the Xe reactivity worth after shutdown, increases at
first because Xe continues to be produced by decay of I, whereas consumption is now reduced
in the absence of Xe destruction by neutron absorption. After a while, however, the Xe concen-
tration reaches a maximum, starts decreasing, and eventually approaches zero. This behaviour
is shown in Fig. 4, which shows the Xe reactivity worth after shutdown from full power.
Figure 4 135Xe reactivity worth after shutdown

Because the Xe reactivity (or load) increases after shutdown, a reactor that was critical at the
time of shutdown subsequently becomes sub-critical and cannot be restarted until the Xe load
drops back to a value close to its steady-state level. The time during which the reactor cannot
be restarted due to increased Xe load after shutdown is known as reactor “dead time”. Given
the Xe half-life of approximately nine hours, the reactor dead time, which spans several half-
lives, is of the order of 1.5–2 days. Some reactors have systems to compensate for some of the
shutdown Xe load, in the form of reactivity devices that are inserted in the core during normal
Reactor Dynamics 37
steady-state reactor operation. As Xe builds up after shutdown, removal of these devices can
counterbalance the Xe reactivity load, enabling the reactor to be brought to critical and re-
started. Adjuster rods in the CANDU reactor can serve this purpose, but only up to 30 minutes
after shutdown. Beyond 30 minutes, the Xe load becomes larger than the adjuster-rod reactiv-
ity worth. Because the Xe load increases with the neutron flux, Xe-poison dead time generally
affects only high-power reactors.
8 Reactivity Coefficients and Feedback

Macroscopic cross sections can change as a consequence of various parameters, and in turn,
these changes induce changes in keff and hence in reactivity. The usual parameters that influ-
ence reactivity are:
 fuel temperature
 coolant temperature
 moderator temperature
 coolant density.
Changes in reactivity induced by changes in any such parameter are referred to as the reactivity
effect of the respective parameter. For example, the reactivity change induced by a change in
fuel temperature is called the fuel-temperature reactivity effect. The derivative of the reactivity
with respect to any of the parameters, with the others kept constant (i.e., the partial derivative),
is called the reactivity coefficient of that parameter. To illustrate this, assume that all parame-
ters are kept constant with the exception of one, e.g., fuel temperature, which is varied. As-
sume further that reactivity is plotted as a function of the variable parameter, in this case fuel
temperature. The plot in question would be a plot of  (T f ) . If a certain fuel temperature T f 0 is
taken as a reference, then the effect on reactivity of deviations from T f 0 can be calculated,
namely  (T f )   (T f )   (T f 0 ) .  (T f ) is called the fuel-temperature reactivity effect. The
d  (T f )
derivative of the reactivity with respect to the fuel temperature, namely  T f (T f )  , is
dT f
called the fuel-temperature coefficient. Of course, because reactivity also depends on other
parameters, it becomes clear that this derivative should be a partial derivative. In general, if the
reactivity depends on several parameters,
   ( p1 ,... pn ) , (171)
then the reactivity coefficient with respect to a parameter pi is defined as the partial derivative
of the reactivity with respect to that parameter:
 ( p1 ,... pn )
 pi 
pi . (172)
Of course, in reality, several parameters may vary simultaneously, and their variations may be
impossible to decouple. For example, the moderator density also varies when the moderator
temperature varies, due to thermal expansion. In this case, the single-parameter reactivity
coefficients are somewhat artificial, and additional combined reactivity coefficients can be
defined, which are of more practical use. For example, in the case of simultaneous variation of
both moderator temperature Tm and moderator density dm, a much more useful quantity would
be:
d   (Tm , d m )  (Tm , d m ) dd m
 Tm ( dm )   
dTm Tm d m dTm , (173)
dd m
where the derivative is specified by the thermal expansion law.
dTm
Such coefficients are called combined reactivity coefficients. One very useful combined coeffi-
cient of this kind is the power coefficient of reactivity (PCR), which is defined as:
d n
 ( p1... pn ) dpi
PCR  
dP i 1 pi dP , (174)
where the pi are all the parameters which change with power and on which the reactivity
ultimately depends, such as fuel temperature, coolant temperature, coolant density, and so on.
Because the core parameters depend on the flux level and because they influence the reactivity,
which in turn influences the flux, the reactivity coefficients are said to express the feedback
which describes the connection between the flux level and the cross-section values. When such
feedback is accounted for during a transient by recalculating the cross sections and the reactiv-
ity as functions of the flux level (and hence of power level), it is said that kinetics calculations
with feedback, or dynamic calculations, are being performed.
This sub-section will close with a presentation of a few alternate expressions for reactivity
coefficients. If reactivity is expressed using the effective multiplication constant, then the
reactivity coefficient can be expressed as follows:
d ( p) d  1  1 dk ( p)
P   1    2
dp dp  k ( p)  k ( p) dp
. (175)
For reactors close to critical ( k  1 ), this can be processed to yield:
1 dk ( p ) 1 dk ( p )
P  
k 2 ( p ) dp k ( p ) dp . (176)
The last form can also be expressed as:
 lnk ( p)
1 dk ( p) d
k ( p) dp dp . (177)
The last two expressions are often used to calculate reactivity coefficients.
9 CANDU-Specific Features
Given the presentation of the basic concepts of reactor kinetics in previous sections, this section
will be devoted to presenting how some of these concepts apply to CANDU reactors.
Reactor Dynamics 39
9.1 Photo-Neutrons: Additional Delayed-Neutron Groups

CANDU reactors are heavy-water cooled and moderated. In such reactors, neutrons can be
produced by the interaction of gamma rays (with a minimum energy of 2.22 MeV) with deute-
rium:
2
1 D    11H  n . (178)
Because gamma rays can be emitted by fission products with certain delays, the process is very
similar to that through which a “true” delayed neutron is emitted by an emitter which is the
daughter of a precursor. However, note than not all gamma rays will interact by photo-neutron
emission. Effective precursor concentrations can be defined for the photo-neutrons, such that
the photo-neutron production rate density is equal to:
S pn   pnC pn
. (179)
The term effective photo-neutron precursor concentration is used because the effective concen-
trations must also account for the fact that not all emitted gamma rays will result in the produc-
tion of a photo-neutron and that the fraction of photo-neutrons emitted depends on the
geometrical arrangement of the core lattice. Photo-neutron precursors can be grouped by their
decay constant, similarly to “real” precursors. It is customary to use 11 photo-neutron groups,
for a total of 17 delayed-neutron groups. Once the photo-neutron groups have been defined,
photo-neutrons are treated no differently than regular delayed neutrons in the kinetics calcula-
tions.
9.2 Values of Kinetics Parameters in CANDU Reactors: Comparison with LWR

and Fast Reactors
Deuterium has a much smaller neutron-absorption cross section than hydrogen. Consequently,
CANDU reactors have a better thermalized spectrum and hence a much longer generation time
(and lifetime) than light-water reactors and even longer compared to fast reactors. The typical
generation time of a CANDU core is approximately 1 ms, compared to 0.05 ms for an LWR core.
This makes CANDU transients “slower” than LWR transients. Reactivities close in value to the
delayed-neutron fraction induce less peak power transients in a CANDU core than in an LWR
core. The difference is even larger when comparison is made with fast reactors.
9.3 CANDU Reactivity Effects

CANDU reactivity effects depend on the specific characteristics of the CANDU lattice. Of par-
ticular interest is the coolant density effect, also known as the coolant void reactivity effect,
which is absent in other types of reactors which do not separate the coolant from the modera-
tor. Plots of the various effects are shown in Figs. 5 to 8 for fresh fuel.
Figure 5 CANDU fuel-temperature effect

The decrease in reactivity with increased fuel temperature is due primarily to an increase in
resonance absorption due to Doppler resonance broadening.
Figure 6 CANDU coolant-temperature effect
Reactor Dynamics 41
Figure 7 CANDU moderator-temperature effect
Figure 8 CANDU coolant-density effect

It is apparent that reactivity decreases with coolant density, which means that it increases with
void fraction. This effect occurs primarily because when coolant is lost, effective moderation is
still possible due to the moderator in the calandria vessel. Overall, the reduced slowing-down in
the coolant leads to reduced resonance absorption and an increased fast fission rate. Both
these phenomena increase reactivity when coolant is lost. As fuel burns, a mitigating factor
appears in the form of the low-lying Pu fission resonance, which begins to play a role as Pu is
created. Reduced upscattering in the coolant reduces fissions in the low-lying Pu resonance and
hence reduces reactivity, although not enough to make it negative.
9.4 CANDU Reactivity Devices

As in any reactor, power in CANDU reactors is controlled by controlling reactivity. In turn,
reactivity is controlled by means of reactivity devices which can be inserted into or removed
from the core. By inserting or removing reactivity devices from the core, the absorption rate is
varied; hence, reactivity can be varied, and power can be increased or decreased, or the reactor
can be completely shut down. Reactivity devices in CANDU reactors come under the control
either of the Reactor Regulating System (RRS) or of one of the two independent shutdown
systems (SDS1 and SDS2).
For a CANDU 6, the reactivity devices under the control of the RRS are as follows:
 14 liquid-zone-control compartments (H2O-filled)
 21 adjuster rods
 4 mechanical control absorbers
 moderator “poison”.
The reactivity devices under the control of the shutdown systems are:
SDS-1: 28 cadmium shutoff rods which fall into the core from above
SDS-2: high-pressure Gd-poison injection into the moderator through six horizontally oriented
nozzles.
The reactivity worth of these reactivity devices is shown in Table 2.
Table 2 CANDU reactivity device worth
Total Reactivity Maximum Reactivity

Function Device
Worth (mk) Rate (mk/s)
14 liquid zone con-
Control 7 ± 0.14
trollers
Control 21 adjusters 15 ± 0.10
4 mechanical control ± 0.075 (driving)
Control 10
absorbers -3.5 (dropping)
Control moderator poison – -0.01 (extracting)
Safety 28 shut-off units 80 -50

6 poison-injection
Safety > 300 -50
nozzles
Reactor Dynamics 43

Chapter 5 relies on knowledge acquired from Chapter 4, Reactor Statics and indirectly on
knowledge from Chapter 3, Nuclear Processes and Neutron Physics.
11 Problems
1. 1020 nuclei of 235U undergo fission with a delayed neutron yield of 0.0125 and a delayed
neutron fraction of 0.005.
a) What is the total neutron yield?
b) How many precursors are produced?
c) How many emitters are eventually produced?
2. A radioactive waste site consists of two cylindrical tanks that contain liquid waste in the
form of fissile material in an aqueous solution. The tanks are in the form of cubes with the
side equal to 1m. The neutronic properties of the radioactive waste are: SA=0.00100cm-1,
D=1cm, nu=2.5, SF=0.00158 cm-1, v=2200 m/s. Assume all fission neutrons are prompt.
a) Calculate the reactivity, generation time and neutron life time for one of the tanks.
b) The site manager decides to save space and money, by storing the content of both tanks
in a larger cubical tank, with side 3 2m . Calculate the reactivity, generation time and
neutron life time for the new tank. Comment on the result.
3. A thin foil made of a mixture of isotopes one of which is fissile has a fission macroscopic
cross section equal to 0.001 cm-1. When fissioning, the fissile isotope produces two types of
precursors: one with yield 0.05 and a half-life of one minute, and another with yield 0.03
and a half-life of two minutes. The sample is subjected to a pulse of neutrons at t=0, and to
another pulse of neutrons at t=90s. The first pulse has a fluence of 108 n/cm2, and the
second pulse has a fluence of 5x107 n/cm2. What is the total number of precursors at t=5
minutes?
Note: Assume that the number of nuclei that react with neutrons (through fission or oth-
erwise) is negligible compared to the original number of nuclei present in the sample, and
that neutrons emitted from the sample do not interact in the sample.
4. A thin-foil of fissile material is irradiated uniformly in a neutron flux of 108 n/cm2/s, starting
at t=0. There are 1022 fissile nuclei in the sample and the fission microscopic cross section is
2000b. There is only one group of delayed neutrons. The total neutron yield per fission is
2.33. The delayed neutron fraction is 0.001 and the half-life of precursors is one minute.
The number of neutrons emitted by the sample per second is measured with a neutron de-
tector.
a) What does the detector indicate 20 minutes after the start of the irradiation?
b) At 25 minutes the irradiation stops. What does the detector indicate two minutes later?
Note: Assume that the number of nuclei that react with neutrons (through fission or
otherwise) is negligible compared to the original number of nuclei present in the sam-
ple, and that neutrons emitted from the sample do not interact in the sample.
5. Consider a homogeneous nuclear reactor for which all neutrons are born prompt. The
reactor is cubical, with side equal to 4m. The neutronic parameters of the reactor are:
D  1cm
 a  0.003cm 1
  2.5
v  2200m / s
The reactor is initially critical, operating at 3000 MW fission power. The energy liberated per
fission is approximately 200MeV. The extrapolated size of the reactor can be approximated
by its physical size.
a) Find  f .
b) Calculate the volume-integrated flux in the reactor.
c) Calculate total neutron population in the reactor.
d) Calculate the neutron generation time and life time.
6. Consider a slab homogeneous reactor (infinite in the y and z directions and finite in the x
direction) extending from -2m to 2m in the x direction, and with the following parameters:
D  1cm
 a  0.003cm 1
  2.5
Assume that the physical length and the extrapolated length of the reactor can be approx-
imated to be equal.
The reactor is initially critical.
a) Find the fission cross section.
b) A control plate, 1 cm thick (and extending to infinity in the y and z directions, just like
the reactor) is introduced at x=1m. The neutronic parameters of the control plate are:
D  1cm
 a  0.01cm 1
 f  0cm 1
c) Assuming the control plate is thin-enough so that the unperturbed flux across it can be
assumed to be constant and equal to the value at the center of the plate, calculate the
reactivity of the reactor after the introduction of the control plate.
d) Note: When a plate is inserted in the reactor, its material displaces (that is replaces) the
unperturbed reactor material at the position of the plate.
Reactor Dynamics 45
7. Consider a reactor with 6 delayed neutron groups, with the following parameters.
v  2200 m / s
  0.001 s
  0.000
  2.5
1  0.0010
 2  0.0012
3  0.0008
 4  0.0002
5  0.0011
 6  0.0014
1  0.001 s 1
2  0.005 s 1
3  0.010 s 1
4  0.015 s 1
5  0.050 s 1
6  0.100 s 1
The reactor operates in steady state at 2000 MW. The energy per fission is approximately
200MeV.
What is the total delayed neutron fraction?
What is the total delayed neutron yield?
What is the precursor population for each of the groups 1 to 6?
What is the total delayed neutron source?
8. Consider a homogeneous nuclear reactor with one delayed neutron group. The reactor is
cubical, with side equal to 4m. The neutronic parameters of the reactor are:
D  1cm
 a  0.003cm1
 f  0.00136cm1
  2.5
v  2200m / s
  0.005
  0.2s 1
Assume the extrapolated size of the reactor equals its physical size.
a) Calculate keff.
b) The reactor is maintained subcritical by the addition of 10B, which has a microscopic
cross section of approximately 4000b. ( 1 barn  1024 cm2 ). If k  0.980 , what is the
number density of Boron atoms?
c) An external neutron source is introduced into the reactor at point b). The neutron bal-
ance (point kinetics) equations are thus written:

n  n  C  S


C  n   C

where n is the total neutron population, and S is the strength of the external source (neu-
trons/s).
If S  10 6 neutrons , what is the equilibrium (steady-state) neutron population? What is the
s
equilibrium precursor population?
d) At t=0, the external neutron source is removed from the reactor. What are the neutron
population and the precursor population 5 seconds after the removal of the external
source?
9. Consider a reactor with six delayed neutron groups, with the following parameters:
  0.001s
  0.005
The reactor is initially operating at a steady-state power of 1000MW. A control rod that was
initially in the core is accidentally ejected at t=0, yielding a 2mk reactivity increase. What is
the reactor power immediately after the rod ejection? Use the prompt jump approxima-
tion.
10. Consider a reactor with one delayed neutron group, with the following parameters:
  0.001s
  0.005
  0.02s 1
The reactor is initially in steady state operation at a power of 1000MW. A control rod with a
reactivity worth of 2 mk is inserted in the reactor at t=0. At t=2s, a second, identical, control
rod is inserted. What is the reactor power at t=4s?
Notes:
Use the prompt jump approximation.
The prompt jump approximation is also valid when the reactor is not initially critical.
12 References and Further Reading

[Lamarsh2001] Lamarsh, J. R. and Baratta, A. J., Introduction to Nuclear Engineering (third
edition). Prentice-Hall, 2001.
[Lamarsh2002] Lamarsh, J. R., Introduction to Nuclear Reactor Theory. American Nuclear Soci-
ety, La Grange Park, Illinois, 2002.
[Ott1985] Ott, K. O., Neuhold, R. J., Nuclear Reactor Dynamics. American Nuclear Society,
Lagrange Park, IL, 1985.
[Rose1991] P. F. Rose (ed.), ENDF/B-VI Summary Documentation. Cross Section Evaluation
Working Group, Report BNL-NCS-17541 (ENDF-201), 1991.
[Rozon1998] Rozon, D., Nuclear Reactor Kinetics. Polytechnic International Press, Montreal, QC,
1998.
Reactor Dynamics 47
[Stacey1970] Stacey, W. M., Space-Time Nuclear Reactor Kinetics, Academic Press, 1970.
13 Acknowledgements
ments during the development of this Chapter. Their feedback has much improved it. Of
course the responsibility for any errors or omissions lies entirely with the authors.
Marv Gold
Bruce Wilkin
1
CHAPTER 6
Thermal-Hydraulic Design
Prepared by
Dr. Nikola K. Popov
Summary
This chapter covers the thermal-hydraulic design of nuclear power plants with a focus on the
primary and secondary sides of the nuclear steam supply system. This chapter covers the
following topics: evolution of the reactor thermal-hydraulic system; key design requirements for
the heat transport system; thermal-hydraulic design principles and margins; design details of
the primary and secondary heat transport systems; fundamentals of two-phase flow;
fundamentals of heat transfer and fluid flow in the reactor heat transport system; other related
topics.
©UNENE, all rights reserved. For educational use only, no assumed liability. Thermal-Hydraulic Design – December 2015
Table of Contents
1 Introduction........................................................................................................................... 10
1.1 Overview....................................................................................................................... 10
1.2 Learning outcomes........................................................................................................ 12
1.3 Summary of relationship to other chapters ................................................................... 12
1.4 Thermal-hydraulic design ............................................................................................. 12
2 Reactor Types ....................................................................................................................... 14
2.1 CANDU reactor design................................................................................................. 14
2.1.1 Reactor core and calandria vessel ............................................................................. 17
2.1.2 Primary heat transport system design ....................................................................... 19
2.1.3 Steam generators....................................................................................................... 20
2.1.4 Pressurizer................................................................................................................. 21
2.1.5 Primary pumps .......................................................................................................... 22
2.1.6 Primary heat transport piping.................................................................................... 22
2.1.7 Secondary heat transport system design ................................................................... 22
2.1.8 Turbine...................................................................................................................... 23
2.1.9 Condenser ................................................................................................................. 23
2.1.10 Heat exchangers and pumps.................................................................................. 23
2.2 Problems ....................................................................................................................... 23
3 CANDU Thermal-Hydraulic Design Evolution ................................................................... 25
3.1 CANDU reactor evolution ............................................................................................ 25
3.1.2 Steam generators....................................................................................................... 27
3.1.3 Heat transport pumps ................................................................................................ 28
3.1.4 Reactor core .............................................................................................................. 28
3.1.5 Reduction in radiation exposure ............................................................................... 29
3.2 CANDU reactor types................................................................................................... 30
3.2.1 Nuclear Power Demonstration station ...................................................................... 30
3.2.2 Douglas Point............................................................................................................ 31
3.2.3 Pickering A and B..................................................................................................... 32
3.2.4 Bruce A and B........................................................................................................... 33
3.2.5 CANDU 6 ................................................................................................................. 34
3.2.6 Darlington ................................................................................................................. 34
3.2.7 Advanced CANDU designs ...................................................................................... 35
3.3 Non-PHWR CANDU designs....................................................................................... 37
3.3.1 CANDU-BLW .......................................................................................................... 37
3.3.2 CANDU-OCR........................................................................................................... 38
3.4 Problems ....................................................................................................................... 39
4 Thermal-Hydraulic Design Requirements ............................................................................ 40
4.1 Fuel requirements.......................................................................................................... 40
4.1.1 Metallic fuels ............................................................................................................ 42
4.1.2 Ceramic fuels ............................................................................................................ 42
4.1.3 Dispersion fuels ........................................................................................................ 43
4.2 General fuel sheath (cladding) requirements ................................................................ 44
4.2.2 Zirconium.................................................................................................................. 45
4.3 Reactor coolant requirements ....................................................................................... 46
4.3.1 Ordinary water and heavy water coolants................................................................. 46
Thermal-Hydraulic Design 3
4.3.2 Gaseous coolants....................................................................................................... 48

4.3.3 Liquid metal coolants................................................................................................ 48
4.4 Moderator requirements................................................................................................ 48
4.5 Control material requirements ...................................................................................... 50
4.5.1 Hafnium .................................................................................................................... 50
4.5.2 Cadmium alloys ........................................................................................................ 50
4.5.3 Rare-earth oxides ...................................................................................................... 50
4.5.4 Gadolinium nitrate .................................................................................................... 51
4.5.5 Boron-containing materials....................................................................................... 51
4.6 HTS design requirements and engineering considerations........................................... 52
4.7 Problems ....................................................................................................................... 54
5 Thermal-Hydraulic Design Limits and Margins................................................................... 55
5.1 Heat generation parameter definitions .......................................................................... 55
5.2 Thermal design limits ................................................................................................... 55
5.3 Thermal design margins................................................................................................ 57
5.4 Problems ....................................................................................................................... 58
6 Thermal-Hydraulic Design Fundamentals ............................................................................ 59
6.1 Two-phase flow fundamentals...................................................................................... 59
6.1.1 Key parameters and definitions ................................................................................ 59
6.1.2 Non-dimensional numbers ........................................................................................ 62
6.1.3 Flow patterns............................................................................................................. 65
6.1.4 The boiling process ................................................................................................... 70
6.1.5 Two-phase flow models............................................................................................ 74
6.1.6 Two-phase flow in fuel bundles................................................................................ 75
6.1.7 Problems ................................................................................................................... 77
6.2 Thermodynamics of nuclear energy conversion ........................................................... 78
6.2.1 Definitions................................................................................................................. 78
6.2.2 First law of thermodynamics .................................................................................... 81
6.2.3 Reactor cycles ........................................................................................................... 83
6.2.4 Second law of thermodynamics ................................................................................ 86
6.2.5 Reactor power cycle.................................................................................................. 88
6.2.6 Improvements in reactor cycles ................................................................................ 93
6.2.7 Entropy and the laws of thermodynamics............................................................... 102
6.2.8 Problems ................................................................................................................. 104
7 Heat Transfer and Fluid Flow Design................................................................................. 106
7.1 Heat transfer in the primary heat transfer system ....................................................... 106
7.2 Primary pumps ............................................................................................................ 107
7.3 Steam generator heat balance...................................................................................... 111
7.3.1 Steam generator without a pre-heater region .......................................................... 113
7.3.2 Approximate solution to steam generator without a pre-heater region................... 114
7.3.3 Calculation of main HTS parameters...................................................................... 116
7.3.4 Steam generator with pre-heater (simplified analytical solution)........................... 118
7.3.5 Steam generator with pre-heater (numerical solution)............................................ 123
7.3.6 Problems ................................................................................................................. 125
7.4 Heat transfer in the fuel elements ............................................................................... 126
7.4.1 General heat conduction equation........................................................................... 126
7.4.2 Temperature distribution in the radial direction ..................................................... 127

7.4.3 Temperature distribution in the axial direction....................................................... 132
7.4.4 Axial distribution of quality.................................................................................... 136
7.4.5 Thermal conductivity and fuel element material properties ................................... 137
7.4.6 Problems ................................................................................................................. 141
7.5 Fluid flow fundamentals ............................................................................................. 145
7.5.1 Introduction to pressure losses................................................................................ 145
7.5.2 Pressure drop conservation equations..................................................................... 147
7.5.3 Correlations for calculating pressure losses............................................................ 149
7.5.4 Flow Instability ....................................................................................................... 166
7.5.5 Problems ................................................................................................................. 174
7.6 Heat transfer between the fuel elements and the coolant............................................ 176
7.6.1 Heat transfer regimes .............................................................................................. 176
7.6.2 Convective heat transfer in turbulent forced flow .................................................. 177
7.6.3 Sub-cooled boiling heat transfer in forced flow ..................................................... 178
7.6.4 Critical heat flux (CHF) .......................................................................................... 180
7.6.5 Transition and film boiling heat transfer ................................................................ 203
7.6.6 Single-phase heat transfer to vapour....................................................................... 217
7.6.7 Problems ................................................................................................................. 218
7.7 Critical flow ................................................................................................................ 221
7.8 Water hammer............................................................................................................. 221
7.8.1 Types of water hammer .......................................................................................... 222
7.8.2 Analytical models and computer codes for water hammer analysis....................... 226
7.8.3 Diagnostics and assessment of water hammer........................................................ 227
7.8.4 Prevention, mitigation, and accommodation of water hammer .............................. 228
7.8.5 Problems ................................................................................................................. 230
7.9 Natural circulation ...................................................................................................... 231
7.9.1 Natural circulation phenomenon............................................................................. 231
7.9.2 CANDU natural circulation .................................................................................... 235
7.9.3 CANDU natural circulation experimental studies .................................................. 240
7.9.4 Natural circulation calculations and predictions..................................................... 241
7.9.5 Problems ................................................................................................................. 243
8 References........................................................................................................................... 244
9 Further Reading .................................................................................................................. 247
10 Glossary .............................................................................................................................. 248
11 Nomenclature...................................................................................................................... 249
12 Acknowledgements............................................................................................................. 250
13 APPENDIX A – Reactor Types.......................................................................................... 251
13.1 General evolution of reactor designs .......................................................................... 251
13.2 Pressurized water reactors........................................................................................... 252
13.2.1 General information ............................................................................................ 252
13.2.2 Reactor vessel ..................................................................................................... 254
13.2.3 Steam generators................................................................................................. 256
13.2.4 Primary pumps .................................................................................................... 256
13.2.5 Pressurizer........................................................................................................... 257
13.2.6 Reactor refuelling and reactivity control ............................................................ 257
13.2.7 Fuel assembly and fuel pin design ...................................................................... 258

13.2.8 Reactor core design............................................................................................. 262
13.2.9 WWER reactor design ........................................................................................ 262
13.3 Boiling water reactors ................................................................................................. 263
13.4 Gas-cooled reactors..................................................................................................... 267
13.4.1 Magnox reactors.................................................................................................. 267
13.4.2 AGR reactors ...................................................................................................... 268
13.4.3 HTGCR Reactors ................................................................................................ 269
13.5 Channel-type reactors ................................................................................................. 269
13.5.1 SGHWR reactor .................................................................................................. 269
13.5.2 CANDU reactor .................................................................................................. 270
13.5.3 RBMK reactor..................................................................................................... 270
13.6 Fast breeder reactors ................................................................................................... 272
List of Figures
Figure 1 Heat engine concepts ..................................................................................................... 13
Figure 2 Typical CANDU plant ....................................................................................................... 15
Figure 3 CANDU 6 reactor cooling loops....................................................................................... 16
Figure 4 CANDU reactor, fuel channel, and fuel bundle .............................................................. 18
Figure 5 CANDU calandria vessel and reactor vault ..................................................................... 18
Figure 6 CANDU primary heat transport system .......................................................................... 20
Figure 7 Typical steam generator design...................................................................................... 21
Figure 8 Typical pressurizer design............................................................................................... 21
Figure 9 Typical primary pump design.......................................................................................... 22
Figure 10 CANDU steam generator design evolution................................................................... 28
Figure 11 Nuclear Power Demonstration (NPD) heat transport system ...................................... 31
Figure 12 Douglas Point heat transport system ........................................................................... 32
Figure 13 Pickering A heat transport system................................................................................ 32
Figure 14 Bruce heat transport system ........................................................................................ 33
Figure 15 Darlington heat transport system................................................................................. 34
Figure 16 ACR-1000 heat transport system.................................................................................. 36
Figure 17 CANDU BLW heat transport system ............................................................................. 37
Figure 18 CANDU OCR heat transport system.............................................................................. 38
Figure 19 Fuel rod design types .................................................................................................... 41
Figure 20 Critical heat flux ratio.................................................................................................... 56
Figure 21 Thermal margins ........................................................................................................... 58
Figure 22 Relationship between quality and void fraction [COL1972]......................................... 62
Figure 23 Flow patterns in vertical two-phase flow ..................................................................... 66
Figure 24 Flow patterns in horizontal two-phase flow................................................................. 66
Figure 25 Flow patterns and heat transfer regimes in horizontal heated tubes.......................... 67
Figure 26 Flow patterns and heat transfer regimes in vertical heated tubes .............................. 68
Figure 27 Flow pattern map in vertical flow [HET1982]............................................................... 69
Figure 28 3-D representation of boiling surfaces ......................................................................... 71
Figure 29 Boiling curve.................................................................................................................. 71

Figure 30 Parametric trends of the boiling curve ......................................................................... 72
Figure 31 Quality effects on the boiling curve.............................................................................. 73
Figure 32 Heat transfer regimes in a boiling horizontal channel ................................................. 73
Figure 33 Details of sub-cooled boiling phases in horizontal flow ............................................... 74
Figure 34 Effect of tube diameter on quality and void fraction ................................................... 76
Figure 35 Effect of bundle geometry on quality and void fraction .............................................. 76
Figure 36 Change of thermodynamic states in a system.............................................................. 78
Figure 37 Concept of parameters of thermodynamic state in piston geometry ......................... 79
Figure 38 Ideal Carnot cycle.......................................................................................................... 84
Figure 39 The concept of entropy in closed cycles....................................................................... 87
Figure 40 Carnot cycle for steam .................................................................................................. 89
Figure 41 Steam-water diagrams used in thermodynamic analyzes (a, b, c)............................... 90
Figure 42 Simplified Rankine cycle in a nuclear power plant ....................................................... 91
Figure 43 Typical impact of turbine outlet pressure on cycle thermodynamic efficiency ........... 92
Figure 44 Typical impact of turbine inlet pressure on cycle thermodynamic efficiency.............. 93
Figure 45 Rankine cycle with superheating .................................................................................. 94
Figure 46 Rankine cycle with superheating and reheating........................................................... 95
Figure 47 Rankine cycle with regeneration circuit ....................................................................... 96
Figure 48 Rankine cycle with open feedwater reheating............................................................. 97
Figure 49 Rankine cycle with multiple open reheaters ................................................................ 97
Figure 50 Rankine cycle with closed feedwater reheating........................................................... 98
Figure 51 Rankine cycle with multiple closed reheaters .............................................................. 99
Figure 52 Rankine cycle with dryer............................................................................................. 100
Figure 53 Real Rankine cycle with losses.................................................................................... 101
Figure 54 Typical advanced PWR secondary cooling system ..................................................... 101
Figure 55 Typical CANDU 6 secondary cooling system............................................................... 102
Figure 56 Reactor cooling systems ............................................................................................. 106
Figure 57 Primary pump flow diagram ....................................................................................... 107
Figure 58 Primary pump operating quadrants ........................................................................... 109
Figure 59 Typical primary pump characteristics......................................................................... 110
Figure 60 Typical shutdown pump characteristics ..................................................................... 111
Figure 61 Reactor steam generator loop.................................................................................... 112
Figure 62 Steam generator simplified heat duty diagram.......................................................... 113
Figure 63 Steam generator heat duty diagram without pre-heating......................................... 113
Figure 64 Process for calculating primary loop parameters....................................................... 116
Figure 65 Steam generator heat duty diagram with pre-heating .............................................. 119
Figure 66 Reactor power versus heat-duty diagram of the SG .................................................. 123
Figure 67 Steam generator numerical calculation concept........................................................ 124
Figure 68 Nodalization concept for steam generator numerical calculation............................. 124
Figure 69 Cross section of a cylindrical fuel pin.......................................................................... 127
Figure 70 Radial temperature distribution in a fuel element..................................................... 131
Figure 71 Radial temperature gradient per unit thickness in a fuel pin..................................... 132
Figure 72 Temperature distribution in the axial direction in a fuel element ............................. 135
Figure 73 Temperature distribution in the fuel cladding and the fuel element ........................ 136
Figure 74 Thermal conductivity in a fuel pin .............................................................................. 138
Figure 75 Impact of oxygen-to-metal ratio on the heat conduction coefficient in fuel............. 138
Figure 76 Heat capacity of a fuel pin .......................................................................................... 139
Figure 77 Gap conductance in a fuel pin .................................................................................... 139
Figure 78 Fuel aging effect on temperature in a fuel pin ........................................................... 140
Figure 79 Impact of gap thickness and linear power on heat conductance............................... 140
Figure 80 Asymmetric temperature profiles in a fuel pin .......................................................... 141
Figure 81 Force-momentum balance in a control volume ......................................................... 147
Figure 82 Moody chart for friction factors ................................................................................. 152
Figure 83 Martinelli-Nelson two-phase multiplier in separated flow ........................................ 154
Figure 84 Single-phase pressure distribution over a square-edged orifice................................ 156
Figure 85 Two-phase pressure distribution over a square-edged orifice .................................. 156
Figure 86 Bundle correction factor (37-element hexagonal bundle) ......................................... 157
Figure 87 Eccentricity effect ....................................................................................................... 158
Figure 88 Eccentricity correction factor ..................................................................................... 158
Figure 89 Bundle junction misalignment effect.......................................................................... 160
Figure 90 Typical single-phase pressure distribution over misaligned bundles......................... 160
Figure 91 Typical pressure drop in misaligned fuel bundles ...................................................... 161
Figure 92 Typical misaligned junction pressure drop signatures ............................................... 161
Figure 93 Typical two-phase pressure distribution over misaligned bundles............................ 162
Figure 94 Pressure gradient along a flow channel in boiling...................................................... 163
Figure 95 Example of abrupt area changes and orifice .............................................................. 164
Figure 96 Pressure drop behaviour with abrupt area changes and orifice ................................ 164
Figure 97 Pressure drop components......................................................................................... 165
Figure 98 Concept of channel flow instability ............................................................................ 168
Figure 99 Channel flow instability under low heat flux .............................................................. 169
Figure 100 Impact of different pump curves on channel flow instability .................................. 170
Figure 101 Definition of channel flow excursion instability in a single channel ........................ 171
Figure 102 Principles of flow excursion instability in two parallel channels.............................. 173
Figure 103 Results of flow excursion instability in two parallel channels ................................. 174
Figure 104 Concepts of CHF, dryout, and burnout..................................................................... 181
Figure 105 Transitions among CHF mechanisms ........................................................................ 182
Figure 106 Pool boiling CHF mechanisms................................................................................... 183
Figure 107 Flow boiling CHF mechanisms .................................................................................. 183
Figure 108 Effect of inlet temperature and mass flux ................................................................ 186
Figure 109 Effect of outlet pressure in pool boiling ................................................................... 186
Figure 110 Effect of outlet pressure on flow boiling .................................................................. 187
Figure 111 Effect of heated length on local CHF ........................................................................ 187
Figure 112 Effect of quality on local CHF.................................................................................... 188
Figure 113 Effect of tube diameter on local CHF........................................................................ 188
Figure 114 Concept of operating margins .................................................................................. 194
Figure 115 Critical channel power .............................................................................................. 195
Figure 116 CHF margin definitions.............................................................................................. 196
Figure 117 CHF margin evaluation methods (a, b, c).................................................................. 198

Figure 118 CANDU bundle experimental simulator ................................................................... 200
Figure 119 Axial flux distribution in full-scale bundle test ......................................................... 200
Figure 120 CHF variation along a heated surface....................................................................... 200
Figure 121 Different types of spacer devices for CHF robustness.............................................. 201
Figure 122 CHF behaviour around fuel bundle spacer devices .................................................. 201
Figure 123 Types of transition boiling and film .......................................................................... 204
Figure 124 Concept of transition boiling in pool boiling ............................................................ 205
Figure 125 Minimum film pool boiling – Taylor instability......................................................... 206
Figure 126 Film boiling types ...................................................................................................... 207
Figure 127 Pool film boiling mechanisms with horizontal heated wall...................................... 208
Figure 128 Pool film boiling patterns.......................................................................................... 208
Figure 129 Forced convection film boiling patterns................................................................... 209
Figure 130 Film boiling heat transfer temperature and velocity profiles .................................. 210
Figure 131 Film boiling heat transfer mechanisms..................................................................... 211
Figure 132 Bundle dry patches in the dryout plane ................................................................... 215
Figure 133 Bundle circumferential temperature profiles........................................................... 216
Figure 134 Rapid valve operation ............................................................................................... 222
Figure 135 Flow of sub-cooled water with steam condensing in vertical pipe (water cannon) 223
Figure 136 Counter-current flow of steam and water in a horizontal pipe ............................... 223
Figure 137 Pressurized water flowing in a vertical pipe filled with steam................................. 224
Figure 138 Hot water entering a lower-pressure line ................................................................ 225
Figure 139 Steam-driven water slug in piping system................................................................ 225
Figure 140 Filling of voided piping system.................................................................................. 226
Figure 141 Principles of natural circulation ................................................................................ 231
Figure 142 Various types of natural circulation.......................................................................... 232
Figure 143 Natural circulation in CANDU ................................................................................... 236
Figure 144 Onset of bi-directional flow in CANDU ..................................................................... 238
Figure 145 Criteria for flow reversal in a CANDU core ............................................................... 238
Figure 146 Intermittent buoyancy-induced flow in CANDU....................................................... 239
Figure 147 Typical temperature histories in a CANDU channel in natural circulation............... 240
Figure 148 Reactor generations.................................................................................................. 251
Figure 149 PWR primary reactor cooling system ....................................................................... 254
Figure 150 Typical PWR reactor vessel ....................................................................................... 255
Figure 151 Typical PWR square fuel assembly design ................................................................ 258
Figure 152 Typical PWR fuel design............................................................................................ 259
Figure 153 Typical PWR fuel assembly support spacer plates ................................................... 260
Figure 154 Typical PWR reactor control sites............................................................................. 260
Figure 155 Typical PWR control assembly (cluster) design ........................................................ 261
Figure 156 Typical PWR control-rod design................................................................................ 261
Figure 157 Russian WWWR reactor design ................................................................................ 263
Figure 158 BWR reactor design .................................................................................................. 264
Figure 159 Typical BWR reactor vessel coolant circulation........................................................ 265
Figure 160 BWR reactor fuel assemblies in the core.................................................................. 265
Figure 161 BWR fuel and control assemblies ............................................................................. 266

Figure 162 Typical Magnox reactor design................................................................................. 267
Figure 163 Typical AGR reactor design ....................................................................................... 268
Figure 164 Typical HTGCR reactor design................................................................................... 269
Figure 165 SGHWR reactor ......................................................................................................... 270
Figure 166 RBMK reactor design ................................................................................................ 271
Figure 167 Fast reactor design.................................................................................................... 273
List of Tables
Table 1 Typical fuel characteristics for key reactor types ............................................................ 26
Table 2 CANDU main process parameters and features .............................................................. 27
Table 3 Evolution of D2O content in the core and of power in the fuel channel ......................... 29
Table 4 Summary of uranium fuel characteristics [POP2014]...................................................... 41
Table 5 Summary of fuel cladding characteristics [POP2014]...................................................... 44
Table 6 Summary of coolant characteristics [POP2014] .............................................................. 47
Table 7 Summary of moderator characteristics [POP2014] ......................................................... 49
Table 8 Factors affecting pressure drop ..................................................................................... 146
Table 9 CANDU distribution of pressure drop in the primary loop ............................................ 146
Table 10 Separate CHF effects in fuel bundles ........................................................................... 185
Table 11 Sample of CHF look-up table........................................................................................ 191
Table 12 Sample of PDO look-up table [IAEA2001] .................................................................... 214
Table 13 Typical PWR reactor parameters ................................................................................. 253
1 Introduction
Section 1 provides an introduction to thermal-hydraulic design. It defines the expectations and
learning outcomes for this chapter and indicates the relationship of this chapter to other
chapters in this textbook.
The objective of this chapter is to describe the generic thermal-hydraulic design of nuclear
reactors under normal operating conditions, with a specific focus on design details of CANDU
reactors.
This chapter covers the thermal-hydraulic design of a CANDU nuclear power reactor, with
general comparisons to other reactor types and designs. Thermal-hydraulic design covers the
reactor primary and secondary heat transport systems. In fact, the primary heat transfer
design defines the maximum power levels (globally and locally) that can be safely generated in
the reactor core and thus defines the design characteristics of many systems and components,
such as the reactor core physics and the fuel design.
1.1 Overview
This section describes the objectives, principles, and methodologies of reactor thermal-
hydraulic design. The thermal-hydraulic design of the reactor process systems that are
required to transport heat energy away from the nuclear reactor source and transform this
heat energy into useful work (generally electrical energy) are the focus of nuclear engineering
and of this chapter.
Section 2 presents the principles of reactor design, with a focus on CANDU reactor design.
Designs of other reactor types are described in Appendix A, in which the focus is on providing a
historical perspective on reactor thermal-hydraulic and systems design and on pressurized
water reactors (PWRs) and boiling water reactors (BWRs).
Section 3 covers the design evolution of the CANDU reactor, including a general description of
the overall design of the CANDU heat transport system and the design and evolution of the
main components such as primary pumps, steam generators, and the reactor core.
Section 4 defines the thermal-hydraulic design requirements, including fuel cladding (fuel
sheath), coolant, fuel, moderator materials, and control materials. Reactor core component
materials are discussed and component requirements assessed. This section also discusses
various fuel-coolant-moderator arrangements, their optimization, and their performance
within the reactor design. Advantages and disadvantages of all the variations are discussed and
possible solutions suggested. Finally, the section provides general requirements for the
thermal-hydraulic design process.
Section 5 discusses reactor thermal-hydraulic design limits from the perspective of various
reactor designs. It explains the concepts of reactor thermal margins and their application to
reactor design assessment. Reactor thermal margins are an important parameter in reactor
thermal-hydraulic design because they provide assurance that the heat generated by the fuel is
removed from the reactor core under all possible operating conditions.
Section 6 covers thermal-hydraulic design fundamentals. The first part of this section presents
the fundamentals of single- and two-phase flow and heat transfer. Two-phase flow and heat
transfer present a number of thermal-hydraulic design challenges, which are explained and
discussed in this section. In addition, this section covers the thermodynamics of the reactor
primary and secondary heat transport systems. The concepts of thermodynamic laws and their
application to reactor design are presented. The concept of reactor thermodynamic efficiency
is defined and its application to reactor performance assessment explained. Various secondary
system designs are discussed, and a description of the secondary side components, such as
steam turbines, steam condensers, feedwater systems and pre-heaters, and feedwater pumps,
is provided.
Section 7 is the key section in this chapter because it describes the design of reactor heat
transfer and fluid flow. This section outlines primary heat transport system behaviour,
describes the various mathematical models, and discusses the most important characteristics
of the primary heat transport system. The design and operation of the primary pumps is
described. The design and operation of the steam generators is covered in detail because this
component connects the primary and secondary heat transport systems, and therefore
understanding its behaviour is essential for understanding overall reactor thermal-hydraulic
behaviour. Flow stability in single-channel and parallel-channel instability situations is
explained and its relevance to reactor and pump operation presented.
Also presented in Section 7 are heat transfer in the fuel elements and their heat transfer
behaviour and operation in the reactor. Various topics are covered, such as fuel pellet cladding
gap heat conduction, variability of heat conductivity in the fuel with temperature, and the
influence of other important parameters. Fluid flow fundamentals are also covered, including
calculation of pressure drop in the primary heat transport system under single- and two-phase
operating conditions, calculation of flow resistance and its impact, and other important
aspects.
Heat transfer between fuel and coolant is also discussed in Section 7, including heat transfer
regimes in single-phase and two-phase operation, with a particular focus on boiling heat
transfer in a CANDU fuel channel. The concept of critical heat flux is defined and discussed,
along with various critical heat flux approaches, experimental data, and prediction methods.
The impact of critical heat flux on reactor thermal margins is discussed and methods for
improvement identified. An important part of the critical heat flux prediction model is the
look-up table, which is explained and its application described. Part of this section is devoted
to post-critical heat flux heat transfer, i.e., transition boiling and film boiling. Various heat
transfer modes are discussed, with particular attention to CANDU fuel bundles. Most of the
important heat transfer correlations are listed and explained and their application discussed.
Finally, the last few sub-sections of Section 7 are devoted to special topics in reactor thermal-
hydraulic design. One section explains the critical flow phenomenon and its relevance to safety
analysis, as well as the water hammer phenomenon. It provides insights on the risk from the
water hammer hazard and provides high-level information on preventing this phenomenon in
reactor design.
The last part of Section 7 covers natural circulation, which is an important phenomenon
because it provides assurance that reactor decay heat will be removed from the core if forced-
flow cooling is lost. This section describes the natural circulation phenomenon and focusses on
its application in CANDU reactors. The CANDU-specific phenomena of core cooling in the
absence of forced flow and of intermittent buoyancy-induced flow in a CANDU fuel channel are
covered.
1.2 Learning outcomes

The goal of this chapter is to help the reader to develop:
 An understanding of the principles and concepts of thermal-hydraulic design of nuclear
reactors.
 An understanding of the evolution of the CANDU thermal-hydraulic design.
 Knowledge of general design principles, with a focus on the main parameters, design
limits, and thermal-hydraulic margins by which the thermal-hydraulic design is
characterized.
 An understanding of the design concepts and key features of the primary and secondary
heat transport systems.
 Knowledge of the details of heat transfer and fluid flow and the heat transfer
methodology used in the thermal hydraulics of the reactor heat transport system.
 A basic knowledge of unique CANDU thermal-hydraulic features and behaviour.
The reader needs to be generally familiar with other topics and disciplines related to thermal
hydraulics. In addition, it is assumed that the reader has a basic understanding of fluid flow
and heat transfer phenomena.
1.3 Summary of relationship to other chapters

Thermal-hydraulic design is fundamental to reactor design and behaviour. Therefore, a good
understanding of thermal-hydraulic topics is required to understand the other chapters in this
textbook.
This chapter is tightly connected to the chapters covering core physics, fuel design, and safety
design and analysis. Furthermore, this chapter provides the knowledge base for understanding
Chapter 7 on thermal-hydraulic analysis.
1.4 Thermal-hydraulic design

Nuclear reactor thermal-hydraulic design is concerned with designing the process systems
required to transport heat (energy) from the nuclear reactor core (i.e., fuel) and to transform
this heat energy into useful work (i.e., electrical energy) in the turbine-generator unit.
Thermal-hydraulic design is associated with a number of interrelated systems, including the
reactor core, the heat transport system, steam generators, turbines, the pressure control
system, the coolant inventory control system, and the power control systems. In addition, a
number of components are important, including valves, pumps, pipes, vessels, and heat
exchangers, and a number of engineering and scientific disciplines, including reactor physics,
heat transfer, fluid mechanics, thermodynamics, chemistry, metallurgy, control, and stress
analysis.
The most important role of reactor design from a safety perspective is to ensure that:
that
 the fission reaction can be achieved and controlled (addressed by the physical
physic core
design and fuel design);
 heat can be removed from the core to tthe he ultimate heat sink (addressed by the
thermal-hydraulic
hydraulic design);
 the reactor operation can be monitored (addressed by the instrumentation
instrumentatio and control
design), and
 the radioactive material can be contained within the reactor facility (addressed by the
radiation protection and civil engineering design). Among
mong the most important aspects
of reactor engineering with respect to safe operation is the heat transfer and fluid flow
(i.e., the thermal-hydraulic
ydraulic design)
design).
The basic concept of a nuclear power plant is presented in Figure 1; it is similar
lar to that of any
other thermal power plant. The overall objective is to produce useful shaft work using a
thermodynamic heat engine (a turbine) with a heat source (a reactor) and a heat sink (a lake,
the sea, or the atmosphere). Following this basic concept,
ncept, a number of variations of the
nuclear power plant have been designed.
A number of inter-related
related systems and components are integrated in into nuclear power plant
design and interact with each other
other. Understanding their limitations
ions and characteristics is an
essential part of reactor design
design. Consequently, the process designer must appreciate the
characteristics and limitations of all the major components and systems to carry out the
detailed design of a particular system
system, i.e., to make intelligent choices. Design is, after all, the
process of constraining the possible alternatives (in reaching a design objective) down to one
choice. The overall goal is to provide an effective process within the context of the whole
operation. This means that the system must perform its process function safely and efficiently
ef
at a reasonable cost. The interplay of these key concepts and systems constitutes
constitute “the design
process”.
The final arbitrators in resolving the conflicting demands of each ssubsystem
ubsystem are:
are adequate level
of safety, low overall cost, material limits (temperature, mechanical stress, erosion, corrosion,
etc.), regulations, past experience
experience, standardized design requirements, and quality assurance
(QA).
Figure 1 Heat engine concepts
©UNENE, all rights reserved. For educational use only, no assumed liability. Thermal-Hydraulic
Hydraulic Design – December 2015
2 Reactor Types
A number of reactor types are operating around the world, and many more have been
designed in the past 50+ years of nuclear energy utilization, but never built. Many of these
reactor types use very different physical and thermal-hydraulic concepts; however, only a
handful have been able to ensure commercial viability over the past 50 years.
A number of these reactor types have proven to be commercially and economically
competitive, the most successful of which include [ELW1990]:
 Pressurized water reactors (PWRs)
 Boiling water reactors (BWRs)
 Gas cooled reactors
 Channel type reactors.
Although a detailed discussion of these reactor types is beyond the scope of this text, the
general evolution of reactor design and several of these designs are discussed in Appendix A.
The CANDU reactor design is a channel-type reactor, and a general overview, as relevant to the
topic of thermal hydraulics, is presented in Section 2.1.
2.1 CANDU reactor design

The CANDU reactor has been designed and built by the Canadian nuclear industry [AECL1978,
AECL1997], based on research and development undertaken by AECL, and has been
successfully operating for 50 years in Canada and internationally. This is an important reactor
type because, unlike other power reactors, it uses natural rather than enriched uranium and
therefore offers much flexibility for countries wanting to enter into nuclear programs and fuel
fabrication without requiring expensive uranium enrichment services. This section provides a
brief description of the thermal hydraulics of CANDU reactors. Other sections in this chapter
provide more detailed information about the thermal-hydraulic design of the CANDU reactor
and its other systems and components.
Figure 2 Typical CANDU plant
Figure 2 provides a schematic diagram of a typical CANDU heat transport system.

Figure 3a provides a detailed diagram of the CANDU 6 primary heat transport system (blue
solid lines) and the shutdown system (green dashed lines) [AECL1981]. Figure 3b provides a 3D
view of the CANDU 6 cooling system layout in the reactor building (the nuclear island)
[AECL2005].
A CANDU nuclear steam supply system’s power production process starts like that of any other
nuclear steam supply system, with controlled fission in the reactor core.
a) CANDU 6 primary and shutdown cooling loops
b) CANDU 6 cooling loops in containment – 3D view

Figure 3 CANDU 6 reactor cooling loops
2.1.1 Reactor core and calandria vessel

The core in a CANDU reactor is horizontal (Figure 4), with reactor channels in the core
containing the reactor fuel and heavy water coolant, whereas the heavy water moderator is in
the calandria vessel surrounding the reactor core [AECL2009a, AECL2010]. The reactor is made
up of a stainless-steel horizontal cylinder, the calandria, closed at each end by end shields that
support the horizontal fuel channels spanning the calandria and provide personnel shielding. In
the CANDU 6 design, the calandria is housed in and supported by a light water-filled, steel-lined
concrete structure, the reactor vault, which provides thermal shielding (Figure 5). In the
Darlington/Bruce CANDU reactor design, the calandria vessel is housed in a steel shield tank
assembly, which performs a similar function to the CANDU 6 calandria vault. The calandria
contains heavy water (D2O) moderator at low temperature and pressure, reactivity control
mechanisms, and several hundred fuel channels.
Figure 4a) Typical CANDU reactor and heat transport system
Figure 4b) Typical CANDU fuel channel
Figure 4c) Typical CANDU fuel bundle

Figure 4 CANDU reactor, fuel channel, and fuel bundle
Neutrons produced by nuclear fission are moderated (slowed) by the D2O in the calandria. The
moderator D2O is circulated through systems that cool and purify it and control the
concentrations of the soluble neutron absorbers used to adjust reactivity.
The fuel channels are also shown in Figure 4c. Each fuel channel supports 12 fuel bundles in
the reactor core (13 in the Darlington/Bruce CANDU design). The fuel channel assembly
includes a zirconium alloy pressure tube, a zirconium calandria tube, stainless steel end-fittings
at each end, and four spacers that maintain separation of the pressure and calandria tubes.
Each pressure tube is thermally insulated from the cool, low-pressure moderator by the CO2-
filled gas annulus between the pressure tube and the concentric calandria tube.
The CANDU fuel bundle typically consists of 37 elements (although advanced 43-element
designs have been qualified, and 28-element designs are still in service at Pickering station)
arranged in circular rings, as shown in Figure 4c.
Figure 5 CANDU calandria vessel and reactor vault
Each fuel element consists of natural uranium in the form of cylindrical pellets of sintered
uranium dioxide contained in a zircaloy-4 sheath closed at each end by an end cap. The fuel
elements are held together by end plates at each end to form the fuel bundle. The required
fuel element separation is maintained by spacers brazed to the fuel elements at the transverse
mid-plane. The outer fuel elements have bearing pads brazed to the outer surface to support
the fuel bundle in the pressure tube.
The CANDU reactor assembly, shown in Figure 5, includes the fuel channels contained in and
supported by the calandria. Each end shield consists of an inner and an outer tube sheet joined
by lattice tubes at each fuel channel location and a peripheral shell. The inner spaces of the
end shields are filled with steel balls and are light water-cooled. The fuel channels, supported
by the end shields, are located on a square lattice pitch. The calandria is filled with heavy
water moderator at low temperature and pressure.
2.1.2 Primary heat transport system design

The CANDU primary heat transport system consists of primary, secondary, and tertiary loops,
as shown in Figure 2 [AECL2009a]. There are several variations of the CANDU heat transport
system design, the latest being the Pickering, Bruce/Darlington, and CANDU 6 designs. The
CANDU 6 heat transport system design is described in the following sections.
The heat transport system (HTS) circulates pressurized D2O coolant through the fuel channels
to remove the heat produced by fission in the nuclear fuel. The coolant transports the heat to
steam generators, where it is transferred to light water to produce steam to drive the turbine.
Two HTS coolant loops (one in the Bruce design) are provided in CANDU reactors (Figure 6).
Each loop has one inlet and one outlet header, as well as one primary pump and one steam
generator at each end of the reactor core. D2O is fed to each fuel channel through individual
inlet feeder pipes from the inlet headers and is returned from each channel through individual
outlet feeder pipes to the outlet headers. Each heat transport system loop is arranged in a
“figure-of-eight”, with the coolant making two passes in opposite directions through the core
during each complete circuit and the pumps in each loop operating in series. The coolant flow
in adjacent fuel channels is in opposite directions.
Figure 6 CANDU primary heat transport system

In most CANDU reactors, the pressure in the heat transport system is controlled by a
pressurizer connected to the outlet headers at one end of the reactor. Valves provide isolation
between the two loops and the pressurizer in the event of a loss-of-coolant accident.
Figure 6 provides a view of the “figure-of-eight” design of the CANDU heat transport system.
On the right side, this figure shows the CANDU heat transport system to scale, with all feeders
connected to the reactor core. Coolant is transported from the outlet headers to the steam
generator by a large steam generator inlet pipe. Primary coolant from the steam generator
outlet is transferred to the primary pump by means of a large pump suction pipe and from the
pump to the inlet header by means of large pump discharge pipes.
The main components of the reactor primary heat transport system, in addition to the reactor,
are the steam generators, primary pumps, connecting piping to each channel, distribution
headers, large piping connecting these to the pumps and the steam generators, and the
pressurizer. These components are described in the next few sections.
2.1.3 Steam generators

The CANDU steam generators consist of an inverted U-tube bundle within a cylindrical shell.
Heavy water coolant passes through the U-tubes [AECL2009a, AECL2010]. The steam
generators include an integral pre-heater on the secondary side of the U-tube outlet section
and integral steam-separating equipment in the steam drum above the U-tube bundle. A
typical steam generator structure is shown in Figure 7.
Figure 7 Typical steam generator design Figure 8 Typical pressurizer design

Operation and design details and calculations for the steam generator are provided in Section
7.3. The steam generator plays an important role in reactor heat transport system operation
because it connects the reactor operation with the turbine-generator operation.
2.1.4 Pressurizer
The pressure in the reactor primary coolant system is maintained at a controlled level by a
pressurizer. Figure 8 shows a view of a typical pressurizer [AECL2009a, AECL2010]. The
pressurizer contains steam in the upper section of its cylinder and water in the lower section.
The pressurizer is connected to the primary loop through a surge nozzle at the bottom.
Heaters are provided at the bottom of the pressurizer internals, and a spray nozzle, relief
nozzle, and safety nozzle are installed at the top of the pressurizer head.
A “positive surge” of water from the primary loop because of increasing loop pressure is
compensated for by injecting cold water from the top of the pressurizer, which condenses the
steam in the upper portion and thus reduces system pressure.
A “negative surge” of water empties the pressurizer, reducing steam pressure at the top of the
pressurizer and thus loop pressure. In this situation, the electrical heaters at the bottom of the
pressurizer are automatically activated, converting a portion of the water into steam and
resulting in a loop pressure increase. By performing these sequences (i.e., creating steam when
the loop pressure is too low or decreasing steam when the loop pressure is too high), the
pressurizer maintains loop pressure within a certain design range and also ensures smooth
pressure changes in the primary loop.
2.1.5 Primary pumps

The primary pumps used in the CANDU heat transport system are vertical, centrifugal motor-
driven pumps with a single suction and a double discharge [AECL2009a, AECL2010]. As shown
in Figure 9, the pump impeller is at the bottom of the pump, and the pump shaft extends
upward to the pump motor, passing through a number of pump seals and holding the pump
flywheel.
Figure 9 Typical primary pump design

Cooling of the reactor fuel in the event of electrical power supply interruption is maintained by
the rotational momentum of the heat transport pumps during reactor power rundown and by
natural convection flow after the pumps have stopped. More information on pump function,
design, and operation is provided in later sections of this chapter.
2.1.6 Primary heat transport piping

The CANDU reactor contains a relatively large number of pipes, called feeders, and manifolds,
called headers, in the primary heat transport system, which are used to distribute coolant to
the fuel channels in the core. Although these components have important functions and are
mentioned in this chapter, a detailed discussion is beyond the scope of this textbook. Note
that feeders are unique to reactor designs with fuel channels and provide a number of design
advantages, but are also vulnerable to certain types of accidents and aging effects, as
mentioned in other chapters.
2.1.7 Secondary heat transport system design

The NPP secondary heat transport system transfers the generated energy from the primary
closed circuit to the secondary, where the heat energy is transferred into mechanical energy of
rotation in the turbine and then into electrical energy by the electric generator. The main
components of the secondary heat transport system are the steam turbine, condenser, heat
exchangers, feedwater pumps, valves, and piping; these are covered in the next few sections.
2.1.8 Turbine
The CANDU steam turbine is typically a tandem compound unit, directly coupled to an
electrical generator by a single shaft. It consists of one double-flow high-pressure cylinder
followed by external moisture separators, five steam reheaters, and three double-flow low-
pressure cylinders. The turbine is designed to operate with saturated inlet steam. The turbine
system includes main steam stop valves, governor valves, reheat intercept valves, and
emergency stop valves. All these valves close automatically in the event of a turbine protection
system trip.
In the following sections, more details are provided about turbine operation, efficiency, and
other relevant parameters.
2.1.9 Condenser
The turbine condenser consists of three separate shells. Each shell is connected to one of the
three low-pressure turbine exhausts. Steam from the turbine flows into the shell, where it is
condensed by flowing over a tube bundle assembly through which cooling water is pumped.
The condenser cooling water typically consists of a once-through circuit that uses water from
an ocean, lake, or river or is connected to cooling towers. The condensed steam collects in a
tank at the bottom of the condenser called the “hot well”. A vacuum system is provided to
remove air and other non-condensable gases from the condenser shell. The condenser is
designed to accept turbine bypass steam to permit reduction of reactor power from 100% to
70% if the turbine is unavailable.
2.1.10 Heat exchangers and pumps

On its return to the steam generators, condensate from the turbine condenser is pumped
through the feedwater heating system. Typically, it first passes through three low-pressure
feedwater heater units, each of which contains two heaters fed by independent regenerative
lines. (This arrangement permits maintenance work to be carried out on the heaters with only
a small effect on turbine generator output.) Two of the heater units incorporate drain cooling
sections and the third a separate drain cooling stage. Next, the feedwater enters a deaerator,
where dissolved oxygen is removed. From the deaerator, the feedwater is pumped to the
steam generators through two high-pressure feedwater heaters, each incorporating drain
cooling sections (see Figure 55).
Several stages of feedwater pumps are installed to raise the pressure from the condenser
pressure (vacuum) of 4–6 kPa to the steam generator pressure of 4.7 MPa. More information
on heat exchanger design and calculations is provided in later sections of this chapter.
2.2 Problems
1. Name and describe the function of the main components of the CANDU primary heat
transport system.
2. Describe the main components of the pressurizer in a CANDU reactor, with a detailed
explanation of the method it uses to control the primary heat transport system
pressure.
3. Provide a detailed description of the steam generator function, with specific reference
to its role in the relationship between the primary and secondary heat transport
systems. Comment on the relationship of these systems with the overall size of the
steam generators.
3 CANDU Thermal-Hydraulic Design Evolution

To obtain a better understanding of CANDU design concepts, one needs to understand how the
design has evolved over the past 50 years [AECL1978, AECL1997]. This section provides
information focussing on the reasons why certain evolutionary steps were taken and why
certain design solutions were abandoned, based on operating experience and on evolving
regulatory requirements. The design evolution is covered from two perspectives: general
evolution resulting from nuclear operating and regulatory experience around the world, and
CANDU-specific operating and regulatory experience in Canada.
3.1 CANDU reactor evolution

The evolution of the heat transport system is of primary importance in understanding the
evolution of power reactor technology. The primary heat transfer system contains the reactor
core and the fuel and is therefore an important link between the reactor thermal-hydraulics,
physics, and fuel.
The evolution of the heat transport system, which is described in this section, captures the
design evolution of the reactor core, the fuel and fuel channels in the CANDU reactor, the type
and number of primary heat transfer loops, and the primary pumps and steam generators.
Also covered are the evolution of the systems and components on the secondary side of the
heat transfer.
The CANDU design had its beginnings in the early 1950s, with preliminary engineering studies
on a 20 MWe and a 200 MWe plant [AECL1997]. The design concepts were based on
experimental confirmation at the ZEEP, NRX, and NRU experimental reactors at the AECL Chalk
River Laboratory facilities. These studies eventually culminated in commitments to the
construction of NPD and Douglas Point. NPD began operating in 1962 and Douglas Point in
1966. At the same time, commitments were made to construct Pickering in 1964 and Bruce in
1969. The 1970s witnessed the excellent operating performance of Pickering and Bruce and
further commitments to construct the Gentilly-2 (Quebec), Embalse (Cordoba, Argentina),
Point Lepreau (New Brunswick), Wolsong (Korea), Pickering B (Ontario), Bruce B (Ontario), and
Darlington (Ontario) plants.
In most cases, successive plants have meant an increase in reactor power output. Evolutionary
developments have been undertaken to fit the requirements of stricter safety goals, higher
ratings and sizes, new regulations, better reliability and maintainability, and lower costs. These
evolutionary changes have been introduced in the course of engineering parallel reactor
projects with overlapping construction schedules—circumstances which provide close contact
with the practical realities of economics, manufacturing, construction activities, and
performance in plant commissioning. Features for one project furnished alternative concepts
for other plants on the drawing board at that time, and the experience gained in first
application yielded a sound basis for re-use in succeeding projects. Thus, the experience
gained in NPD, Douglas Point, Gentilly-1, and KANUPP contributed to Pickering and Bruce. In
turn, all these plants contributed to the CANDU 6 design (i.e., Gentilly-2, Point Lepreau,
Wolsong, Cernavoda, Embalse, and Qinshan). The evolutionary changes that have taken place
are discussed in the following sections.
Table 1 provides a general comparison of key design features of other reactor types with those
of CANDU [GAR1999, POP2014, POP2015]. Table 2 provides information on the evolution of
various key components in CANDU heat transport systems.
Table 1 Typical fuel characteristics for key reactor types
Characteristic BWR PWR AGR LMFBR CANDU
Moderator H2O H2O Graphite - D2O
Coolant H2O H2O CO2 Molten Salt D2O
Neutron Energy Thermal Thermal Thermal Fast Thermal
Fuel Enriched UO2 Enriched UO2 Enriched UO2 PuO2/UO2 Natural UO2
Fuel Geometry Cylindrical Cylindrical pellet Cylindrical pellet Cylindrical pellet Cylindrical pellet
pellet in clad in clad tube in clad tube in clad tube in clad tube
tube
Fuel Assembly Up to 10 x 10 Up to 17 x 17 Concentric Hexagonal rod 37-element fuel

rod array rod array circles array bundles (typ.)
3.1.1 Primary heat transport system

The evolution of the CANDU design has involved a continuing quest for higher reliability, better
equipment maintainability, and reduced radiation doses to operating staff. This has been
manifested in a dramatic reduction in the number of components, as shown in Table 2. For
example, NPD had approximately 100 valves per MW in the nuclear steam supply system. This
was reduced to less than 1 valve per MW in the Bruce, Gentilly-2, and Darlington designs. In
addition, there are no valves in the large HTS main piping (past the Pickering design). The
number of steam generators has gone from 12 in Pickering, to 8 in Bruce, to 4 in the CANDU 6
and Darlington designs. Other evolutionary changes have included improvements in sub-
cooling margins, increased gross and channel flowrates, and increases in system pressure and
temperature. All materials in the heat transport circuit are now specified for very low levels of
cobalt to minimize radiation fields, thus improving CANDU radiation protection robustness.
Table 2 CANDU main process parameters and features
Parameter Douglas Point Pickering Bruce/Darlington CANDU 6

Power Output MWe
210 515 750 / 850 700
Number of Channels 306 390 480 380
Number of Pumps 10 16 4 4
Pump Type Vertical Vertical Vertical Centrifugal Vertical Centrifugal

Centrifugal Single Centrifugal Single Stage Single Stage
Stage Single Stage
Power per Pump kW 600 1170 > 8250 > 5250
Pump Code BPVC Sect. VIII BPVC Sect. VIII BPVC Sect. III Class 1 BPVC Sect. III Class
1
Pump Seismic None None DBE Category ‘A’ DBE Category ‘A’
Classification
Number of Steam 80 12 4 or 8 4
Generators
SG Power MW/boiler 2.5 ~ 45 ~ 95 150
SG Material M-400 M-400 I-600 I-800
Number of SG Tubes 196 2600 > 4200 3550

Steam generator size has been generally limited by the industrial capability to produce the
generators. Figure 10 [GAR1999] shows the evolution of the steam generator in terms of its
size and power. The power of the Darlington steam generators is close to 800 MW. Current
CANDU 6 plant designs typically have four steam generators.
Monel was used as the tubing material for Douglas Point, RAPP, KANUPP, and Pickering. This
material has proven quite satisfactory for the non-boiling coolant conditions of those plants.
Inconel 600 was used in NPD and in Bruce. This is a more costly material than Monel; however,
its corrosion resistance in a boiling environment (as in Bruce) is much superior. Currently,
Incoloy 800 is used in all 600-MW class CANDU 6 operating reactors. This material is more or
less equal in most respects to Inconel 600, has greater resistance to intergranular attack, and is
somewhat lower in cost.
Figure 10 CANDU steam generator design evolution
3.1.3 Heat transport pumps

Pump-motor sets have retained essentially the same configuration in all CANDU stations, i.e.,
vertical electric motor-driven, centrifugal, volute-type casing, one radial guide bearing in the
pump with pumped fluid as lubricant, a tilting pad-type guide and double-acting thrust bearing
in the motor, and mechanical shaft seals. Table 2 provides information on the evolution of the
primary pump [GAR1999].
Maintainability has been improved by providing interchangeable sub-assemblies. Appropriate
shielding placement has made it possible to change a pump motor on Bruce while the reactor
continues to operate at 60%–70% power.
There has been a trend away from solid rotor flywheels (Douglas Point to Gentilly-2) to
additional packages of rotor laminations located just outboard of the main rotor (Point
Lepreau, Bruce 'B'). This manner of fabrication eliminates the requirement for in-service
inspection because it is highly unlikely that a defect could grow from one lamination to
another.
Regulatory requirements for pumps have grown from minimal at the beginning to the present
time, at which the pump pressure boundary is considered in the same way as for nuclear
pressure vessels (ASME Section III, Class I). Consequently, non-destructive examination (NDE)
and quality assurance requirements have increased considerably.
3.1.4 Reactor core

In 1955, a detailed design of a demonstration natural uranium reactor was carried out. It was
called the Nuclear Power Demonstration (NPD) and was based on a vertical pressure vessel
concept [AECL1997]. In 1957, this was changed to a horizontal pressure tube configuration—a
configuration which has remained in subsequent heavy water-cooled reactors. The horizontal
configuration aided the on-line fuelling scheme by making double-ended fuelling feasible. It
also permitted the use of vertical safety control rods, which did not interfere with the pressure
tubes and feeders.
Reactor core evolutionary changes have been intended to achieve:
a) large increases in core rating with the minimum increase in reactor size (higher power
density enabling reduction in capital cost);
b) reduction in shop fabrication costs through simplification and standardization;
c) reduction in field assembly and shortened construction schedules through more shop
fabrication and modularization.
Table 3 Evolution of D2O content in the core and of power in the fuel channel
Reactor Design D2O in Core per MW MW Thermal per Metre

Thermal [m3/MWth] Length of Fuel Channel
[MWth/m]
NPD 0.410 0.163
Douglas Point 0.169 0.453
KANUPP 0.182 0.443
Pickering A 0.157 0.752
Bruce A & B 0.112 0.881
Gentilly-2 0.105 0.931
The major impact of higher power densities on capital costs is to reduce the heavy water
inventory. The amount of heavy water in the reactor core per MW produced in the reactor is
listed in Table 3. Higher power densities required more MWs of power produced per metre
length of fuel channel.
3.1.5 Reduction in radiation exposure

Canada has accepted the recommendations made by the International Commission on
Radiological Protection (ICRP) on maximum permitted doses for occupationally exposed
persons.
Note that severe accident doses are different from these normal operating doses and that
discussion is ongoing about these doses, especially under the influence of the accident in
Fukushima, Japan in 2011.
The major factors affecting the radiation dose accumulated by a worker are:
a) Amount of equipment.
b) Frequency of failure.
c) Time required for repair, service, and inspection.
d) Radiation conditions (fields and airborne concentrations).
Because radiation dose is proportional to the product of these four factors, a reduction in any
factor will reduce the dose received. The following general classification of some solutions in
the design stage has emerged:
1) Avoid adding equipment.
2) Eliminate equipment and remove unnecessary redundancy.
3) Simplify equipment.
4) Provide necessary equipment of high reliability.
5) Relocate equipment to lower radiation fields.
6) Eliminate materials such as cobalt, which could become highly radioactive.
7) Provide better chemical control and purification.
8) Arrange for quick removal for shop maintenance.
9) Extend interval between maintenance periods.
10) Reduce in-situ maintenance times.
11) Provide adequate space around equipment.
12) Provide adequate shielding so that maintenance can take place in low fields.
3.2 CANDU reactor types

Over the past 60 years, AECL in Canada has designed many types of CANDU reactors, some of
which were built and operated, and some of which were never built or were abandoned. This
section provides a brief overview of these past CANDU designs, with a view to understanding
the evolution of the CANDU heat transport system and its thermal-hydraulic design
[AECL1997].
3.2.1 Nuclear Power Demonstration station

Figure 11 shows the simplified HTS schematic for the Nuclear Power Demonstration (NPD)
CANDU design [GAR1999, POP2014]. This was the first reactor design prepared by AECL. The
HTS circuit contained in-line isolating valves for maintenance purposes. Pump reliability was
enhanced by using three 50%-capacity pumps with check valves to prevent reverse flow
through the non-operating pump. The check valves were placed at the pump discharge, of
course, rather than at the suction to meet net positive suction head (NPSH) requirements. The
66 inlet and 66 outlet feeders at each end of the core terminated in a reactor inlet and a
reactor outlet header respectively. Hence, bidirectional channel flow was used to limit spatial
reactivity feedback. Channel flow was trimmed to match the radial power distribution by
inserting an orifice plate into the inlet end fittings. All feeders were of the same diameter.
Pump flywheels were used to match the power rundown during a Class IV power failure to
ensure adequate fuel cooling, as in all CANDU stations. Boilers were placed above the core to
enhance thermo-siphoning. Feed and bleed provided pressure and inventory control.
Figure 11 Nuclear Power Demonstration (NPD) heat transport system

The NPD nuclear station had some significant design features that were quite different from
other CANDU stations. There was only one set of inlet and outlet headers, with relatively long
connecting pipes running from one side to the other of the reactor vessel. The end fittings of
the reactor fuel channels did not have shield plugs, leading to a large holdup of heavy water in
this region. The core itself consisted of two fuel bundle types. The central region had 19-
element bundles, and the outer region had 7-element bundles.
Another difference from later CANDU reactors is that the steam generator was a horizontal U-
tube vessel with the steam drum situated above and connected to the steam generator by a
series of four-inch risers and downcomers.
3.2.2 Douglas Point

The Douglas Point power plant was the first station to incorporate features typical of later
CANDU designs [GAR1999, POP2014]. Figure 12 shows a simplified HTS schematic for the
Douglas Point design. This station used the “figure-of-eight” loop layout. This configuration
had the advantage of reducing D2O holdup and pressure drop by eliminating the long piping
runs to the far end of the core, which were inherent in the NPD design. This introduced the
possibility of east-west (loop end-to-end) imbalances. Redundancy in pumps was required to
achieve adequate reliability. As in NPD, bidirectional channel flow, check valves at pump
discharges, and isolation valves were used. Channel flow was trimmed to match the radial
power distribution using different feeder sizes or orifice plates in inlet feeders and shield plugs.
Figure 12 Douglas Point heat transport system
3.2.3 Pickering A and B

The Pickering stations have a similar thermal-hydraulic design to Douglas Point. The Pickering
heat transport system is shown in Figure 13 [OPG2000]. Power output was increased to 540
MWe, and two loops were used to reduce the rate of blowdown in the event of a loss-of-
coolant accident (LOCA).
Figure 13 Pickering A heat transport system

A loop interconnect was provided to reduce loop-to-loop imbalance. Manufacturing limits on
steam generators and pumps led to the use of 12 operating steam generators and 12 operating
pumps with four reserve pumps. Component isolation was still possible, but check valves were
eliminated because of the leakage and poor reliability experienced at the Douglas Point plant.
Trimmed channel flow was achieved by varying feeder sizes and inlet feeder orifice plates. The
Pickering fuel bundle has 28 fuel elements and is the only CANDU design currently operating
with 28-element fuel.
3.2.4 Bruce A and B

Figure 14 shows the simplified schematic of the Bruce HTS system [BPR2000]. It shows a
significant difference in layout compared to the Pickering station. For Bruce (and later stations,
such as CANDU 6 and Darlington), the reliability experience gained from previous plants
justified the elimination of stand-by pumps. For radiation protection and maintenance
reasons, valves were eliminated. Manufacturing now permitted larger components, and
therefore eight steam generators and four pumps were included. Channel flow was not
trimmed to meet power distribution capability as in all other CANDUs. A constant radial flow
distribution was maintained by changing feeder sizes to account for geometry and feeder
length differences. As in all CANDU designs, fuel channel flow velocity was limited to 10 m/s
due to concerns about fretting of the fuel bundle and pressure tubes. The Bruce channel
design had thirteen 37-element fuel bundles located in the fuel channel, but the core length
was kept the same. Therefore, at each end of the core, half a fuel bundle protruded out of the
reactor core region.
Figure 14 Bruce heat transport system
3.2.5 CANDU 6
The CANDU 6 has been discussed in previous and following sections. The figure-of-eight loop
HTS design was adopted as of the Pickering design [AECL2009a, AECL2010]. However, also as in
the Bruce design, fewer components were used. Increased confidence in knowledge of two-
phase flow led to the use of boiling under normal conditions in the HTS. Erosion and corrosion
concerns at the steam generator inlet limited the quality to 4.5% at this position or nominally
4% at the reactor outlet header (ROH). Erosion/corrosion concerns also limited single- and
two-phase coolant flow velocities to a maximum value of 15.25 m/s to 16.75 m/s. The
presence of boiling required a surge tank or pressurizer to accommodate the larger shrink and
swell during transients. The pressurizer was used for pressure control (using heaters and steam
bleed valves), whereas inventory control remained as feed and bleed. This is the same
approach as in the Bruce design because the Bruce design is nominally single-phase. A heat
transport system schematic of the CANDU 6 design is shown in Figure 3.
3.2.6 Darlington
The HTS schematic for the Darlington design, shown in Figure 15, is similar to the CANDU 6
design. The Darlington reactor core design is similar to the Bruce reactor core design (480 fuel
channels with 13 bundles per channel) [OPG2002]. The Darlington HTS process conditions
were chosen to be very close to those of the CANDU 6 because that was the state of the art at
that time. An optimization program showed that higher tube pressures, higher qualities, and
higher velocities were economical. However, state-of-the-art engineering limits on pressure
tubes, qualities, and velocities forced the optimization to stop at these limits, which were the
same as for the CANDU 6 design.
Figure 15 Darlington heat transport system
The HTS for Darlington was designed by Ontario Hydro with design support from AECL. AECL
retained responsibility for the primary HTS between the headers (RIH, feeders, end-fittings, fuel
channels, and ROH), whereas Ontario Hydro assumed design responsibility for the rest of the
system.
The Darlington HTS design has four inlet headers, two on each side of the calandria vessel, and
two outlet headers, one on each side of the calandria vessel. This configuration enables mixing
in the outlet headers between the two loops, a feature that is useful for balancing the coolant
parameters in both loops. However, it also exhibits some weaknesses in achieving isolation of
the intact loop when the other loop is exposed to a break and a large LOCA situation.
3.2.7 Advanced CANDU designs

The advanced CANDU designs responded to continuing emphasis on safety, reliability, and
maintainability (R&M), quality assurance (QA), reduction in radiation doses, standardization,
modularization, and capital cost reduction. The excellent performance record of Pickering and
Bruce was maintained through a vigorous R&D program and a common-sense approach to QA.
Along with maintaining the above features, some of the high-level design directions were:
a) Reduction of capital cost;
b) Minimization of all areas of cost, from engineering to fabrication, construction, and
commissioning; and
c) Shortening of the overall construction schedule.
Heat transport system and process designs also reflected the evolution in the state of the art,
notably in the following areas:
1) Critical heat flux;
2) Erosion/corrosion velocity limits;
3) Single- and two-phase pressure drop and heat transfer correlations;
4) Thermo-siphoning;
5) Safety guidelines and requirements;
6) Stability aspects of two-phase flows;
7) Two-phase pump performance requirements;
8) Pump seals;
9) Process modelling (e.g., pressurizer, headers, boilers);
10) Creep of fuel channels;
11) Fuel design (fretting, hydraulic characteristics);
12) Power output and other constraints as required by clients;
13) Feeder sizing criteria.
3.2.7.1 CANDU 3, CANDU 9, and ACR-700

The CANDU 3, CANDU 9, and ACR-700 conceptual and preliminary designs were completed in
the 1990s and 2000s to meet a market need for small low-cost power reactors, large natural
uranium-fuelled reactors, and customer-specific needs respectively. Each of these reactor
types had different heat transport designs; however, no stations were built using these reactor
designs, and they were abandoned by AECL.
3.2.7.2 ACR-1000
The ACR-1000 conceptual and preliminary design was completed at the end of the 2000s;
however, no station was built using this reactor design. Figure 16 shows the key features of the
primary ACR-1000 heat transport system [AECL2009b]. The ACR-1000 reactor uses light water
as coolant with a low level of fuel enrichment (around 2.1%). This combination, along with
certain fuel design changes, resulted in negative void reactivity. The fuel design consisted of a
43-element bundle (CANFLEX) with low-enriched uranium fuel and with the central and
intermediate rings of larger diameter than the outer two rings. The central element did not
have fuel in it, and the outer two rings had pins with a certain percentage of neutron absorber
to achieve negative void reactivity. The fuel had target burn-up as high as 20,000 MWd/t.
Reactor power was about 3200 MWt with 520 fuel channels in the core.
Figure 16 ACR-1000 heat transport system

The heat transport system has two figure-of-eight loops with four steam generators, with four
pumps, two pumps, and two steam generators on each side of the loop. Although this reactor
design has a negative void reactivity and from this perspective does not need to have two
primary loops to reduce the power pulse in a large LOCA, it had two loops for a different
reason. Because the reactor power was high, in addition to other reasons, having one loop (as
in ACR-700) would require very large steam generators and primary pumps. Hence, splitting
the core into two loops brought the scale of the steam generators and primary pumps closer to
those readily available on the market.
The reactor pressure was high (about 13 MPa), along with high inlet and outlet temperatures,
resulting in better thermal efficiency than earlier models.
3.2.7.3 Enhanced CANDU 6
The Enhanced CANDU 6 (EC6) design followed the CANDU 6 design, but included a number of
design changes to enhance safety, operability, maintainability, constructability, and economics
[AECL2009a, AECL2010]. Most of these changes were made in the reactor safety systems, with
few in the thermal-hydraulic design. The EC6 heat transport system basically followed the
CANDU 6 design, which has been covered in previous sections.
3.3 Non-PHWR CANDU designs

A number of CANDU-generic reactor designs were completed that either did not use heavy
water coolant or were horizontal. These designs were never used in any CANDU stations, but
they present a particular set of design directions that were considered in the past. These
designs were later abandoned for various reasons, eventually focussing AECL’s design effort on
the horizontal heavy water reactor, known as CANDU, which was fuelled, moderated, and
cooled in a horizontal direction. Some of these early designs are covered in the following
sections.
3.3.1 CANDU-BLW
This was the second version of the basic CANDU concept to reach the prototype reactor stage
(the 250 MWe Gentilly-1 plant) [AECL1997]. Its major difference from the others lay in the
choice of coolant: boiling light (ordinary) water, hence its name, BLW. Its reactor coolant and
turbine systems were fundamentally the same as those of the BWR described earlier, i.e., a
direct cycle was used.
Figure 17 CANDU BLW heat transport system

For this version, a vertical orientation was chosen. A number of detailed considerations
relating to the boiling coolant led to this choice.
Figure 17 provides a schematic illustration of the design. Ordinary water is pumped to the
bottom of each fuel channel through an individual feeder pipe. As the water passes upwards
and absorbs heat from the fuel, a fraction (~18%) is evaporated to steam. The resulting
steam/water mixture then flows to a conventional steam drum where the steam and water are
separated. The steam then flows to the turbine, and the water, mixed with incoming
feedwater in the drum, flows down to the circulating pumps, completing the cycle.
The British developed a similar version, called the SGHWR (steam-generating heavy water
reactor). A 100 MWe prototype was built. It differed from the Gentilly-1 design in that it used
slightly enriched fuel. This enabled lower-purity heavy water moderator to be used, reducing
capital costs. The fuelling costs were, however, somewhat higher.
In Japan, another similar version, called the FUGEN reactor, was developed and operated for
about 20 years. The moderator cooling system was similar to the conventional CANDU version.
3.3.2 CANDU-OCR
A third version of the basic CANDU concept used an organic fluid as the coolant [AECL1997]. It
would have been similar to the PHWR concept, except that the boilers would likely be of the
“once-through” type, with some steam superheating provided. This was made possible
because the coolant temperature at the reactor outlet could be ~100°C higher than with heavy
water cooling.
Figure 18 CANDU OCR heat transport system

The Whiteshell Reactor-1 (WR-1) experimental reactor at the AECL Whiteshell establishment
used this concept, except that the heat was used only for building heating, i.e., no turbine was
provided. The reactor operated under coolant conditions which were the same as those used
in a commercial power plant.
Figure 18 shows a design drawing of the CANDU-OCR reactor with the key features explained
above. This reactor was never built because this concept was abandoned due to certain
weaknesses that were encountered, such as fire protection with coolant that is naturally
flammable, aging of the coolant (which required constant replacement and hence had a
negative impact on reactor availability (associated outage frequency)), and radiation protection
of the removed aged coolant.
One of the benefits of using organic coolant is that it is a mixture of several components and
therefore does not boil completely at a certain temperature. This helps to avoid going through
a two-phase region at one pressure and temperature, thus avoiding problems with critical heat
flux. Organic coolant also helps to achieve much higher coolant temperatures in the core
without the need to elevate the core pressure (as would be necessary for water). This
improves thermal efficiency significantly and makes the mechanical design easier and “lighter”
because of lower core pressures.
3.4 Problems
1. Describe and contrast the main similarities and differences between a CANDU primary
heat transport system and a PWR primary heat transport system design.
2. List three different CANDU reactor designs and explain the differences between each of
their heat transport system designs.
3. Identify four impacts of radiation doses on workers and list the general solutions that
have emerged at the design stage to mitigate their impacts.
4. Explain the benefits of using organic coolant in the CANDU-OCR design.
4 Thermal-Hydraulic Design Requirements

This section covers power reactor design requirements with a focus on the thermal-hydraulic
design of the reactor core [GAR1999, POP2014]. However, it should be kept in mind that the
thermal-hydraulic design is tightly linked to and interfaces with other aspects of reactor design,
particularly core physics and fuel design. Therefore, in various sections throughout this
chapter, these interfaces are mentioned and covered to the extent necessary to obtain a better
understanding of thermal-hydraulic design.
Relevant core configurations for various reactor types are covered, with a specific focus on the
CANDU design. Requirements for coolant, fuel, moderator, and key materials used for reactor
control and construction are covered, typical materials used for these components discussed,
options considered, and analysis provided.
When discussing fuel, coolant, and moderator requirements, note that neutron economy is
repeatedly mentioned as an important parameter [GLA1967]. This is true even for enriched
uranium reactors because the amount of fissile isotope enrichment, and hence the cost of the
fuel, is very sensitive to the neutron economy of the reactor. This is particularly so because
uranium enrichment is very costly because it involves an isotope separation rather than a
chemical separation process. In addition to cost, the fuel enrichment process involves a
number of safeguards, regulations, and rules, which make this process politically sensitive.
4.1 Fuel requirements

In all commercial power reactors, fuel is used in solid form and in various geometries such as
solid rods (cylindrical), plates, spheres, or annular configurations (see Figure 19). The key
parameter governing fuel design, with respect to thermal hydraulics, is the external surface
area-to-volume ratio. Good heat transfer to the coolant medium is promoted by high values of
this ratio, whereas low fuel manufacturing costs and good neutron economy generally are
promoted by low values of this ratio. This presents a “classical” optimization problem during
reactor design. Therefore, from this perspective, the annular fuel geometry best meets this
requirement; plate fuel is next, and cylindrical fuel is the worst. However, in most power
reactors in operation today, cylindrical fuel rods are used, primarily due to lower
manufacturing costs.
Figure 19 Fuel rod design types

Fuel requirements include (but are not limited to):
a) Low cost – both material and fabrication.
b) Good neutron economy.
c) Good corrosion resistance to coolant.
d) Physical stability under effects of radiation, temperature, and pressure.
Typical uranium-based fuel materials that have been considered or used include:
1) Uranium metal.
2) Uranium / other material alloy.
3) Ceramic uranium dioxide.
4) Uranium carbide.
5) Uranium silicide.
A summary of uranium fuel characteristics is presented in Table 4 [POP2014].
Table 4 Summary of uranium fuel characteristics [POP2014]
U Fuel Cost Neutron Corrosion Physical

Form Economy Stability
U Metal Lowest OK Poor Poor (swelling)
U Alloy Higher Lower OK OK
UO2 Higher Lower Excellent Excellent. High

T.
UC Lower than UO2<UC<U Good, except Good. High T.

UO2 Metal against water
US ~UC ~UC Good even ~UC

with water
Among uranium-based fuels, uranium metal is generally lowest in manufacturing cost and
highest in neutron economy, the latter because of its high density and the absence of other
neutron-absorbing elements. However, it has poor corrosion resistance to most coolants,
which is important in the event of fuel cladding failure. Its geometric stability in reactor use is
poor, primarily because of the swelling effects of fission products with specific volume greater
than that of the parent uranium.
Large quantities of alloying agents, such as zirconium, can be used, which effectively solve the
geometric stability and coolant corrosion problems. Unfortunately, both cost and neutron
economy suffer.
Uranium dioxide is the form in which uranium fuel is used in the vast majority of today’s power
reactors. It is somewhat more expensive to manufacture and less neutron-economical than
uranium metal because of its lower density, but possesses excellent corrosion resistance to
most coolants and a high degree of geometric stability. Being a ceramic, it can withstand high
operating temperatures, which has been the deciding factor.
Uranium carbide may be attractive as a fuel for future reactor designs. It is relatively
inexpensive to manufacture (comparable to UO2) and has somewhat better neutron economy
than UO2 (because of its higher density), but not as good as that of uranium metal. It has good
corrosion resistance against many coolants, but unfortunately not to water. Its dimensional
stability is good, and it can operate at high temperatures.
Uranium silicide is a more recent development with most of the advantages of uranium carbide
and in addition adequate resistance to corrosion by water coolants.
4.1.1 Metallic fuels

Unalloyed uranium metal is a very poor reactor fuel because it exhibits substantial growth
under irradiation. Highly irradiated specimens have been known to show axial growth equal to
more than 60% of the original sample length.
Uranium metal is highly chemically reactive and can be used at high temperatures with only a
few coolants, e.g., carbon dioxide and helium. Low-alloy uranium fuel (with a small amount of
alloying material) has been demonstrated to increase corrosion resistance to high-temperature
water and to improve irradiation stability. Good results have been obtained with U3Si (U-3.8
wt.% Si). Irradiation stability is somewhat improved, and corrosion resistance is about 500
times greater than that of unalloyed uranium. More satisfactory fuels can be obtained by
addition of alloying materials, for example, 10–20 wt.% Mo or Nb alloys containing 10%–13.5%
molybdenum; this option has appeared most promising.
Fuels that are highly satisfactory from both the irradiation-damage and water-corrosion
standpoints can be obtained with large additions of alloying material. The only high-alloy
uranium fuels in current power-reactor application are alloys of uranium and zirconium
because of the good neutron economy features of zirconium.
4.1.2 Ceramic fuels

Uranium dioxide is the most widely used of all reactor materials [POP2014]. It has an
irradiation stability far superior to that of metallic fuels and can withstand the high burn-up
required for economical power-reactor application. When properly prepared, it exhibits
excellent resistance to high-temperature water or sodium.
Uranium dioxide is obtained from the gaseous UF6 product of the diffusion plant. Hydrolysis of
UF6 produces UO2F2, which is reacted in a dilute ammonia solution to form a precipitate of
ammonium diuranate. This precipitate is calcined to UO3, which is then reduced to UO2 by
hydrogen at about 800°C. The oxide is produced as a fine powder, which needs to be pressed
into a high-density compact form followed by sintering at high temperature into cylindrical
pellets.
The thermal conductivity of uranium dioxide is quite low. At the high power output required in
power-reactor service, this leads to very high temperature gradients across the fuel element.
The resulting thermal stresses generally lead to radial cracking of the pellets during operation.
However, this cracking does not appear to cause any deterioration in fuel performance,
provided that the pellets are suitably restrained by cladding. Fuel element centreline melting
does remain a concern and imposes an important safety margin criterion.
The chief limitation on UO2 fuel performance is the swelling caused by gaseous fission products.
At low and moderate burn-up, the swelling is slight and roughly linear with burn-up. Above a
critical burn-up, the swelling increases markedly, and continued exposure of the fuel leads to
unacceptable dimensional changes. The critical burn-up is primarily a function of fuel density, a
value of about 17x103 MWd/tHE (heavy elements) being obtained with fuel of 97% oxide
density and a value on the order of 42x103 MWd/tHE with 93% density fuel.
Thorium dioxide (ThO2) behaves similarly to UO2 under irradiation. Thermal stress-induced
cracking is also observed in ThO2 fuels.
Carbide fuels are intended primarily for use in fast-breeder reactors (where they provide better
breeding ratios). Because these reactors operate on the U-Pu cycle, the fuel consists of mixed
uranium and plutonium oxides. Thermal conductivity of UC is substantially better than that of
UO2. Hence, considerably higher specific powers are possible without risk of centreline
melting. However, at higher temperatures, the swelling rate becomes excessive. The higher
thermal conductivity of carbide fuel leads to a much lower Doppler coefficient than that
obtained with an oxide core (the Doppler effect is the change in frequency of a wave for an
observer moving relative to its source). Because the Doppler coefficient in the fuel is the main
component of negative reactivity feedback in most accident situations, designing a safe
carbide-fuelled core may be more difficult than designing an oxide core.
Nitride fuels have also been considered for use in fast reactor fuel.
4.1.3 Dispersion fuels

In a dispersion-type fuel, particles of fissile material are embedded in a metallic ceramic matrix
[POP2014]. Such fuels can generally withstand significantly higher burn-up than alloy fuels. If
the fuel particles are separated sufficiently, the areas damaged by fission fragments will not
overlap, and a continuous metal phase will remain.
UO2 and PuO2 can be dispersed in aluminum, stainless steel, and zirconium alloys. The low-
temperature limitation on the use of aluminum eliminates these dispersions from
consideration for use in power reactors. Although stainless steel suffers from no such
limitation, poor neutron economy eliminates it from consideration in commercial reactors.
Dispersion of UO2 in zirconium alloys forms highly satisfactory fuels if highly enriched uranium
is used. Fuel performance is still limited by fuel swelling.
The most important dispersion fuel in use today is the dispersion of mixed uranium-thorium
carbides or oxides in graphite. These dispersions are used in high-temperature gas-cooled
reactors. Dispersion is achieved by multiple layers of pyrolytic carbon.
4.2 General fuel sheath (cladding) requirements

During fission, new isotopes of a wide variety of elements are produced. Many of these remain
radioactive for a significant time after they are generated and hence constitute a potential
radiation hazard to plant operators and the public at large. It is therefore clearly desirable to
keep these fission products “bottled up” within the fuel where they are generated. This is the
primary function of the fuel sheath (or cladding in light water reactor technology).
The fuel cladding takes the form of an impervious “skin” or “shell” which encloses the fuel
material. The most important requirements that cladding material should meet are [GAR1999,
POP2014]:
a) Corrosion resistance to coolant.
b) Mechanical durability.
c) High operating temperature capability (high melting temperature).
d) Good neutron economy.
e) Low cost, both material and fabrication.
f) Impermeability to fission products.
The following cladding materials have been considered or are currently used in operating
reactors:
1. Aluminum.
2. Magnesium (Magnox).
3. Stainless steel.
4. Ceramics.
5. Zirconium (used in CANDU reactors).
A summary of fuel cladding characteristics is presented in Table 5.
Table 5 Summary of fuel cladding characteristics [POP2014]
Cladding Corrosion Mechanical High T Neutron Cost FP
Type Resistance Durability Capability Economy Containment
Al Good, Low Low Good Low Good
except at
high T
Mg ~Al, Ok for ~ Al > Al ~ Al > Al ~ Al
CO2
Stainless Good Good Good Poor Good Good
Steel
Zr OK OK OK Excellent High Good
Ceramic Good Brittle Excellent OK OK OK
4.2.1 Aluminum, magnesium, stainless steel, and ceramic

Aluminum and its alloys possess many attractive properties such as low cost, easy fabrication,
high ductility (important in preventing cladding failures), good neutron economy, and
impermeability to fission products. Their major disadvantages for power reactor use are poor
mechanical properties at high temperatures and poor high-temperature corrosion resistance
with most coolants. Because the latter are temperature-dependent, aluminum alloys are
widely used in research reactor fuels where cladding operating temperatures are low, but are
not currently used in power reactors.
Magnesium alloys are similar to aluminum alloys in most regards. An alloy called “Magnox”
has, however, better high-temperature properties and adequate corrosion resistance to permit
its use in some CO2-cooled power reactors (i.e., Magnox reactors, used in the past and
described in Appendix A).
Stainless steel is a very attractive material in all major regards except for its poor neutron
economy. It has been and still is used in a number of reactors where its poor neutron economy
is somewhat less important because enriched uranium fuel is used. Stainless steels are used as
cladding in reactors where high-temperature service is needed. This type of cladding was used
in the British high-temperature, CO2-cooled reactors. Both gas-cooled and sodium-cooled fast-
reactor designs use stainless-steel cladding.
Most cladding materials in current use are metals, although ceramic-type materials have seen
limited use in certain applications.
4.2.2 Zirconium
Zirconium, in its various low-alloy forms, is by far the most common cladding material in
current use. It is the primary material type used for CANDU reactor fuel cladding. Despite its
relatively high base material cost, it combines to a large degree all the desirable cladding
properties for use with most coolants.
The primary advantage of zirconium as a cladding material is its very low cross section for
thermal neutrons, which greatly improves neutron economy. Zirconium has very good water
corrosion resistance at high temperatures. Adding tin, iron, and chromium to zirconium greatly
improves its mechanical properties. The best-known zirconium alloys are Zircaloy 2, 3, and 4,
which have found wide application as cladding for power reactor fuel elements.
Zirconium alloys are unsuitable for use at very high temperatures, even though the melting
point of zirconium is 1852°C. At 862°C, zirconium transforms from a close-packed hexagonal
structure to one that is body-centred cubic, and it is necessary to stay below this phase change.
At these temperatures, a reaction with UO2 can occur at the cladding inner interface of the
zirconium.
Zirconium alloys exhibit significant creep at the temperatures and stresses typical of PWR
reactor design. Creep rates increase markedly with temperature and are accelerated by
reactor irradiation.
At high temperatures (beginning at above 800°C), zirconium chemically reacts with steam to
release hydrogen in an exothermic reaction. This reaction must be considered in evaluating
any LOCA event that could expose the fuel elements to steam. Reaction of an appreciable
fraction of the cladding could add significantly to the severity of the accident. Moreover, the
mechanical properties of ZrO2 are not as favourable as those of zirconium.
The last point to be made about zirconium cladding is its cost. Zirconium tubes are more
expensive than stainless-steel tubes. Fabrication costs are also higher because all welding must
be done in an inert atmosphere. However, the decrease in fuel costs more than offsets the
increased material and fabrication costs.
Most power reactors in operation today use zirconium alloy as cladding material.
4.3 Reactor coolant requirements

The purpose of reactor coolant is to transport the heat generated in the reactor fuel either to
the turbine (direct-cycle reactor) or to intermediate heat exchangers (indirect-cycle reactor).
The coolants may be liquids, two-phase liquid/vapour mixtures, or gases.
Reactor coolant must meet the following general requirements [POP2014]:
a) High heat capacity
b) Good heat transfer properties
c) Low neutron absorption
d) Low neutron activation
e) Low operating pressure at high operating temperature
f) Non-corrosive to fuel cladding and coolant system
g) Low cost.
The following coolant materials have been considered or are currently used in operating
reactors:
1. CO2 gas
2. Helium
3. Ordinary water
4. Heavy water
5. Organic fluid
6. Molten salt
7. Liquid metal.
The main commercial power reactor coolant types are described briefly in the following
sections.
The CANDU reactor design is the only reactor design that uses heavy water as both a
moderator and a coolant and is fuelled by natural uranium fuel.
A summary of coolant characteristics is presented in Table 6.
4.3.1 Ordinary water and heavy water coolants

Of the candidate HTS liquid coolants, ordinary (light) water is by far the most commonly used.
It is inexpensive, has excellent heat transfer properties, and is adequately non-corrosive to the
zirconium alloys used for fuel cladding and reactor structural components and to ferritic or
austenitic steel coolant system materials. Its disadvantages include only moderate neutron
economy and relatively high vapour pressure under coolant temperatures at power reactor HTS
conditions. It is activated by neutrons in the reactor core, but this activity dies away rapidly,
permitting reasonable access to the coolant system for shutdown maintenance. A further
disadvantage is that water transports system corrosion products, permitting them to be
activated in the reactor core. These activated corrosion products then create shutdown
radiation fields in the coolant system. The water coolant may be used as a liquid in an indirect-
cycle system or may be permitted to boil, producing steam in a direct-cycle system.
Table 6 Summary of coolant characteristics [POP2014]
Corrosi
Coolant
Neutro
Activat
Capaci
Econo
Coeff.
Coeff.
Heat
Cost
Type
ion
my
HT
HT
ve
ty
CO2 Gas < He n
Good OK, except Low Low Low Low
high T
He Higher Good Good, if Low Low Low Low
pure
H2O Very Low Moderate OK High Excellent Excellent Yes, short
T1/2
D2O High Excellent OK High Excellent Excellent Yes, short
T1/2
Organic Moderate H2O < organic Excellent High Excellent Excellent None
< D2O
Liquid High Moderate Select High Excellent Excellent Yes, long
Metal Materials T1/2
Heavy water is used as both coolant and moderator in CANDU-type reactors. Its outstanding
advantage is much better neutron economy than ordinary water. Its primary disadvantage is
its high cost, plus a somewhat higher tendency to activate into tritium. Otherwise, its
properties are similar to those of ordinary water.
The relatively ionizing irradiation (including γ’s released because of neutron-radiative capture)
causes decomposition (radiolysis) of water. Decomposition occurs in both H2O and D2O by the
same mechanism. The rate of gas evolution (radiolysis) is proportional to radiation flux and
decreases with increasing temperature.
Radiolysis of water is an important phenomenon that must be considered in analysis of
accident events (as well as in normal operation) in a power reactor. Hydrogen is produced by
radiolysis of water discharged from any break and can accumulate in the reactor building.
Radiolysis of discharged water can occur because of γ-radiation from the sources within the
discharged water (high contribution) or from the reactor fuel (low contribution). Above a
certain concentration, hydrogen will burn at an explosive rate, creating a high risk to reactor
components and building structures (e.g., in the Fukushima accident). The accumulation of
hydrogen produced by radiolysis of discharged water in a LOCA event can be controlled by
passive autocatalytic recombiners.
Water circulated through a reactor core exhibits appreciable induced radioactivity. The γ
activity induced is primarily due to the 7.4 s half-life 16N produced by fast-neutron interaction
with 16O. Additional activity may be introduced by activation of dissolved impurities and
dissolved or suspended corrosion products. Hence, water purification is an important process
that is continuously performed by dedicated systems.
4.3.2 Gaseous coolants

Two common gaseous coolants are in use: CO2 and helium. CO2 has the advantages of low
cost, low neutron activation (important in minimizing radiation fields from the coolant system),
high allowable operating temperatures, good neutron economy, and for gases, relatively good
heat transfer properties at moderate coolant pressures. At very high temperatures, CO2 tends
to corrode neutron-economical fuel cladding materials and the graphite moderator used in
most gas-cooled reactors. Its chief drawback, as for all gases, is its poor heat transfer
properties relative to liquids. As a result, coolant pumping power requirements tend to be very
high, particularly if high reactor power densities are to be achieved (which is desirable to
minimize reactor capital costs).
A good gaseous coolant should have a low neutron-absorption cross section, high heat
capacity, and high thermal conductivity.
All gaseous coolants have the disadvantage that a significant fraction of the plant energy
output must be used to circulate the coolant. Furthermore, the low heat transfer coefficient
requires a large heat transfer area. This may require a larger core or the use of extended
surfaces (fins).
4.3.3 Liquid metal coolants

Certain liquid metals can be used as coolants. Of these, only sodium and a sodium-potassium
eutectic called NaK have achieved significant use. Their advantages include excellent heat
transfer properties and very low vapour pressures at high temperatures. Fuel cladding and
coolant system materials require careful selection to avoid “corrosion”. Their chief
disadvantages include incompatibility with water (the turbine working fluid), relatively high
neutron absorption, a relatively high melting point (leading to coolant system trace heating
requirements), and high coolant activation with sustained radiation fields after reactor
shutdown.
Liquid metal coolants are important only for fast reactors because these metals are poor
neutron moderators. Sodium is one of the most suitable liquid-metal coolants.
Sodium becomes radioactive when 24Na is formed by neutron capture. This radioisotope has a
15-h half-life and emits gamma rays of 1.37 and 2.75 MeV [GAR1999]. Hence, shielding of the
cooling system is necessary. Care must be taken to ensure a leak-tight system. Sodium is
chemically reactive and will burn on exposure to air, evolving the oxide as dense smoke.
Furthermore, it reacts violently with water, producing NaOH and hydrogen gas.
4.4 Moderator requirements

Most currently operating power reactors are of the thermal type, i.e., the energy of the
neutrons causing fission is in the thermal range. Because the neutrons produced by fission
have very high energies, they must be slowed down, or “thermalized”. The medium used for
this is called the moderator. It is deployed as a continuous medium surrounding the fuel
“cells”. The fuel cells form a geometric pattern, called the reactor “lattice”. The optimum
spacing between these fuel cells is a function of several variables, including the mass of fuel per
cell, the mean free path of the neutrons being thermalized, the degree to which the moderator
wastefully absorbs neutrons, and the cost of the moderating medium. A summary of
moderator characteristics is presented in Table 7 [POP2014].
The best moderator is something that is the same size as a neutron, i.e., the hydrogen atom,
1
H1. However, hydrogen does absorb neutrons as well. The deuterium atom, 2H1, at twice the
mass of hydrogen, is almost as good a slowing-down agent, but because it already has an extra
neutron in the nucleus, it has a very low absorption cross section. Therefore, deuterium is a far
better moderator overall than hydrogen. By using deuterium in the form of heavy water,
natural uranium can be used as a fuel. If ordinary water is used, the fuel must be enriched in
235
U.
A good moderator has a high scattering cross section, a low absorption cross section, and slows
down the neutron in the least number of collisions (high lethargy, ξ).
The desirable properties of moderators are:
a) High moderating efficiency
b) Low neutron absorption
c) Resistance to irradiation and corrosion
d) Low cost, including material, manufacture, and installation.
The following materials have been considered or are currently used as moderators in operating
reactors:
1. Graphite
2. Ordinary water
3. Heavy water.
Graphite has been widely used as a moderator for power reactors. The carbon atom is
relatively “light”, graphite is relatively inexpensive, and carbon is a relatively weak absorber of
neutrons. Nevertheless, the carbon atom is sufficiently large, leading to relatively long neutron
mean free paths for thermalization, that graphite-moderated reactors tend to be large.
Furthermore, the relatively large amount of graphite required leads to significant neutron
wastage through absorption.
Table 7 Summary of moderator characteristics [POP2014]
Moderator Cost Neutron Moderator Irradiation Activation Mean Free
Type Economy Efficiency Stability Path
Graphite OK H2O < Medium Excellent Irrelevant Long
Graphite <
D2O
H2O Very Low Moderate Low Excellent Good Small
D2O High Excellent Highest Excellent Good Medium
Ordinary water is a much more efficient moderator in terms of the neutron mean free path for
thermalization because of its hydrogen atoms. It is also very inexpensive. Unfortunately,
however, hydrogen also has a significant “appetite” for absorbing thermal neutrons, which
hurts neutron economy. Most of the reactors operating in the world use ordinary water as
coolant and moderator.
Heavy water is almost as good as ordinary water in terms of neutron mean free path because
the deuterium atoms (which replace the hydrogen atoms in ordinary water) are relatively
“light”. Its outstanding advantage relative to ordinary water is that it has a very small
“appetite” for absorbing neutrons. Hence, it promotes a high level of neutron economy. Its
major disadvantages are its high cost and the possibility of its activation into tritium.
4.5 Control material requirements

Reactor control is most commonly accomplished by moving control rods or control mechanisms
in the reactor core or by injecting liquid containing material (poisons) with high neutron-
absorption cross section in the thermal and near-thermal range [POP2014] into the coolant
and/or moderator.
4.5.1 Hafnium
Hafnium is one of the best control-rod materials for water-cooled reactors. It is found together
with zirconium and is created as a by-product of the separation process for zirconium. Hafnium
is chemically similar to zirconium and shows the same high resistance to corrosion by high-
temperature water.
Hafnium consists of four isotopes, each of which has an appreciable neutron absorption cross
section. Note that the capture of neutrons in one isotope leads to the production of the next
higher isotope, which is also an effective absorber. Therefore, hafnium remains an effective
poison for a significant time.
4.5.2 Cadmium alloys

By alloying cadmium, which has a high thermal-absorption cross section, with silver and
indium, which have high resonance absorption, a highly effective absorber is produced. The
alloy is typically composed of 80% Ag, 15% In, and 5% Cd, can be readily fabricated, and has
adequate strength at water-reactor temperatures.
These alloys have exhibited moderate resistance to corrosion by hot water after plating with
nickel. Bonding of the base to the nickel was ensured by heating the control rod to an elevated
temperature before reactor exposure. Corrosion under reactor conditions was good, but in
subsequent use of the alloy, it was encapsulated in stainless-steel tubes so that direct contact
with the coolant was eliminated.
Cadmium is typically used in several types of CANDU reactivity mechanism designs. For
example, the shut-off rods in CANDU reactors are composed of a thin layer of tubular cadmium
sandwiched between stainless steel layers.
4.5.3 Rare-earth oxides

Several of the rare earths (e.g., samarium, europium, and gadolinium) have both high thermal-
neutron-absorption cross sections and significant resonances in the epithermal region. The
oxides, the only chemical form in which the rare earths have been considered, can be formed
into refractory ceramic pellets. Because the oxides hydrate rapidly in hot water with attendant
swelling, their use in water-cooled reactors could lead to difficulties. They may, however, be
suitable for use in gas-cooled reactors. Dispersions of rare-earth oxides in metals such as
stainless steel have been prepared. Such dispersions, when suitably clad, can be used as water-
cooled reactor-control rods because in the event of a cladding fracture, only the oxide particles
on the dispersion surface would become hydrated.
4.5.4 Gadolinium nitrate

The CANDU reactor uses a gadolinium nitrate (GdNO3) solution in water as a reactor control
(shutdown) poison. Tanks full of concentrated GdNO3 solution are maintained under pressure
by helium gas. The solution can be injected into the moderator system at a reactor shutdown
trip signal, which provides a fast and effective reactor shutdown mechanism. To restart the
reactor, the gadolinium nitrate is removed from the moderator by a purification system.
Because the poison is introduced into a low-temperature, low-pressure moderator, there are
no corrosion-related concerns.
4.5.5 Boron-containing materials

The very high neutron absorption cross section of 10B and the low cost of boron have led to
wide use of boron-containing materials in thermal-reactor control rods and burnable poisons.
Unalloyed metallic boron is not suitable for control-rod use.
Boron alloyed or dispersed in stainless steel forms inter-metallic compounds with iron, nickel,
and chromium in the metal matrix. The result is a major decrease in ductility. Acceptable
materials can be obtained by limiting the boron concentration to 2–3 wt.% B. The boron-
stainless steel materials have adequate corrosion resistance in water-cooled reactors.
The performance of boron-stainless steel materials is limited because of the 10B (n, α) reaction,
which leads to severe swelling localized at the surface of the element (due to the short travel
distance of the alpha particles).
Boron carbide (B4C) is of much greater interest than elemental boron. Boron carbide can be
formed into pellets and effective control elements produced by placing these within stainless-
steel tubes. In high-temperature gas-cooled reactor designs, control elements can also be
produced by dispersing B4C in graphite.
In addition to its use in control elements, boron has been used (in PWRs) to control reactivity
changes with fuel burn-up by dissolving boric acid in the coolant. Immediately after refuelling,
enough boric acid is added to the coolant so that the reactor is just critical with all control rods
nearly completely withdrawn. As burn-up proceeds, the boric acid concentration in the coolant
is reduced to maintain criticality exactly (“chemical shim control”). However, one significant
disadvantage of boric acid in coolant is that it is corrosive to heat transport system and reactor
components (i.e., Davis-Besse pressure vessel head corrosion) and must be carefully monitored
and mitigated.
Boron may also be used to compensate for the change in reactivity with fuel lifetime through
“burnable poison”. In this scheme, a small amount of boron is incorporated in the fuel or into
special burnable poison rods to reduce the beginning-of-life reactivity. Boron burn-up causes a
reactivity increase that partially compensates for the reactivity decrease due to fuel burn-up
and accumulation of fission products. Stainless-steel-boron alloys and dispersions can be used
successfully for this purpose because the boron burn-up in such rods can be kept low.
Alternatively, pellets of boron-silicate glass encapsulated in hollow stainless-steel tubes may be
used.
4.6 HTS design requirements and engineering considerations

This section provides a discussion of design requirements and engineering considerations for
the heat transport system and associated systems, which transfer fission heat to the reactor
coolant to produce steam in the steam generators.
This sub-section will discuss some of the thermal-hydraulic features that characterize the
CANDU system.
The main objectives of the heat transport system are to provide heat transfer at high thermal
efficiency and to enable the maximum amount of energy to be extracted from the fuel without
surpassing safe operating limits.
The requirements for such a system can be summarized as follows:
a) Due to the decay heat produced by the fuel even when the reactor is shut down,
continuous coolant flow must be provided. This leads to requirements for pumps,
pump flywheels, stand-by cooling, thermo-siphoning, and various other system
functions.
b) Costs should be minimized with due regard for other requirements. This usually leads
to trade-offs between, for example, heavy water (D2O) costs, pumping power costs,
equipment and piping size and costs, and layout and engineering constraints.
c) Layout should minimize radiation exposure and maximize maintainability and
accessibility within other constraints.
d) Provision must be made for pressure and inventory control of the heat transfer system.
Excessively high pressure could damage fluid boundaries (pipes, etc.). Low pressure
could lead to high coolant voiding and possible fuel damage as well as pump damage
from cavitation. Low inventory jeopardizes coolant circulation and pressure control.
e) The system must be sufficiently reliable because downtime leads to high replacement
energy costs, high person-rem exposure, and high repair costs.
f) The design should provide high process efficiency.
g) The system should exhibit ease of constructability to reduce the initial cost and time of
construction and to enhance maintainability.
h) The system should meet, and preferably surpass, all safety and licensing requirements.
Various coolants can be used in the CANDU design to achieve these objectives and
requirements.
Any nuclear station design involves trade-offs in design features to achieve the lowest-cost
power within safety limits.
Another nuclear consideration is that the coolant should have a low induced radioactivity.
Both H2O and D2O produce 16N and 19O, which emit γ’s in the 6–7 MeV range. This leads to
reduced accessibility and maintainability while on power. The short half-life (<1 minute) allows
shutdown accessibility. Tritium, abbreviated as 3H or T, has a 12-year half-life and represents a
major dose challenge for the CANDU station. Because tritium is a β-emitter, the problem is one
of leakage, leading to possible absorption or ingestion by humans. Organic coolant has very
little induced reactivity and aids in ease of operation and accessibility.
Coolants should also be stable in a radiation environment. At the high system pressure of the
heat transport systems of H2O and D2O, radiolysis is not a problem.
The choice of coolant also depends on other factors such as pumping power, heat capacity,
heat transfer coefficients, flow rates, pressure drop, boiling point, freezing point, corrosion,
flammability, thermal stability, and cost.
Water (both D2O and H2O) is an attractive heat transport fluid because it offers a good balance
of the considerations described above. In addition, water requires less pumping power for a
given amount of heat removal.
For the Bruce reactors (which generate about 750 MWe), approximately 24 MWs of pumping
power are required for each reactor. This represents over 2% of the electrical power generated
[GAR1999]. Because a MW saved here by reducing pumping power is gained as electrical
output, considerable emphasis is placed on lowering pumping power.
Limiting flow rates for coolant water, in addition to ensuring adequate heat transfer rates from
the fuel and to the steam generator tubes, depend on many factors such as temperature, the
presence of boiling, water chemistry, geometry, and flow regime. Fretting considerations have
led to a 10 m/sec limit on fuel channel velocity in single-phase water. Erosion/corrosion
considerations have led to limits of 4.3 to 6.1 m/s in the steam generator tubes and 16.8 m/s in
heat transport piping.
The fuel distribution in the coolant is designed to maximize the fuel surface-to-volume ratio so
that the largest heat transfer surface can be exposed to the coolant for maximum heat transfer
without drying out the fuel surface. However, if this approach is carried to extremes, the fuel
volume in the core will be less than the optimum, and parasitic neutron absorption by the
sheath will increase. Present designs use 37 or 28 elements in a fuel bundle.
Use of boiling in the coolant permits higher heat transfer due to the high heat transfer
coefficient of nucleate boiling, but introduces challenges to the attempt to minimize pressure
drop and pump size, as well as other two-phase flow effects.
Ideally, the coolant temperature should be as high as possible for maximum overall thermal
efficiency. Therefore, a high-boiling-point, low-vapour-pressure liquid is desirable so that the
heat transport system can operate at the lowest possible pressure. This reduces the thickness
of the pressure boundary and hence is important for reducing parasitic burn-up in the core.
Organic coolant is far superior to water from this point of view.
In the case of organic coolant, the secondary-side H2O pressure is higher than the primary-side
pressure. Hence, boiler tube leaks will cause a water leak into the primary coolant system.
Corrosion of heat transport system materials must be minimized because of possible

deterioration, flow restrictions, and contamination with active isotopes.
The CANDU-PHW heat transport system has water coolant, low-cobalt carbon steel piping,
stainless-steel end fittings, Zircaloy pressure tubes, and Monel or Incoloy steam generator
tubes. A pH of 10.2 to 10.8 is maintained by lithium hydroxide and hydrogen gas added to keep
the dissolved oxygen content low to help minimize corrosion. The intent is to produce and
maintain a continuous and adherent film of magnetite on all carbon steel surfaces. Corrosion
with organic coolant is a lesser problem and can be controlled by degassing, using N2 cover gas,
and using a dechlorinator system.
No flammability or thermal stability problems exist with water (except for the possible Zr-water
reaction producing H2 during a LOCA).
The cost of D2O is high, making it the more expensive coolant. This contributes to a high capital
cost for the CANDU-PHW, but a low operating cost due to the efficient use of natural U.
4.7 Problems
1. Identify and explain the primary reasons that natural uranium ceramic is used as the
traditional fuel in the CANDU reactor design.
2. Identify and explain the primary reasons that zirconium and zirconium alloys are used as
the fuel cladding material in the CANDU reactor design.
3. Identify and explain several advantages and drawbacks of using heavy water as a
moderator and coolant in the CANDU reactor design.
4. Summarize and explain the main heat transport system design requirements for the
CANDU reactor design, as related to the main objectives of the system.
5 Thermal-Hydraulic Design Limits and Margins

This section covers the basic design principles of power reactor design, including reactor
thermal-hydraulic characteristics, along with an overview of these for different reactor types,
but with an emphasis on the CANDU design. It also covers the energy production and transfer
parameters, the thermal design limits, and the thermal design margins, with a discussion on the
reactor thermal-hydraulic figure of merit.
5.1 Heat generation parameter definitions

This section covers the reactor power production parameters generally used in reactor
engineering. Energy production in a nuclear reactor is expressed in a variety of terms that are
used by various disciplines.
The volumetric energy (heat) generation rate, q″′ (kW/m3), is mostly used by reactor physicists
because they deal with the fission reaction rate that results in a volumetric energy generation
rate in the fuel. This term does not provide any information about heat transfer in the core,
nor does it indicate any level of core margin.
The surface heat flux, q″ (kW/m2), is most important to the core thermal-hydraulic designer
and is directly related to reactor thermal margins. The thermal-hydraulic engineer deals with
heat transfer from the fuel element surface to the coolant.
The linear heat generation rate or power rating, q′ (kW/m), is important to both the
thermal/fuel designer and the metallurgical designer because it expresses fuel performance
characteristics in terms of linear power rating. Fuel design, testing, and qualification are
usually described using this parameter.
The rate of energy generation per fuel element, q (kW), is useful in expressing heat generated
separately in each fuel element, which is used in modelling heat transfer phenomena in the
core.
The core power, Q (kW), is used in overall calculations of core energy output. Usually, for
marketing purposes, this core power is expressed in terms of electrical power delivered by the
plant to the grid, or occasionally in terms of net electric power accounting for plant internal
energy consumption. However, for thermal-hydraulic design purposes, reactor power is
expressed in terms of reactor thermal power. Roughly speaking, reactor electric power is equal
to reactor thermal power divided by reactor thermal efficiency.
The core power density, Q″′ (kW/m3), is used as a figure of merit for core thermal performance.
This power level accounts for various materials in the core and thus provides information on
reactor core compactness and effectiveness.
The core specific power, Q/(mass of fissionable atoms) (i.e., kW/kg), is used as a figure of merit
for core thermal performance from the nuclear physics perspective.
5.2 Thermal design limits

Core thermal design limits are expressed in terms of several parameters, including heat flux
from the fuel elements to the coolant, fuel centreline maximum temperature in the core, fuel
cladding maximum temperature in the core, bundle maximum power, and channel maximum
power.
Figure 20 presents the important concept of reactor thermal limits with respect to the critical
heat flux (CHF). The definition, assessment, and calculation of the CHF are covered in the
following sections. Reactor thermal-hydraulic design must ensure that the CHF is not reached
in any fuel element or bundle in the reactor core. The CHF is the heat flux at the surface of the
fuel element that results in a sudden change in heat transfer regime from liquid in good contact
with the heated surfaces, to loss of local liquid contact with surfaces because of vapour
blanketing the fuel-element surface. This phenomenon leads to severe reduction of the heat
transfer coefficient, and for heat flux-controlled surfaces (like fuel rods in a reactor), in a
significant increase in fuel temperature as well as damage to or failure of fuel sheaths.
Figure 20 Critical heat flux ratio

Figure 20 shows a typical variation of the heat flux and the critical heat flux along a fuel
element. For simplicity, the reactor heat flux in the core is assumed to be represented by a
cosine function that peaks at the centre of the core. As will be seen later, the CHF follows a
decreasing curve from the fuel channel inlet to the fuel channel outlet. In Figure 20, the ratio
of CHF and local heat flux is also illustrated, which is one way to assess how close the reactor is
to the occurrence of CHF anywhere in the core.
The key objective of thermal-hydraulic reactor core design, for a desired core power, is to
ensure that the heat flux in all fuel elements in the core will remain below the CHF. The CHF is
among the important parameters in core thermal-hydraulic design. Note that local power in
the core will vary with position along the fuel channel, including the end flux peaking at bundle
ends, and from channel to channel in the radial direction of the calandria.
Fuel melting and cladding melting are other important factors and are dealt with in later
sections of this chapter.
CANDU reactor design-related CHF phenomena are further described in Section 7.6.4.
5.3 Thermal design margins

This section defines thermal design margins in terms of heat generation rate and the resulting
heat flux from the fuel to the coolant. Variation of the reactor parameters in the reactor core is
taken into account to define the core average quantities, maximum hot channel quantities, and
maximum fuel element quantities that are used in the design space to optimize reactor design
options. The heat generation rate during normal operation and the postulated design basis
events define the key design limit for which power reactors need to be designed with a certain
margin.
Design margins are applied to parameters of primary importance to ensure that the reactor can
be safely operated without exceeding important limits. These important operating parameters
are assumed to be inside a so-called safe operating envelope, which defines the range of values
that a particular parameter can take on during reactor operation. The left side of Figure 21
shows a general definition of the safe operating envelope, identifying the optimal operating
range, the lower and upper alarm range, the lower and upper buffer zone, and the range where
reaching the failure point will be immanent for any of the important thermal margin
parameters.
Some of the key operating parameters that are usually considered by thermal-hydraulic
designers are:
 Coolant temperature
 Coolant pressure
 Critical heat flux
 Fuel maximum temperature (centreline)
 Fuel cladding maximum temperature
 Bundle maximum power
 Channel maximum power
 Reactor total thermal power.
The thermal design margins function as the most important parameter for demonstrating the
robustness, operability, and flexibility of a given reactor design and serve as the main standard
of comparison for the safety level of various reactor types. One of the key thermal design
margin parameters is the ratio of the critical heat flux to the operating surface heat flux, which
must satisfy a reference allowable value that meets economic as well as regulatory
requirements.
Figure 21 Thermal margins

The required core thermal margins are shown on the right side of Figure 21. Variations in
neutron flux and the resulting power generation rate in the core in the radial and axial
directions are represented by the corresponding peaking factors. These must be added to the
nominal value (the core average value) to obtain the nominal peak steady-state value of the
margin parameter. Then uncertainties in the calculated parameters and other engineering
uncertainties are “stacked” on top of the peaking factors to arrive at the maximum peak
steady-state condition.
During transients, the appropriate overpower factor must be stacked on top of the maximum
peak steady-state power to obtain the limit value for design transients. On top of this limit,
there should be a margin for all uncertainties in calculating the margin parameter values and
the uncertainties in monitoring global and local reactor parameters. This brings the margin
parameter to the failure limit.
The left side of Figure 21 indicates that if reactor operation drifts to the lower or upper alarm
point, the reactor control systems will provide corrections to certain reactor parameters to
bring the reactor operation back within the safe operating envelope. Beyond the safe
operating envelope is the safe design envelope. If any of the reactor key safety parameters
drifts into this buffer zone, a reactor trip may occur and shut down the reactor. The reactor
trip parameters have their own margins, which are required to manage reactor shutdown
adequately and to ensure safe reactor shutdown while preventing spurious trips.
5.4 Problems
1. Explain the relationship among the reactor core thermal parameters: reactor core
power, core volumetric generation rate, heat flux, critical heat flux, and linear power.
2. Draw a diagram of core thermal margins to explain the relationship among the various
elements that must be considered to ensure that appropriate margins are available.
6 Thermal-Hydraulic Design Fundamentals

This section covers the fundamental aspects of thermal-hydraulic design, heat transfer, and
fluid flow in the reactor primary and secondary cooling system. It also provides the
fundamentals of two-phase flow and the thermodynamics of the nuclear energy conversion
process.
6.1 Two-phase flow fundamentals

Heat transfer and fluid flow with boiling water play an important role in nuclear reactor
thermal hydraulics. This section discusses two-phase flow definitions and basic theory. It
covers flow regimes, experimental observations and data, and application to reactor thermal
hydraulics. It also provides a brief description of various two-phase flow models.
6.1.1 Key parameters and definitions

Two-phase flow is encountered in many engineering systems in the chemical, process, power
generation, and petroleum industries, including oil-gas pipelines, boilers, heat exchangers,
refrigeration equipment, and evaporators, as well as in nuclear reactor applications. Fully
understanding the implications and models of two-phase flow is important for proper design of
reactor cooling systems and for modelling fluid flow and heat transfer in the reactor under
different conditions. Two-phase flow has a profound impact on the ability to remove heat from
the nuclear reactor, which is the primary objective of thermal-hydraulic design.
As a simple approximation, two-phase flow can be treated as an extension of single-phase flow.
However, in reality, two-phase flow can be very complex due to uncertainty in various
interfacial parameters. To simplify the situation, often experimentally obtained correlations
are used in design calculations.
Two-phase flow is simultaneous flow of any two phases (liquid-gas/vapour, solid-gas, or liquid-
solid) of a single substance. Examples include reactor fuel channels and steam generators. This
phenomenon is also referred to as “single-component two-phase flow”. However, two-phase
flow can consist of two components, and in this case, it is referred to as simultaneous flow of
the liquid and gas phases of two substances, for example in oil-gas pipelines. This is also
referred to as “two-component two-phase flow” [HET1982].
The primary parameters used in two-phase flow modelling are:
 Thermal: thermal power, temperature, heat flux, etc.
 Hydraulic: pressure, mass flow rate, fluid temperature, pressure drop, etc.
 Geometric: flow and heated areas, hydraulic and heated equivalent diameters, etc.
Along with these primary parameters, in two-phase flow analysis, the following calculated
parameters are commonly used:
 Mass flux, heat flux
 Mass quality, equilibrium quality, thermodynamic quality
 Void fraction.
In addition, two-phase flow calculations require fluid property information such as density,
viscosity, enthalpy, thermal conductivity, and heat capacity, which are functions of the primary
fluid parameters listed above.
Void fraction is the ratio of the cross-sectional area occupied by vapour and gaseous phases to
the total flow area of a pipe [WAL1969]. The opposite of void fraction is the liquid fraction, as
given by Eq. (1):
Ag Af
 ; (1   )  . (1)
A A
Equation (2) defines mass quality as the ratio of vapour mass flow to total mass flow. The
opposite is the liquid quality:
Wg Wg Wf Wf
x  ; (1  x)   . (2)
W W f  Wg W W f  Wg
Mass flux is the mass flow rate per unit flow area and is given by Eq. (3):
W 
G    . (3)
A 
Gas and liquid mass fluxes are defined using the steam quality, as in Eq. (4):
Gg  G x; Gf  G (1 x) . (4)
Volumetric flux (usually referred to as superficial velocity) is the volumetric flow rate over total
flow area, as in Eq. (5) [WAL1969]:
Q
j . (5)
A
The corresponding vapour and liquid volumetric fluxes are given by relationships similar to the
void fraction, as in Eq. (6):
Qg Qf
jg  ; jf  . (6)
A A
Using the relationships defined in the previous equations, the vapour and liquid phase
velocities can be expressed as in Eqs. (7) and (8):
Wg Qg Gx
vg    , (7)
 g Ag Ag g 
Wf Qf G (1  x )
vf    . (8)
 f Af Af  f (1   )
These equations assume that no relative motion exists between the two phases. However, if a
two-phase mixture is moving, the vapour, because of its buoyancy, density, and different
resistance characteristics, tends to move at a higher velocity than the liquid. In a homogeneous
system, a slip ratio S is defined as equal to one in the absence of flow or in homogeneous flow
and greater than one in non-homogeneous two-phase systems. S is defined as the ratio of the
average velocity of the vapour to that of the liquid. Hence, the slip ratio is defined as the ratio
between the vapour velocity and the liquid velocity, as in Eq. (9) [WAL1969]:
 g Wg  f A f  x    f 1 .
S       (9)
 f W f  g Ag  1  x    g   
Experimental data or theoretical correlations for S covering all possible operating and design
variables do not exist. In boiling-water reactor studies, values for S may be estimated from
data that closely approach those of interest. To do this, a certain amount of individual
judgment is necessary. Otherwise, experimental values of S under conditions similar to a
particular design must be obtained. This procedure is usually expensive and time-consuming,
but may be necessary in some cases.
The importance of obtaining accurate values of S may best be emphasized by the following
discussion. One step in the core channel design procedure is to set a maximum value of alpha
at the channel exit. This is usually determined from nuclear (moderation) considerations. A
corresponding value of x at the selected S is then determined from the above equations. The
latter determines the heat generated in the fuel channel. In design, the usual procedure is to
assume a constant value of S along the length of the fuel channel. This simplification may
introduce error into the results. However, S has been observed to be fairly constant over most
of the channel length, indicating this assumption to be a good one.
The quality of the flowing liquid-vapour mixture can be expressed using Eq. (2) in terms of
liquid and vapour mass flow rates. There is an equivalent quality of a stationary liquid-vapour
mixture, which can be expressed by a similar relationship involving liquid and vapour mass.
The mass flow quality varies between 0% and 100%, i.e., no vapour to 100% vapour in the flow.
Note that mass quality does not carry information about the thermal state of the fluid.
However, where appropriate in experimental or theoretical work, if the two phases are in
thermal equilibrium (at the same enthalpy or temperature), mass quality can be called
equilibrium mass quality.
Often in thermal-hydraulic experiments or calculations, thermodynamic quality is used. It is
expressed in terms of a ratio of enthalpies, as in Eq. (10), where hm is the mixture enthalpy and
hg and hf are vapour and liquid saturation enthalpies. Thermodynamic enthalpy can be
negative when the liquid is below saturation condition (sub-cooled) or greater than one when
steam is superheated. From Eq. (10), thermodynamic quality does not carry information about
the flows (velocities) of the two phases. Usually, thermodynamic quality is expressed for a
mixture at hydrodynamic equilibrium (i.e., both phases travelling at the same velocity) and is
sometimes referred to simply as equilibrium quality [COL1972, DEL1981]:
hm  h f
xth  . (10)
hg  h f
It is very important to understand the relationship between mass flow quality and void fraction.
There is no meaningful relationship between thermodynamic quality and void fraction because
void fraction refers to flow geometry, whereas thermodynamic quality refers to the thermal
properties of the phases. The relationship between mass flow quality and void fraction is
obtained by manipulating the above equations and can be expressed as in Eq. (11) [WAL1969]:
x  vg 1
  . (11)
1  x  v f  x  vg  1 x   v f 
1   
 x   vg 
Figure 22 shows the relationship between void fraction and mass flow quality. As expected, at
or above the critical pressure of 22.12 MPa, the void fraction and quality are identical because
the liquid and vapour phases no longer exist for water regardless of pressure and temperature.
Further examination of Figure 22 shows the following:
1. For constant quality, void fraction decreases with pressure.
2. For any pressure, dα/dx decreases with quality.
3. At low quality values, dα/dx increases as pressure decreases and becomes very large at
low pressure.
4. At atmospheric pressure, a low quality (about 2%) generates almost 100% void fraction
because of the low vapour density.
Figure 22 Relationship between quality and void fraction [COL1972]

For CANDU fuel channel exit conditions, because the pressure is approximately 10 MPa and the
quality is typically 2%–4%, the void fraction can be as high as 30%.
6.1.2 Non-dimensional numbers

Dimensionless numbers in fluid mechanics are an important tool to capture various parameters
and phenomena into a relationship that is applicable to a wide range of parameter values. This
section covers the most important dimensionless numbers that are used throughout this
chapter and are important in analyzing mass, momentum, and heat transfer. The most
important dimensionless numbers are listed in the following sections.
6.1.2.1 Reynolds number
The Reynolds number (Re) is a dimensionless number that measures the ratio of inertial to
viscous forces and consequently quantifies the relative importance of these two types of forces
under a given flow condition. The Reynolds number can be defined for a number of different
situations where a fluid is in motion relative to a surface. These definitions generally include
the fluid properties of density and viscosity, plus a velocity and a characteristic length or
characteristic dimension. This dimension is a matter of convention; for example, a radius or a
diameter is equally valid for spheres or circles, but one is chosen by convention. For flow in a
pipe or a sphere moving in a fluid, the internal diameter is generally used today. Other shapes
such as rectangular pipes or non-spherical objects have an equivalent diameter defined. The
velocity may also be a matter of convention in some circumstances, notably stirred vessels.
With these conventions, the Reynolds number is defined by Eq. (12) [HET1982, ELW1978]:
  v2 L2   v  L v  L
Re    , (12)
 v L  
where:
 v is the mean velocity of the object relative to the fluid (m/s)
 L is a characteristic linear dimension (the length travelled by the fluid; the hydraulic
diameter when dealing with river systems) (m)
 µ is the fluid dynamic viscosity (Pa·s or N·s/m² or kg/(m·s))
 υ is the kinematic viscosity (υ = µ/ρ) (m²/s)
 ρ is the fluid density (kg/m³).
For flow in a pipe, the Reynolds number is given by Eq. (13) [WAL1969]:
  v  DH v  DH Q DH
Re    , (13)
   A
where:
 DH is the hydraulic diameter of the pipe, i.e., its characteristic length travelled (m)
 Q is the volumetric flow rate (m3/s)
 A is the pipe cross-sectional area (m²).
For shapes such as square, rectangular, or annular ducts where the height and width are
comparable, the characteristic dimension for internal flow situations is taken to be the
hydraulic diameter DH as defined by Eq. (14):
4A
DH  , (14)
P
where A is the cross-sectional area and P is the wetted perimeter. The wetted perimeter for a
channel is the total perimeter of all channel walls that are in contact with the flow. This means
that the length of the channel exposed to air or steam is not included in the wetted perimeter.
6.1.2.2 Nusselt number
In heat transfer at a boundary (surface) within a fluid, the Nusselt number (Nu) is the ratio of
convective to conductive heat transfer across (normal to) the boundary, as in Eq. (15)
[HET1982, ELW1978]. In this context, convection includes both advection and conduction.
Named after Wilhelm Nusselt, it is a dimensionless number. The conductive component is
measured under the same conditions as the heat convection, but with a (hypothetically)
stagnant (or motionless) fluid.
A Nusselt number close to one, meaning that convection and conduction are of similar
magnitude, is characteristic of “slug flow” or laminar flow. A larger Nusselt number
corresponds to more active convection, with turbulent flow typically in the 100–1000 range.
The convection and conduction heat flows are parallel to each other and to the surface normal
of the boundary surface and are all perpendicular to the mean fluid flow in the simple case:
Convective heat transfer h  L
NuL   , (15)
Conductive hetatransfer k f
where:
 L is the characteristic length (m)
 kf is the thermal conductivity of the fluid (kW/m °K)
 h is the convective heat transfer coefficient (kW/m2 °K).
The characteristic length should be selected in the direction of growth (or thickness) of the
boundary layer. Some examples of characteristic length are the outer diameter of a cylinder in
(external) cross flow (perpendicular to the cylinder axis), the length of a vertical plate
undergoing natural convection, or the diameter of a sphere. For complex shapes, the length
may be defined as the volume of the fluid body divided by the surface area. The thermal
conductivity of the fluid is typically (but not always) evaluated at the film temperature, which
for engineering purposes may be calculated as the mean (average) of the bulk fluid
temperature and the wall surface temperature.
For relations defined as a local Nusselt number, the characteristic length should be taken as the
distance from the surface boundary to the local point of interest. However, to obtain an
average Nusselt number, this relation must be integrated over the entire characteristic length.
Typically, for free convection, the average Nusselt number is expressed as a function of the
Rayleigh number and the Prandtl number as: Nu=f(Ra, Pr) [HET1982]. For forced convection,
the Nusselt number is generally a function of the Reynolds number and the Prandtl number, or
Nu=f(Re, Pr). Empirical correlations are available for a wide variety of geometries that express
the Nusselt number in the forms described above.
6.1.2.3 Prandtl number
The Prandtl number (Pr) is a dimensionless number that represents the ratio of momentum
diffusivity (kinematic viscosity) to thermal diffusivity [HET1982, ELW1978]. It is named after
the German physicist Ludwig Prandtl and is defined by Eq. (16):
 viscous diffisuion rate c p 

Pr    , (16)
 thermal diffusion rate k
where:
 υ is the kinematic viscosity, υ=µ/ρ (m2/s)
 α is the thermal diffusivity, α=k/(ρcp) (m2/s)
 µ is the dynamic viscosity (Pa s = N s/m2)
 k is the thermal conductivity (W/(m K))
 cp is the specific heat (J/(kg K))
 ρ is the density (kg/m3).
Note that whereas the Reynolds number is subscripted with a length scale variable, the Prandtl
number contains no such length scale in its definition and is dependent only on the fluid and
the fluid state. As such, the Prandtl number is often found in property tables alongside other
properties such as viscosity and thermal conductivity.
6.1.2.4 Weber number
The Weber number (We) is a dimensionless number in fluid mechanics that is often useful in
analyzing fluid flows with an interface between two different fluids, especially for multiphase
flows with strongly curved surfaces. It can be thought of as a measure of the relative
importance of the fluid’s inertia compared to its surface tension [HET1982, ELW1978]. The
quantity is useful in analyzing thin film flows and the formation of droplets and bubbles. It is
named after Moritz Weber and can be expressed by Eq. (17) [HET1982, ELW1978]:
fluid inertia   v2  l
We   , (17)
surfacetension 
where
 ρ is the density (kg/m3 )
 v is the mean velocity of the object relative to the fluid (m/s)
 l is the characteristic length, typically the droplet diameter (m)
 σ is the surface tension (N/m).
6.1.3 Flow patterns

When a liquid is vaporized in a heated tube or a heated channel, the liquid and vapour
generated take on a variety of configurations known as flow patterns. The particular flow
pattern depends on pressure, flow, heat flux, entrance conditions (i.e., local phase distribution
at the inlet), and channel geometry; however, the strongest influences on the flow pattern are
the phase mass flow rates or velocities. Because the name given to a flow pattern is largely
subjective, many terms exist in the literature which purport to describe the various possible
phase distributions. The sequence of flow patterns generally encountered in vertical upwards
concurrent flow for increasing levels of thermodynamic quality is shown below. As steam mass
flow increases (i.e., velocity increases) and steam temperature increases, the fluid flow
patterns change from single-phase

phase sub-cooled liquid to single-phase
phase steam flow, passing
through the flow regimes as indicated from left to right on the diagram below. Indications are
provided in this diagram of the values of thermodynamic quality x for various flow regimes.
Single- Single- Bubbly

ubbly Slug Churn Wispy Single-phase Single-phase
phase phase flow flow flow annular saturated superheated
sub-cooled saturated flow steam flow steam flow
liquid flow liquid flow
(x < 0) (x = 0) (0 < x < 1) (x = 1) (x > 1)
Figure 23 shows the flow patterns for vertical flow. These flow patter
pattern
ns are generally
symmetrical across the pipe cross section.
In horizontal flow, the flow patterns are similar. However, in horizontal flow, gravity has a
different effect, and depending on the mass flow rate in the tube or channel, stratified flow
may occur with bubbly flow, slug flow, and wispy annular flow patterns.. The flow patternspatt
typical of horizontal two-phase
phase flow are shown in Figure 24. A cross-sectionalal view of various
flow patterns in horizontal flow is shown in Figure 25,, in which gravity plays a significant role,
so that the combination of gravity and drag force at the liquid
liquid-vapour
vapour interface can result in
various cross-sectional
sectional flow patterns.
Figure 23 Flow patterns in vertical two

two- Figure 24 Flow patterns in horizontal
phase flow two-phase flow
a) Flow regimes in a horizontal heated channel
b) Evolution of flow regimes in a horizontal tube
Figure 25 Flow patterns and heat transfer regimes in horizontal heated tubes
The distribution of phases inside a confined area strongly depends on:
 Liquid and vapour velocities.
 Fuel channel geometry and fuel bundle geometry, particularly in reactors where fuel
channels are interconnected in the radial direction, like PWRs and BWRs.
 Surface (wall) heating, which influences near-wall flow patterns, resulting in an internal
void gradient.
 Appendages, which homogenize flow patterns at downstream locations; the patterns
transition back to a basic pattern at locations farther away.
Figure 26 shows the flow patterns and heat transfer regimes for two different situations in
vertical flow [DEL1981]. On the left side, a reactor channel is shown with the following
parameters: high liquid mass flow rate, high heat flux from the walls, and high water sub-
cooling, which is typical of PWRs. To the left of the reactor channels on the first vertical line,
corresponding flow patterns are shown for this situation, whereas the leftmost line shows the
heat transfer regimes for these flow patterns. Starting from bottom to top, the following pairs
of flow pattern and heat transfer regime (note that “film” on this figure refers to steam film
next to the heated walls) are typically observed:
a) Single-phase liquid – single-phase forced convection to liquid;

b) Bubbly flow – nucleate boiling;
c) Inverted annular flow – inverted annular flow film boiling;
d) Dispersed droplet flow – dispersed flow film boiling; and
e) Single-phase vapour – single-phase forced convection to vapour.
Note that in Figure 26, the point where the wall loses contact with the liquid is shown at the
transition between bubbly flow and inverted annular flow.
Figure 26 Flow patterns and heat transfer regimes in vertical heated tubes
On the right side of Figure 26, a reactor channel is shown with the following parameters:
low liquid mass flow rate, medium heat flux from the walls, and saturated water; this is
typical of BWRs. To the right of the reactor channel on the first vertical line, corresponding
flow patterns are shown for this situation, whereas the rightmost line shows the heat
transfer regimes for these flow patterns. Starting from bottom to top, the following pairs of
flow pattern and heat transfer regime (note that “film” in this figure refers to steam film
next to the heated walls) can be observed:
a) Single-phase liquid – single-phase forced convection to liquid;
b) Bubbly flow and slug flow – nucleate boiling;
c) Annular flow – forced convective evaporation;
d) Dispersed droplet flow – dispersed flow film boiling;
e) Single-phase vapour – single-phase forced convection to vapour.
Figure 27 Flow pattern map in vertical flow [HET1982]
Note that in Figure 26, the point where the wall loses contact with the liquid is shown at the
transition between annular flow and dispersed droplet flow. Figure 26 indicates that the flow
patterns at the entry and exit of the fuel channel are similar for the two cases. The difference
is how and where the wall and the liquid lose contact. In addition, the amount of water or the
void fraction percentage in the channel does not necessarily determine the contact of liquid
with the heated walls. Therefore, as long as the walls are in contact with liquid water, heat
transfer is good, and the amount of water in the channel is not important.
Two-phase flow parameters change significantly from one flow pattern to another, in particular
the interfacial area between the two phases, and this has a significant impact on the exchange
of mass, momentum, and heat between phases. In principle, the degree of thermal and
hydraulic disequilibrium between phases has a significant impact on the sustainability of
certain flow patterns. Therefore, certain flow patterns, like churn flow, exhibit instability and
fast transitions. Various two-phase flow parameters are affected differently by flow patterns,
and hence various correlations and models are needed to capture phenomena for each flow
pattern. This implies the need for well-defined and predictable flow patterns in two-phase
flow modelling.
Many scientists and researchers in recent decades have tried to correlate flow patterns using
selected parameters.
Figure 27 shows such an attempt for a vertical flow regime. In this figure, the flow patterns are
defined as a function of gas and liquid momentum fluxes. In this example, the flow patterns
are correlated by several dimensionless number groups.
6.1.4 The boiling process

Because water as the working fluid in LWRs and CANDUs undergoes a change of phase during
heat transfer from the nuclear reactor primary to the secondary cooling system, it is important
to understand the boiling process, its phenomena, and its parametric trends.
As liquid is being heated to saturation, with additional heating, boiling starts. Hence, boiling
initiation is strongly dependent on pressure, i.e., the saturation temperature. There are two
distinctive types of boiling: one is referred to as pool boiling, and the other is forced convective
boiling. Vapour formation can take the following forms:
 at the liquid surface to the open environment, when the liquid is superheated;
 homogeneous (internal heating within the fluid at the molecular level); and
 heterogeneous (requires nucleation sites in the fluid or at the heated surface).
Heterogeneous boiling from nucleation sites is a complex boiling mechanism that depends on a
number of forces acting on a vapour bubble, including dynamic forces, buoyancy, and surface
tension forces. Much research has been devoted to bubble formation at nucleation sites and is
presented in a number of two-phase flow textbooks [TON1965, TON1975]. Further discussion
of bubble dynamics and formation at nucleation sites is beyond the scope of this textbook.
Figure 28 shows a three-dimensional representation of the two-phase boiling water region as a
function of heat flux, temperature, and thermodynamic quality. This figure indicates different
flow and heat transfer regimes along the path of changing thermodynamic quality. At high
water sub-cooling, the critical heat flux is high, but it diminishes as quality increases. The
definition of critical heat flux and its impact on heat transfer in the two-phase region are
discussed in Section 7.7.
The boiling process can be analyzed in two distinct situations: pool boiling and forced
convective boiling. Pool boiling refers to conditions typical of vessels with unlimited volume
and under near-atmospheric pressure. Forced convective boiling refers to boiling in enclosed
piping and vessels in which pressure may reach very high values, and in which flow is ensured
by an operating pump.
Figure 29 shows a pool boiling process in a diagram relating heat flux to the degree of wall
temperature superheating (i.e., the temperature difference between the wall temperature and
the saturation temperature). This figure shows a cross section of the 3-D curve in Figure 28 at
low quality. Three heat transfer regimes can be seen: nucleate (or sub-cooled) boiling,
transition boiling, and film boiling. These three regimes are separated by two important points:
the critical heat flux, and the minimum film boiling temperature. The minimum film boiling
temperature, sometimes called the Leidenfrost temperature, represents the maximum
temperature at which wetting of the heated surface is possible.
Figure 29 also indicates in which direction the process evolves from left to right and vice versa.
If heat flux is increasing, then after the CHF point is reached, the surface becomes detached
from liquid contact. Beyond this point, heat transfer by liquid evaporation (latent heat of
vaporization) is severely reduced. Therefore, because heat transfer to liquid is severely
reduced, but heat flux continues to be delivered from internal regions of the wall to the wall
surface, the wall surface temperature significantly increases, as represented by the horizontal
dashed line on the diagram from CHF to the film boiling curve (sometimes referred to as the
heat flux-controlled region). This significant increase in wall surface temperature can lead to
surface damage and melting, which is very troubling in a nuclear reactor where the fuel rod
cladding is a major barrier to fission product release.
Figure 28 3-D representation of boiling surfaces

On the other hand, as the heat flux is reduced from the region of film boiling back to the CHF
point, moving along the bottom part of the curve, it reaches the minimum film boiling
temperature. From this point, the heat flux starts to increase in the transition boiling region,
during which the heated surface becomes intermittently wetted and dried while the wall
temperature is reduced. Once the CHF point is reached by moving from right to left in the
diagram, the wall surface is in continued and sustainable contact with the liquid coolant, and
heat flux is reduced below the CHF value. This path is sometimes referred to as the
temperature control part of the boiling curve.
Figure 29 Boiling curve
Figure 29 shows a boiling curve with heat transfer regimes clearly marked. This figure
represents one cross section of Figure 28. It shows the heat flux below saturation (in the sub-
cooled region), where forced convection to liquid occurs. The transition point between forced
convection to liquid and nucleate boiling is called the onset of nucleate boiling and designates
the conditions under which the first bubbles start to grow on the wall surface.
Figure 30 shows the parametric trends on the boiling curve [LEU2004, POP2014]. As the mass
flux increases in an annular flow regime, the minimum wetting temperature increases, and the
boiling curve flattens out. As the mass flux increases in forced convective sub-cooled or low-
quality boiling, the CHF and the minimum wetting temperature increase, but the boiling curve
retains its shape.
Figure 30 Parametric trends of the boiling curve

A similar effect is seen when increasing the degree of liquid sub-cooling; the CHF point and the
minimum wetting temperature increase, but the boiling curve retains its shape. The effect of
quality is more complex and is shown in Figure 31.
Figure 31 [LEU2004, POP2013] shows the impact of thermodynamic quality on the boiling
curve. For quality equal to or greater than one, the boiling curve is transformed into a straight
line, and therefore the CHF concept is not relevant, but the wall temperature increases as the
quality increases above one. For saturated boiling (x = 0), the boiling curve has the usual S-
shape. As quality increases above zero, the CHF significantly decreases. Moreover, if the
quality drops further below zero (i.e., the liquid becomes more sub-cooled), the CHF also
increases, and the boiling curve becomes steeper in the nucleate boiling region.
Figure 31 Quality effects on the boiling curve

Figure 32 shows the flow patterns and heat transfer regimes in a boiling channel. In the top
part of the figure, the temperature distributions along the heated channel in the coolant and at
the wall surface are shown. It is possible to recognize the flow patterns that are most likely to
occur given the temperature difference between the coolant and the wall. Note that nucleate
boiling starts after the wall surface temperature has risen above the water saturation
temperature by a certain amount. Nucleate boiling continues in areas C through F, and then at
the point of dryout, when the CHF has been reached, the wall surface temperature exhibits a
rapid increase as wall cooling decreases due to loss of contact between the wall and the liquid.
Figure 32 Heat transfer regimes in a boiling horizontal channel
Figure 33 provides model details on the transition from forced convection to liquid and
nucleate boiling. It also shows the point of initiation of nucleate boiling, when the first bubbles
appear on the wall surface. At the point of net vapour generation, the bubbles can detach
from the walls and move to the bulk liquid. A further increase in heat flux at the wall surface
happens when the liquid temperature reaches saturation, which creates an opportunity for
bubbles to become detached from the wall in great numbers and move with the liquid flow,
leading to significant phase separation and disequilibrium.
Figure 33 Details of sub-cooled boiling phases in horizontal flow

Three main types of models are used to analyze two-phase flow:
 “homogeneous” flow models,
 “separated” flow models, and
 “flow pattern” or “drift flux” models.
The basic equations are solved within the framework of each of these idealized
representations. To apply these models, it is necessary to know when each should be used and
to be able to predict the transition from one pattern to another.
6.1.5 Two-phase flow models

The homogeneous model and the separated flow model are the two most widely used and
tested models of two-phase flow that are presently available. The homogeneous model
considers two-phase flow as a homogenized mixture (pseudo-fluid) possessing average fluid
properties. It neglects any local effect that flow patterns may have on two-phase flow and
ignores interaction between phases. In other words, the homogenous model considers the
phases to be in hydrodynamic and thermodynamic equilibrium.
The homogeneous model is applicable only if the flow parameters do not vary rapidly and
thermal disequilibrium has little influence. Although the homogeneous model predicts the
dependence of the two-phase friction multiplier on pressure and quality reasonably well, it has
two unsatisfactory features: (a) the friction multiplier is a function of pressure and quality only
and is independent of mass flux; and (b) it generally under-predicts low-quality data. Many
early software codes used in CANDU design, such as SOPHT [SOP1980], used this type of
modelling.
The separated flow model considers the two phases to be segregated into two streams: one of
liquid, and one of vapour. Conservation equations are written separately for each phase, and
interaction between phases is taken into account by constitutive relationships. The basic
equations for the separated flow model are not dependent on the particular flow configuration
adopted. The basic assumptions of the separated flow model in analyzing two-phase pressure
drop are:
a) The velocities of each phase are constant, but not necessarily equal in any given cross
section within the zone occupied by the phase.
b) Thermodynamic equilibrium exists between the two phases.
c) Empirical correlations can be used to relate the two-phase friction multiplier and the
void fraction to the dependent flow variables.
Among the various flow patterns, this model would be expected to be most valid for the
annular flow pattern. Newer versions of the separate flow model can accommodate both
thermodynamic and hydrodynamic disequilibrium between the flow phases.
The drift flux approach satisfactorily accounts for the influence of mass velocity on the void
fraction as seen in the separated flow model, and an empirical expression may be used to
provide the required relationship between the void fraction and the independent flow pattern.
With the separate flow model and the drift flux model, appropriate void fraction calculations
must be performed to ensure model fidelity. Many correlations have been proposed and used
in various models; a listing of these correlations is beyond the scope of this handbook (further
details can be found in [WAL1969]). Very often used as a void fraction correlation is the
homogeneous equation given by Eq. (18):
xa  g
 . (18)
(1  xa )  f  xa  g
Another often-used void fraction correlation is the Armand-Massina correlation [LEU2004],

which is given by the following equation:
(0.833  0.167 xa ) xa  g
 . (19)
(1  xa )  f  xa  g
6.1.6 Two-phase flow in fuel bundles

Two-phase flow in fuel bundles is very complex and requires significant experimental support
to understand important phenomena and develop adequate models and correlations. Further
information about fuel bundle effects will be given in later sections of this chapter.
The CANDU fuel bundle experimental void fraction database includes data for bundles with 3 to
37 fuel pins, in addition to tube and annulus data. The experimental database includes a wide
range of flow conditions with uniformly heated fuel elements or fuel elements with certain
radial and axial heating profiles. Figure 34 illustrates that bundle geometries with 6 and 37 fuel
elements do not significantly affect the relationship between void fraction and quality
[LEU2004, POP2013].
Figure 35 also shows that the impact of the hydraulic diameter of a 7-element bundle does not
have a significant impact on the relationship between void fraction and thermodynamic quality,
except for very low quality values (representing highly sub-cooled conditions) [LEU2004,
POP2013].
Figure 34 Effect of tube diameter on quality and void fraction
Figure 35 Effect of bundle geometry on quality and void fraction
6.1.7 Problems
1. Explain the relationship between the mass quality and the void fraction in two-
phase flow.
2. Explain the difference between the mass quality and the thermodynamic quality in
two-phase flow.
3. Explain the main characteristics of the boiling curve, the main parameters that are
important for two-phase flow calculations, and the main parameters that influence
its shape.
4. Explain the key flow patterns that can occur in vertical and horizontal two-phase
flow in heated channels. Explain how the flow regime maps are developed and used
in reactor thermal-hydraulic design.
5. Explain the possible models used in two-phase flow calculations, their range of
applicability and key features.
6. A BWR reactor operates at a thermal power of 1400 MWt. Water enters the bottom
of the reactor core at 275 °C, and passes through the core at a flow rate of 6050
kg/s. the reactor operates at pressure of 7 MPa. Calculate the steam flow rate
produced in the reactor.
6.2 Thermodynamics of nuclear energy conversion

This section discusses the thermodynamics of energy conversion in nuclear reactor systems,
including a description of the laws of thermodynamics. It provides definitions of basic
thermodynamic laws, followed by a description of reactor power cycles, starting with the
Carnot cycle, and followed by the Rankine cycle. The definition and analysis of cycle thermal
efficiency is discussed, and methods for improving efficiency are described.
6.2.1 Definitions
This section describes the most important concepts and parameters in reactor thermodynamic
analysis. Consider a system undergoing a change from state 1 to state 2, as shown in Figure 36.
Initially, the system is at energy level E1. Then, after a certain amount of energy, Q, is added to
the system and a certain amount of work, W, is performed by the system, the system is
brought to energy level E2. It would be interesting to discover by what parameters this change
is driven and how effective is the energy exchange between the system and its surroundings.
Figure 36 Change of thermodynamic states in a system
6.2.1.1 Work
Figure 37 shows a practical example of the general concept shown in Figure 36. In Figure 37, a
piston is moved from a position 1 to a position 2 by sequentially moving small amounts of
weight applied at the piston stem. Initially, the piston is under pressure p1, with volume V1 and
temperature T1. After moving a certain amount of weight, the piston ends up at pressure p2,
volume V2, and temperature T2. The system volume changed by dV=Aˑdx each time a weight
was removed.
Figure 37 Concept of parameters of thermodynamic state in piston geometry

The infinitesimal amount of work done by the system (shown in Figure 37) is given by Eq. (20)
[SAA1966]:
dW  f dx  p A dx  p dV . (20)
Assuming that the fluid is an ideal gas and that the above process has been performed at a
constant temperature (i.e., is an isothermal process), then the total work performed by weight
removal is given by Eq. (21):
dV V 

2

2
W1 2  p dV  mRT  mRT ln  2  . (21)
V  V1 
1 1
For a constant-pressure process (i.e., an isobaric process), the total work is given by Eq. (22):
W12  p dV  p(V2 V1 ) .

2
(22)
1
For a constant-volume process, because the piston does not move, no work is performed, as
stated by Eq. (23):
W12   p dV  0 .
2
(23)
1
One can use the ideal gas law from thermodynamics of gases, as given by Eq. (24):
p V  m  R  T , (24)
where P is pressure [MPa or N/m2]; V is volume [m3]; m is mass [kg]; T is temperature [°K]; and
R is the gas constant [Nˑm/(kg °K) or J/(kg °K)].
6.2.1.2 Heat
Heat as a form of energy exists in all bodies and is essential for energy transfer. The nature of
heat is associated with the movements or vibration of molecules or atoms at the microscopic
level in any body, solid, liquid, or gas. Each energy transfer process eventually ends up being a
heat transfer process, i.e., it involves transfer of molecular movement or vibration from one
body to another. Even the energy exchanged by radiation eventually ends up being transferred
into molecular movement. Transfer of energy as heat is a microscopic process because it

represents the microscopic motion and interactions of microscopic constituents such as
molecules and photons.
The level of molecular movement in a certain body or system is measured in terms of its
temperature. The higher the level of excitement of the microscopic constituents of a system,
the higher will be its temperature. Hence, a temperature difference between two bodies or
two systems is essential to the heat transfer process because it sets up the potential for energy
transfer.
Often assumptions and simplifications are made in heat transfer processes to analyze and
understand certain aspects more effectively [BIR1960, ROH1985]. One of these is the
assumption of an adiabatic process, which means that during the process, no heat transfer to
and from a given system undergoing adiabatic change is made with any surrounding systems.
Another important assumption is to treat a system as isolated from the surrounding systems.
This means that in addition to no heat transfer with the surrounding systems, there is also no
work exchanged with the surrounding systems.
6.2.1.3 Energy
Energy is a conserved property of a physical system, which cannot be observed directly, but can
be calculated from the system’s state. The energy of a system can present itself in various
forms and is therefore difficult to describe by a comprehensive definition. Often, the best
description of energy is as “the capacity to perform work”.
For the thermal-hydraulic design engineer, three energy components are of most importance
[LEU2004, POP2013]:
 Internal energy, EU [kJ], or specific internal energy eu [kJ/kg] (internal energy is usually
simply expressed in U, or u);
 Kinetic energy, EK [kJ], or specific kinetic energy, ek [kJ/kg]; and
 Potential energy, EP [kJ], or specific potential energy, ep [kJ/kg].
Internal energy describes the thermodynamic energy of system molecules or atoms (referred
to as heat). Kinetic energy is the energy of motion, which is carried by a body or system with a
certain mass. Potential energy is energy possessed by a system when moving from one place
to another in a force field, such as a gravitational field, an electro-magnetic field, or the nuclear
force field in the atom. Most often for a thermal-hydraulic engineer, potential energy denotes
the gravitational potential.
The total energy of a system in thermal-hydraulic terms can be expressed as the sum of the
above components by Eq. (25) or Eq. (26):
E  EU  EK  EP U  EK  EP , (25)
e  eu  ek  ep  u  ek  ep . (26)
The energy of a system depends on the temperature, pressure, velocity, and elevation of the
system with respect to its surroundings.
6.2.1.4 Enthalpy
Enthalpy is a measure of the total energy of a thermodynamic system. It includes the internal
energy, U (or thermodynamic potential) and the volume and pressure (the energy required to
make room for a system change). Enthalpy is useful in describing system thermodynamic
changes because it simplifies the description of energy transfers and system changes. Enthalpy
is a thermodynamic property of a substance.
Consider a system undergoing a change because of addition of heat (energy) to the system at
constant temperature; its total energy change can be described in terms of the system’s
internal energy and the product of the system’s pressure and volume, as in Eq. (27):
2 2 2
 dQ   dU   P  dV .
1 1 1
(27)
The sum of the system’s internal energy and the product of its pressure and volume defines its
enthalpy level, as shown by Eq. (28):
H U  PV . (28)
On a unit mass basis, the specific enthalpy is shown by Eq. (29), where specific heat is q = Q/M,
specific internal energy is u = U/M, specific volume is v = V/M, M is mass, and P is pressure:
huP v. (29)
Enthalpy is a state parameter and can be expressed in terms of other parameters. For sub-
cooled or super-heated conditions, h  f ( P , T ) , and for saturated conditions, h  f ( P ) or
h  f ( T ) , it is expressed in terms of pressure or temperature. In the two-phase region, the
following equation can be used to calculate mixture enthalpy (under homogeneous conditions):
hm  x  hg  (1 x)  hf  hf  x  hfg , (30)
where hg is the enthalpy of the saturated gas, hf is the enthalpy of the saturated liquid, and hfg
is the latent heat of vaporization.
6.2.2 First law of thermodynamics

The first law of thermodynamics is one of the most fundamental laws of nature; it simply states
that the total energy change in a system is equal to the total energy supplied to the system
from surrounding systems minus the energy transferred to the surrounding systems [SAA1966]:
Energy supplied – Energy removed = Change of energy level.
Hence, for a system with a particular mass that remains constant during the process, the first
law of thermodynamics can be expressed as in Eq. (31):
QW E  E2  E1 . (31)

The above equation written per unit mass becomes Eq. (32):
q we  e2 e1 . (32)
With these definitions, it is now possible to define the work performed by a system that goes
through an isobaric process as in Eq. (33):
W   P dV  P(V2 V1) .
2
(33)
1
Using the first law of thermodynamics, as in Eq. (31), and substituting this equation into Eq.
(33), one obtains the following equation:
Q P(V2 V1) U2 U1. (34)

Rearranging the terms in Eq. (34) yields Eq. (35), which clearly shows the advantage of using
enthalpy because it can provide a better expression of the first law of thermodynamics:
Q U2 U1  P(V2 V1)

 (U2  PV2 ) (U1  PV1)  H2  H1 H . (35)
In the case of a system with multiple energy and work inputs and outputs, the first law of
thermodynamics must be applied to the sum of these inputs and outputs, as in Eq. (36). Thus,
Eq. (37) is obtained as an energy balance equation:
E  0  energyinflow  energyoutflow (36)
1
 mV12  mu1  Q  PV
1 1  mgZ 1
2
1
 mV22  mu 2  W  P2V2  mgZ 2 . (37)
2
Converting Eq. (37) in terms of specific parameters yields the following equations in terms of
internal energy (Eq. (38)) and enthalpy (Eq. (39)):
 1   1 
q  w  u2  Pv2  V22  gZ2   u1  Pv1  V12  gZ1  , (38)
 2   2 
 1   1 
q  w  h2  V22  gZ2   h1  V12  gZ1  . (39)
 2   2 
In the case of a turbine, there is no heat addition (q = 0) and no change in elevation for a
horizontal turbine (Z1 = Z2), and therefore:
1 2
w   h1  h2   V1  V22  . (40)
2
For a flow through a horizontal nozzle, there is neither heat addition nor work delivered (q = w
= 0) and no change in elevation (Z1 = Z2), and therefore:
V12  V22  2 h1  h2  . (41)
Finally, there is a special form of the energy balance equation that is very useful to thermal-
hydraulic engineers for calculating fluid parameters in a pipe flow—the Bernoulli equation—in
which neither heat addition nor work is delivered (q = w = 0):
 1 2   1 2 
h1  2 V1  gZ1   h2  2 V2  gZ2   const , (42)
 1 2 
u  Pv  V  gZ   const . (43)
2 
Equation (43) clearly shows that the key fluid parameters for a flow are connected, so that to
add up to a constant, an increase in one parameter must result in a reduction in the other
parameters. For example, considering horizontal flow, an increase in fluid velocity results in a
decrease in fluid pressure, and vice versa.
6.2.3 Reactor cycles

Nuclear reactors directly or indirectly produce steam that performs a thermodynamic cycle,
which is used to convert nuclear energy using the energy conversion chain shown in the
diagram below.
Nuclear fuel Turbine
Steam Generator Electrical
Nuclear fission internal Cooling Fluid rotational
production magnetic field energy
energy energy
Converts the
Converts Converts
Converts fuel Converts kinetic energy Converts
Converts mass kinetic energy steam energy
internal internal of the turbine generator
into kinetic of nuclear into kinetic
energy into energy of the shaft into magnetic
energy of particles into energy
internal coolant into rotation of energy into
nuclear molecular (rotation) of
energy of the steam generator electrical
particles motion of the the turbine
coolant production magnetic energy
fuel mass shaft
poles
The shaded area of the diagram represents the thermodynamic cycle that is implemented with
the objective of delivering work to the electric generator.
This process leads to rejection of a certain part of the thermal energy into the environment
(lake, river, sea, or the atmosphere) to be able to close the cycle; this will be explained in later
sections.
With regard to energy generation, the cycle thermodynamic efficiency, ηt, is defined to provide
a measure of cycle performance:
W
t  . (44)
Qin
6.2.3.1 Carnot cycle
The Carnot cycle is a theoretical thermodynamic cycle that represents the most efficient way of
converting a given amount of thermal energy between two thermal reservoirs into work
(turbine work), or conversely the most efficient way of converting a given amount of work into
creating a temperature difference between two reservoirs (refrigeration).
Figure 38 shows an ideal Carnot cycle in pressure versus volume and temperature versus
entropy diagrams. The Carnot cycle consists of the following processes [SAA1966]:
a) delivering heat to the cycle by a reversible isothermal expansion process (process from
points 1 to 2 in Figure 38);
b) delivering work output by isentropic expansion of gas (process from points 2 to 3 in
Figure 38);
c) rejecting waste heat from the cycle by reversible isothermal compression of gas
(process from points 3 to 4 in Figure 38); and
d) inputting work into the cycle by isentropic compression of gas (process from points 4 to
1 in Figure 38).
Figure 38 Ideal Carnot cycle

To calculate the thermal efficiency of the Carnot cycle, it is necessary to calculate the total heat
delivered and the total work performed and to use these results in Eq. (44).

2
For the process from points 1 to 2 (at constant temperature, U12  m cv dT  0 ):
1
dV V
W12   p dV  m R T1 
2 2
 m R T1 ln 2 . (45)
1 1 V V1
Using the first law of thermodynamics, from Eq. (31), the heat addition can be expressed as:
Q12 W12 U12  0 , (46)
V2
Q12  W12  m R T1 ln . (47)
V1
For the process from points 2 to 3 (with no heat addition, Q23  0 ):

3 3 p3V3  p2V2
W23   p dV   C V  dV  , (48)
2 2   1
U23   m cv dT  m cv (T3  T2 ) .
3
(49)
2
Using the first law of thermodynamics, from Eq. (31), the link between internal energy change
and delivered work is:
Q23 U23 W23  0 . (50)
For the process from points 3 to 4 (at constant temperature, U34  0):
V4
W34  m R T3 ln  Q34 . (51)
V3
For the process from points 4 to 1 (with no heat addition, Q41  0 ):

p1V1  p4V4
W41  , (52)
  1
U41  mcv (T1 T4 ) W41 . (53)
These equations have used the following thermodynamic relationships to calculate the thermal
efficiency:
 For a quasi-static adiabatic process:
 1  1  1  1
TV
2 2  TV
3 3 and TV
4 4  TV
1 1 . (54)
 For an isothermal process
T1 T2 and T3 T4 , (55)
1 1  p2V2
pV 3 3  p4V4 .
and pV (56)
 Volume relation
V2 V3
 . (57)
V1 V4
Hence, the net work done can be obtained by summing up the work in the separate processes
described by Eqs. (45), (48), (51), and (52) and using the relationships provided by Eqs. (54),
(55), (56), and (57):
Wcycle  W12 W23 W34 W41 (58)
V2 p3V3  p2V2 V pV  p V
 m R T1 ln   m R T3 ln 4  1 1 4 4 ,
V1   1 V3   1
V2
Wcycle  mR(T1  T3 ) ln . (59)
V1
Now the cycle thermal efficiency can be obtained from Eq. (44):
net work done Wcycle Wcycle mR(T1  T3 ) ln (V2 / V1 ) T1  T3 T
t       1 3 . (60)
heat input Qin Q12 mR T1 ln (V2 / V1 ) T1 T1
One important observation can be made from Eq. (60): the thermal efficiency of the cycle
depends on the temperatures of the upper and lower thermal reservoirs. Hence, the higher
the temperature at which heat is added to the cycle and the lower the temperature at which
heat is extracted from the cycle, the higher will be the thermal efficiency. Therefore, design
efforts to improve cycle thermal efficiency are usually directed towards increasing the
temperature at which heat is delivered to the cycle. However, it is also evident that this cycle
will operate more efficiently if the lower temperature is reduced.
6.2.4 Second law of thermodynamics

The second law of thermodynamics is a very important law of nature and primarily applies to
thermodynamics. This law indicates that over time, differences in temperature, pressure, and
chemical potential decrease in an isolated physical system, leading eventually, with a time
constant that depends on the system thermodynamic inertia, to a state of thermodynamic
equilibrium.
There are other definitions of this law, depending on the discipline. The above definition is
given in the context of thermodynamics, but this law can also be considered from the
perspectives of quantum mechanics or astrophysics, and in such cases, its definition is adjusted
to the objectives of the corresponding discipline.
The second law of thermodynamics has many implications, of which a few are stated below:
 Carnot’s principle [CAR2015]: The historical origin of the second law of thermodynamics
was in Carnot’s principle. It refers to the Carnot engine, fictively operating in a very
slow mode (quasi-static) so that heat is transferred between two thermal reservoirs
that maintain their internal equilibrium. The Carnot principle, which originated at the
beginning of the 19th Century, is still valid today. It states that, “The efficiency of a
quasi-static or reversible Carnot cycle depends only on the temperature of the two heat
reservoirs, and is independent of the working substance. A Carnot engine operated in
this way is the most efficient possible heat engine using these two temperatures”.
 Clausius’ statement [CLA2015]: “It is impossible to construct a device that executes a
thermodynamic cycle so that the sole effect is to produce a transfer of heat energy from
a body at a low temperature to a body at a high temperature”.
 Kelvin’s statement [KEV2015]: “It is impossible, by means of inanimate material agency,
to derive mechanical effect from any portion of matter by cooling it below the
temperature of the coldest of the surrounding objects”.
 Kelvin-Planck statement [KPL2015]: “It is impossible to construct a device that executes
a thermodynamic cycle, exchanges heat energy with a single reservoir, and produces an
equivalent amount of work”.
 Planck’s principle [PLA2015]: “The internal energy of a closed system is increased by an
isochoric adiabatic process”.
Whichever definition is used for the second law of thermodynamics leads to the basic
conclusion that any time an energy transfer is performed, from thermal to mechanical to
electrical or vice versa, a penalty is paid in terms of increasing the disorder of the isolated
system. In other words, thermal
hermal energy cannot be con
converted
verted into work by a cyclic process with
100% efficiency.
The mathematical formulation of the second law of thermodynamics is best visualized using
Figure 39. In this figure, on the left side
side, a generalized thermodynamic cycle is shown taking
place between a hot reservoir at temperature TH and a cold reservoir at temperature TC. On
the right side, the Carnot cycle diagram is shown between the same boundaries. Therefore,
the area in red, labelled as QC, is the amount of energy exchanged between the system and the
cold reservoir. The area in white, labelled as W, is the amount of work exchanged by the
system and its surroundings. The amount of heat exchanged with the hot reservoir, labelled as
QH, is the sum of the two, QC and W.. If the system is behaving like an engine, the process
moves clockwise around the lo loop,
op, and if it behaves as a refrigerator, it moves counter-
counter
clockwise.
Figure 39 The concept of entropy in closed cycles

From the above, and because the energy addition to the cycle and the energy extraction from it
have opposite signs, it follows that:
TH Q Q Q
  H  C  H  0. (61)
TC QC TC TH
This can be generalized by splitting the general reversible cycle shown in Figure 39 into a
number of infinitesimally small Carnot cycles
cycles. For each small cycle,, noting that common
boundaries cancel, the following re
relation can be obtained:
QC QH
 0. (62)
TC TH
Summing all cycles and performing a circular integration:
Q Q
 T
0   T
0. (63)
Thee above equation also indicates that the line integral Q / T
L
is path-independent
path
[SAA1966]. Because the closed integral is equal to zero

zero, dQ/T must be an exact differential and
must be a state variable, i.e., a property of the state of the material, like internal energy,
pressure, temperature, density (u, P, T, ρ), etc. This differential is defined to be the entropy, S.
Because S is a system property, it is possible to express any equilibrium state in terms of S plus
one other state variable (T, P, or something else):
 In a reversible process (Figure 40), the quantity, δQ/T, from point A to point B is the
same regardless of the path;
 Hence, the entropy, S, can be defined as
BQ
SB  S A   ; (64)
A T
reversible
 Specific entropy, s (J/kg K).

For a reversible process, heat addition and heat extraction from the cycle are defined as:
QH TH (SB SA) , (65)
QC TC (SA  SB) TC (SB SA) . (66)

Using Eqs. (65) and (66), the net work done by the cycle is defined as:
Wcycle  QH QC  TH (SB  SA) TC (SB  SA)  (TH TC )(SB  SA) . (67)
Therefore, the thermal efficiency is defined using Eq. (44) as:

Wcycle (TH  TC ) ( S B  S A ) TH  TC T
t     1 C . (68)
QH TH (S B  S A ) TH TH
The following relation represents the generalization of the entropy definition for reversible and
irreversible processes:
Q Q
 T
 0 or S  
T
. (69)
Hence, for reversible processes, the equality sign applies, whereas for irreversible processes,
i.e., realistic processes, the < sign applies. All real processes are considered irreversible. This
indicates that for closed systems, any process, regardless of the path, results in an increase in
entropy.
6.2.5 Reactor power cycle

Figure 40 shows how an ideal Carnot cycle is applied with water as the working fluid. On the
right side, isothermal addition of energy (heat) to the cycle is achieved between the saturated
water and saturated steam states. Extraction of waste heat from the cycle is carried out
isothermally in the two-phase region.
Figure 40 Carnot cycle for steam

The ideal Carnot cycle is impossible to achieve with real turbo machinery. Therefore, a
variation of the Carnot cycle, called the Rankine cycle, is shown on the left side of Figure 40
[SAA1966]. This cycle performs complete condensation of the steam at the lower thermal
reservoir temperature. Hence, water enters the adiabatic compression in the pump as liquid
water. With this modification, more heat rejection is needed to bring the feedwater to
saturation, avoiding a liquid-vapour mixture in the pump. This reduces thermal efficiency.
Figure 41 shows the steam-water diagrams used by thermal engineers and designers
[WIK2015a]. The first part, Figure 41a, shows the P-v diagram, followed by Figure 41b, the T-s
diagram, and Figure 41c, the h-s diagram. The P-v diagram is useful for representing changes in
volume between different points in the cycle. Note that pressure and temperature are linked
parameters at saturation conditions, and hence in the two-phase area, both join in the same
straight line. The T-s diagram shows temperature changes in the cycle for given entropies. This
diagram is useful for the Carnot and Rankine cycles because heat addition and heat extraction
are performed as isothermal processes, which are straight horizontal lines in this diagram.
According to the ideal Carnot cycle and the ideal Rankine cycle, isentropic expansion and
isentropic compression (no losses) are represented by vertical straight lines, which make the
diagram a very easy way to observe heat changes as square areas. On the other hand, the h-s
diagram is very useful to the turbo machinery designer because each energy change during any
of the processes is shown as a vertical increment on the y-axis, whereas reversibility or
irreversibility of the cycle is easily observable on the x-axis.
Figure 42 shows an ideal Rankine cycle for a nuclear power plant (ideal because irreversibility is
not taken into account). The left side shows the NPP secondary side with the steam generator
and other equipment. Certain points in the diagram are labelled on the reactor sketch with
numbers.
The separate processes of the cycle as labelled on the left side with their position are:
 Heat supplied: area 6-1-2-2A-3-4-5-6
 Heat removed: area 6-1-4-5-6
 Net work done: area 1-2-2A-3-4-1.

The assumptions of this idealized Rankine cycle are:
 Negligible changes in kinetic energy and potential energy;
 All processes are reversible; and
 Negligible pressure and heat losses.
Figure 41 Steam-water diagrams used in thermodynamic analyzes (a, b, c)
Figure 42 Simplified Rankine cycle in a nuclear power plant

The thermodynamic power cycle in reactor systems is similar to the Rankine cycle. As sketched
in Figure 42, the steam generator boils the working fluid (water) isothermally (at saturation
temperature), the turbine expands the fluid and performs shaft work, the condenser extracts
the reject heat and condenses the fluid, and the feedwater pump returns the fluid to the steam
generator at pressure. Of course, the realistic reactor cycle is not reversible, but the principles
of the cycle are the same. The typical cycle used in power plants is called the Rankine cycle.
The T-s and h-s diagrams for an ideal simple Rankine cycle are given in Figure 41 and Figure 42.
In the ideal Rankine cycle in Figure 42, saturated steam (point 3) enters the turbine and
expands isentropically to position 4. At point 4, the wet steam enters the condenser, where
heat is removed until the fluid is condensed to a saturated liquid at point 1. After leaving the
condenser at saturation pressure P2, the fluid is compressed isentropically from pressure P2 to
the boiler pressure P1. The high-pressure liquid enters the boiler at point 2, where the fluid is
first brought to saturation at pressure P2 (point 2A) and then vaporized to saturated steam at
point 3.
In process 1-2 (isentropic compression in the feedwater pump), no heat is exchanged with the
surroundings, i.e., q12 0. Hence, the work input to achieve isentropic compression is given by
Eq. (70):
w12  (h2  h1 )   v dp ~ v1 ( p2  p1 ) .

2
(70)
1
In process 2-2A-3 (heating at constant pressure in the boiler), no work is exchanged with the
surroundings, i.e., w2-3 = 0. Hence, using the h-s diagram, the heat addition to the system is
obtained as a difference of enthalpies, as given by Eq. (71):
q23  qin  h3 h2 . (71)
In process 3-4 (isentropic expansion in the turbine), no heat is exchanged with the
surroundings, i.e., q3-4 = 0. Hence, the work delivered by the turbine is given by Eq. (72):
w34  h3  h4 . (72)
In process 4-1 (heat removal at constant pressure in the condenser), no work is exchanged with
the surroundings, i.e., w4-1 = 0. Hence, the heat extraction from the system is obtained as a
difference of enthalpies, as given in Eq. (73):
q41  h4  h1 . (73)
The net work performed by this cycle can be obtained as:
wcycle  w34  w12  (h3  h4 )  (h2  h1) . (74)
The thermal efficiency of the ideal Rankine cycle is obtained by Eq. (75):
wcycle (h3  h4 )  (h2  h1 ) q2 3  q41
Rankine    . (75)
qin h3  h2 q23
As noted earlier, the thermal efficiency of the ideal simplified Rankine cycle is much less than
that of the Carnot cycle because more heat is wasted to the environment. The performance of
the Rankine cycle can be improved in practice by:
(1) raising the boiler pressure
(2) lowering the exhaust pressure
(3) using superheat
(4) using reheat.
Options (1), (3), and (4) effectively raise the inlet temperature, whereas (2) effectively lowers
the outlet temperature, with a corresponding effect on cycle efficiency.
Figure 43 Typical impact of turbine outlet pressure on cycle thermodynamic efficiency

The condenser pressure is limited by the temperature of the available cooling water, the size
and cost of the condenser, and the size of the vacuum pumps required to deaerate the
condenser. Figure 43 shows the impact of a condenser pressure decrease in a typical PWR
nuclear plant. The practical lower limit on condenser pressure is 3–6 kPa(a). Consequently,
options (1), (3), and (4) are used to achieve increased efficiency [TOD2011].
The increase in boiler pressure results in an increase in the net work performed by the cycle,
with a corresponding decrease in heat rejected. Figure 44 shows the impact of an inlet
pressure increase in a typical PWR nuclear plant [TOD2011]. However, for the indirect power
cycle, the downside of raising the boiler pressure (and the temperature because the steam is
saturated) is that it forces the primary-side temperature to increase to provide sufficient ΔT to
transfer the heat from the primary to the secondary side. This higher primary-side
temperature requires higher primary-side pressure, which in turn requires thicker pressure
vessel walls. In a pressure vessel-type reactor, this can be costly or can lead to reduced plant
reliability or life. In pressure tube reactors, the main drawback is the increased parasitic
neutron absorption in the pressure tube walls, which need to be thicker, resulting in lower fuel
burn-up.
Figure 44 Typical impact of turbine inlet pressure on cycle thermodynamic efficiency
6.2.6 Improvements in reactor cycles

In the following sections, improvements to the Rankine cycle used in power reactors are
described, their usefulness discussed, and their consequences analyzed. The improvements
described in the following sections are used in power reactors with certain variations and
design specifics by various reactor types and vendors. However, they all contribute to higher
NPP thermal efficiency, which is considered as an important marketing parameter.
6.2.6.1 Rankine cycle with superheat
Figure 45 illustrates the Rankine cycle with superheat [LEU2004, POP2013]. Superheat causes a
net increase in the temperature at which heat is received, with a resulting improvement in
cycle efficiency. Another important factor is that the amount of moisture in the fluid leaving
the turbine is reduced, which increases turbine efficiency and reduces turbine blade erosion.
However, when using superheat, one must have a high-temperature heat source or else reduce
boiler pressure. Some recent approaches and ideas include use of traditional gas-fired
superheaters in combination with nuclear power reactors.
Figure 45 Rankine cycle with superheating

The differences between the cycle with superheat and the simple Rankine cycle are:
 Increased net work output (area 2B-3-4-4B-2B)
 Increased heat supply (area 2B-3-4-5B-2B)
 Increased exhaust steam quality from x4B to x4.
The thermal efficiency of the cycle with superheat is given by Eq. (76) and is higher than for the
simple Rankine cycle:
wcycle q23  q41 (h3  h4 )  (h2  h1 )
Superheat    . (76)
qin q23 h3  h2
6.2.6.2 Reheat cycle
The effective temperature of heat addition is increased and the moisture content further
reduced by using reheat in the Rankine cycle. A schematic diagram of the power plant and an
appropriate temperature-entropy diagram are shown in Figure 46 [LEU2004, POP2013]. High-
pressure, superheated steam is expanded in a high-pressure turbine to an intermediate
pressure 4 and the fluid then returned to a second-stage boiler and superheater and reheated
to state 3A. The reheated steam is then expanded in a low-pressure turbine to the final
exhaust pressure 4A. The moisture content of the working fluid is drastically reduced by use of
reheat, and this approach is used in all fossil-fuelled and many nuclear power plants (without
the superheating part).
Figure 46 Rankine cycle with superheating and reheating

The approach used to compute the work and the efficiency of reheat cycles is the same as that
used in the example problem for the simple Rankine cycle. One calculates the work produced
in each turbine separately and the required pumping work. Heat is added to the fluid at two
different stages of the cycle and is given by the difference in enthalpy between states 3 and 2
and states 4 and 3A. Wet steam at point 4 is removed after the high-pressure turbine stage,
reheated at constant pressure to a superheated state, and admitted to the low-pressure
turbine stage. Superheated steam at point 3A is expanded to the design exhaust pressure at
point 4A.
The net work performed by this cycle is given by Eq. (77):
wcycle  w12  w34  w3A4A  (h1  h2 )  (h3  h4 )  (h3A  h4A ) . (77)
The heat supplied to this system is given by Eq. (78):
qin  (h3  h2 )  (h3A  h4 ) . (78)

The thermal efficiency of this cycle is given by Eq. (79):
(h1  h2 )  (h3  h4 )  (h3 A  h4 A )
Reheat  . (79)
(h3  h2 )  (h3 A  h4 )
6.2.6.3 Regeneration cycle
In theory, modifications to the cycle can be made to reduce cycle irreversibility. One of the
principal sources is the sensible heat addition required to bring the boiler feedwater up to
saturation temperature. This is accomplished by using some of the flow through the turbine to
heat the feedwater. To achieve reversibility, a possible setup is shown in Figure 47, where
internal heat from the turbine is used to heat the feedwater in process 2-2A [LEU2004,
POP2013]. This cycle could provide the same thermal efficiency as the Carnot cycle. However,
this is not a practical modification because it is impossible to design a turbine to serve as both a
power-production device and a heat exchanger. Modifications have, however, been designed
to make this design change more practical, one of which is presented in the next section.
Figure 47 Rankine cycle with regeneration circuit
6.2.6.4 Rankine cycle with feedwater heaters
This improvement is based on the regeneration principle, which was briefly covered in the
previous section. This improvement is based on extracting a small amount of steam at an
intermediate pressure from the turbine and using it to heat feedwater.
Two types of feedwater heaters exist in practice:
 Open type, in which direct contact between extracted steam and feedwater is achieved,
resulting in good heat transfer; and
 Closed type, in which only thermal contact between extracted steam and feedwater is
achieved.
For both types, the reheating is usually designed in several stages. The following sections cover
both these types in detail.
6.2.6.5 Rankine cycle with open feedwater heater
Figure 48 shows the Rankine cycle with one open feedwater heater (the left side shows the
loop and the right side the T-s diagram of the cycle) [LEU2004, POP2013].
Figure 48 Rankine cycle with open feedwater reheating

The extraction fraction of steam is calculated from an energy balance on the feedwater heater
using Eq. (80):
m 1  y  h2  m y h6  m  y  1  y   h3 ,
1 y  h2  y h6  h3 . (80)
The pump work is obtained by summing the work from both pumps, as given by Eq. (81):
wpump  w12 1 y   w34 . (81)
The turbine work is obtained by summing up the two parts of the turbine expansion, one with
the full steam flow, and one with the (1-y) portion of the steam flow, as in Eq. (82):
wturbine  w56  1 y  w67 . (82)
The heat supplied to the cycle is not affected by the heater and is given by Eq. (83):
q45  h5  h4 . (83)
Usually, operating NPPs have multiple stages of feedwater heating; an example of a plant with
two open feedwater heaters is shown in Figure 49.
Using the heat balance at both heaters, one can obtain the enthalpies downstream from the
mixing points Z and Y, as shown in Eqs. (84) and (85):
m 1"h1"  m 4h4 m1"h1"  m 4h4
h4'   , (84)
m 1"  m 4 m1  m1'
m 1'h1'   m 1  m1'  h5
h5'  . (85)
m 1
Figure 49 Rankine cycle with multiple open reheaters
The turbine work in this case is obtained by summing up the work at the three parts before and
after the steam extractions, as shown in Eq. (86):
1  h1  h1'    m
wT  m 1'  h1'  h1"   m
1  m  2  h1"  h2  . (86)
6.2.6.6 Rankine cycle with closed feedwater heater
Figure 50 shows the Rankine cycle with one closed feedwater heater (the left side shows the
loop and the right side the T-s diagram of the cycle) [LEU2004, POP2013].
Figure 50 Rankine cycle with closed feedwater reheating

The extraction fraction of steam is calculated from an energy balance on the feedwater heater
using Eq. (87):
1 y h2  y h6  1 y h2 A  y h3 . (87)
The pump work is obtained by summing the work from both pumps, as given by Eq. (88):
wpump  w12 1 y   w34 . (88)
The turbine work is obtained by summing up the two parts of the turbine expansion, one with
the full steam flow, and one with the (1-y) portion of the steam flow, as in Eq. (89):
wturbine  w56  1 y  w67 . (89)
The heat supplied to the cycle is not affected by the heater and is given by Eq. (90):
qin   h5  h2 A 1 y    h5  h4  y . (90)
Usually, operating NPPs have multiple stages of feedwater heating; an example of a plant with
two closed feedwater heaters is shown in Figure 51 [POP2013].
Figure 51 Rankine cycle with multiple closed reheaters

Using the heat balance at both heaters, one can obtain the enthalpies downstream from the
exit points 1 and 2, as shown in Eqs. (91), (92), and (93):
m 1" h1"  m 4 h4  m 1' h6'   m 1'  m 1"  h5
h4'  , (91)
m 1
m 4 h4'   m 1'  m 1"  h5'
h4"  , (92)
m 1
m 1'  h1'  h6   m 1h4"
h7  . (93)
m 1
6.2.6.7 Thermal efficiency moisture separation
The role of the moisture separator between the high-pressure and low-pressure turbines in the
cycle is two-fold:
 to provide feedwater heating, thus improving cycle thermal efficiency; and
 to remove moisture from the steam, enabling dry steam to enter the low-pressure
turbine (thus reducing water droplet erosion of the turbine blades).
Figure 52 shows an example of a moisture separator where, using the principle of the open
heater, the extracted moisture from the separator is mixed with the feedwater, thus
transferring the latent heat of condensation to the feedwater [POP2013].
By defining the parameters at certain points in the T-s diagram, h1"  hf  at p1'  and 1 x1'  m
1
and h1"'  hg  at p1"  and x1' m
1 , and by performing an energy balance at the separator, one can
obtain the enthalpy after the mixing point using Eq. (94):
h1" x1" m 1"  h4 x1"'  m 1  m 1" 
h5  . (94)
m 1
Figure 52 Rankine cycle with dryer
6.2.6.8 Actual Rankine cycle
The real Rankine cycle is not reversible because of various losses that occur in nature and that
must be accounted for. Figure 53 shows the real Rankine cycle and the impact of various losses
[TOD2011]. The left figure shows the pump losses, the middle figure shows the losses in the
steam lines to the turbine, and the right figure shows the losses in the turbine. Some of these
losses are listed below:
 Irreversible frictional losses at various places in the turbo machinery;
 Irreversible heat losses from various piping, vessels, etc.;
 Pump losses, in particular:
o Heat loss to the surroundings;
o Fluid friction with the pump blades;
o Mechanical losses due to friction (bearings, gears, etc.);
 Turbine losses, in particular:
o Pressure drop between the superheater and turbine, which reduces entrance
pressure from P1 to P2;
o Heat loss to the surroundings, which reduces the temperature change from
point b to point c;
o Steam expansion inside the turbine, which is irreversibly adiabatic (non-
isentropic) and hence reduces efficiency and increases steam quality;
o Mechanical friction; and
o Steam bypass outside the turbine blade passages.
Figure 53 Real Rankine cycle with losses
6.2.6.9 Sample PWR and CANDU 6 cycle diagrams
Figure 54 shows a typical secondary side of an advanced PWR reactor [APR2011]. It is clear
that multiple feedwater heaters, one moisture separator, and double reheaters have been
designed into the system. The secondary system operates with two feedwater pumps at two
different pressures.
Figure 54 Typical advanced PWR secondary cooling system
Figure 55 Typical CANDU 6 secondary cooling system

Figure 55 shows a typical CANDU 6 Rankine cycle [AECL2010]. This plant has one double-sided
high-pressure turbine and two double-sided low-pressure turbines. On the low-pressure
turbine, there is one open feedwater heater and three consecutive closed feedwater heaters.
On the high-pressure turbine, there are two closed feedwater heaters. Between the high-
pressure and low-pressure turbines, several stages of moisture separators deliver heating to
the feedwater heaters. The thermal efficiency of such an arrangement is around 35%.
6.2.7 Entropy and the laws of thermodynamics

Entropy is a state parameter and can be expressed in terms of other parameters. For sub-
cooled or superheated conditions, s=f(P,T), and for saturated conditions, s=f(P) or s=f(T), it is
expressed in terms of pressure or temperature. In the two-phase region, the following
equation can be used to calculate mixture entropy (under homogeneous conditions):
sm  x  sg  (1 x)  s f  s f  x  s fg , (95)
where sg is the entropy of the saturated gas, sf is the entropy of the saturated liquid, and sfg is
the entropy change during vaporization.
Entropy is a quantitative measure of the microscopic (molecular) disorder of a system. Entropy
generated through any process in an isolated system cannot be negative. Moreover, entropy is
a non-conserved property (it always increases in isolated systems). In engineering processes,
entropy increase is a measure of irreversibility in these processes (a measure of engineering
system inefficiency or of system losses).
Entropy change is closely associated with heat transfer. In heat transfer from a hotter to a
colder subsystem, the hotter subsystem exhibits a decrease in entropy, whereas the cold
system shows an increase. However, the increase in the cold subsystem is greater, thus making
the entropy change of the combined system positive.
Probably the most important outcome of the entropy changes in certain processes is the fact
that every process ends up with a certain penalty for moving energy in any direction. In other
words, any energy exchange between different systems results in an entropy increase, i.e.,
there is always a certain cost when moving energy from one place to another, or transferring
energy from one form to another. Any conversion of energy in a system from one form to
another or any exchange of work with its surroundings must be accompanied by some
irreversible dissipation of energy to the surroundings.
Entropy can be seen as the ability to expend energy to carry out a task. For example, a fully
charged battery has low entropy, which increases as the battery is used; or a clockwork toy has
low entropy when wound up, which increases as it unwinds.
Processes in nature change in the direction of reaching equilibrium. Entropy reaches a
maximum value when equilibrium is reached. Entropy is statistical in nature because it explains
how phenomena in the micro world affect phenomena in the macro world, capturing the
probability of a number of micro events to cause a macro effect.
In modern thermodynamics, the laws of thermodynamics are known as:
 First law – characterizes the energy balance;
 Second law – characterizes cycle/system efficiency;
 Third law – characterizes entropy at near-zero temperatures; and
 Zeroth law – characterizes thermodynamic equilibrium.
The zeroth law of thermodynamics states that if two bodies are in thermodynamic equilibrium
(at the same temperature) with a third body, then they are in thermodynamic equilibrium with
each other. This is a fundamental law that defines the relationship between different systems
(bodies) operating and interacting with each other. This law should have been labelled as the
first law, but because it was defined later, it was called the zeroth low to avoid confusion.
The first law of thermodynamics states that the change of internal energy in a system is equal
to the amount of heat supplied, minus the amount of work performed by the system on its
surroundings. In other words, it states that energy can be converted from one form to another,
but cannot be created or destroyed, i.e., everything requires energy to do anything. Different
forms of energy must be fed into a system before it can perform any useful work or create any
impact on its surroundings.
The second law of thermodynamics states that the entropy of an isolated system never
decreases because isolated systems spontaneously evolve toward thermodynamic equilibrium,
i.e., the state of maximum entropy. Moreover, it states that a certain cost must always be paid
when moving energy from one place to another or transferring energy from one form to
another. When it appears that a system’s entropy decreases, this generally means that the
system is not isolated from its surroundings and that by scaling up to a larger system, it will be
found that the system is a part of a larger system with lower entropy.
The third law of thermodynamics states that the entropy of a perfect crystal drops to zero
when the temperature of the crystal is at absolute zero. However, according to quantum
theory, because nothing can be considered to be truly at rest, every molecule, atom, or
subatomic particle will always have a minimum amount of energy even when cooled to
absolute zero (the zero point energy). Hence, one may take the third law as stating the
minimum entropy level at a temperature of absolute zero.
6.2.8 Problems
1. Calculate the Rankine cycle thermal efficiency which operates at saturated steam at
pressure of 7 MPa, and temperature in the condenser is 30oC. Calculate the steam
quality at the exit of an ideal turbine following an adiabatic expansion to condenser
pressure for the above conditions, and for the conditions if the turbine has 85%
efficiency. Compare the results with the Carnot cycle.
2. A steam generator in a PWR nuclear power plant produces 3200 t/h saturated steam at
pressure of 5 MPa. The temperature in the condenser is 35oC. Calculate the ideal
turbine power assuming that it operates in Rankine cycle, and the steam quality at the
turbine exit.
3. In a Rankine cycle the working fluid enters the turbine at 4.5 MPa with condenser
temperature of 35°C. Calculate the thermal efficiency of this cycle assuming that:
a. no preheaters are used;
b. one preheater is used to preheat the feed water to 60°C below the saturation
temperature; and
c. the feed water is preheated in 3 preheaters to the same temperature as above.
Given:
a. Turbine inlet pressure, Pi=4.5 MPa
b. Condenser temperature Tc=35 ºC or 308 ºK
c. Feedwater heater outlet temperature, TF=Tsat-60 ºC
Assumptions
a. The Rankine cycle described in the problem statement is for a CANDU 6 reactor
b. Steam entering the turbine is saturated
c. Losses in the turbine and pumping losses are the same in all cases, since it is the
difference in efficiency that is important here assume these losses are negligible.
4. Consider an emergency water tank that is supposed to deliver water to a reactor
following a loss of coolant event. The tank is pressurized by the presence of nitrogen at
1.0 MPa. The water is discharged through a 0.2 m ID pipe. The reservoir has a diameter
of 5 m, and the water level is 15 m above the reactor. Calculate the maximum flow rate
delivered to the reactor if the water is inviscid and reactor pressure is (a) 0.8 MPa, and
(b) 0.2 MPa.
5. Two alternative steam cycles are proposed for a nuclear power plant, as variation of the
Rankine ideal cycle, all three operating between the steam temperature of 293oC and
condensing temperature of 33oC. The first one operates with condensate in saturated
conditions, while the feedwater downstream of the condensate pumps is subcooled.
The steam is saturated at the inlet of the expander. (a) Assuming ideal machinery,
calculate the cycle thermal efficiency and steam rate for each cycle.
6. Water at 100 kPa and 20 °C flows through a smooth, circular pipe of 2 cm ID, and length
of 3 m. If the flow velocity is 3 m/s, calculate the pressure drop along the pipe and the
pumping power necessary to maintain the flow in the pipe.
7. Explain the 1st and 2nd law of thermodynamics and the relationship between these laws.
Provide examples of their application in reactor thermal-hydraulics design. Explain the
role of entropy in reactor thermal-hydraulics.
8. Develop an expression for thermal efficiency for a secondary heat transport system
shown in Figure 55 (CANDU-typical). Show the thermal efficiency in a generic
expression by using generic values of key parameters and by making appropriate
assumptions.
7 Heat Transfer and Fluid Flow Design

7.1 Heat transfer in the primary heat transfer system
The heat balance in the primary heat transport system is covered in this section, including heat
balance in the reactor core and heat balance in a steam generator. An interpretation of the
relationship between the primary and secondary loop parameters is provided, and the role of
the steam generator as a connection between the primary and secondary loops is explained.
Approximations in the steam generator heat transfer model are used to derive trends for
various parameters that affect the steam generator design. Analytical and numerical models
are developed and used to calculate the distribution of key parameters along the primary and
secondary sides of the generator, and the steam generator heat duty diagram is developed and
explained.
Figure 56 shows a simplified diagram of the reactor primary heat transport system. The reactor
is the heat source, because it is in the reactor that heat is generated. The steam generator is
the heat sink, which takes the heat out of the primary heat transport system and transfers it to
the secondary heat transport system. The primary pump serves as the driver of the flow in the
primary heat transport system by removing heat from the heat source (the reactor) and
dumping it into the heat sink (the steam generator).
a) Primary b) Secondary c) Tertiary
Figure 56 Reactor cooling systems
The heat balance in the reactor can be described by a simple algebraic equation, Eq. (96), from
which the reactor heat production can be expressed by Eq. (97):
Wphp,hot Wphp,cold  Q, (96)
Q  Wp  hp ,hot  hp ,cold  , (97)
where: Wp = primary coolant mass flow rate (kg/s)

hp,hot = core exit enthalpy (hot leg), (kJ/kg)
hp,cold = core inlet enthalpy (cold leg) (kJ/kg)
Q = reactor power transferred to the coolant (kJ/s or kW).
7.2 Primary pumps

The primary pumps are the vital component in the reactor heat transport system (HTS). The
primary function of the reactor HTS is to provide continuous heat removal from the reactor
core in normal operation, during transients, and during reactor shutdown, and the primary
pumps are the essential components that support this function. HTS pump start-up and shut-
down are routine operations of the reactor HTS. Pump failures due to loss of power or
inadvertent operator action are relatively common events in all nuclear power plants and are
considered in the safety case for each reactor type.
The primary side flow is determined by a balance between the head generated by the primary
heat transport system pumps (HTS pumps) and the circuit pressure losses due to friction,
losses, and other effects.
Figure 57 shows the pressure drop in the loop as a function of mass flow rate in the loop. The
pump operating point is indicated at the intersection between the pump head curve and the
system flow resistance curve. The system pressure drop curve follows a quadratic law because
it depends on the square of the flow velocity, i.e., the square of the mass flow rate (or mass
flux), as in Eq. (43). The pump head curve may have various shapes depending on the pump
type and design. For “short periods” of time, in transition to the stable operating point, the
pump and the system can temporarily operate outside the stable operating point. In this
respect, a “short period” of time should be considered in relation to the pump and system
hydraulic inertia.
Figure 57 Primary pump flow diagram

Accurate prediction of pump performance includes specification of its head (H), torque (τ),
discharge or volumetric flow rate (Q), and rotor speed (ω). The pump motor, by exerting a
torque on the rotating shaft, provides energy to the impeller, which creates the flow associated
with a head increase from the pump suction to the pump discharge side.
The pump motor provides torque for:
 overcoming frictional resistance and any local losses in the loop (piping, valves, core,
etc.);
 overcoming frictional losses in the pump rotating parts;

 acceleration of fluid in the loop; and
 acceleration of pump rotating parts (including the flywheels).
For a pump, only two parameters can be considered independent among H, Q, τ, and ω. The
other two are determined from the pump characteristics. It is commonly assumed that the
pump steady-state characteristics also hold for transient conditions. The pump characteristics
are described by specific relationships, called homologous relationships) [TOD1990]:
H1 H 2 Q1 Q2 H1 H 2
 ,  ,  . (98)
12 22 1 2 Q12 Q22
Pump manufacturers produce realistic relationships from pump tests, for example, relating the
head to both volumetric flow and pump angular velocity, as given by Eq. (99):
2
H1 Q Q
 a1  a2    a3   . (99)
 2
   
The common approach is to use non-dimensional parameters with respect to the rated pump
performance conditions (rated conditions refer to the rated quantities, which represent the
point of pump best performance as defined by the manufacturer), as given by Eq. (100):
Q H  
 , h ,  ,  . (100)
QR HR R R
These non-dimensional parameters are used to express homologous pump relationships in a
form similar to Eq. (98), in which the actual parameters are expressed in their non-dimensional
forms. Homologous relationships of these forms are used in computer programs describing
pump behaviour.
Figure 58 shows the four possible types, i.e., quadrants, of pump operation. Computer models
must be able to simulate pump performance in any of the four quadrants, and therefore pump
manufacturers must provide information about pump operation in all quadrants [TOD1990].
During a coast-down transient, the pump may pass from the normal pumping region, quadrant
I, through the reverse flow but positive rotation region, quadrant II, to reverse flow and
rotation, quadrant III, unless the rotor is equipped with an anti-reverse ratchet to avoid this
quadrant (such as in most reactor HTS pumps).
Figure 58 Primary pump operating quadrants

In postulated events with breaks at the pump suction or discharge side and different pump
operation control logic, a pump can operate in any of the first three quadrants. Therefore,
pump tests are performed to develop pump homologous relationships for all four possible
operating quadrants.
Pump characteristics under transient flow conditions have been investigated and tested. One
of the most important conditions that must be avoided in operation of reactor HTS pumps is
two-phase flow in the bubbly flow regime. Pump characteristics deteriorate significantly
following initiation of bubble formation in the pump suction. Moreover, prolonged operation
in two-phase flow (the bubbly flow regime) can damage the pump impeller blades.
One of the most important objectives of HTS flow design is to make sure that the pump suction
pressure does not fall below the net positive suction head required (NPSHR) [POP2014]. For a
pump to deliver its rated output, the absolute pressure (including the velocity head V2/2g) of
the fluid at the pump inlet must exceed the vapour pressure by an amount sufficient to
overcome any entrance or frictional losses between the point of entry into the pump and the
impeller. The NPSH is defined as the absolute pressure at the pump inlet expressed in metres
of liquid, plus the velocity head, minus the vapour pressure of the fluid at pumping
temperature, and corrected to the elevation of the pump centreline in the case of horizontal
pumps, or to the entrance of the first-stage impeller for vertical pumps.
NPSHR (required) is determined by the pump manufacturer and is defined as a function of
pump speed and pump capacity. NPSHA (available) represents the energy level of the fluid
over the vapour pressure at the pump inlet and is determined entirely by the system preceding
the pump. Unless NPSHA>NPSHR under any operating condition, some of the fluid will
vaporize in the pump inlet, and vapour bubbles will be carried onto the impeller. These
bubbles will collapse violently at some point downstream of the pump inlet (usually at some
point on the pump impeller blade surface) as the pressure is increased in the pump, producing
very sharp, crackling noises accompanied by physical damage to the pump impeller surface.
This phenomenon is known as cavitation and is highly undesirable (in fact, it represents a micro
water hammer effect on the impeller surface). The net mechanical effect of cavitation is pump
vibration and impeller damage, whereas the net effect on pump performance is loss of pump
efficiency and pump head.
The primary side flow is determined by a balance between the head generated by the primary
pumps and the circuit head losses due to friction:

Ppump  Pcircuit . (101)
Normally, the pump head curve is provided in the form of a polynomial such as Eq. (102):
Ppump  A0  A1 W  A2 W 2  ... . (102)
The hydraulic network losses are proportional to the square of the velocity, i.e., the mass flow
rate, along with a loss coefficient, as shown by Eq. (103):
Pcircuit  K W2 . (103)

Figure 59 and Figure 60 show the pump curves of two pumps in the CANDU HTS; the first is the
primary HTS pump, and the second is the shutdown cooling pump.
These pumps have very different functions. The HTS primary pump operates almost
continuously during the operating lifetime of the plant, and therefore the requirements on this
pump are very demanding. The shutdown pump operates normally when the reactor is shut
down to help remove decay heat from the reactor core. Both diagrams show pump volumetric
flow rate on the x-axis, pump head on the left y-axis, and pump efficiency, pump power, and
NPSHR on the right y-axis.
Figure 59 and Figure 60 show that pump head declines with increasing volumetric flow,
whereas pump power increases [POP2014]. This means that because of momentum balance
considerations, if a pump is required to produce higher volumetric flow, its head will be lower.
Note that for a pump delivering higher volumetric flow, the pump power must increase.
Dotted lines in Figure 59, labelled S1, S2, and S3, show three possible system configurations
that impose different flow resistance.
Figure 59 Typical primary pump characteristics
Figure 60 Typical shutdown pump characteristics

The efficiency curve is usually bell-shaped, and it shows that the highest efficiency is achieved
in a relatively narrow range of volumetric flow rate. Pump efficiency is very important for the
HTS pump because it operates continually over 60 years of plant life, and therefore a small loss
of efficiency can result in a relatively large loss of plant earned value. Efficiency drops very
sharply if the pump volumetric flow rate moves outside a narrow range.
The NPSHR curve is relatively flat over a wide range of volumetric flow rates. In fact, well-
designed pumps should have NPSHR curves as flat as can possibly be obtained. However, when
the volumetric flow rate exceeds a certain value, the NPSHR curve starts to increase fast,
creating an area of pump cavitation risk.
7.3 Steam generator heat balance

The steam generator has a very important role in energy transport from the reactor core to the
turbine and the electrical generator because it connects the primary and secondary loops.
Hence, understanding steam generator behaviour is important in understanding plant
behaviour.
A simplified steam generator heat balance is shown in Figure 61. The feedwater flow WFW
enters the steam generator with enthalpy hFW and exits with enthalpy hSAT. The primary mass
flow rate in the reactor is Wp.
Figure 61 Reactor steam generator loop

The heat transfer from the primary to the secondary steam generator side is expressed by Eq.
(104):
dQ  U SG  Tp  Ts   dA , (104)
where USG is the overall (global) heat transfer coefficient (kJ/m2-°C),

A is the heat transfer area in the steam generator (m2),
Tp is the primary-side temperature (°C), and
Ts is the secondary-side temperature (°C).
The total heat transfer in the steam generator can be obtained by integrating the above
equation over the length of the steam generator tubes and is given by Eq. (105):
Q   dQ   U SG  (T p  Ts )  dA . (105)
Q
Figure 62 shows a general heat-duty diagram for a steam generator with a pre-heater and with
saturated primary coolant entering the steam generator U-tubes (i.e., with a certain
percentage of quality). The y-axis indicates fluid temperatures, and the x-axis provides a
conceptual representation of space in the steam generator. In this diagram, no heat losses or
pressure losses are shown (i.e., an ideal steam generator is assumed).
The primary coolant moves through the U-tubes from right to left in the diagram, starting as
saturated with a certain percentage of quality and becoming sub-cooled as it transfers the heat
to the secondary side. The secondary coolant (feedwater) enters sub-cooled (zero mass
quality, i.e., negative thermodynamic quality) and, as it receives heat from the primary side,
heats up to saturation. Thereafter, the secondary coolant boils off as it receives more heat
through the steam generator. Note that the point at the primary coolant temperature where
the secondary coolant temperature reaches saturation is called the “pinch point” [GAR1999,
POP2014].
Figure 62 Steam generator simplified heat duty diagram

From the above discussion, it is clear that sufficiently good heat transfer is required to achieve
acceptable steam pressure. The steam generator calculation of U-tube length, diameter, and
number (i.e., the area used in heat transfer) is carefully performed so that heat transfer from
the primary to the secondary side is achieved efficiently with as little heat loss as possible, and
so that the economic constraints of manufacturing the steam generator and the system can be
satisfied.
7.3.1 Steam generator without a pre-heater region

To achieve a better understanding of the behaviour and design of the steam generator and the
impact of its various parameters, it is useful to analyze first a simplified version of the steam
generator with no pre-heater, with saturated feedwater entering the steam generator, and
with sub-cooled primary coolant entering the steam generator. This simplified steam
generator is shown in Figure 63 [GAR1999].
Figure 63 Steam generator heat duty diagram without pre-heating

The primary-side heat balance in the steam generator can be obtained by integrating Eq. (105).
Assuming that the global heat transfer coefficient USG is constant along the steam generator,
the primary coolant temperature variation is linear along the U-tube, and by separating
primary and secondary side terms, one obtains Eq.(106):
(Tp,hot  Tp,cold )
Q  U SG ASG U SG ASGTs . (106)
2
Equation (106) can be further simplified into Eq. (107):
T  T p ,cold 
Q  U SG ASG  p , hot  Ts  . (107)
 2 
The temperatures can be replaced by enthalpies using Eq. (108):
h  Cp T  constant . (108)
Using Eqs. (108) and (107), the heat transferred in the steam generator can be given by Eq.
(109):
U SG ASG  h p , hot  h p ,cold 
Q   hs  . (109)
Cp  2 
Using Eq. (96) we find:
U SG ASG  h p , hot  h p ,cold 
Q   hs   W p  ( h p , hot  h p ,cold ) . (110)
Cp  2 
The secondary side steam flow can be calculated by an energy balance on the secondary side of
the steam generator, as given by Eq. (110); using this equation, one can calculate the mass flow
rate at the secondary side, as given by Eq. (111):
Q  Ws  hs  hFW  , (111)
Q
Ws  . (112)
 hs  hFW 
7.3.2 Approximate solution to steam generator without a pre-heater region
Combining Eqs. (97), (101), (110), and (112) yields the set of four equations describing the heat
transfer in the primary and secondary loops, as shown by Eq. (113) [GAR1999, POP2014]:
Q  Wp  hp ,hot  hp ,cold 
U SG ASG  hp ,hot  hp,cold  T T 
Q   hs   U SG ASG  p, hot p,cold  Ts 
Cp  2   2 . (113)
Ppump  Pcircuit
Q  Ws  hs  hFW 
From the first equation in Eq. (113), one can obtain the reactor outlet (hot) enthalpy:
Q
hp ,hot   hp ,cold . (114)
Wp
The second equation in Eq. (113) can be transformed into the following equation by dividing it
by the primary mass flow rate and using Eq. (114), while also neglecting heat losses between
the reactor and the steam generator:
Q U A  Q  U A  Q 
 SG SG   h p , cold  hs   SG SG   T p , cold  Ts  . (115)
Wp C pW p  2W p  W p  2 C p W p 
Rearranging Eq. (115) in terms of outlet enthalpy at the steam generator and outlet
temperature, one can obtain Eq. (116) and Eq. (117):
Q  C pW p 1 QC p Q
h p ,cold      hs    hs , (116)
Wp  U SG ASG 2  U SG ASG 2W p
Q  Wp 1 
T p , cold      Ts . (117)
Wp  U SG ASG 2 C p 
From Eqs. (116) and (117), all the parameters Q, Wp, Cp, ASG, USG, etc., are positive quantities.
Therefore, the reactor inlet enthalpy (and temperature) will rise as the primary flow rate
increases, will rise as the secondary-side temperature (saturation temperature) and enthalpy
increase, and may go up or down as power changes.
Substituting Eq. (116) into Eq. (117),
U SG ASG  h p , hot  h p , cold   T p , hot  T p , cold 
Q   hs   U SG ASG   Ts  . (118)
Cp  2   2 
The enthalpy (and temperature) at the steam generator primary side inlet can be obtained as
follows:
Q  C pW p 1 QC p Q
h p , hot      hs    hs , (119)
Wp  SG SG
U A 2  U A
SG SG 2W p
Q  Wp 1 
T p , hot      Ts . (120)
Wp  U SG ASG 2 C p 
Using Eqs. (116), (117), (119), and (120), one can obtain the following useful relationships
[GAR1999, POP2014]:
h p , hot  h p ,cold Q  C pW p   QC p 
h p , aver      hs     hs , (121)
2 W p  U SG ASG   U SG ASG 
Q  Wp 1  Q Wp Q
T p , hot  T p , aver      T s  Ts  , (122)
Wp  U SG ASG 2C p  W p U SG ASG 2W pC p
Tp,hot  Tp,cold Q Q
Tp,aver  Tp,aver  Tp,cold  Tp,aver  Ts  , (123)
2 2C pWp U SG ASG
Q Q Q Q Q
Tp,hot  Tp,cold  Tp,hot  Ts   Tp,cold  Ts   . (124)
WpC p U SG ASG 2WpC p U SG ASG 2WpC p
7.3.3 Calculation of main HTS parameters

From an HTS designer perspective, the key question is how to design the primary heat
transport system. As long as the primary flow rate, or the reactor inlet and outlet enthalpies,
have not yet been defined, it is necessary to make an assumption about one of these to be able
to calculate the other parameters when the desired reactor power is given.
Given a rough estimate of flow for calculating USG (which is not a strong function of flow
because Wp is large and flow is turbulent), Taver can be calculated, as well as an estimate of the
spread of Tp,aver (±Q/2WpCp). This gives a good first estimate of temperatures and enthalpies.
With the above enthalpy, temperature, and density estimates, the circuit losses can be
calculated and compared with available pump head and flow. The flow estimate can be
updated and the whole procedure repeated until convergence is reached.
Figure 64 Process for calculating primary loop parameters

Figure 64 illustrates the process used to calculate the primary heat transport system
parameters [GAR1999, POP2014]. Note that first the reactor power, the steam generator, and
the outlet header pressure and saturation enthalpy are defined. Then the primary flow rate is
assumed. This makes it possible to define the steam generator global heat transfer coefficient
and the primary fluid heat capacity. Then using Eqs. (116) and (119), the reactor inlet and
outlet enthalpies can be calculated. Once these are defined, the next step is to define other
primary fluid state parameters from the state laws, such as the fluid density for the inlet and
outlet reactor enthalpy. After these parameters are defined, it is necessary to check whether
the assumed USG has converged. If not, USG is updated, and another round of these calculations
is performed.
After convergence on USG is reached, the pump head and primary loop resistance can be
calculated because the primary flow rate has been defined. The last step is to check whether
the pump head and the loop resistance match. If not, then the whole process shown in Figure
64 must be repeated.
_____________________________________________________________________________
Problem 7.3.3.1
To illustrate that this process leads to convergence, a sample calculation will be performed
using typical CANDU 6 parameters as given by Eq. (125) [GAR1999, POP2014].
Q  2000  MWth 
W  8000  kg / s 
Ts  265 C   hs  1150  kJ / kg 
C p  5  kJ / kg  C  . (125)
U SG  5  kJ / s  C  m 2   C W p p / U SG ASG   0.625
ASG  3200  m 2 
Equation (116) can be expressed in terms of primary flow rate, as shown by Eq. (123):
Q
Wp  . (126)
 C Q 
2 p  hs  hp ,cold 
 U SG ASG 
From the above equation, one can define the infinitesimal change in the primary flow rate with
the reactor inlet enthalpy as follows:
Wp Q Q 2W p2
   . (127)
 h p ,cold  C pQ 
2
 Q 
2
Q
2  hs  h p ,cold  2 
 U SG ASG   2W p 
Substituting the parameters for CANDU 6 from Eq. (125) into Eq. (127),
2
  kg  
2  8000   
 Wp   s   kg 2 
  64  . (128)
 hp ,cold 6  kJ   kJ  s 
2 10 kW  
 s 
If Eq. (128) is expressed in terms of the initial values of primary flow rate and reactor inlet
enthalpy, the result gives the change in the primary flow rate as a function of reactor inlet
enthalpy with respect to the initial value (a guessed value):

W 
 p 
 W p ,o 
    W p  hp ,cold ,o  64  1181 25  9.0 . (129)
 h   hp ,cold Wp ,o 8000
  p ,cold 
 hp ,cold ,o 
 
Therefore, the change in primary flow rate will be nine times the change in reactor inlet
enthalpy:
 W   h p ,cold 
 p
W   9.0     . (130)
 p ,o   h p , cold , o 
Hence, if we were to guess the reactor inlet enthalpy, and our guess were only 25% wrong, the
primary flow rate would be in error by about 225%. This indicates that the process involving
assuming reactor inlet enthalpy first is inherently divergent. However, if we were to guess the
primary flow rate, and if we were in error by about 25%, the calculated reactor inlet enthalpy
would be in error by only 2.8%. Therefore, this process is inherently convergent. This confirms
that the process for primary heat transport system design shown in Figure 64 is appropriate.
End of Problem 7.3.3.1.________________________________________________________
7.3.4 Steam generator with pre-heater (simplified analytical solution)

Figure 65 shows a simplified heat-duty diagram for a steam generator with a pre-heater for a
fully sub-cooled primary side [GAR1999, POP2014]. This figure shows that the secondary-side
temperature changes from TFW to Ts (i.e., from feedwater to saturation temperature). The
steam generator design may include a section that is targeted to perform feedwater pre-
heating, or it may not have a specific section. Whatever the design, steam generator pre-
heating will always be relevant because the feedwater enters the steam generator in a sub-
cooled state and needs to be heated up to saturation before the steam generator can boil it
and generate steam.
Figure 65 Steam generator heat duty diagram with pre-heating

The reactor heat balance equation is given by:
Q  W p C p Tp ,hot  Tp ,cold  , (131)
where Q is reactor power (or steam generator power), Cp is D2O heat capacity (assumed
constant), Tp,hot is ROH temperature, and Tp,cold is RIH temperature.
The elementary amount of heat transferred from the primary to the secondary side of the
steam generator can be captured using the Fourier law:
dQ  U SG Tp  Ts  dASG , (132)
where USG is the overall heat transfer coefficient through the SG piping, dASG is the incremental
heat transfer area from the primary to the secondary side, Tp is the D2O temperature in the SG
tubes, and Ts is the H2O temperature in the SG shell.
As shown in Figure 65, the secondary side of the steam generator can be split into two parts,
the pre-heater part and the boiler part. Therefore, the heat transferred from the primary to
the secondary side can also be divided into two parts, as shown in Eq. (133):
Q  QPREHEATER  QBOILER . (133)
The fraction of the steam generator area involved in pre-heating is indicated by γ. The pre-
heater part of the steam generator brings the feedwater to saturation. The primary-side
temperature at which the secondary side is brought to saturation is denoted as TPINCH.
Using Eq. (113), the total heat exchanged between the primary and secondary sides in the
steam generator can be shown by means of the following relation [GAR1999, POP2014]:
T  TPINCH Ts  TFW   T p , hot  TPINCH 
Q   U SG ASG  p ,cold    1   U SG ASG   Ts  , (134)
 2 2   2 
where Ts is the saturation temperature at the secondary side and TFW is the feedwater
temperature at the inlet of the steam generator secondary side.
The pinch temperature can be expressed in terms of steam generator parameters using Eq.
(124):
TPINCH  Tp,cold   Tp,hot  Tp,cold   Tp,cold  

Q
. (135)
C pWp
By substituting Eq. (135) into Eq. (134), one can obtain the heat transfer in the steam generator
as expressed by the following equation:
 Tp,hot  Tp,cold   
Q  USG ASG    Ts  Ts  TFW   . (136)
 2  2 
Recall the relationship between the reactor inlet and outlet temperature from Eq. (124):
Q
Tp,hot   Tp,cold . (137)
WpCp
The reactor inlet and outlet temperatures can be expressed by the following relationship:
Q  W pC p 1 
T p ,cold      Ts  TFW   Ts , (138)
W p C p  U SG ASG 2  2
Q  W pC p 1 
T p , hot      Ts  TFW   Ts . (139)
W pC p  U SG AG 2  2
By examining Eq. (138), one can evaluate the impact of reactor power on the variation of
reactor inlet temperature. The saturation temperature Ts at the secondary side does not
depend on reactor power if the pressure on the secondary side is kept constant. The variation
in feedwater temperature, TFW, can be assumed to be linearly dependent on reactor power.
The primary flow rate Wp, the heat capacity CP, the area of the steam generator ASG, and the
steam generator global heat transfer coefficient USG exhibit second-order variation with reactor
power Q.
To define the dependence of reactor inlet temperature on reactor power, it is necessary to
evaluate the impact of the remaining parameter, i.e., the steam generator pre-heater area γ,
on reactor power.
The heat addition to the secondary side of the steam generator in the pre-heater area can be
calculated as:
QPREHEATER  WsCp,s Ts  TFW  . (140)
The heat extraction from the primary side of the steam generator can be calculated from Eq.
(134) as follows:
 Tp,cold  TPINCH   Ts  TFW 

QPREHEATER   USG ASG     . (141)
 2   2 
Combining Eqs. (140) and (141) leads to the following expression for the steam generator pinch
area [GAR1999, POP2014]:
Ws C p ,s Ts  TFW 
 . (142)
 Tp,cold  TPINCH  T  T  
U SG ASG   s FW 
 2 2 
The most strongly variable parameter in Eq. (142) is the feedwater flow WFW. The heat balance
on the secondary side given by Eq. (140) can be rearranged in terms of enthalpies as follows:
Q Wh
s s WFW hFW . (143)
Because Ws ≈ WFW, it is appropriate to assume that hs ≈ constant and hFW ≈ constant.
Therefore, feedwater flow and steam generator pre-heater area are proportional to reactor
power, i.e., WFW ∞ Q and γ ∞ Q.
It is normally desirable to describe the pre-heater area in terms of a ratio between the pre-
heater area at 100% power and at current power, as follows:
 100Q
 , (144)
Q100
where γ100 is the pre-heating fraction at 100% full power and Q100 = Q at 100% full power.
By substituting Eq. (144) into Eqs. (138) and (139), the reactor inlet and outlet temperatures
can be obtained as follows:
Q  W pC p 1  Q
T p ,cold      100 Ts  TFW   Ts
W pC p  U SG ASG 2  2Q100
Q  Q100  W p C p 1  
      100 Ts  TFW    Ts , (145)
Q100  W p C p  U SG ASG 2  2 
Q  Q100  W pC p 1  
T p , hot       100 T s  T FW    T s . (146)
Q100  W p C p  U SG ASG 2  2 
It is clear that the reactor inlet and outlet temperatures are equal to the steam saturation
temperature with corrections to capture the primary-side and secondary-side effects. Both
these effects are roughly proportional to reactor power. Hence, at 0% full reactor power,
Tp,cold=Ts and Tp,hot=Ts.
_____________________________________________________________________________
Problem 7.3.4.1
To illustrate the above analysis and the magnitude of the correction terms, a sample CANDU 6
assessment can be performed with the following given parameters:
 Reactor power Q100 = 2.064×106 [kJ/s]
 Steam generator global heat transfer coefficient: USG = 4.5 [kJ/sec °C m2]
 Steam generator area: ASG = 3200 [m2] per steam generator (12,800 [m2] total for a
four-loop reactor)
 Primary flow rate: Wp = 8250 [kg/s]
 Heat capacity of primary loop: Cp = 4.25 [kJ/kg °C]
 Steam generator pre-heating portion: γ100 = 0.15 at 100% full power
 Feedwater temperature: TFW = 177 [°C] at 100% full power
 Secondary-side temperature (at saturation): TS = 260 [°C]
Using the above parameters, the reactor inlet temperature can be calculated using Eq. (145):
Q Q
Tp,cold  58.86  0.60  0.50  6.22  260  6.32  6.22  260 . (147)
Q100 Q100
Hence, even at 100% full power, the net correction to the reactor inlet temperature Tp,cold is
less than 0.1°C. Even allowing for large variations in USG and other parameters, the effect on
Tp,cold is expected to be small over the full power range. This has been confirmed by detailed
calculations. The reactor outlet temperature can be calculated using Eq. (146):
Q Q
Tp,hot  64.75  6.22  260   58.53  260 . (148)
Q100 Q100
At 100% full power, the reactor outlet temperature To is 318.5 [°C], which is greater than the
saturation temperature of 310 [°C] at 10 [MPa]. Hence, the assumption of no boiling at the
reactor outlet header is not correct. Therefore, the calculation must be repeated with an
estimate of the amount of boiling using the following expression:
 xh fg 
Q  Q subcool  Qboil  W p C p  T p , sat  T p ,cold     W p C p T p , sat  T p ,cold   W p xh fg . (149)
 C p 
The power level at which boiling starts is given by Eq. (146), into which the CANDU 6
parameters from Eq. (148) are introduced to obtain the following expression:
Tp , hot  260 310  260 Q 50
    0.854 . (150)
58.53 53.53 Q100 53.53
This means that boiling starts in a CANDU 6 core when the reactor power is at 85.4% of full
power. In other words, it means that out of 100% reactor power, 85.4% is used to heat the
primary fluid to saturation and 14.6% is used to boil some of the primary fluid.
If this information is used in Eq. (149), the following expression can be obtained for the reactor
power split in the primary heat transport system:
Q Q Q WC W xh
 subcool  boil  p p Tp , sat  Tp ,cold   p fg  0.854  0.146 . (151)
Q100 Q100 Q100 Q100 Q100
Hence, using just the boiling part of Eq. (151), the steam quality at the reactor outlet header
can be calculated as follows:
0.146 Q100 0.146  2.064  10 6  kJ / s 
x   0.045 . (152)
W p h fg 8250[ kg / s ]  800[ kJ / kg ]
In the above calculation of the steam quality in the reactor outlet header as 4.5%, it was
assumed that the value of 85.4% of the total reactor power to onset of boiling as calculated by
Eq. (152) remains valid as the power goes up beyond the onset of boiling. Hence, it follows
that the parameters that determine the onset of boiling, Wp, Cp, USG, ASG, Ts, and TFW, have been
assumed not to change significantly when boiling starts in the primary heat transport system.
This is only approximately true: Wp and USG are affected by the presence of two-phase flow.
However, the approximation used in the above calculation is good enough to illustrate the
point.
End of Problem 7.3.4.1_________________________________________________________
Enough information is now available to sketch out the heat duty diagram as a function of
power. Figure 66 presents the impact of reactor power on the heat-duty diagram of the steam
generator {GAR1999, POP2014].
Figure 66 Reactor power versus heat-duty diagram of the SG

It is clear from the figure that the increase of reactor power impacts the primary side in such a
way that the primary fluid temperature curve becomes steeper, eventually reaching saturation
and starting to boil. The higher the reactor power, the greater will be the quantity of the
primary fluid that will be converted to steam, and the higher will be the quality of the primary
coolant in the reactor outlet header.
The reactor power also affects the secondary side. The lower the reactor power, the larger will
be the pre-heater part of the steam generator, i.e., more of the steam generator heat transfer
area will be engaged in pre-heating feedwater, and less will be available to boil the secondary
fluid.
7.3.5 Steam generator with pre-heater (numerical solution)

Consider the steam generator to be a counter-current heat exchanger, with the simplified heat-
duty diagram shown in Figure 67. For any small segment dz of the heat exchanger, according to
the figure, the heat transferred from the primary to the secondary side can be calculated using
the Fourier law:
dQ  U SG Tp  Ts  dASG . (153)
The primary and secondary temperatures will vary with position in the steam generator.
Energy balance on the primary and secondary sides gives the following expressions:
dQ  WP dhP , (154)
dQ  WS dhS . (155)
The minus sign indicates a heat flow from the primary to the secondary side, i.e., the enthalpy
of the primary fluid decreases, and the enthalpy of the secondary fluid increases.
For a single phase on either side, one can relate enthalpy with temperature as follows:
dh=CP⋅dT.
Therefore, the change in fluid temperatures along the steam generator primary and secondary
sides can be expressed as:
U SG dASG T p  Ts 
dT p   , (156)
C p , pW p
U SG dASG T p  Ts 
dTs  . (157)
C p , sW s
The steam generator area can be conveniently divided into N segments, each described as
ΔASG=ASG/N. The numerical algorithm for calculating the fluid temperatures simply starts at
one end of the heat exchanger with known or assumed temperatures and flows at that
boundary and repeatedly applies Eqs. (156) and (157) as it moves forward to the other end of
the heat exchanger. The principle of the calculation is shown in Figure 68. If the calculations
start at the cold end (where feedwater enters and primary fluid exits), single-phase flow on
both sides can be ensured.
At each successive nodal point, the fluid temperatures can be calculated by the following
expressions [GAR1999, POP2014]:
Tp,i 1  Tp,i 
U SG ASG
C p, p NWp
Tp,i  Ts,i  , (158)
Ts ,i 1  Ts ,i  SG SG Tp ,i  Ts ,i  .
U A
(159)
C p ,s NWs
At each nodal point i, the calculated temperature should be compared to the saturation point.
Once the saturation temperature is reached, the secondary-side temperature remains at the
saturation temperature.
Figure 67 Steam generator numerical Figure 68 Nodalization concept for steam

calculation concept generator numerical calculation
The result is the temperature profiles for both the primary and secondary sides for the given
flow rates, area, heat transfer coefficient, and cold side temperatures. Equation (153) is used
to accumulate the total heat transferred by the heat exchanger. If the heat transferred is
above or below the desired heat transfer, then iteration is required. Typically, it is desirable to
know the primary-side temperature profile for a given set of secondary-side conditions, given
the primary-side flow and a given steam generator geometry and heat transfer coefficient. In
this case, the primary side inlet temperature is varied until the target Q is met; once again, the
primary side “floats” on the secondary side.
The effect of power is seen through Ws, which is proportional to Q. For low Q, Ts will rise
rapidly to the saturation temperature. This decreases the effective temperature difference
between the primary and secondary sides. Hence, Tp will not rise as quickly as in the high-
power case.
7.3.6 Problems
1. Explain the difference in the design of steam generators that have an area designed for
preheating of feed water and these that do not. Explain how feed water heating is
achieved in both types of design. Explain the role of the pinch point in both types of
design.
2. Using the values provided in Problem 7.3.4.1, draw a heat duty diagram of a
corresponding steam generator following the format provided in Figure 66.
3. Explain the process for calculating main parameters of the primary heat transport system.
4. Using the values provided in Problem 7.3.4.1, and using the steam generator heat
balance template equations, calculate analytically the steam generator parameters.
Determine the effect of steam generator fouling on the steam generator performance.
Determine the effect of power level on the steam generator performance.
5. Using the values provided in Problem 7.3.4.1, and using the steam generator heat balance
equations use numerical technique to calculate steam generator parameters. Determine
the effect of steam generator fouling on the steam generator performance. Determine the
effect of power level on the steam generator performance. Determine the impact of the
numerical calculation increments on the precision of the calculations.
7.4 Heat transfer in the fuel elements

Heat transfer in the fuel, cladding, and coolant is very important from the perspective of
adequate thermal design of the reactor core. The most important objective of the thermal
designer is to ensure that the heat generated in the reactor at any reactor state is continuously
removed from the core within adequate thermal margins.
This section describes the calculation of temperature profiles in the radial and axial directions
in the fuel, cladding, and coolant. It also discusses the key parameters and their influence on
primary heat transport system design. Note that coolant mixing within the bundle geometry is
beyond the scope of this text.
7.4.1 General heat conduction equation

The interface between the fuel and the coolant is centrally important to reactor design because
it limits the heat flux that can be removed from the fuel to the coolant, thus limiting the reactor
power output.
For a solid body, the general thermal energy balance equation can be written as follows
[BIR1960, TOD2011]:
    e   
 t
   
 ,
dV   q ''' r , t dV   q " r , t ndS (160)
V V S
where ρ is the material density, e is the internal energy, V is the volume, S is the surface area,
q''' is the volumetric heat generation, q'' is the heat flux, and n is the unit vector on the surface.
The term on the left-hand side is the heat storage term, which represents the net change in
energy level. The first term on the right-hand side of the equation is the heat generation term,
which describes the heat generated at any time and any volume in the solid body. The second
term on the right-hand side shows the net energy exchange from the body to the surroundings.
The right-hand side shows the net energy exchange from the body to the surroundings. Gauss’
Law connects the surface integral to a volume integral of a body that is enclosed by a given
surface [BIR1960]:
  
q  
 " r , t n dS    q " r 
, t dV . (161)
S V
Using Gauss’ Law and Eq. (161), the surface integral in the energy balance equation, Eq. (160),
can be converted to a volume integral, following which the volume integral can be dropped for
all terms in the equation, and the internal energy e can be replaced by the temperature T times
the heat capacity c:
  cT   
t
   
 q ''' r, t q" r, t . (162)
In addition, a relation is needed to specify the heat flux in terms of temperature. In a solid
body, Fourier’s law of thermal conduction is applicable and can be expressed as follows
[ELW1978]:
 
   
q " r, t  k  T r, t . (163)
Hence, substituting Eq. (163) into Eq. (162) leads to an equation describing the temperature
distribution in a solid body in which heat generation and conduction take place:
  cT   
t
    
 q ''' r, t  kT r, t  , (164)
where the solid body density ρ is in [kg/m3]; the specific heat capacity c is in [J/(kg oK]; the heat
conductivity k is in [J/(m oK sec)]; the heat flux q’’ is in [J/(m2 sec) = W/m2]; the heat generation
rate q’’’ is in [J/(m3 sec) = W/m3]; the temperature T is in [oK]; and α is defined as (k/ρc)
[m2/sec].
7.4.2 Temperature distribution in the radial direction

Consider a typical cylindrical fuel pin composed of a fuel element surrounded by metal
cladding, as shown in Figure 69. The fuel pellets are in the fuel element and are surrounded by
a small gap between the fuel and the cladding, which offers substantial resistance to heat
transfer. The flowing coolant surrounds the pin. This discussion will consider the fuel, gap,
cladding, and coolant separately to develop the temperature profile in each material. Then the
resulting equations will be combined to yield the full fuel-to-coolant temperature profile. It is
sufficient for the present purposes to focus on the steady state to provide an illustration of a
fuel temperature distribution and the separate contributions of thermal resistance through
various materials. The steady-state version of Eq. (164) will not have the term on the left-hand
side, meaning that the heat generated in the fuel is equal to the heat removed by the coolant:
Figure 69 Cross section of a cylindrical fuel pin

 
    
 k f T r, t   q ''' r, t . (165)
For a cylindrical fuel pellet, Eq. (165) must be expanded into cylindrical coordinates:
1d dT 
kf r   q '''  r  . (166)
r dr  dr 
Equation (166) must be integrated twice; after the first integration, it is transformed into the
following equation:
dT r2
kf r   q '''  r  . (167)
dr 2
The constant of integration is zero because the temperature gradient at r=0 is zero. The
thermal conductivity, k, is a strong function of T in fuel, and therefore its variation needs to be
taken into account in the integration process. However, in the first step, its average value k f
can be assumed to be within the range of fuel temperature variation. Hence, the second
subsequent integration results in the following relation:
rF2
 k f T dT  k f TF  To    q "'  r  ,
TF
(168)
To 4
where the subscript 0 indicates the centre point and the subscript F indicates the fuel pellet
radius. Because T=To at r=0, the constant of integration is again zero. Finally, with the
assumption of average fuel conductivity in the fuel pellet, the temperature difference across
the fuel pellet has the following form:
rF2 q'
T fuel  To  TF  q "'  , (169)
4k f 4 k f
where q' rF q" is the linear power density. Note that by using linear power density, the
2
same ∆T is obtained for a given q' regardless of the fuel radius.

_____________________________________________________________________________________________
Problem 7.4.2.1
Calculate for UO2 ceramic fuel the radial temperature difference in the fuel. For UO2 ceramic
fuel, kF is typically 0.02–0.03 [W/cm°K]. At q'=500 [W/cm], ∆Tfuel in the fuel pellet is about
1400°C.
End of Problem 7.4.2.1________________________________________________________
In the gap between the fuel pellet and the cladding, no heat is generated. Therefore, for
steady-state conditions, Eq. (164) transforms into the following equation:
1 d  dT 
 kg r   0. (170)
r dr  dr 
This equation can be readily integrated once to yield the following equation:
dT
kg r  const . (171)
dr
The constant of integration is determined by considering the heat flux q'' at the fuel-gap
interface:
 dT  q'
k g    q"  . (172)
 dr  r rF 2 rF
Substituting Eq. (172) into Eq. (171) yields the following equation:
dT q' dT q'
kg r   kg  . (173)
dr 2 dr 2 r
Assuming that the gap conductivity does not change significantly with temperature and
performing integration, the temperature difference in the gap can be obtained as:
q '  rF  tG 
k g  TGAP  k g TF  TC   ln  . (174)
2  rF 
The subscript C indicates the gap-cladding interface. The boundary condition T=TC at r=rF+tG is
incorporated into this solution. Finally, the gap temperature difference is obtained as:
q' r t 
 TGAP  ln  F G . (175)
2 k G  rF 
Because the natural logarithm of a small number can be approximated by the number itself,
i.e., ln(1+x) ≈ x [GAR1999, POP2014], the gap temperature difference is:
q '  tG 
 TGAP    . (176)
2 rF  k G 
_____________________________________________________________________________________________
Problem 7.4.2.2
Calculate for UO2 fuel the radial temperature difference in the fuel-cladding gap. The gap
conductivity kG is about 0.002 W/cm°K, but it varies considerably with the amount of fission
product gases. For a gap thickness of 0.005 cm, ∆TGAP is about 300°C for q' of 500 W/cm.
End of Problem 7.4.2.2________________________________________________________
Because with irradiation the fuel will swell to touch the cladding at certain points (because the
surfaces have a certain degree of roughness), an effective heat transfer coefficient, hG, is used:
hG  TGAP   q " . (177)
Therefore, taking into account the relationship between heat flux and the linear power
q '  rF2q" , the temperature difference in the gap can be obtained as follows:
q'
TGAP  . (178)
2 rF hG
_____________________________________________________________________________________________
Problem 7.4.2.3
Calculate for UO2 fuel the radial temperature difference in the fuel-cladding gap. A heat
transfer coefficient of 0.5–1.1 W/cm2°K gives a ∆TGAP less than 300°C.
End of Problem 7.4.2.3________________________________________________________
For the gap region described above, the steady-state equation for the cladding region, in which
there is no heat generation, is similar:
1 d  dT 
 kc r   0 . (179)
r dr  dr 
This equation is solved in the same manner as for the gap to yield:
q '  rF  tG  tC  q'  r t t 
k c  TCLAD  k c TC  TS   ln     TCLAD  ln  F G C  . (180)
2  rF  tG  2 k c  rF  tG 
The subscript S indicates the cladding-coolant surface interface. The boundary condition T=TS
at r=rF+tG+tC is incorporated into this solution.
Because the natural logarithm of a small number can be approximated by the number itself,
i.e., ln(1+x) ≈ x [GAR1999, POP2014], the gap temperature difference is:
q'  tC 
TCLAD   . (181)
2  rF  tG   kC 
_____________________________________________________________________________________________
Problem 7.4.2.4
Calculate for UO2 fuel the radial temperature difference in the fuel cladding. The cladding
conductivity kC is about 0.11 W/cm°K, giving a ∆TCLAD of about 80°C for a q' of 500 W/cm.
End of Problem 7.4.2.4________________________________________________________
The heat flux from the cladding to the coolant is determined as:
q "  hS TS TFL  , (182)
where TFL is the bulk temperature of the coolant fluid. Hence, the temperature drop from the
cladding surface to the bulk fluid temperature is:
q'
TCOOL  . (183)
2 hS  rF  tC  tG 
_____________________________________________________________________________________________
Problem 7.4.2.5
Calculate for UO2 fuel the radial temperature difference in the cladding-to-coolant interface. A
heat transfer coefficient of ~4.5 W/cm2°K gives a ∆TCOOL of about 10°C–20°C.
End of Problem 7.4.2.5________________________________________________________
By summing up Eqs. (169), (178), (181), and (183), the overall temperature difference between
the fuel centreline and the bulk coolant can be obtained according to the following equation:
q'  1 1 t G  tC 1 
To  TFL       . (184)
2  2 k f h r
G F k C  rF  t G  hS  rF  t G  t 
C 
The expression in brackets in Eq. (184) is the thermal resistance across the fuel element:
 1 1 tG  tC 1 
Rth       . (185)
 2 k f h r
G F k C  rF  t G  hS  rF  t G  t 
C 
Hence, the centreline fuel temperature is greater than the coolant temperature by an amount
that depends on the amount of heat generated and on various resistances to heat flow. For a
given fuel design, most of the parameters are fixed under normal operation. The one
exception is the heat transfer coefficient to the coolant, hS, which is covered in Section 7.6.
Figure 70 shows a radial temperature distribution across the fuel element for two linear power
rates [POP2014]. Linear power level 45 [kW/m] represents full-power operation of the fuel
element in the middle of the core. Linear power level 15 [kW/m] represents operation at full
power at the core ends, or at a certain decay power level shortly after reactor shutdown. This
figure shows that the highest temperature difference occurs in the fuel pellet (above 1400°C at
full power) because UO2 is a poor heat conductor, but its melting temperature is approximately
2800°C–2900°C. The temperature difference in the gap is significant considering that the gap is
very thin. The cladding and the water are very good heat conductors, and therefore the
temperature difference is relatively small.
Figure 70 Radial temperature distribution in a fuel element

The fuel element design must ensure that given the coolant pressure and saturation
temperature, the fuel pellet centreline temperature avoids exceeding the melting temperature
by a good margin. This objective also applies to the temperature transients in design basis
events.
Figure 71 shows the temperature gradient for various fuel element materials in absolute terms
and per unit thickness [POP2014]. This figure clearly indicates that the highest relative
contribution is from the fuel-cladding gap and the second highest from the fuel pellet.
Figure 71 Radial temperature gradient per unit thickness in a fuel pin
7.4.3 Temperature distribution in the axial direction

The fuel and coolant temperature distributions depend on heat transfer in the radial direction
at each position in the channel axial direction. The general heat conduction and convection
equations for axial heat transfer are covered in Chapter 7.
The amount of heat generated along the fuel channel follows the neutron flux and fission
reaction distribution along the fuel channel. Assuming that the reactor core does not use any
techniques for balancing (i.e., flattening) the neutron flux and power generation in the axial
direction, the natural axial variation of the neutron flux is given by the following relation
[POP2014]:
z 
  z    max cos  , (186)
 Le 
where Le is the effective core length and Φmax is the maximum neutron flux in the middle of the
core.
Axial heat generation follows the neutron flux variation and is given by the following equation:
z 
q "'  z    q "' max cos  , (187)
 Le 
where (q’’’)max is the heat flux at the middle of the fuel element in the axial direction.
Consider a heat balance for a differential section of a fuel element of length dz, fuel cross
section Af [m2], and mass flow rate w [kg/s]. In a steady-state situation, the sensible heat gain
by the passing coolant (assuming no phase change) is equal to the heat generated in the
differential fuel element (cp is the specific heat capacity at a given pressure [kJ/kg oC]):
w cpdTc  q"' Af dz . (188)
The axial variation of the fluid temperature Tc can be obtained by substituting Eq. (187) into Eq.
(188) and integrating [ELW1978]:
z
 z 
dTc   q "' max A f dz , (189)
Tf
wc p   cos 
Tf 1
 Le / 2  Le 
 q "' max A f Le    z   q "' max V f    z 

Tc  Tc ,in  1  sin     Tc ,in  1  sin   , (190)
 wc p   Le    wc p   Le  
where Vf [m3] is the volume of the fluid portion of the fuel element. The subscript “in”
designates the entry point in the channel. In this equation, the equivalent core length is
approximated by the actual core length, i.e., L=Le.
The coolant temperature is measured in the middle of the fuel element for z=0 and has the
following form:
 q "' max Af L
. (191)
Tc , mid  Tc ,in 
 wc p
The exit temperature of the coolant from this channel can be obtained by integrating Eq. (189)
to the channel exit (labelled as 2, i.e., at z=+L/2):
2  q "' max A f L
T f ,out  T f ,in  . (192)
 wc p
This equation shows that the coolant temperature rise in the channel will be greater for higher
heat generation rates, larger fuel elements, and longer channels and smaller for higher mass
flow rates.
The heat transferred between the cladding and the coolant at any location z along the fuel
channel, per unit area of cladding surface in contact with the coolant, is given by h(TSH-Tc),
where h is the heat transfer coefficient. This coefficient can be assumed constant along the fuel
channel with no boiling in the channel. Hence, the heat balance between the fuel cladding and
the coolant can be expressed by the following equation:
z 
hCSH TSH  Tc  dz   q "' max Af cos   dz , (193)
 L 
where CSH [m] is the fuel cladding circumference, TSH [oC] is the cladding temperature, and h
[kJ/m2 oC] is the heat transfer coefficient. Substituting Eq. (190) into Eq. (193) yields the
cladding temperature variation along the fuel channel:
 q "' max A f L    z    q "' max A f z . (194)
TSH  Tc ,in  1  sin  L    hC cos  
 wc p    SH  L 
The temperature variation at the fuel pellet surface along the fuel channel and the
temperature of the fuel pellet centreline along the fuel channel can be obtained by means of
the following relation:
TSH  Tc z 
  q "'max Af cos   . (195)
Rx L  L 
Hence, for fuel pellet surface temperature, the following equations can be developed:
T f , S  Tc  z 
  q "' max Af cos   , (196)
R f ,S L  L 
   z 
T f ,S  Tc  R f ,S L  q "' max Af cos    , (197)
  L 
 1 t G  tC 1 
R f , S     . (198)
 hG rF k C  rF  tG  hS  rF  tG  t SH  
The fuel pellet centreline temperature variation can be described by the following equations:
T f ,CL  Tc z 
  q "' max A f cos  , (199)
R f ,CL L  L 
   z 
T f ,CL  Tc  R f ,CL L  q "' max Af cos    , (200)
  L 
 1 1 tG  tC 1 
R f ,CL      . (201)
 2 k f hG rF k C  rF  tG  hS  rF  tG  t SH  
Substituting Eq. (190) into Eq. (200) yields the fuel centreline temperature:
 q "'max
Af L    z  z .
T f ,CL  Tc ,in  1  sin      q "'  max A f LR f ,CL cos   (202)
 wc p   L   L 
The location of the maximum cladding temperature can be obtained by differentiating Eq. (194)
and equating it to zero, i.e., dTSH/dz=0. Hence:
 hC SH L  H
cot 1  wc p RSH  ,
L (203)
z SH ,m  tan 1  
   wc  
 p 
1 1
RSH   . (204)
hAf hCSH L
The location of the maximum fuel pellet centreline temperature can be obtained by
differentiating Eq. (202) and equating it to zero, i.e., dTf,CL/dz=0. Hence:
 cot 1  wc p R f ,CL  .
L
z f ,CL ,m  (205)

The maximum cladding temperature can be found by substituting Eq. (203) into Eq. (194):
1  1   2 
TSH ,m  Tc ,in   q "'max Af LRSH   , where   wcpRSH . (206)
  
The maximum cladding temperature can be found by substituting Eq. (205) into Eq. (202):
1  1   2 
T f ,CL ,m  Tc ,in   q "'max Af LR   , where    wcp Rf ,CL . (207)
  
Figure 72 Temperature distribution in the axial direction in a fuel element

Figure 72 shows the temperature distribution in the coolant and the cladding plotted on the
same graph with the heat flux distribution along the core [ELW1978]. The heat flux distribution
is assumed to have a cosine distribution, which is more pronounced in a core in which no
measures have been taken to flatten the neutron flux. In this case, the variations in the coolant
and cladding temperatures are more pronounced. Note that the coolant temperature curve is
symmetrical. It has a small gradient at the entry and exit parts of the core, where the heat flux
is lower. The highest coolant temperature gradient is in the middle part of the core, where the
heat flux is substantial. Of course, different core heat flux distributions can have a significant
impact on the temperature profile. Moreover, reactivity devices and fuel channel orificing (i.e.,
channel-to-channel variation of flow rates) can also have significant local effects on the
temperature profile.
Figure 73 shows a temperature distribution that includes the coolant (Tc), fuel sheath (TSH), fuel
pellet surface (Tf,S), and fuel pellet centreline (Tf,CL) temperature profiles. The positions of the
maximum values of TSH, Tf,S, and Tf,CL occur at ZSH,m, Zf,S,m, and Zf,CL,m, closer to the fuel element
mid-plane [ELW1978].
Figure 73 Temperature distribution in the fuel cladding and the fuel element
The reason that the maximum fuel centreline temperature is closer to the fuel mid-plane is that
the radial heat flow through the fuel and cladding at any section along the fuel channel is
proportional to the heat flux at that location. The temperature differences (TSH ̶ Tc), (Tf,S ̶ TSH),
and (Tf,CL ̶ Tf,S) generally follow cosine-shaped functions according to the heat flux cosine profile.
The maxima of the cladding and fuel temperatures are affected by the heat flux maximum (at
the fuel mid-plane) and the fluid maximum (at the fuel channel exit). Because the cladding is
close to the fluid, the impact of coolant temperature is strong, and hence its maximum is pulled
away from the mid-plane. Because the fuel centreline temperature occurs inside the fuel,
where the heat is generated, its maximum is pulled toward the mid-plane, where the heat flux
maximum is located.
7.4.4 Axial distribution of quality

If the coolant boils in a fuel channel, flow quality changes along the channel. It is important in
this situation to calculate the change in flow quality and the length of the channel where
boiling occurs. The enthalpy of the two-phase coolant is defined using the following relation:
h  hc,sat  xhfg , (208)
where x is the flow quality in the two-phase mixture, hc,sat is the saturated liquid enthalpy, and
hfg is the latent heat of vaporization. The heat balance over the differential part of the fuel
channel yields the following equation:
 L /2
w f  h  z   hcold    q '( z )dz . (209)
 L/ 2
In the above equation, hcold is the inlet enthalpy in the channel. If the axial position where the
coolant starts to boil (the point where h(z)=hSAT) is defined as ZBB, the heat balance equation,
Eq. (209), transforms into the following equation:
z
w f  h  z   hc , sat    q '( z ) dz  x ( z ) w f h fg . (210)
Z BB
Using the above equation, the flow quality of the two-phase mixture can be expressed by the
following equation:
z
1
x( z ) 
w f h fg 
Z BB
q '( z ) dz . (211)
7.4.5 Thermal conductivity and fuel element material properties

In developing Eq. (166), it has been assumed that the fuel thermal conductivity can be
approximated by an average value that will apply in the range of parameters used in the
present calculations. This section covers the effect of various factors on thermal conductivity
and other thermal parameters that must be considered when developing a model to calculate
fuel-element temperature profiles.
Many factors affect UO2 fuel thermal conductivity. The major factors are temperature,
porosity, oxygen-to-metal atomic ratio, PuO2 content, pellet cracking, and fuel burn-up.
Figure 74 shows the temperature dependence of UO2 thermal conductivity [ELW1978].
Experimental measurements have shown that thermal conductivity decreases with increasing
fuel temperature (i.e., has a negative slope) until about 1750°C, after which it slowly increases.
However, the slope is very steep from 0°C–500°C and very shallow thereafter. Nevertheless,
the impact of temperature must be taken into account when calculating fuel temperature
profiles, which is routinely done in thermal-hydraulic computer programs. Figure 74 also shows
the integral of the thermal conductivity, which monotonically rises with temperature; this is the
value to use in solving Eq. (170).
When using experimental values of thermal conductivity, one must be aware of the
measurement uncertainties. Note that these uncertainties are significant, particularly at higher
temperatures.
Figure 74 Thermal conductivity in a fuel pin

The oxide fuel is produced by sintering pressed powdered UO2 or mixed oxide at high
temperature. Under such conditions, for any material with a given density, usually only around
90% of the maximum possible (theoretical) density of the solid can be produced.
The thermal conductivity of a solid usually decreases with increasing presence of voids within
the structure. Therefore, to maximize thermal conductivity, fuel porosity must be minimized.
However, during fuel irradiation, fission gases are produced, resulting in increased internal
pressures that may swell and deform the fuel. Hence, a certain degree of porosity is desirable
to accommodate the fission gases and limit the possibility of swelling. This is very important
for reactor cores that have high power density and therefore produce more gases per unit
volume.
The oxygen-to-metal atomic ratio of UO2 and PuO2 can vary from the theoretical
(stoichiometric) value of 2 because of fuel irradiation and burn-up. A reduction or increase in
the oxygen-to-metal ratio results in a reduction of thermal conductivity, as can be clearly seen
in Figure 75 for UO2 and PuO2 mixtures [TOD2011].
Figure 75 Impact of oxygen-to-metal ratio on the heat conduction coefficient in fuel

The thermal conductivity of mixed fuel decreases as the plutonium oxide content increases.
Therefore, as fuel burns up and ages and plutonium content increases, thermal conductivity
decreases.
Figure 76 shows the impact of fuel temperature on fuel heat capacity for UO2 and PuO2 fuels
[TOF2011]. It is evident that heat capacity increases with temperature and decreases with an
increase in plutonium content.
Figure 76 Heat capacity of a fuel pin

The properties of the pellet-to-cladding gap are very important. As shown in Figure 71, its
contribution to thermal resistance is substantial. The gap consists of an annular space occupied
by gas. Initially, its composition is fill gas, which is usually an inert gas such as helium. With
burn-up, the gas composition changes as the gap is penetrated by gaseous fission products
such as xenon and krypton, resulting in a gap pressure increase.
Moreover, the fuel pellet cracks with irradiation, leading to circumferential deformation of the
gap and fuel pellet swelling. All this causes a reduction of thermal resistance in the gap and
increases its thermal conductance with burn-up. Figure 77 shows the effect of burn-up on
thermal conductance [TOD2011].
Figure 77 Gap conductance in a fuel pin

The left side of Figure 78 shows the temperature profile in the gap at nominal geometry (fresh
fuel). Note that near the walls, there is a mild temperature rise towards the fuel pellet and a
reduction towards the cladding because of the smaller number of molecules moving near the
walls (the boundary layer effect). The right side of Figure 78 shows the effect of the
deformations that occur with fuel burn-up (aging) because of swelling and cracking of the fuel
pellet and cladding. After a certain amount of irradiation, a number of peaks and valleys will
develop at the pellet and cladding surfaces. These may eventually produce contact between
fuel and cladding, improving thermal conductivity.
Figure 78 Fuel aging effect on temperature in a fuel pin

Figure 79 shows the typical impact of fuel aging on gap conductance for different linear heat
ratings as a function of initial gap thickness [TOD2011]. Note that better thermal conductance
is achieved for smaller gap thickness and for higher linear power rates, as could be expected
from the above explanation. Therefore, as gap thickness becomes smaller with fuel burn-up,
heat transfer is improved. In addition, this figure shows that using a smaller gap thickness for
fresh fuel improves heat transfer.
Figure 79 Impact of gap thickness and linear power on heat conductance

The temperature profiles described so far in this section were derived based on the assumption
of radial symmetry across the fuel rod. This assumption was a useful way to simplify the
derivation of temperature profiles across the fuel element. This assumption is acceptable for
vertical reactor cores, such as in LWRs.
However, for horizontal reactor cores, such as CANDU, the assumption of radial symmetry
across the fuel rod is not applicable for certain flow regimes. This is relevant for situations in
which bubbly flow develops in the upper part of the fuel channel, or even more importantly,
when stratification develops (e.g., under accident conditions) in the fuel channel.
Figure 80 Asymmetric temperature profiles in a fuel pin
In these situations, the heat flux to the upper part of the fuel element could be smaller than to
the lower part. This results in a skewed (asymmetric) temperature profile, as shown in Figure
80. In such a case, the maximum fuel pellet centreline temperature will be skewed towards the
side with the smaller heat flux. This could also have an important impact on the cladding
temperature, and it is particularly relevant to some postulated accidents in which channel
stratification occurs.
7.4.6 Problems
1. A PWR reactor produces 2440 MWt with the following core information: UO2 pellet
diameter 0.96 cm; Zr clad inside diameter 0.98 cm; Zr clad outside diameter 1.12 cm;
active fuel length 3.7 m. The core contains 220 fuel bundles each with 14 x 14 fuel pins
(but 20 positions are used for control elements). The fission energy is distributed
uniformly across the fuel pin. and TB=300°C, kUO2=0.050 W/cm°C, hcool=1.8 W/cm2 °C,
hgap=0.57 W/cm2 °C], kclad=0.12 W/cm°C. Calculate core averaged fuel pin temperature
profile.
2. The core of a BWP reactor consists of 764 fuel assemblies, each containing a square
array of 49 fuel rods on a 1.9 cm pitch. The fuel rods are 4.4 m long, but fuelled only
along over 3.6 m of their length. The outside diameter of the fuel rods is 1.4 cm, the
cladding thickness is 0.8 mm, and the fuel pellets are 1.2 cm, thus leaving a gap of 0.2
mm. The UO2 has average density of 10.3 g/cm3. The radius of the core is 2.4 m, and the
reactor is designed to produce 3300 MWt. The peak to average power density is 2.6.
Calculate (a) the maximum heat flux in kW/m2; (b) the temperature distribution at the
location of maximum heat flux ignoring the gap between fuel pellet and cladding; and (c)
the temperature distribution at the location of maximum heat flux taking into account the
gap.
3. For PWR cylindrical solid fuel pellet operating ta a heat flux of 1.7 MW/m2 and surface
temperature of 400ºC, calculate the maximum temperature in the pellet for two assumed
values of heat conduction: 3 W/mºC, and k = 1 + 3e-0.0005T where T is ºC. The UO2 pellet
diameter is 10 mm, and density is 95% of the theoretical density.
4. A CANDU channel has a typical geometry of 37 fuel elements per bundle, 12 bundles
per channel, with power, pressure and inlet and outlet temperature. Assume for this
channel linear power rating of 60 kW/m; burnup at the time of peak rating of 50
MWh/kgU; sheath-to-coolant heat transfer coefficient 40 kW/m2K; pellet-to-sheath heat
transfer coefficient of 60 kW/m2K; and pellet density of 10.6 Mg/m3. Calculate coolant,
cladding and fuel surface and centreline temperatures in radial direction for each bundle
in axial direction assuming typical power distribution along the channel with a typical
power peaking factor. Draw appropriate diagrams showing the results.
5. A fuel pin with pellet radius of 4.6 mm, clad inner radius of 4.89 mm, and outer radius of
5.46 mm, calculate the maximum liner power that can be obtained from the pellet such
that the mass average temperature in the fuel does not exceed 1200 ºC. Take the bulk
fluid temperature at 307.5 ºC, and the coolant heat transfer coefficient of 28.4 kW/m2 ºC.
In the gap consider only conduction heat transfer. Fuel conductivity is 3 W/m2 ºC; clad
conductivity is 18.69 W/m ºC; and helium gas conductivity is 0.277 W/m ºC.
6. Using analytical technique develop equations for calculating temperature distribution in
radial direction in the fuel element (fuel pellets, gap, cladding and coolant) calculate
temperature distribution in radial direction at heat loading of 40 kW/m. For the thermal
conductivity in the fuel pellets assume the distribution shown in the diagram below. To
solve the problem approximate the thermal conductivity by a suitably selected quadratic
relationship.
Compare the results with the case assuming constant average heat conductivity in the
fuel.
7. Using a numerical technique calculate temperature distribution in radial direction in the

fuel element (fuel pellets, gap, cladding and coolant) calculate temperature distribution in
radial direction at heat loading of 40 kW/m. For the thermal conductivity in the fuel
pellets assume the distribution shown in the diagram above. Compare the results with the
case assuming constant average heat conductivity in the fuel (0.03 W/cmoC).
8. Calculate axial temperature distribution of fuel centreline, fuel surface, cladding and
coolant in axial direction for a typical CANDU fuel assuming maximum fuel heat load of
50 kW/m, and average fuel load of 25 kW/m. Develop relations for and calculate the
positions of maximum temperatures. As per figure below, as a base case assume (1)
cosine heat flux distribution, (2) constant heat flux distribution, and (3) asymmetrical
linear heat flux distribution (with maximum at bundle 9), and compare the maximum
temperatures and their locations in the core.
9. Calculate axial temperature distribution of fuel centreline, fuel surface, cladding and
coolant in axial direction for a typical CANDU fuel assuming maximum fuel heat load of
50 kW/m, and average fuel load of 25 kW/m. Develop relations for and calculate the
positions of maximum temperatures. As per figure below, as a base case assume (1)
cosine heat flux distribution, (2) linearly increasing heat flux distribution to the channel
exit, and (3) linearly decreasing heat flux distribution to the channel exit, and compare
the maximum temperatures and their locations in the core.
10. Using parameters of a typical CANDU channel, calculate the axial temperature
distribution in a fuel channel. Identify the key parameters that affect the calculations,
and investigate the sensitivity of the results to these key parameters.
7.5 Fluid flow fundamentals

The flow of coolant in the reactor primary heat transfer system is of key importance for cooling
the reactor core under all operational situations, during accidents, and during reactor
shutdown. This section provides information on pressure drop calculations and discusses the
special topic of flow instabilities.
Design of the primary pump is of critical importance to ensure a robust reactor thermal-
hydraulic design. The sources of various friction and local pressure losses around the primary
heat transport system are described and compared between PWR and CANDU reactors.
Calculation methods used to determine pressure losses around the loop are provided, along
with an explanation of the empirical correlations used to calculate pressure losses in single- and
two-phase flows in the CANDU primary heat transport system.
The reactor core has a complex geometry, particularly the CANDU fuel channels and the HTS
headers and feeders. Moreover, the CANDU reactor primary heat transport system operates
partially in two-phase flow. This section describes the type and nature of the experimental
tests that were performed to define a number of parameters in the pressure drop calculation.
7.5.1 Introduction to pressure losses

In the nuclear industry, pressure drop calculations are important to determine the appropriate
flow rate in the core and the fuel channels, thereby ensuring appropriate heat removal from
the core under all possible operating modes. Because almost all reactors under some
operating modes operate with two-phase flow, pressure drop calculations in this situation
become complex, directly affecting power output and requiring substantial R&D support.
Two-phase pressure drop depends on many parameters, including geometric configuration,
mass and volume fractions of the phases, pressure, fluid properties, mass flux, duct orientation,
flow direction, and particularly flow patterns. Furthermore, two-phase flow is usually
accompanied by heat transfer and changes of phase in multi-component flow. To take into
account these many parameters, various empirical correlations have been developed and are
used for pressure drop calculations. Mechanistic models also vary in complexity and fidelity,
and often they are based on many assumptions to make them workable in a given parameter
range.
It is necessary to consider the pressure loss associated with every component in the primary
heat transport system. This includes distributed pressure loss due to friction; local pressure
losses due to sudden variations in shape, flow area, direction, etc.; and fuel string-related
losses and pressure losses due to acceleration (because of flow area variation or fluid density
change) and elevation differences (gravity term).
Table 8 shows various components of pressure losses in the reactor heat transport systems in
terms of geometry, fluid status, flow nature, flow pattern, flow direction, operating conditions,
and driving force [IAEA2001]. Each of these situations introduces major differences in the type
and complexity of pressure loss calculations.
Pressure losses must be characterized for a number of components with different geometries.
In addition to frictional pressure losses in the system, Table 9 shows the most important
pressure losses in the CANDU reactor core that are specific to CANDU reactors [IAEA2001].
Table 8 Factors affecting pressure drop
Geometry Basic shapes: Circular pipe, rectangular channel, annulus, etc.

Other shapes: rod bundle, spacer, valve, orifice, plenum, header, pump,
etc.
Fluid status Single-phase
Two-phase – one, two, and multiple components.
Flow nature Laminar
Turbulent
Flow Bubbly, slug, annular, etc.
patterns
Flow Vertical up-flow
direction Inclined flow
Horizontal flow
Operating Steady-state
Conditions Transient
Driving force Forced convection
Natural convection
Table 9 CANDU distribution of pressure drop in the primary loop
Pressure Drop in Primary Loop

Entry loss from steam generator outlet pipes to header
Header to feeder entry loss
Inlet feeder bends
Inlet grayloc
Inlet grayloc to liner tube entry
Liner tube to channel entry
Fuel locator
Junction between two bundles
Channel to liner tube entry
Liner tube to outlet grayloc
Outlet grayloc
Outlet feeder bends
Feeder to header entry loss
Header to steam generator inlet pipe entry
The objective of calculating pressure losses with adequate precision is to optimize pump
capacity and power requirements, as well as to determine coolant flow rate in the primary
circuit, local conditions in bundles and sub-channels, and flow rates across parallel
interconnected sub-channels in fuel bundles.
7.5.2 Pressure drop conservation equations

The pressure losses in the primary heat transport system are represented by direct or indirect
terms in the conservation equations:
• Mass-balance (continuity) equation;
• Momentum-balance equation; and
• Energy-balance equation.
For design calculations, usually steady-state flow in a channel of uniform flow area in the axial
direction is considered. Moreover, negligible variation of fluid properties over the calculation
domain is assumed.
The momentum conservation equation is essential to calculate pressure losses along flow
paths. The other conservation equations are covered in Chapter 7. The momentum equation
can be written for homogeneous or separated flow conditions. Figure 81 shows a calculation
domain (control volume) in an inclined flow situation. The forces acting on this control volume
are indicated in the figure. Momentum balance over this control volume in fact means a
balance of all forces acting on the control volume.
Figure 81 Force-momentum balance in a control volume

Homogeneous flow model
The momentum balance equation over the control volume using the homogeneous flow
assumption is given by Eq. (212) [BIR1960, WAL1969]:

  dP  d
  P   P  dz  z  dA    W  zdS    Gum   zdA    H g  sin    z  dA , (212)
A S A
dz A
where P is the pressure, τw is the wall friction stress, G is the mass flux, θ is the inclination
angle, g is the gravity acceleration, dz is the differential axial distance along the channel, δz is
the infinitesimal axial distance in the channel, um is the mixture velocity, S is the channel
circumference, A is the area of the flow channel, and ρH is the homogeneous mixture density as
a function of flow quality, given by Eq. (213) [WAL1969]:
1 x   1  x   g
 e  . (213)
H  g  
Equation (212) can be transformed into the following equation after performing some algebraic
simplifications and removing the integrals:
dP S d  G2 
  W      H g sin  . (214)
dz A dz   H 
By substituting the fluid density from Eq. (213), the pressure drop equation can be obtained in
the following form:
dP S d  xe 1  xe  
  W  G 2      H g sin  . (215)
dz A dz   g   
Equation (215) contains three pressure drop components on the right side: friction,
acceleration, and gravity:
dP   dP   dP   dP   .
          (216)
dz   dz  f  dz  a  dz  e 
The negative sign in the above equation indicates that pressure decreases along the flow path.
Separated flow model
In a separated flow model, both flow phases are treated separately, and therefore Eq. (212)
takes on the following form:
  
dP
  P   P  dz  z dA    W   z  dS  
d
dz
 Gu  Gg ug   zdA   H g  sin    z  dA . (217)
A S A A
The mixture density can be assumed to be a function of the two-phase void fraction as given by
the following relation:
TP    g  (1)   . (218)
By substituting Eq. (218) into Eq. (217), the pressure drop equation can be expressed as (in the
same form as Eq. (215)):
dP S d
  W   1    G u     G g u g      g  1        g  sin  . (219)
dz A dz 
Further developing the acceleration term using the relationship between the mixture and
phasic mass flux leads to the following expression for the separated flow pressure drop:
dP S d  x a2 (1  x a ) 2 
  W  G 2        g  1        g  sin  . (220)
dz A dz     g (1   )     
The frictional pressure loss occurs between the fluid and the channel wall. It depends primarily
on the fluid velocity, the tube diameter, and the fluid viscosity.
The acceleration pressure losses occur because of the change in fluid momentum between
different locations in the channel. They can be significant in channels with varying flow area
and fluid temperature.
The gravity-induced pressure loss in fact occurs because of a change in elevation, i.e., a change
of hydrostatic head. Therefore, it is relevant in vertical channels, particularly in piston flow. In
single-phase closed loops, the hydrostatic head in sections with upward flow cancels with that
in sections with downward flow.
Local pressure losses can result from many different flow obstructions such as valves, orifices,
bundle junctions, and appendages. In addition, changes in flow direction in elbows, T-
junctions, etc., can result in pressure losses. Significant pressure losses occur with sudden
changes in flow area such as sudden contractions, expansions, and orifices.
Using a term for local pressure losses as explained above, Eq. (216) can be written in the
following form:
dP   dP   dP   dP   dP   .
            (221)
dz   dz  f  dz  a  dz  e  dz  l 
7.5.3 Correlations for calculating pressure losses

Equation (221) has four terms for which correlations must be developed. Before describing
these correlations, it is necessary to define them.
Frictional pressure loss
The frictional pressure loss for a single-phase liquid flow is defined by the following relation
[IAEA2001]:
f  L  G2 f  L W 2
p f   , (222)
2 Dhy  Dhy 2  A2
where L is the pipe length, G is the mass flux, Dhy is the hydraulic diameter (equal to four times
the flow area/the wetted perimeter), ρl is the liquid density, W is the mass flow rate, A is the
pipe cross section, and f is the friction factor. The frictional pressure drop occurs all along the
length and is hence referred to as distributed pressure drop. This equation is applicable for
single-phase and homogeneous two-phase flows, although the methods of calculating the
friction factor, f, and the density, ρ, differ in the two cases. Pressure drops across tubes,
rectangular channels, bare annuli, and a bare rod bundle (i.e., without spacers) are examples of
this component. The challenge in the above equation is to define the appropriate friction
factor f for the given flow conditions. This process is described in the following sub-sections.
Acceleration pressure loss
This reversible component of pressure loss is caused by a change in flow area or density.
Expansion, contraction, and fluid flow through a heated section are examples. The acceleration
pressure loss due to flow area changes can be expressed as [IAEA2001]:
Pa  G 2
v  
1  Ar  W 2
. (223)

2 A02 l
The term Δvl describes the velocity difference between two points in the channel, Ar is the ratio
of the smaller to the larger flow area, Ao is the smaller flow area, and φ is equal to 1 for single-
phase flow. For two-phase flow, φ is described by the following relation [IAEA2001]:
 x3 1  x   
3
 g l 
  2 2  2 2   . (224)
  g
  g 1      x  l  1  x    g 
The acceleration pressure loss due to density change for single-phase and two-phase flows can
be expressed as:
 1   1  
pa  G 2       . (225)
   m o     m i

 
For single-phase flows, this component is negligible, but it can be significant in two-phase
flows. For two-phase flow, the above equation can be used with ρm given by [IAEA2001]:
1  x2 1 x 
2
  . (226)
 m   g   l 1    
To evaluate acceleration pressure loss due to density change, accurate prediction of fluid
density is necessary. For single-phase flow, fluid density can be predicted reasonably well using
established relationships for thermo-physical properties of fluids. For two-phase flow, it is
necessary to predict the void fraction, and in turn the acceleration pressure loss, accurately to
determine density. Hence, the correlation for void fraction must be chosen carefully.
Elevation pressure loss
This reversible component of pressure loss is caused by the difference in elevation and can be
expressed as:
pe    g    l 1      g  sin   z . (227)
Local pressure losses

These are the localized irreversible pressure loss components caused by changes in flow
geometry and flow direction. Pressure drops across valves, elbows, tees, and spacers are
examples. The local pressure loss (i.e., the form loss) is defined by the following relation
[IAEA2001]:
G2
pl  K l , (228)
2 
where Kl is the local form loss coefficient, which has different correlations for different
geometries, single-phase flows, and two-phase flows. The challenge in the above equation is to
define the local friction factor coefficient Kl properly. This topic is covered in the following sub-
sections.
7.5.3.1 Frictional pressure drop correlations in single-phase flow
For friction pressure loss correlations, circular pipe, annuli, rectangular channels, and rod
clusters are the most common applications. For local pressure loss correlations, spacers, top
and bottom tie plates, and locations of flow area changes such as constrictions, expansions,
bends, tees, and valves are the most common applications. For CANDU-type fuel bundles, the
degree of misalignment of two adjacent fuel bundles is also important in estimating the
pressure drop. In addition, in-core effects like radiation-induced creep, blister formation,
swelling, and corrosion are also important factors affecting pressure drop, but are not dealt
with here.
Figure 82 shows the Moody chart for friction factor as a function of Reynolds number and pipe
roughness [MOO1944, IDE1996]. This logarithmic diagram is the basis for calculating friction
factors in single-phase flow in pipes. On the left side is the area of laminar flow, represented
by a single straight line. Towards the right side, a group of lines is shown that indicate higher
friction factors for higher values of pipe roughness and slowly declining friction factors for
higher Reynolds numbers.
Figure 82 Moody chart for friction factors

For fully developed laminar flow, the friction factor is given by [IAEA2001]:
64
f  , (229)
Re
which is valid for Reynolds numbers less than 2000, as given by Eq. (12). This correlation is
shown by a straight line on the left side of Figure 82.
For turbulent flow in smooth pipes, several friction factor correlations have been proposed and
are used. A few commonly used correlations for smooth pipe are given below.
The Blasius equation [AECL2001] has the following form:
f  0.316  Re 0.25 , (230)
which is valid in the range of 3000 ≤ Re ≤ 105.
The following equation, which is valid in the range of 3000 ≤ Re ≤ 105, is also often used for
design [IAEA2001]:
f  0.184  Re 0.2 . (231)
Colebrook-White (1938) proposed the following equation, which is valid for both smooth and
rough pipes over the whole range of Reynolds numbers above 3000 [IAEA2001]:
1   / Dtube 2.51 
 2  log   . (232)
f tube  3.7 Re f rube 
 
This correlation is often used in computer codes, but it requires several iterations because it is
not explicit in terms of the friction factor f. Many other correlations attempt to capture
different effects such as different parameter ranges or heat transfer effects. These include
changes in near-wall velocity gradient due to fluid density changes, bubble formation, sharp
variations due to liquid-film thinning, and liquid-surface or vapour-surface contact. These all
create different wall heating effects that have a significant impact on the friction factor, even
before two-phase flow occurs.
7.5.3.2 Frictional pressure drop correlations in two-phase flow
Many two-phase flow pressure drop correlations can be found in the literature. These
correlations can be classified into the following four general categories: (1) empirical
correlations based on the homogeneous model; (2) empirical correlations based on the two-
phase friction multiplier concept; (3) direct empirical models; and (4) flow pattern-specific
models.
In addition, computer codes based on two-fluid or three-fluid models require correlations to
partition wall friction among the fluids and to determine interfacial friction correlations.
The most commonly used approach is the two-phase multiplier approach, which is described in
this section. In this case, the two-phase pressure drop is calculated by multiplying the single-
phase pressure drop by a two-phase frictional multiplier. Equation (233) exemplifies this
approach:
pf ,TP  L20pf ,L0 L20  L2 1  x 
2 b
or ∆‫݌‬௙,்௉ = ߶௅ଶ ∆‫݌‬௙,௅, , (233)
where pf ,L is the single-phase pressure drop based on only single-phase liquid in the channel
and p f ,L 0 is the single-phase pressure drop based on total flow as single-phase liquid in the
channel. Assuming f  a Reb , the relationship between  L20 and L2 can be expressed as
L20  L2 1  x 
2 b
.
The two-phase frictional multipliers L or  L 0 are empirical factors based on experimental

2 2
data and are expressed in the form of graphs or correlations that typically depend on quality
and pressure. A mass-flux effect is observed primarily at low flows (and hence is flow-regime-
dependent). Surface heating has a strong impact (a near-wall effect) on two-phase multipliers
in tubes and annuli, but not in bundles (due to compensating effects).
The correlation for a homogeneous two-phase frictional multiplier in its simplest form can be
expressed as:
b
f      g     
  TP   1  xe  
2
    . (234)
f  TP   
   TP 
L0
 g
The two-phase viscosity correlation can have various forms with different degrees of
complexity. The simplest correlation was introduced by Cicchitti et al. [IAEA2001]:
TP  xa  g  1 xa   . (235)
In general, the following definitions of two-phase frictional multipliers are often used in non-
homogeneous two-phase flow:
 dp / dz TPF ,  dp / dz TPF , (236)
 LO
2
 GO
2

 dp / dz  LO  dp / dz GO
 dp / dz TPF ,  dp / dz TPF , (237)
 L2   G2 
 dp / dz  L  dp / dz G
where the denominators refer to the single-phase pressure gradient for flow in the same duct,
with mass flow rates corresponding to the mixture flow rate in the case of LO
2
and GO , and
2
individual phases in the case of L and G . Among these, LO is the most popular frictional
2 2 2
multiplier for steam-water flow. Figure 83 shows a popular form of this two-phase friction
multiplier suggested by Nelson-Martinelli [COL1972]. The frictional multiplier is shown as a
function of mass quality and pressure. It rises sharply when two-phase flow is initiated (even
for very small mass qualities), and its value reaches 10 times the single-phase value even at low
quality. Moreover, the friction multipliers are higher at low pressure than at high pressure due
to the difference in volumetric and mass flow rates (as shown in Figure 22).
Figure 83 Martinelli-Nelson two-phase multiplier in separated flow

For diabatic two-phase flow (i.e., flow with heat transfer), the quality, void fraction, flow
pattern, and other properties change along the heated section. To calculate the pressure drop
in such cases, two approaches are usually followed. In the first approach, which is most
commonly used, the average LO
2
is calculated as:
L
1
 2
LO   LO
2
( z ) dz . (238)
L0
The approach can be used in cases where LO

2
( z) is a function that can be integrated.
Numerical integration is used in other cases.
In the second approach, the heated section is subdivided into a number of small segments.
Based on average conditions (i.e., xi, αi, and flow pattern) in that segment, the pressure drop is
calculated as in adiabatic two-phase flow using one of the models described previously.
7.5.3.3 Local pressure drop correlations
The single-phase local pressure losses are given in Eq. (228). The two-phase local pressure
losses are normally obtained using the two-phase multiplier approach. In this case, the two-
phase local pressure loss is given by the following relation [IAEA2001]:
Pl ,TP  l2,LO Pl ,SP . (239)
A simple form of the two-phase local loss multiplier can be obtained by the following relation:
   g 
l2,LO  1  xa  
   . (240)
 g 
Sudden contraction
Single-phase loss coefficients for sudden contraction can be calculated based on flow-area
ratio, according to the following equation [IDE1996]:
3/ 4
 A 
K contr  0.5  1  f  , (241)
 A0 
where Af and Ao are smaller and larger pipe cross-sectional areas respectively.
In general, the irreversible pressure two-phase pressure drop due to a flow-area change is
estimated from the knowledge of single-phase loss coefficient using an appropriate model.
Sudden expansion
In the case of sudden expansion, the following correlation is used, which is based on the ratio
of area changes [IDE1996]:
2
 A 
K exp  1  f  . (242)
 A0 
Fitzsimmons (1964) provides the following equation to calculate the pressure change across an
abrupt expansion:
G 2 Ar2  l 2  1 1    2 1 1  
p   x     1  x      , (243)
L   g  1 Ar  2     1  1  Ar 1   2   
where subscripts 1 and 2 refer respectively to the upstream and downstream locations of the
abrupt expansion and Ar = Af / A0.
Fittings and bends
The single-phase pressure drop due to bends and fittings can be calculated using the
appropriate loss coefficients from Idelchik [IDE1996]. For bends, the loss coefficient depends
on the bend angle, i.e., on the ratio of pipe diameter to bend radius. Because pressure drop is
dependent on piping component geometry, other specific configurations (i.e., end-fitting flow
characterization) must be experimentally tested to determine the specific pressure drop
correlations.
Orifices
The pressure loss through an orifice can be calculated using the following equation [IDE1996]:
W02 2
porf  forf LO . (244)
2
The loss coefficient in an orifice depends on the orifice geometry, particularly at the orifice
entry. An extensive description of loss coefficients with different orifices in single-phase flow is
provided by Idelchik [IDE1996]. For separated two-phase flow (stratified) at an orifice, Beattie
[IAEA2001] obtained the following expression for LO
2
:
0.8 0.2
      
2
 1  x  L  1 1  x  L G  1 . (245)
 G     L L 
LO

Figure 84 Single-phase pressure distribution Figure 85 Two-phase pressure distribution

over a square-edged orifice over a square-edged orifice
Figure 84 shows a diagram of a pressure loss in a square-edged orifice in a single-phase liquid
flow with 40% blocked area (a blunt-edged orifice). The pressure falls sharply over the orifice
because of the sudden increase in velocity (according to the Bernoulli equation, Eq. (43)). Due
to turbulence downstream of the orifice, unrecoverable losses occur because mechanical
energy is dissipated into thermal energy. Therefore, downstream of the orifice, where the pipe
diameter is the same as that upstream of the orifice, the pressure is lower than it would have
been without the orifice.
Figure 85 shows the two-phase pressure distribution over the same orifice as in Figure 84. The
pressure drop slope is much higher in two-phase flow than in single-phase flow, which signifies
an increase in frictional pressure drop. In addition, the unrecoverable losses are much higher,
and the pressure recovery occurs over a longer pipe section (pressure recovery is “smoother”).
Grid spacers and tie plates
Because of the variation and complexity of their geometry, it is extremely difficult to establish a
pressure loss coefficient correlation of general validity for grid spacers. However, calculation
methods that are reasonably accurate for design purposes can be achieved. To determine
pressure drop across spacers more precisely, experimental studies are required.
Generally, tie plates are used at the ends of rod cluster fuel elements and structurally join all
the fuel pins. Unlike spacers, the flow areas on the downstream and upstream sides of tie
plates are different. Moreover, they are generally located in the unheated portion of the
bundle. An approximate calculation for design purposes can be made using the contraction
and expansion model for local pressure losses. In addition, the friction losses over the
thickness of the tie plates can be calculated using the hydraulic diameter concept. For two-
phase pressure losses, the homogeneous or slip model described above can be used.
7.5.3.4 Pressure losses in CANDU fuel bundles
Several short bundles are stacked end-to-end in CANDU-type PHWRs instead of the long single
fuel bundle used in PWRs and BWRs. Due to this basic difference in design concept, some of
the issues and geometries are unique to the CANDU design.
Figure 86 Bundle correction factor (37-element hexagonal bundle)

The pressure losses in CANDU bundles are calculated based on the hydraulic-equivalent
diameter approach using a tube-based equation. Several correction factors are used, including
a correction for geometric effects (the differences between tubes and bundles); a correction
for eccentricity effects (differences between concentric and eccentric bundles) in crept
channels; a correction for channel shape effect (converging and diverging channels) in crept
channels; and a correction for surface heating effect.
Figure 87 Eccentricity effect

Figure 86 shows a sample of the fuel bundle effect in a hexagonal fuel bundle [LEU2004]. The
correction factor is given as a function of the pitch-to-rod-diameter ratio. It is evident from this
figure that the correction factor can be greater than or less than one.
Figure 88 Eccentricity correction factor

Figure 87 shows [LEU2004] on the left side a cross section of the fresh CANDU fuel bundle in an
uncrept channel, and on the right side the same bundle in a crept channel, in which the
pressure tube diameter has increased after a number of years of irradiation. It is evident that
the eccentricity of the irradiated fuel bundle has significantly increased. The effect of
eccentricity increase in an annular channel on the friction correction factor is shown in Figure
88. This figure shows that the correction factor for eccentricity decreases from one to a certain
value that depends on the degree of eccentricity and on the ratio of the inner and outer rod
diameters. Pressure tube creep and development of significant bundle eccentricity lead more
fluid to flow at the top of the bundle where the resistance to flow is lower, thus leaving the
bundle with reduced fluid flow. This phenomenon also results in development of a non-
uniform velocity distribution in the fuel channel.
The following empirical correlation was proposed by Snoek and Ahmad [IAEA2001] for a
friction factor based on experiments on a six-metre long electrically heated horizontal 37-rod
cluster (this correlation has not been validated or accepted for a CANDU fuel bundle):
f  0.05052  Re 0.05719 for 108,000≤Re≤418,000. (246)
The following empirical correlations were proposed by Venkat Raj [IAEA2001] based on a set of
experiments with prototype horizontal 37-rod clusters for PHWRs with a CANDU-type spacer.
They include the junction pressure drop (these correlations have not been validated or
accepted for a CANDU fuel bundle):
f  0.22  Re 0.163 for 10,000≤Re≤140,000, (247)
f  0.108  Re 0.108 for 140,000≤Re≤500,000.
However, in addition to the above friction factor, in a CANDU fuel bundle, a number of local
pressure losses must be taken into account. The total CANDU bundle pressure losses must
include: (1) friction; (2) local resistances from bundle junctions, spacers, buttons (43-element
bundle), and bearing pad planes; (3) acceleration; and (4) end fittings.
Various authors have attempted to combine the above effects into a single pressure loss
correlation. The friction and form losses have been combined into a bundle loss coefficient
given by the following relation [LEU2004, IAEA2001]:
G 2  f bundle Lbundle  G2
. (248)
 Psp ,bundle  K sp ,bundle   K junction  K appendage 
2    Dhy  2
 
The best approach for CANDU fuel bundles is to perform experimental measurements of the
pressure losses in a CANDU fuel channel experimental apparatus, i.e., in a so-called fuel bundle
or fuel channel flow characterization test. Once pressure losses have been precisely measured
for a fuel channel design, they can be tabulated or correlated in a convenient way for use by
designers and analysts.
Figure 89 Bundle junction misalignment effect

One of the important contributors to pressure losses in CANDU fuel bundles is fuel bundle
misalignment. Figure 89 shows two CANDU fuel channels that visually illustrate this
phenomenon [LEU2004]. The channel on the left has all fuel bundles aligned, so that each fuel
element is behind the previous fuel element upstream. This perfect alignment achieves the
lowest possible pressure loss from local resistance at the bundle end plates. However, in
practice, this perfect alignment is not possible because fuelling machines cannot ensure that
bundles are always inserted in a channel in the same position. In practice, as fuel bundles are
repeatedly inserted into the fuel channel, this misalignment develops a random distribution, as
shown on the right side of Figure 89. Pressure loss measurements are therefore performed
with various degrees of misalignment to cover all possible situations that can arise in practice.
Figure 90 Typical single-phase pressure distribution over misaligned bundles
Figure 90 shows a comparison of typical pressure drops in single-phase flow with three bundles
with a misalignment of 45o compared to fully aligned bundles. This figure clearly shows that at
each bundle junction, an abrupt local pressure loss can be measured. If the bundles are
misaligned to a greater extent, the abrupt pressure drop at the junction will be larger.
Figure 91 Typical pressure drop in misaligned fuel bundles
Figure 92 Typical misaligned junction pressure drop signatures

Figure 91 shows a CANDU fuel bundle flow characterization test in which actual pressure
measurements were collected along the fuel bundle [LEU2004]. It is evident that each change
of flow area along the fuel bundle, such as bundle end plates, spacers, buttons, and bearing
pads, creates a clear pressure loss of magnitude proportional to the change in flow area. It is
obvious from this figure that the pressure drop variation is symmetrical within a 45 degree
cycle.
Figure 92 shows a pressure drop signature from misaligned fuel bundles [LEU2004]. It is clear
that the degree of misalignment has a direct impact on the pressure drop measurements. The
effect is more pronounced in the case of two-phase flow, in which case the difference in the
measured pressure drops is much higher, both in absolute value and in variability with the level
of misalignment.
Figure 93 Typical two-phase pressure distribution over misaligned bundles

Figure 93 shows a typical pressure drop effect for misaligned fuel bundles in two-phase flow.
This figure should be compared to Figure 90, which shows the same test in single-phase flow.
Again, the absolute value of the pressure drop is higher in the case of two-phase flow, and the
pressure losses at the bundle end plates are also higher.
7.5.3.5 Use of pressure drop measurements to define the onset of significant void (OSV)
Because pressure drop measurements show such distinctive differences in pressure drop
between single-phase and two-phase cases, it has been noted that a sudden increase in
pressure drop at a certain location in a thermal-hydraulic network can be used as an indication
of the initiation of bulk boiling (i.e., the onset of significant void (OSV)).
Figure 94 shows pressure drop measurements for a crept fuel channel in which three different
power rates were applied, resulting in different channel boiling outcomes [LEU2004, POP2014].
The coolant inlet was on the left side and the outlet on the right side. The reference case (dark
blue line) is the case of sub-cooled conditions, with outlet thermodynamic quality negative and
nominal channel power. When channel power was increased to greater than 200%, the
pressure drop increased significantly downstream of the ninth fuel bundle because in this area
boiling was initiated, resulting in slightly positive channel outlet thermodynamic quality. When
the channel power was increased to three times nominal power, the outlet thermodynamic
quality further increased, and the pressure drop as well. Note that when boiling starts at a
certain location downstream (in this case, around the eighth fuel bundle), the pressure drop
starts to increase, which indirectly indicates the point where OSV occurs in this channel.
Hence, pressure drop measurements can reveal where boiling occurs in a fuel channel.
Figure 94 Pressure gradient along a flow channel in boiling

The Saha-Zuber correlation [REL2010, CAT1998, CAT2000, CAT2005] is used to calculate the
OSV point in tubes and has the following form:
q  D  C pf
xOSV  0.0022 for Peclet number (GDCp/k) < 70,000, (249)
k f  H fg
q
xOSV  154 for Peclet number ≥ 70,000. (250)
k f  H fg
A modified Saha-Zuber correlation for fuel bundles is based on empirical coefficients using full-
scale bundle data.
7.5.3.6 Pressure drop in pipes and orifices
The information covered in the previous sections can now be used to describe pressure drop
behaviour in a channel with various changes in geometry. Figure 95 shows such an example, in
which fluid comes from a pipe on the left side, flows through a pipe with a different diameter
and an orifice, and then into a much larger pipe on the right side. Hence, in this example,
major abrupt changes in flow cross-sectional area can be observed.
Figure 95 Example of abrupt area changes and orifice
Figure 96 Pressure drop behaviour with abrupt area changes and orifice
The pressure drop diagram for this case is shown in Figure 96. At the transition point between
the pipe on the left side (diameter D1) and the middle pipe (diameter D2), i.e., at the sudden
contraction point, the velocity increases to V2. As it accelerates, the fluid goes through a vena
contracta with eddy flow around the pipe corners. The recoverable pressure loss resulting
 
from the fluid acceleration is equal to  V22 V12 / 2 , and the unrecoverable local loss (sudden
contraction) resulting from the turbulence around the vena contracta is K e  V2 / 2 . The
2
velocity in the vena contracta is higher than the velocity appropriate for the pipe with
equivalent diameter D2. Therefore, the pressure in the vena contracta decreases further
around the area occupied by the vena contracta. Downstream of the vena contracta, the
pressure recovers where the influence of the vena contracta terminates, and the fluid velocity
assumes the value appropriate for a pipe with diameter D2.
Following the sudden contraction, the flow goes through an orifice with diameter Do, through
another vena contracta, and experiences further unrecoverable pressure loss.
On the other side, where the sudden expansion from the pipe into the larger pipe (diameter D3)
occurs, the reverse happens. The velocity slows down in the larger right-hand pipe from V2 to
V3. Note that velocity V3 is different from V1 because of the different pipe size. At the sudden
expansion, the unrecoverable local loss resulting from the turbulence around the vena
contracta is KeV3 / 2, and the recoverable deceleration pressure gain is  V32  V22 / 2 .
2
 
Note that the slope of the pressure drop curve in the three pipes is different because of the
different hydraulic diameters of the three pipes.
Having defined all the pressure loss terms in a piping configuration, it is interesting to compare
their absolute values and to obtain an impression of their relative contribution to the overall
pressure drop. This is done in Figure 97, in which typical pressure drop components are shown
for a fuel channel in a typical BWR reactor [TOD2011].
Figure 97 shows that for low mass fluxes, the relative contributions of the friction, acceleration,
and gravity terms are similar. However, with increased mass flux, the relative contribution of
the friction component significantly outweighs the other components. Note that the friction
component in the above figure includes both distributed and local losses.
Figure 97 Pressure drop components
7.5.4 Flow Instability
7.5.4.1 Fundamentals of flow instability
Flow instabilities are of particular importance for the reactor core. Under certain flow and heat
generation conditions, flow instabilities can cause significant changes in the fluid flow
distribution among certain channel groups in the core, thus exposing parts of the core to
insufficient cooling and consequent overheating, which can cause fuel damage. Moreover,
certain types of flow instabilities can cause pump malfunctions if the pump is not properly
selected. This section explains the nature of flow instabilities in the reactor core, provides
information on various types of instabilities, and focusses particularly on the Ledinegg
instability [TOD2011].
Understanding flow instabilities is instrumental to obtaining a good understanding of the
thermal-hydraulic system behaviour of nuclear reactor heat transport systems. A fluid flow in a
heated channel with a sub-cooled inlet condition undergoes large volume changes in a non-
uniform manner. Because the thermal and hydraulic properties of the flow change
continuously along the channel, the flow at any axial point in the channel can never become
fully developed thermally or hydrodynamically. Because the flow is not in equilibrium, the flow
properties fluctuate upstream and downstream of the point considered, often leading to
instability. Flow instabilities are of different types depending on the system configuration and
operating conditions. Flow oscillations can dramatically reduce the ability of the coolant to
remove heat from the reactor core. On the basis of primary features such as oscillation
periods, amplitudes, and relationships between pressure drop and flow rate, flow instabilities
have been classified into several types. Flow instabilities in a heated channel are usually
caused by two-phase flow and density wave effects. These instabilities are primarily classified
as static or dynamic. Some static and dynamic instabilities occur particularly during start-up
conditions.
The scale of instability can vary from macroscopic, involving the whole heat transport system,
to microscopic, occurring locally in some component or part of the thermal-hydraulic network.
In terms of the driving process, instabilities can be grouped into static or dynamic.
A “static instability” occurs when a small perturbation from the original steady-state flow leads
to a new stable operating condition that is not close to the original state. The mechanism and
the threshold conditions can be predicted using steady-state system characteristics. The
pressure drop characteristics, nucleation properties, and flow regime transitions of a flow
channel play an important part in characterizing static instabilities. Critical heat flux (CHF),
which limits the heat transfer capability of boiling systems, is influenced by static instabilities.
Ledinegg instability, flow pattern transition instability, geysering, chugging, and vapour burst
are also categorized as static instabilities.
Dynamic instability is caused by dynamic interaction among flow rate, pressure drop, and void
fraction. The mechanism involves the propagation of disturbances by pressure and void or
density waves. Density wave oscillations (DWOs), parallel channel instability, and pressure
drop oscillations (PDOs) are in this group.
Ledinegg instability or flow-excursive instability is characterized by a sudden change in the flow
rate to a lower value or a flow reversal. This instability will be further explained in this section.
Flow pattern transition instability is caused mainly by the different pressure drop
characteristics of various flow patterns. For example, if bubbly flow changes to annular flow
due to large void generation, the channel pressure drop decreases. For constant pressure drop
in a heated channel, when a bubbly flow changes to an annular flow pattern, the flow rate
increases, leading to decreased vapour generation. This makes the flow pattern revert to a
bubbly or slug flow pattern, and the cycle is repeated.
Geysering occurs at low power and low circulation flow rates. It occurs due to bubble
formation, growth, and subsequent collapse during start-up. In vertical channels during start-
up, with high inlet sub-cooling, voids are generated, and a large slug of bubbles is formed,
which grows due to the decrease in hydrostatic pressure head as it moves toward the exit. The
vapour then mixes with the liquid in the sub-cooled riser and becomes condensed there. The
condensed liquid re-enters the channel and restores non-boiling conditions. This process
repeats periodically and causes flow oscillations.
Vapour burst instability occurs due to sudden vaporization of the liquid phase with a rapid
decrease in mixture density. For example, a very clean and smooth heated surface may require
high wall superheat for nucleation. The fluid adjacent to the surface is highly superheated, and
vapour generation is rapid when nucleation starts. This in turn ejects liquid from the heated
channel. Rapid vaporization cools the surface, and the cooler liquid keeps vaporization
suppressed until the wall temperature reaches the required nucleation superheat, and the
process repeats. Vapour burst instabilities are observed during the reflood phase during re-
emergence core cooling of a reactor. Other CHF-related local instabilities are discussed in
Section 7.6.
7.5.4.2 Ledinegg flow instability
Ledinegg instability or flow-excursive instability is characterized by a sudden change in the flow

rate to a lower value or a flow reversal. This happens when the slope of the channel demand
pressure drop vs. flow rate curve (the internal characteristics of the channel) becomes
algebraically smaller than that of the loop supply pressure drop vs. flow rate curve (the external
characteristics of the channel, or the pump head curve). Physically, this behaviour exists when
the pressure drop decreases with increasing flow.
Figure 98 Concept of channel flow instability

Figure 98 shows conceptually the phenomenon of Ledinegg instability [TOD2011]. In this
figure, the y-axis represents the pressure drop and the x-axis the mass flow rate. The figure
represents a fuel channel in which the fluid mass flow rate and the heating from the walls are
selected so that part of the channel is in sub-cooled flow and part of the channel is boiling. The
right curve with a shallower slope represents the channel pressure losses for liquid flow. This
curve shows the pressure loss as the square of the mass flow rate (i.e., of the fluid velocity). If
the channel did not have heating, this curve would have ended at point O’ on the y-axis,
representing the hydrostatic head of the stagnant water in the channel.
The left curve shows the pressure drop in the channel when the flow is only gas. This curve is
steeper for a given mass flow rate because the difference between the densities of water and
gas is almost three orders of magnitude. Therefore, a given mass flow rate of gas results in a
much higher volumetric flow of gas compared to liquid, and therefore the velocity of the gas
and its associated pressure drop for a given volumetric flow are much greater than those of
liquid. The gas pressure drop curve ends almost at the zero point of the y-axis because the
hydrostatic pressure of gas is negligible compared to that of liquid.
Figure 99 Channel flow instability under low heat flux

With the above explanation, it is possible to analyze the case with heating in the channel and
the occurrence of two-phase flow, shown by the thicker solid line in Figure 98. When the mass
flow rate is reduced to a point when two-phase flow occurs (consistent with the OSV
explanation in Section 7.5.3.5), two phenomena occur:
 the volumetric flow starts to increase for the same mass flow, and therefore fluid
velocity is higher for the same mass flow rate; and
 the friction factor becomes higher in two-phase flow (see Section 7.5.3.2).
The net result is that as the mass flow rate decreases, more of the liquid is converted to steam,
resulting in higher volumetric flow and higher velocity, and hence the pressure drop increases.
This process continues until all liquid is converted to steam, at which point the solid line joins
the dashed line representing gas flow. With further reduction of flow rate beyond this point,
the pressure drop simply follows the quadratic law of pressure drop vs. velocity, and no more
boiling occurs.
Figure 99 shows a comparison of the system pressure drop curve with heat flux as shown in
Figure 98, but with a lower heat flux (the lower curve in Figure 99) [TOD2011]. When the heat
flux is lower, boiling starts at the lower mass flow rate rather than the higher, except that in
this figure, the flow rate is much smaller. This is shown in Figure 99 by having point B occur at
much lower mass flow rate. If the heat flux and hence the mass flow rate at boiling inception is
very low, then a couple of competing phenomena can create the situation shown by the
bottom curve in Figure 99. In this case, because boiling starts at a very low mass flow rate,
conversion of liquid into steam results in two phenomena: (1) increase of volumetric flow and
hence of liquid velocity, and (2) reduction of hydrostatic head.
Figure 100 Impact of different pump curves on channel flow instability

After boiling inception at point B at the lower curve, reduction of the hydrostatic head
temporarily dominates and creates a reduction in pressure drop. With further decrease in
mass flow rate, a large increase in velocity occurs as more of the liquid is transformed into
steam, resulting in a pressure drop increase.
Figure 100 shows the external pressure head generated by a pump (dashed lines)
superimposed on the system pressure drop curve (solid line) [TOD2011]. It is important to
observe in this figure that there are three cases of external pump head. The one on the top has
one intersection point 1, at which the pump can sustain a stable operating point with liquid
water. In addition, the curve at the bottom has one intersection point 3, at which a blower can
have a stable operating point with gas (or steam). The curve in the middle intersects the
system pressure drop curve at three points (2, 2’, and 2”), and all three are possible operating
points for this pump (or blower). At point 2’, the pressure head device must be a pump, and at
point 2”, it must be a blower. At operating point 2, neither a pump nor a blower can operate
well because the fluid is in two-phase flow, in which case conventional turbo machinery cannot
operate efficiently. Nevertheless, although this operation cannot be achieved by the same
device, it is assumed for the moment that it works, and this case and the consequences of this
“pump” operation will be further examined.
Figure 101 presents a view of flow excursion instability, continuing the explanation started in
Figure 100 [TOD2011]. As explained previously, the unstable operating point is point 1. In this
figure, various external pump head curves are plotted through point 1.
Case 1 (positive displacement pump with vertical pump head) represents a pump with a
delivery flow that does not change with pump head. This pump is inherently stable, and
although it rests on the part of the system curve with negative slope, it intersects the system
loss curve at only one point (point 1).
Figure 101 Definition of channel flow excursion instability in a single channel

Case 2 (constant head) represents a situation in which the head from an external pump is
delivered to a channel, and hence its curve intersects the system curve at three locations.
However, in this case, the pump head is constant irrespective of the flow rate in the channel.
This situation can lead to excursion instability because the curve intersects the system loss
curve at three points (1, 2, and 3); this will be explained further.
Case 3 (low variable head) represents a centrifugal pump for which the head and flow are
related through a polynomial function. This pump head curve also intersects the system loss
curve at three points and also leads to Ledinegg instability.
Case 4 (high variable head) represents a pump that is well throttled, so that a small flow change
results in a large pump head change. This is a typical situation in operating BWRs, which have
well-throttled external circuit pumps. Because the slope of the curve is very steep, the pump
head curve intersects the system loss curve at only one point (point 1), and therefore this
pump operates in an inherently stable manner regardless of where it intersects the system
curve.
From the above examples, a criterion for entering Ledinegg instability can be defined by the
following inequality:
  Psystem    Pext 
 . (251)
w w
This inequality shows that if the pump head curve is Δpext>Δpsystem, the pump operation is
inherently stable in such a system.
By examining Figure 101 carefully, it is clear that the source of this instability is the negative
slope of the system loss curve. This means that if the pump operates at point 1, any small
disturbance of the pump flow or head will lead to instability.
For example, at point 1, if pump flow decreases by an infinitesimal amount δW, the pressure
drop in the system will increase by δ(Δp) (the left upper green triangle). This will make the
pump further reduce its flow delivery, which in turn will make the pressure drop further
increase. Therefore, step by step, the pump operating point will move through points 2’ and 2”
to point 2.
Inversely, if at point 1, the pump flow increases by an infinitesimal amount δW, the pressure
drop in the system will decrease by δ(Δp) (the right lower green triangle). This will make the
pump further increase its flow delivery, which in turn will make the system pressure drop
further decrease. Therefore, step by step, the pump operating point will move through points
3’ and 3” to point 3.
Figure 101 clearly shows that while the pump operating point moves from one point to
another, the pump head and the system losses will be in equilibrium for a period of time that
depends on the system and the pump hydraulic inertia. However, given enough time, the
possible operating points must be at the intersection of the two curves. However, the only
sustainable stable operating point must be the intersection point of the two curves that
satisfies Eq. (251).
Thermal-hydraulic designers need to examine pump curves and system loss curves carefully to
ensure that pump head shape and slope satisfy Eq. (251), thereby ensuring that this type of
instability will not occur.
7.5.4.3 Parallel channel flow instability
One aspect of Ledinegg instability (i.e., excursive instability) is the situation that can arise with
parallel channels operating under a common pressure drop [POP2014]. This situation can
occur in many thermal-hydraulic networks or components. For example, this may occur in a
reactor core with two types of channels under the same pressure drop, or in a CANDU fuel
channel with crept pressure tubes, in which the flow is split into two flow paths, one in the
eccentricity above the fuel bundle, and one through the fuel bundle (see Section 7.5.3.4)
Figure 102 Principles of flow excursion instability in two parallel channels

Figure 102 shows the principles of a two-parallel-channel configuration under the same
pressure drop, which is vulnerable to excursive instability. In this configuration, each channel
has its own pressure loss curve, but different heat flux, total power, and flow rate. Therefore,
each channel will enter the two-phase flow boiling condition at a different flow rate. The first
channel that enters two-phase flow will experience an increase in channel pressure drop, which
will help divert a portion of its flow to the other channel that remains in single-phase liquid
flow. In other words, the channel that first enters two-phase flow will experience successive
flow reductions and will eventually overheat as it loses its cooling.
Figure 103 shows an actual calculation of flow instability in a parallel channel configuration in a
core with two types of fuel channels using a computer code (a similar geometry to that in
Figure 102) [POP2000]. In Figure 103, the mass flow rate is on the x-axis and the pressure drop
on the y-axis. Line 1 represents the total pressure drop across both channels; Lines 2 and 3
represent the pressure drop in each channel, and Line 4 provides the percentage of flow split.
It is apparent in this figure that the flow split between the channels changes as the total flow to
the two-channel loop decreases. Clearly, the two channels behave differently below a total
flow of 0.2 kg/s. At a pressure drop of about 12 kPa, the flow starts to divert from one channel
(Line 3) to the other channel (Line 2). Line 3 indicates an increase in pressure drop due to
boiling, whereas Line 2 indicates an increase in pressure drop due to higher liquid flow.
Figure 103 Results of flow excursion instability in two parallel channels
7.5.5 Problems
1. List and explain the most important factors that have significant impact on the pressure
drop in the CANDU primary heat transport system. Rank the components in the
CANDU primary heat transport system that contribute to the pressure losses in the
system. Explain how reactor aging affects pressure drop in the main heat transport
system.
2. Explain the components of pressure drop in horizontal channels, and provide information
on the parameters that influence their value. Explain the approach used to calculate the
pressure drop components in two-phase flow.
3. A mixture of water and steam at mass quality of 30% and temperature of 400oC flows
through a 10 m long pipe with diameter of 4 inches. The mass flux in the pipe is 2000
kg/m2 s. Calculate the pressure drop along this pipe using the most appropriate two-
phase flow multiplier.
4. Water flows through an insulated horizontal pipeline 100 m long (no heat losses). Water
enters the pipe at temperature of 250oC, and velocity of 3 m/s. Assuming that the pipe is
made of carbon steel of commercial quality, calculate the pump power that is required to
maintain this flow.
5. Two-phase water-steam mixture at 10% quality flows in thermal equilibrium through a
horizontal pipe 6 m long with diameter of 18 mm. The water pressure is 8 MPa.
Calculate the two-phase pressure gradient in this pipe, and the total pressure difference
between the inlet and the outlet using the Friedel correlation provided below. The total
mass flux is 2000 kg/m2 s.
3.24 FH   f go
2o  E  We0.035 E  (1  x ) 2  x 2 F  x 0.78 (1  x)0.224
Fr 0.0454
 g f o
0.91
   g   g 
0.19 0.7
Gm2 Gm2
H      1   Fr  We 
 g gDe  m2  m
       
 x 1 x 
m     for homogeneous model
 g 
  
64
f jo  if Re j  1055
Re j
  Re j 
f jo   0.86859  n  for Re j  1055
  1.964  n Re j  3.8215  
 
Where: x is quality, ρ is density in (kg/m3) (g – gas, and l – liquid), fgo and flo are Darcy
friction factors, μ is dynamic viscosity in (Pa s), (g – gas, and l – liquid), σ is surface
tension in (N/m), and g is acceleration due to gravity (=9.81 m/s2).
6. Explain the cause and the impact of the Ledinegg flow instability in vertical channels.
Explain the importance of this type instability for the process of pump selection.
7. Explain the cause and the impact of flow excursion instability in parallel heated channels,
including the applicability and methods for avoiding or mitigating this type of instability.
7.6 Heat transfer between the fuel elements and the coolant
Section 6.1 provides the fundamentals of two-phase flow, which are needed to understand the
heat transfer regimes in the reactor core. Section 7 provides the basis for heat transfer in the
reactor primary heat transport system. Section 7.4 provides details on heat conduction in the
fuel elements. This section now focuses on heat transfer from the fuel to the coolant.
Heat removal from the reactor core is an important issue in reactor design and operation. The
main objective of heat transfer analysis for single-phase flow in the reactor is to determine the
temperature field in a coolant channel such that the reactor operating temperatures are within
the specified limits, including the rate of heat transfer to and from a surface or object. Because
reactor power densities are typically much higher than for other conventional heat sources, the
heat removal rate from any given reactor core coolant channel is quite large. Heat transfer to
the coolant in single phase therefore requires a coolant with a large heat capacity.
The boiling curve shown in Figure 29 provides a good illustration of the boiling phenomenon
and the boiling crisis in the fuel channel, which have an important impact on reactor thermal
margins. This section discusses the wall heat flux in various heat transfer regimes and the
critical heat flux (CHF) in the channel and discusses their importance and the differences
between PWR and CANDU reactors. Appropriate diagrams explaining the variation of the
critical heat flux with key parameters are shown and discussed. In addition, CHF correlations
and calculation methods are described. The CHF is a critical parameter in reactor core thermal-
hydraulic design because it sets the primary limitations on reactor thermal margins.
7.6.1 Heat transfer regimes

Various flow and heat transfer regimes that are relevant to heat transfer in the reactor core
under various operating modes are illustrated in Figure 32. The heat transfer correlations are
important in reactor thermal-hydraulic design because they provide the tools for a designer to
calculate various parameters of interest in the design process and to verify that the design can
meet the proposed objectives and requirements, primarily the capability to remove generated
heat from the core during all possible operating modes.
The heat transfer from any solid surface to a coolant is given by Newton’s Law in terms of heat
flux:
q"  h (Ts Tb ) , (252)

where Ts is the surface temperature, Tb is the bulk temperature of the coolant, and h (W/m2 K)
is the heat transfer coefficient. The heat transfer coefficient is dependent on coolant
properties, flow parameters such as the velocity field, the temperature fields in the solid
surface and the coolant, and solid surface condition and geometry.
A considerable amount of experimental and theoretical research has been published on the
heat transfer coefficient for turbulent flow in channels. Although most correlations have been
developed based on circular pipes, with appropriate corrective factors, they can be applied to
various geometries.
For fully developed turbulent flow of non-metallic fluid, the general heat transfer correlation
can be expressed as [ELW1978]:
d
 
Nu  C  Re Pr  w  ,
a b
(253)
  
where µw is the fluid viscosity at the wall temperature and a, b, c, d are constants that depend
on fluid and channel geometry.
Heat transport with phase change such as in boiling or condensation is an efficient method to
transfer heat because latent heat per unit mass is very large compared with sensible heat. For
single-component fluids, the interface temperature difference involved for heat transfer in
evaporation and condensation is relatively small. However, when more than one component is
present in a system, the temperature difference can be higher. An example is condensation of
vapours in the presence of non-condensable gases. The two-phase heat transfer phenomena
relevant to reactors include pool boiling, evaporation in vertical or horizontal fuel channels, and
condensation inside or outside tubes.
As indicated in Figure 32, many flow and heat transfer regimes are possible in reactor channels
in two-phase flow, depending on many parameters. In addition, fuel and channel geometry,
heat transport conditions, characteristics, and materials have a significant influence on the type
and magnitude of heat transfer from fuel to coolant. As a result, in the literature, many heat
transfer correlations have been proposed, adopted, and used by various designers and
analysts. This section provides a brief overview of the most important correlations, focussing
on the critical heat flux correlations because of their significance for reactor thermal margins.
7.6.2 Convective heat transfer in turbulent forced flow

The convective heat transfer coefficient in turbulent forced flow of non-metallic fluids follows
the form given by Eq. (253). The Prandtl number, given by Eq. (16), for non-metallic fluids such
as light and heavy water is close to unity. For non-metallic fluids, the Nusselt number given by
Eq. (15) is insensitive to wall surface conditions. Moreover, the Nusselt number varies little in
the case of constant heat flux and constant mass flux.
Even for this relatively well-known situation, many heat transfer correlation variations have
been proposed by different authors. The best-known is the Dittus-Boelter correlation, which is
given by the following relation [ELW1978]:
Nu  0.023  Re 0.8 Pr 0.4 . (254)
Hence, using the Nusselt number, the heat transfer correlation can be obtained as follows
(using a simplified version for plate geometry):
hc  0.023  k f L  Re0.8 Pr 0.4 . (255)
This is the best-known correlation used for single-phase convective heat transfer, as shown in
the first part of the left side of Figure 29. In this correlation, the physical properties in
dimensionless numbers are evaluated at the bulk temperature of the fluid. The correlation
does not take into account the effect of temperature variation from the tube wall to the bulk
fluid. Clearly, for laminar flow, this correlation does not apply, and other correlations are used.
For forced convection in fuel bundles, the following correlation was proposed by Weisman
[ELW1978]:
N u  C  R e 0 .8 P r 1/ 3 , (256)
where the constant C has the following form for square and triangular lattices respectively (S is
the fuel rod pitch, and D is the fuel rod diameter):
S S
C  0.042  0.024, 1.1   1.3
D D
. (257)
S S
C  0.026  0.006, 1.1   1.5,
D D
7.6.3 Sub-cooled boiling heat transfer in forced flow

Heat transfer from the walls to the liquid at relatively high heat flux, starting from the sub-
cooled condition at the entrance of a channel, brings CANDU fuel cladding walls to and beyond
saturation somewhere downstream in the reactor fuel channel. When the liquid layer next to
the wall and adjacent to the wall becomes superheated, bubble nucleation sites are activated.
This point is denoted in Figure 29 as the onset of sub-cooled boiling. Bubbles are generated at
the wall cavities by the heat transferred from the wall as latent heat of evaporation. As soon as
they depart from the wall, they encounter sub-cooled bulk liquid and condense, releasing the
latent heat of vaporization to the bulk liquid. This process helps enhance heat transfer from
the wall, partly by evaporation of the liquid and partly by enhancing turbulence near the wall.
Note that heat transfer by latent heat of evaporation is much more effective than heat transfer
by convective heating of liquid.
The point at which sub-cooled nucleate boiling starts is defined by the following criterion
proposed by Bergles and Rohsenow [AHM2009]:
2.3/ p 0.0234
qW'' , SB  1082 p1.156 1.799 Tw  Tsat SB  . (258)
 
Equation (258) is valid for pressures between 0.1 and 13.6 MPa, and p is in bar, temperatures in
°C, and qW,SB in W/m2.
For the nucleate boiling heat transfer regime, the best-known and most widely used correlation
is the Chen correlation [CAT2005, ELW1978, TOD2011]. This correlation calculates the heat
transfer coefficient as a sum of the convective and nucleate boiling parts as follows:
hSB  hC  hNB . (259)
The convective heat transfer is calculated using a variation of the Dittus-Boelter correlation
given by Eq. (255):
 G 1  x) De  
0.8
kf
hc  0.023   Pr 0.4 F, (260)
 f De
 
1
F  1.0 for  0.1
X tt
0.736 . (261)
 1  1
F  2.35  0.213   for  0.1
 X tt  X tt
The Lockhart-Martinelli parameter Xtt is defined as:
0.5
 l   g 
0.9 0.1
1  x 
      . (262)
X tt  (1  x)   g   l 
The nucleate boiling part of the sub-cooled boiling correlation is defined as:
S  0.00122    T 0.24  p 0.75 c 0.45
p l
0.49 0.79
kl
hNB  . (263)
 0.5 hlg0.24  l0.29  g0.24
The parameter S is defined as follows:

1
S where Re  Rel F 1.25 . (264)
1  2.53 10 6 Re1.17
The Chen correlation is valid for pressures of 0.17–3.5 MPa, liquid inlet velocities of 0.06–4.5
m/s, and heat fluxes up to 2.4 MW/m2.
When the bulk liquid temperature increases to a certain point, as shown in Figure 29,
conditions for onset of significant void, i.e., bulk boiling, are established. The criteria that
determine the onset of bulk boiling are defined by various correlations, of which the best-
known is the Saha-Zuber correlation provided in Eqs. (249) and (250), which is expressed in
terms of quality at the OSV point. However, it is useful to express Eqs. (249) and (250) in terms
of bulk fluid temperature to estimate the level of coolant sub-cooling at the point of OSV.
Figure 32 shows the sub-cooled boiling phases, with OSV labelled as the onset of net vapour
generation, which is upstream of the point at which bulk saturation is reached.
Equations (249) and (250) can be transformed into a relation that defines the wall temperature
at the onset of bulk boiling:
q " De
Tbulk ,OSV  Tsat  0.0022 for Pe  7 104
kl
, (265)
q"
Tbulk ,OSV  Tsat  154 for Pe  7 104
G  cp
where the Peclet number is defined according to the following relation:

G  De c p
Pe  . (266)
kl
The Saha-Zuber correlation is valid for pressures of 0.1–13.8 MPa, mass fluxes of 95–2760
kg/m2s, and heat fluxes of 0.28–1.89 MW/m2.
7.6.4 Critical heat flux (CHF)

When the wall surface temperature is further increased beyond the point that results in sub-
cooled boiling, as shown in Figure 29, it may reach a value that results in a heat flux referred to
as critical heat flux (CHF). This is the maximum heat flux at which continuous liquid contact in a
tube or fuel channel is maintained. Further increase in the wall heat flux will result in an
increase in wall temperature to values that result in a sustained vapour film on the wall surface
(along the horizontal dashed line in Figure 29). Usually, for reactors designed to operate with
liquid water in contact with the fuel cladding, this large increase in cladding temperature could
result in melting and destruction of the cladding. The primary objective of thermal-hydraulic
design is to ensure that this does not happen.
7.6.4.1 Outline of CHF phenomena
The CHF is a complex thermal-hydraulic phenomenon that varies with a large number of
parameters. This section explains the CHF phenomenon for various combinations of key
parameters and provides a discussion of CHF parametric trends. In addition, heat transfer
correlations for pre-CHF heat transfer regimes are provided and discussed [IAEA2001,
TOD2011, BER1981, DEL1981, COL1972, LWU2004].
Development of the CHF correlation requires a significant amount of experimental testing
because of the complex reactor core fuel geometry in PWRs as well as in CANDUs. This section
also provides a brief explanation of the experimental tests and their use to develop CHF
correlations.
In forced convective boiling, the boiling crisis occurs when the heat flux is raised to such a high
level that the heated surface can no longer support continuous liquid contact. This heat flux is
referred to as the critical heat flux (CHF). It is characterized either by a sudden rise in surface
temperature caused by blanketing of the heated surface by a stable vapour layer, or by small
surface temperature spikes corresponding to the appearance and disappearance of dry
patches. The CHF normally limits the amount of power transferred, both in nuclear fuel
bundles and in conventional boilers. Failure of the heated surface may occur once the CHF is
exceeded. This is especially true for highly sub-cooled CHF conditions. At high flows and
positive dryout qualities, the post-dryout heat transfer is reasonably effective in keeping the
heated surface temperatures at moderate levels, and operation under dryout conditions may
be sustained safely for some time, particularly in BWRs or CANDUs.
Figure 104 Concepts of CHF, dryout, and burnout

Figure 104 [COL1972] is of key importance in understanding the various mechanisms and the
definition of the CHF phenomenon, as well as the consequences of CHF occurrence in a reactor
fuel channel (note that this figure is not to scale, values are not shown, and the intent is to
illustrate CHF conceptually for various coolant qualities).
Figure 104 shows the basic concept of CHF occurrence and the terms used to describe the
various stages and types of the CHF phenomenon. On the y-axis, this figure shows the relative
value of the heat flux as given, whereas on the x-axis, the thermodynamic quality is provided,
starting from sub-cooled conditions (negative values), moving to values between 0 and 1 (two-
phase flow), and then to values beyond 1 (superheated steam). The straight solid line with a
negative slope intersecting the x-axis at xe=1 is the locus of CHF. Above this straight line, a line
is shown which is labelled as the burnout locus—this is the wall heat flux that can destroy the
cladding surface under various conditions. On the left side, under sub-cooled conditions,
various heat transfer regimes are illustrated moving from bottom to top in the figure:
convective heat transfer at the bottom (no boiling), sub-cooled nucleate boiling, and sub-
cooled film boiling. In the two-phase region in the middle of the diagram, various heat transfer
regimes are shown: two-phase forced convection, saturated bulk boiling, saturated film boiling,
and liquid-deficient saturated boiling. On the right side of the diagram, convective heat
transfer to steam is shown. These regimes are also illustrated in Figure 26 and Figure 33.
With respect to the diagram in Figure 104, the following terms can be defined. They are related
to the CHF, but have a slightly different meaning:
 Boiling crisis. Refers to sub-cooled or saturated boiling when the amount of steam
produced by boiling surpasses the amount of steam that can be removed from the wall,
thus creating a situation in which heat removal from the wall is limited.
 CHF. Refers to the general condition of limited heat flux in sub-cooled or saturated
boiling; it can include all terms shown on this list.

 Dryout. Refers to conditions at the fuel cladding surface when continuing contact with
liquid water can no longer be maintained. Note that under dryout conditions, there may
still be abundant liquid water in the pipe or fuel channel (so that the void fraction may be
very low), but the cladding surface may still be insulated from the water by a stable
steam film.
 Departure from nucleate boiling (DNB). This term refers to the CHF that occurs under
significantly sub-cooled conditions (as in PWRs). In this case, CHF happens when the fuel
cladding surface can no longer support sub-cooled boiling.
 Burnout. This term refers to the destruction of the cladding surface as it begins to melt.
This can happen very quickly under sub-cooled conditions after CHF is reached. At
saturated conditions (for higher equilibrium quality or void fraction), as shown in Figure
104, burnout can occur only when the heat flux is greater than the CHF (note that the
burnout curve moves further away from the CHF curve in Figure 104 as quality rises,
particularly for qualities x > 0).
Figure 105 [POP2014] shows a different CHF mechanism in a reactor fuel channel
superimposed on the CHF diagram from Figure 104. In this figure, heat flux is indicated on
the y-axis and thermo-dynamic quality on the x-axis. There are four distinctive CHF
mechanism types: nucleation induced under highly sub-cooled conditions, a bubbly layer
induced under sub-cooled conditions, entrainment-induced film depletion at low quality, and
deposition-controlled film depletion at very high quality.
Figure 105 Transitions among CHF mechanisms

Like other thermal-hydraulic phenomena, CHF occurs differently in pool boiling than in
forced-flow boiling. In pool boiling, no forced flow is provided to the liquid container. In this
situation, CHF usually occurs because of either counter-current flow (Helmholtz) instability or
micro-layer evaporation (for highly sub-cooled conditions only).
Figure 106 Pool boiling CHF mechanisms

The left side of Figure 106 shows the counter-current flow instability (i.e., the Helmholtz
instability) [LEU2004]. In this case, the heated surface is covered by a rising vapour column
with counter-current liquid jets flowing downwards due to gravity to compensate for liquid
removal by evaporation. Ultimately, at very high heat flux levels (vapour removal rates), the
relative velocity between liquid and vapour can be so high that an unstable flow situation is
created in which due to interfacial friction, liquid jets cannot penetrate downwards because of
the high flow rate of rising bubbles, resulting in a CHF condition. The right side of Figure 106
shows a micro-bubble developed in a cavity on the heated surface. As the bubble grows, a
microlayer of thin liquid film develops below the bubble. This microlayer is unstable because of
the large heat flux and can break and leave the surface uncovered, i.e., can form a dry patch,
leading to CHF.
Figure 107 Flow boiling CHF mechanisms

In forced-flow boiling, CHF usually occurs because of one of the following three phenomena:
collapse of the liquid sub-layer, bubble crowding, or film depletion. Figure 107 shows these
three phenomena [LEU2004]. On the left side of this figure, the collapse of the liquid sub-layer
occurs with highly sub-cooled liquid. CHF occurrence is due to the spreading of a dry patch
following microlayer evaporation under a bubble and coalescence of adjacent bubbles. The
occurrence of CHF in this case depends on local surface heat flux and flow conditions; it is not
affected by upstream heat flux distribution.
Bubble crowding is shown in the middle of Figure 107 for moderate sub-cooling or saturated
conditions and high flow rate. The bubble population density near the heated surface
increases with increasing heat flux, and a so-called bubble boundary layer often forms a short
distance away from the surface. If this layer is sufficiently thick, it can impede the flow of
coolant to the heated surface. This in turn leads to a further increase in the bubble population
until the wall becomes so hot that a vapour patch forms over the heated surface. This type of
boiling crisis is also characterized by a fast rise in heated surface temperature (fast dryout).
At the right side of Figure 107, the film depletion type of CHF is illustrated, which is possible
with a relatively low mass flux of saturated liquid in high void fraction flow. In the annular
dispersed flow regime (high void fraction and high volumetric flow), the liquid will be in the
form of a liquid film covering the walls and entrained droplets moving at a higher velocity in the
pipe core. Continuous thinning of the liquid film will take place due to the combined effect of
entrainment and evaporation. Near the dryout location, the liquid film becomes very thin, and
due to the absence of roll waves (which normally occur at higher liquid film flow rates),
entrainment is suppressed. If the net droplet deposition rate does not balance the evaporation
rate, the liquid film must break down. The temperature rise accompanying this film breakdown
is usually moderate (stable dryout).
Related to the above explanation of dryout resulting from a high void fraction, high flow rate
situation in a tube is CHF dryout in a fuel channel. This situation is typical of CANDU reactors in
which towards the end of the fuel channels, saturated liquid with significant void fraction and
mass flux flows through a densely packed fuel channel. Because the two-phase flow is very
well mixed, dryout occurs gradually, and fuel is not exposed to sustained loss of contact with
liquid, i.e., slow dryout occurs. During a slow dryout, the heated surface does not experience
the usual dryout temperature excursions; instead, a gradual increase in surface temperature
with power is observed. A slow dryout is usually encountered in flow regimes where the
phases are distributed homogeneously, such as froth flow or highly dispersed annular flow at
high mass velocities and void fractions. Under these conditions, liquid-wall interaction is
significant, limiting the temperature rise with dryout.
Calculations based on cooling by vapour flow indicate only that post-CHF temperatures are
below the minimum film boiling (Leidenfrost) temperature; hence, depositing droplets may wet
the surface, increasing the heat transfer coefficient. In a CANDU reactor, because the coolant
is close to saturation over most of the channel and at relatively high mass flux and void fraction
downstream of the rector mid-plane, the most probable CHF mechanism is slow dryout.
Therefore, in a CANDU reactor, the consequences of reaching CHF are not as serious as in a
highly sub-cooled PWR because the fuel surface can still be well cooled by passing steam and
droplets.
7.6.4.2 CHF parametric trends
As explained above, CHF is a complex phenomenon that depends on many parameters and can
be caused by different phenomena. Table 10 shows various parameters that affect the type
and intensity of CHF in fuel bundles. Discussion of all these effects is beyond the scope of this
textbook; details can be found in the literature [IAEA2001].
Table 10 Separate CHF effects in fuel bundles
GENERAL DETAILS OF SEPARATE EFFECTS
Global Flow Area Effects  n-rod bundle where n>>3 and all sub-channels are identical except for
corners or cold-wall-adjacent sub-channels (e.g., square or triangular arrays
of sub-channels)
 n-rods where n>>3 and adjacent sub-channels are generally not equal (e.g.,
37-rod bundle geometries inside round tubes)
Sub-Channel Effects  Sub-channel size/shape (similarity to tube)
 Cold wall effect
 Distorted sub-channels (due to bowing, cladding stain, PT creep)
 Misaligned bundles (CANDU case)
Length Effects  Similar to appendage effects
Spacer/Bundle  Mixing grids
Appendage Effects  Attached spacers/ bearing pads/ endplates (CANDU)
Flow Orientation Effects  Vertically upward
 Vertically downward
 Horizontal
Axial/Radial Flux  Axial flux distribution (flux peaking/global flux distribution)
Distribution Effects  Radial flux distribution (global RFD effect, cold wall effect, flux tilt across an
element)
Flow Parameter Effects  Mass flow (including zero flow or boiling/flow stagnation case)
Transient Effects  Power/flow/pressure transients
 Combined transients
Effect of Fluid Type  Light water
 Heavy water
 Modelling fluids (Freon) in conjunction with a CHF fluid-to-fluid modelling
technique
A number of studies have been conducted by various researchers to investigate the effects of
various parameters on CHF. An overview of the parametric trends is provided in this section.
These parametric trends are provided for two different CHF approaches.
The approach based on fuel channel inlet conditions is very often used in CHF R&D because of
the convenience of using inlet conditions because these are easily measured and maintained in
CHF experiments. The other method is based on local conditions, i.e., the conditions locally
where CHF occurs. This approach is used in computer-based algorithms because a computer
program can easily calculate local conditions at the CHF location.
The discussion of parametric trends provided here must be related to the general trends
described in Section 6.1, particularly those illustrated in Figure 30 and Figure 31.
The most influential parameters affecting CHF are the fluid pressure, mass flux, level of sub-
cooling, and hydraulic diameter [IAEA2001]:
CHF = f (P,G, X DO Dhy ) . (267)
The general parametric trends on CHF based on inlet conditions are as follows:
1. Heated length has a very high impact on CHF because it affects the amount of heat
transferred to the fluid; hence, CHF strongly decreases with increase of heated length
and quality.
2. CHF increases with an increase in inlet sub-cooling (a decrease in inlet fluid
temperatures).
3. CHF decreases with a decrease in outlet pressure (decrease of channel pressure).
4. For sub-cooled conditions, CHF increases with an increase in mass flow.
5. For saturated conditions, CHF increases with a decrease in mass flow.
6. CHF decreases with a decrease in tube diameter (which decreases the level of mixing
and turbulence).
The general parametric trends related to CHF and based on local conditions are as follows:
1. The effect of heated length on local CHF is negligible.
2. Local CHF decreases with an increase in thermodynamic quality.
3. For sub-cooled conditions, local CHF increases with an increase in mass flow.
4. For saturated conditions, local CHF increases with a decrease in mass flow.
5. Local CHF increases with an increase in outlet pressure.
6. Local CHF decreases with an increase in channel diameter.
Figure 108 Effect of inlet temperature and Figure 109 Effect of outlet pressure in pool
mass flux boiling
Figure 108 shows the relationship between CHF (y-axis) and inlet sub-cooling (x-axis in enthalpy
difference below saturation) for several mass flux values in a straight heated tube [LEU2004].
This figure is drawn in terms of channel inlet conditions. Note that CHF increases with mass
flux and with the degree of sub-cooling. In other words, reactors that have more highly sub-
cooled coolant and higher mass flux can use higher heat flux before reaching CHF. This is an
important fact that demonstrates the strategy of operating PWRs.
In general, pressure has two key effects on heat transfer and CHF, one resulting from a change
in the saturation temperature, and the other from a change in the latent heat of vaporization.
As pressure increases, the saturation temperature increases, but the latent heat of vaporization
decreases. As a result, CHF increases with pressure, as indicated in Figure 109 [TOD2011].
However, as the pressure approaches the critical pressure, the latent heat of vaporization
decreases to zero, reducing the CHF.
Figure 110 shows the effect of coolant outlet pressure (x-axis) for several sub-cooling levels for
forced flow in a heated tube [LEU2004]. This figure is also drawn for a given level of channel
inlet sub-cooling. Again, this figure confirms that higher inlet sub-cooling results in higher CHF.
The impact of outlet pressure must be analyzed in combination with the channel inlet pressure,
quality, and volumetric flux distribution along the channel (the channel length is 3.658 m).
Figure 111, Figure 112, and Figure 113 show CHF parametric trends for local conditions in
heated pipes [LEU2004].
The effect of heated length is shown in Figure 111. It is evident that CHF decreases with
increasing outlet quality in the channel. Moreover, as the heated length of the channel
increases, the CHF value decreases significantly. Hence, for a very short heated length, CHF can
be five times larger than for long heated lengths. This is another important observation in
terms of CANDU reactor design. Because in CANDU the fuel bundles are relatively short, the
bundle endplates and other bundle appendages can significantly improve the CHF value
(further explanations in other sections below).
Figure 110 Effect of outlet pressure on Figure 111 Effect of heated length on local
flow boiling CHF
Figure 112 shows the effect of mass flux under local conditions [LEU2004]. Similar behaviour
can be observed with regard to the impact of sub-cooling on heat flux (for negative
thermodynamic qualities). For saturated conditions, the effect of mass flux is not as important
and should be analyzed in combination with volumetric flux.
Figure 112 Effect of quality on local CHF Figure 113 Effect of tube diameter on
local CHF
Figure 113 shows, for local conditions, the impact of pipe diameter on CHF for various outlet
thermodynamic qualities [LEU2004]. It is evident that larger hydraulic diameters provide a
higher value of CHF for a given quality. Moreover, this figure confirms that for a given pipe
diameter, CHF decreases with increasing thermodynamic quality.
7.6.4.3 CHF prediction methods

Because of the many possible fuel bundle geometries, the wide range of possible flow
conditions, and the various flux distributions for advanced water-cooled reactors, it is
impossible to predict the CHF for all cases using a single CHF prediction method and achieve a
reasonable degree of accuracy. Over 400 correlations for CHF in tubes are currently in
existence [IAEA2001]. The present proliferation of correlations illustrates the complex state of
the art in predicting the CHF phenomenon, even for a simple geometry at steady-state flow
conditions. The complexity of predicting the CHF increases significantly for fuel bundle
geometries during severe transients, when additional parameters characterizing the transient
are required. This demonstrates the need to categorize the important CHF-controlling
parameters and their ranges of interest.
CHF prediction methods can be conveniently grouped into the following categories:
 Analytical models,
 Empirical correlations, and
 Look-up table methods.
Analytical CHF models
Analytical CHF models are based on physical mechanisms and satisfy the conservation
equations [IAEA2001]. They generally require a two-fluid model approach, but occasionally
must use a three-field approach (e.g., dispersed annular flow). Although these models have
been improved significantly and usually predict the correct asymptotic trends, the evaluation
process is complex and time-consuming. Furthermore, because of our limited understanding of
the mechanisms involved and the lack of measurements of interfacial parameters, the models
are still less accurate than empirical correlations over the range of their database. In addition,
the analytical models have limited application because they can be used only in situations that
resemble the conditions upon which the model assumptions are based.
The annular film dryout model is based on a mass balance on the liquid film in annular flow and
postulates that CHF corresponds to depletion of this liquid film. Equations for droplet
entrainment and deposition have been proposed. The model provides reasonable predictions
of CHF for annular flow at medium to high pressures and flows and void fractions exceeding
50%.
The bubbly layer model postulates that CHF first occurs in the lower quality regime when the
bubble layer covering the heated surface becomes so thick and saturated with bubbles that
liquid mixing between the heated surface and the cooler core liquid becomes insufficient.
The Helmholtz instability model is applicable to pool boiling, where the boiling crisis is reached
when the flow of vapour leaving the heated surface is so large that it prevents a sufficient
amount of liquid from reaching the surface to maintain the heated surface in the wet
condition. The phenomenon that limits the inflow of liquid is the Helmholtz instability, which
occurs when a counter-current flow of vapour and liquid becomes unstable.
Empirical CHF prediction methods
Empirical CHF prediction methods can be subdivided into those based on inlet conditions and
those based on local cross-sectional average (CSA) conditions [IAEA2001].
The empirical correlations based on inlet conditions (overall power correlations) are all in the
form of empirical correlations based on CSA inlet conditions (P, G, Tin, or ΔHin) and usually
assume the “overall power” hypothesis. This hypothesis states that, for a given geometry and
set of inlet conditions, the critical power (the power corresponding to the first occurrence of
CHF for that geometry) is independent of axial or radial heat flux distribution. This assumption
permits the use of CHF correlations derived from uniformly heated bundle data to predict
dryout power in non-uniformly heated bundles of identical geometry (i.e., identical cross
section and heated length). This technique is a reasonable one for obtaining a first estimate of
dryout power; it gives reasonable estimates of dryout power in the annular flow regime for
symmetric flux profiles and form factors close to unity. However, it is not recommended for
form factors significantly different from unity. This approach can also be used to predict the
critical power of fuel channels with a fixed cross section, heated length, axial flux distribution
(AFD), and radial flux distribution (RFD), irrespective of the form factor. If the experimental
AFD and RFD represent the worst flux shapes from a CHF point of view, then the empirical
correlations can be used for lower-bound predictions. The overall power correlations have
generally good prediction accuracy. The best examples of this group are the Bowring
correlation for tubes and the EPRI-2 correlation for fuel bundles [IAEA2001, TON1996,
DEL1981, BER1981, CAT2005].
The overall power correlations have limitations, including applicability only to specific
geometries, heated lengths, heat-flux profiles, and the range of conditions in the database.
Hence, any variation of these parameters (i.e., extrapolation outside the database) may affect
CHF prediction and hence predicted reactor, channel, and bundle power. Moreover, these
correlations do not exhibit the correct asymptotic and parametric trends, and the axial and
radial CHF locations cannot be predicted.
The local CHF correlations satisfy the local-conditions hypothesis, which states that the local
CHF is dependent only on local conditions and not on upstream history. In principle, the local-
conditions hypothesis is sound if it is based on true local cross-sectional average conditions.
The reference formulation of local correlations is based on the linear trend of CHF with critical
quality (transformed from the linear trend of critical power with inlet sub-cooling). Additional
terms are included for separate effects (such as axial and radial heat-flux distributions). The
local correlations must be used together with the heat-balance equation to determine critical
power and CHF location (which requires iteration). The best examples of these correlations are
the Biasi correlation for tubes and the Becker correlation for fuel bundles [IAEA2001, TON1996,
DEL1981, BER1981, CAT2005].
An important disadvantage of local correlations is a larger prediction scatter than with overall
power correlations. They are valid only for specific geometries, heat-flux profiles, and the
range of conditions in the database; any variation of these parameters may affect local flow
and enthalpy distributions and hence CHF. As with any correlations of this type, extrapolation
of the applicability range may result in incorrect predictions. Moreover, these correlations
have incorrect asymptotic and parametric trends, and radial CHF location cannot be predicted.
CHF table look-up method
Because most empirical correlations and analytical models have a limited range of application,
a more general technique is needed. This approach is based on developing a table of measured
values for different geometries and flow conditions and supplementing them in terms of key
parameters that affect CHF. This approach has become known as a look-up table method. The
look-up table includes normalized CHF data banks for reference tubes, triangular-array fuel
bundles, and CANDU fuel bundles of natural-uranium fuel in a nominal channel. The look-up
table is developed using generalized correlations and trends, and then experimental data are
superimposed to improve accuracy and ensure that parametric trends are captured. The look-
up table can be statistically extended to table matrix conditions.
The empirical bundle CHF prediction method requires an extensive database with an adequate
range of the parameters of interest. It is suitable for design calculations and for reactor power
evaluations.
The sub-channel codes predict enthalpy and flow in each fuel sub-channel. They require a sub-
channel CHF prediction method (tube-based) and a spacer mixing/enhancement model.
The enthalpy imbalance approach uses enthalpy imbalance (in terms of thermodynamic
quality) between the critical sub-channel and the bundle cross-sectional average values. It is
used in the tube-data-based CHF prediction method with a modified thermodynamic quality to
account for the enthalpy imbalance.
Currently, CANDU 6 design and safety analysis are based on the look-up table prediction
method. This method is applicable to other geometries or flow conditions through
modification factors embedded in the table. The present CHF values used in design analysis are
included in the look-up table for uniformly heated, vertical tubes of 8 mm inside diameter,
cooled with an upward flow of water. They cover the widest range of flow conditions (all
possible CHF regimes) and exhibit correct asymptotic and parametric trends (smooth transition
between various CHF regimes). In addition, detailed prediction uncertainty is available at sub-
region levels. To this point November 13 early afternoon
A sample of the CANDU look-up table is shown in Table 11 [LEU2004, IAEA2001], where CHF
values are provided in a table format as a function of pressure, mass flux, and quality. For
values of these parameters that fall between the values given in the table, an interpolation is
made to calculate the appropriate CHF value.
Table 11 Sample of CHF look-up table
PRESSURE MASS FLUX QUALITY
-2 -1
(kPa) (kg m s ) -0.5 -0.4 -0.3 -0.2 ... ... 0.6 0.7 0.8 0.9
... ... ... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ... ...
7000 2500 10882 9986 8709 7496 ... ... 261 204 99 51
7000 3000 11730 10850 9620 8170 ... ... 346 263 112 52
7000 3500 12535 11558 10344 8740 ... ... 409 317 135 57
7000 4000 13317 12216 10929 9320 ... ... 470 317 136 58
7000 4500 14070 12839 11469 9769 ... ... 492 317 137 59
7000 5000 14792 13465 11954 10124 ... ... 521 326 138 63
7000 5500 15509 14000 12474 10713 ... ... 582 348 153 70
7000 6000 16208 14521 12931 11464 ... ... 655 379 179 84
7000 6500 16875 15091 13336 12214 ... ... 725 422 210 99
7000 7000 17529 15640 13763 12432 ... ... 795 476 243 116
7000 7500 18170 16174 14182 12682 ... ... 866 530 277 132
7000 8000 18806 16673 14610 12995 ... ... 936 581 312 149
8000 ... ...

0 5101 4795 4484 4175 542 456 423 364
8000 50 5714 5359 5025 4744 ... ... 1183 980 938 720
8000 100 6229 5834 5487 5235 ... ... 1678 1457 1449 1050
8000 300 6685 6179 5792 5654 ... ... 2144 1877 1538 1083
8000 500 6958 6354 5920 5763 ... ... 2068 1858 1427 917
... ... ... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ... ...
Compared to other available prediction methods, the look-up table approach has the following
advantages: (i) greater accuracy; (ii) wider range of application; (iii) correct asymptotic trend;
(iv) requires less computing time; and (v) can be easily updated if additional data become
available. Applying the tables to transient heat transfer in bundles requires adjustment factors
to correct for geometry, flux shape, and possibly transient effects. Here the advantages of the
tabular technique (wide range of application, greater accuracy, and greater computing
efficiency) are particularly important to the user.
The look-up table approach has some disadvantages: (i) it is a purely empirical prediction
method, and hence it does not properly reflect the physics, and (ii) it could introduce
erroneous trends if the underlying database were subject to experimental errors.
7.6.4.3.1 CHF correlations in CANDU bundles

Several CHF correlations have been developed for CANDU bundle geometry. These are mainly
based on full-scale bundle data and are used for critical channel power calculations.
An example of a general form of a flux-corrected local CHF correlation for uncrept channels is
given by [LEU2004, IAEA2001]:
a7
 qlocal 

CHFlocal  a1 P G  a4 P G xcr
a2 a3 a5 a6
   . (268)
 qaverage 
An example of a general form of boiling-length-average CHF correlation for uncrept and crept
channels is given by [LEU2004, IAEA2001]:
  qlocal  
b8

CHFBLA  b1P G  b4 P G xcr
b2 b3 b5 b6
 1  b7 
 q   . (269)
  average  
The values of the parameters in the above equations are of a proprietary nature.
7.6.4.3.2 CHF correlations in fuel assemblies

Providing precise CHF predictions for the fuel bundles of advanced water-cooled reactors is a
very difficult task. Advanced water-cooled reactors designs includes a variety of bundle
geometries, as well as a variety of element spacer designs.
The basis of almost any generic bundle prediction method is a tube CHF prediction method
because (i) parametric trends with pressure, mass flux, and quality are similar in tubes and in
bundles, and (ii) tube CHF prediction methods are generally used in sub-channel codes to
predict CHF in bundles.
One of the well-known correlations for PWR fuel bundles is the Westinghouse and Tong
[IAEA2001, TON1996] correlation, which can be given by the following expression:
l
C  q "( z ')  eC (l  z ') dz '
4.23 106 1  xe (l )
7.9
q ''
q''
 cr , F  0
, C [m 1 ] , (270)
q "(l ) 1  e C l

cr , DNB 1.72
F G
qcr"   2.022  0.06238  p    0.1722  0.001427  p  e 18.177 0.5987 px e 

 0.1484  1.596  xe  0.1729  xe xe   2.32  G  3271 1.157  0.869  x e  . (271)
0.2664  0.837  e124Dh  0.8258  0.0003413  hl  hin  
" 2
The critical heat flux at the DNB, qcr , DNB (kW / m ) , is at the local heat flux (for non-uniform
heat flux), l (m) is the distance to the DNB, and xe is the local steam thermodynamic quality.
The above correlation is valid for the following parameter ranges: pressures of 5.5–13.8 (MPa),
mass fluxes of 1350–6789 (kg/m2s), hydraulic diameters of 0.005–0.0178 (m), thermodynamic
qualities of -0.15–0.15, and channel lengths of 0.254–3.658 (m).
Another useful correlation for PWR fuel bundles is given by Bowring [IAEA2001, TON1996]:
A  B  hlg x 2.317   h fg D  G / 4   F1 D G
qcr"  , A , B ,
C 1  0.0143  F2  D G 1/2
4
(272)
0.077  F3 D  G
C , pR  0.145  p
1  0.347  F4  G /1356 
n

F1  p18.942
R exp  20.89 1  pR    0.197 /1.917 
F2  F1 /  p 1.316
R exp  2.444 1  pR    0.309 /1.309   For pR 1MPa , (273)

F3  p 17.023
R exp 16.658 1  pR    0.667 /1.667 
F4  F3 p1.649
R
F1  pR0.368 exp 0.648 1  pR  

F2  F1 / pR0.448 exp 0.245 1  pR    For pR 1MPa . (274)
F3  p 0.219
R
F4  F3 p1.649
R
The above correlation is valid for channel lengths of 0.15–3.7 (m), pressures of 0.2–19 (MPa),
and mass fluxes of 136–18,600 (kg/m2s).
The CANDU fuel CHF correlation is based on the CHF look-up table and is given by the following
expression (note that CANDU 6 uses the look-up table method, whereas the Bruce and
Darlington designs use correlations [IAEA2001]):
CHF  CHFTABLEK1K2K3K4K5K6K7K8K9 , (275)

where K1 to K9 are modification factors to account for sub-channel specific effects (e.g.,
element gap size, sub-channel equivalent diameter, adjacent heated/unheated surfaces,
upstream spacers, etc.) and are defined as: K1 – tube diameter factor; K2 – bundle geometry
factor; K3 – spacer-effect factor; K4 – heated-length factor; K5 – axial-flux-shape factor; K6 –
circumferential-flux factor; K7 – horizontal flow factor; K8 – low-flow factor; K9 – transient-effect
factor.
7.6.4.4 CHF margin evaluation
CHF evaluation in the reactor core is of primary importance for estimating the reactor
operating margins with adequate precision and a good understanding of uncertainties. Figure
114 shows the concept of reactor operating margin evaluation as a function of reactor
operation (aging) [LEU2004, POP2014].
Figure 114 Concept of operating margins

This diagram assumes that the nominal power will remain unchanged during the reactor’s life.
Above the full nominal power are the detector drift margin and the operator’s trip
margin/fuelling/operating perturbation margin, which are also assumed to remain unchanged
during the reactor life. On top of these are the error and uncertainties margin and the trip
imperfection margin, which are also assumed to remain unchanged during the reactor life.
However, the excess margin that is embedded in the reactor core design, and which represents
all aging mechanisms having an impact on the reactor power margin, such as pressure-tube
creep, continually decreases during the reactor life (i.e., the pressure-tube life). When all these
margins are evaluated and added on top of the full nominal power, the resulting maximum
power should still be less than the reactor critical power at which CHF occurs in some fuel
channels.
From the above discussion, it is clear that CHF prediction is an important part of reactor
thermal power margin evaluation. Hence, CHF evaluation and application focusses on the
following concerns:
i. To set the operating power with a comfortable margin to avoid CHF occurrence, in
terms of the following parameters: (a) minimum CHF ratio (MCHFR) at constant
pressure, mass flux, and critical quality; (b) minimum CHF power ratio (MCHFPR) at
constant pressure, mass flux, and inlet fluid temperature; or (c) minimum power ratio
(MCPR) at constant pressure, pump characteristic, and inlet fluid temperature
ii. To evaluate the maximum sheath temperature during LOCA, LOFA, or LORA, which is
usually concerned with the first (initial) CHF location;
iii. To evaluate the thermal-hydraulic and neutronic responses to CHF occurrence in a
reactor core (trip set points), based on knowledge of how CHF spreads in the reactor
core; best-estimate predictions of average CHF and/or area in dryout as a function of
power.
For reactor transient (accident) analyzes, it is important to determine sheath and fuel
temperatures. In this case, CHF is considered as the reference point for post-dryout analyzes.
The set of transient analyzes includes the following scenarios: loss of regulation, loss of flow,
loss of Class IV power, loss of coolant (small and large breaks), etc. More details are provided
in Chapter 16.
The following terminology is used when evaluating CHF margins:
1) Dryout (or critical heat flux, CHF) occurs when the fuel sheath can no longer maintain
continuous liquid contact. Current licensing criteria are no burnout or dryout at any
location in the fuel channel (details depend on whether the first or second trip is
considered in the analysis).
2) Critical power (CP) (also known as critical channel power, CCP) is the power
corresponding to the first CHF occurrence (at constant pump head). Determining this
requires knowledge of various disciplines (physics, fuel, fuel channel, thermal-
hydraulics, etc.).
3) Critical power ratio (CPR) is the ratio of critical power to operating power.
4) Regional overpower protection (ROP) or neutron overpower protection (NOP) is aimed
at preventing burnout in any fuel assembly during a slow LOR event. It includes full-
core analysis of all possible scenarios and establishment of trip set points for various
detectors.
Figure 115 shows the concept of critical channel power (CPR) used in the CANDU CHF margin
methodology, with channel flow shown on the y-axis and channel power on the x-axis
[IAEA2001, LEU2004, POP2014]. In this diagram, a critical parameter is any parameter
associated with the critical heat flux (CHF). Below the diagram, a CANDU fuel channel is shown,
with 12 fuel bundles and associated feeders connecting the channel to the reactor inlet and
outlet header.
The channel hydraulic curve (blue line) determines the level of channel power that can be
removed by a certain flow rate. Therefore, as power rises, the mass quality will also rise (two-
phase flow behaviour), which would result in a reduction in the mass flow rate that can be
pushed through the channel. To this point November 14 afternoon
Figure 115 Critical channel power
The CHF curve (red line) shows the CHF values for certain combinations of channel power and
flow rate. As explained in previous sections, at higher flow rates, the channel power at CHF will
also rise (with the square of the flow rate). The intersection of these two curves shows the
actual point at which CHF will be reached in the channel for a given flow rate and power.
Hence, the intersection point defines the critical flow and the critical power at CHF. The left
part of the diagram for which power is less than the critical power is defined as the pre-CHF
region, and the right side of the diagram is the post-CHF region.
Figure 116 CHF margin definitions

The concept of CHF margin is explained in Figure 116 [IAEA2001, POP2014]. The top diagram
on this figure shows heat flux on the y-axis and thermodynamic quality on the x-axis. The red
solid curve shows a prediction of CHF as a function of thermodynamic quality. Of course,
thermodynamic quality depends on channel power among other parameters, and therefore in
turn, the shape and value of the CHF curve depend on channel power.
The solid black curve shows local heat flux as a function of quality, i.e., axial position in the
channel as quality changes along the channel. Normally, under nominal conditions, the heat
flux is less than the CHF, and therefore the black curve does not touch the red curve for any
quality value. However, if the channel power is increased, the local heat flux along the channel
also increases. At a certain value of the channel power, the local heat flux will touch the CHF at
a certain position, and this position represents the power value that causes CHF (red dashed
line). On the bottom diagram, the CHF ratio is shown, defined by the ratio of the critical heat
flux (CHF) to the flux corresponding to the nominal operating power (NOP):
CHF
CHFRATIO  . (276)
qlocal
The above relation assumes constant inlet thermodynamic quality, constant outlet pressure,
and mass flux.
Figure 117 shows a comparison of CHF margin evaluation methods: the CHF ratio, the CHF
power ratio, and the critical power ratio [IAEA2001, POP2014]. The above explanation along
with Eq. (276) is illustrated on the left side of Figure 117. The MCHFR method is easy to
understand and implement. However, it does not provide a clear account of reactor power
margins because it uses heat flux instead of channel power, which makes it cumbersome to use
for power reactor applications (however, it is often used for research reactor applications).
The CHFPR method is shown on the right side of Figure 117 [IAEA2001, POP2014]. In this
method, rather than a ratio of channel heat fluxes, a ratio of channel powers is used that
represent a certain value of operating heat flux and CHF. This method provides a better
understanding of channel and reactor power margins with respect to CHF. However, it does
not provide a clear account of margins because as channel power rises to reach the CHF line,
the CHF line will also change position because it indirectly depends on channel power.
The most useful method in CANDU reactor applications is the critical power (CPR) method. The
principles of this method are explained above and illustrated in Figure 115 [IAEA2001,
POP2014]. The critical power ratio is defined by the following relation:
Critical Power
CPRATIO = . (277)
Operating Power
This relation holds for constant inlet fluid temperature and pressure and a fixed pump curve.
As is evident from this diagram, channel flow at point NOP on the hydraulic curve corresponds
to NOP power. Keeping the flow at point NOP, the corresponding heat flux that would have
resulted in CHF is at point CHF1 (the top blue point on the CHF curve), corresponding to the
critical power. CP1. The flow at point CHF (lower blue point on the CHF curve) corresponds to
the channel power. CP2.
Hence, the critical power ratio is given by the relation:
CP2
CPR 
NOP . (278)
The margin to CHF is shown as the distance between points NOP and CP2, as can be observed
on the right side of Figure 117.
a) Critical heat flux ratio b) Critical heat flux power ratio
c) Critical power ratio
Figure 117 CHF margin evaluation methods (a, b, c)

There are a number of uncertainties in the development and application of the CHF prediction
methods and CHF margin methods covered in this section. These uncertainties need to be well
understood, classified, and evaluated and then taken into account in reactor applications. The
uncertainties can be broadly grouped into the following categories:
i. Uncertainties in CHF measurements using bundle simulators. These uncertainties can
result from errors in flow, power, pressure, and inlet temperature measurements that
are caused by (a) inadequate CHF detection methods, (b) variations in flux distribution
across/along the bundle, and (c) geometric tolerances;
ii. Prediction uncertainties in CHF correlations or sub-channel codes;
iii. Uncertainties in reactor conditions, including (a) reactor flow, pressure, and
temperature (calculated from system safety codes) and (b) reactor power
measurements (from detectors); and
iv. Uncertainties in extrapolation to in-reactor conditions resulting from (a) electrical vs.
nuclear heating; (b) flux tilts across elements, the bundle, and the core; and reactor
aging effects (creep, fouling of pipes, etc.).
7.6.4.5 CHF bundle effects – R&D testing and application
Extensive R&D support is required to gain a better understanding of CHF phenomena, including
support for development of CHF correlations and look-up tables and development of computer
tools. Any changes in design or operating parameters trigger a need for supporting R&D to
provide evidence of the applicability of the current CHF models. The general strategy in
performing supporting CHF R&D is to conduct it in an incremental fashion to minimize the cost
and optimize the work scope.
CHF experiments are usually based on constant flow conditions (inlet or outlet pressure, inlet
mass flow rate, and inlet-fluid temperature) and constant geometric factors (hydraulic
diameter and heated length) in each test series. The primary interest in these CHF experiments
is measurement of fuel power at CHF. However, determining CHF locations (both axially and
radially) is also an objective. The other measurements of interest are circumferential CHF
location, subsequent CHF occurrences (CHF spreading, or dry patch spreading), and pressure
drops along a channel experiencing CHF.
A number of CHF experiments are performed in a simple geometry consisting of one fuel
element simulator. In such case, the experimental setup consists of a pipe in which heating is
provided by direct electrical current flowing through the pipe wall and coolant flowing past the
pipe outside surface. Another option is to use indirect heating by a coil embedded in the pipe
interior. through which electrical current is supplied. In either case, the heat flux from the pipe
wall to the fluid is directly proportional to the strength of the electrical current. Axial power
profiles can be obtained by varying the pipe wall thickness in direct heating, or by changing the
number of coils per unit length in the case of indirect heating.
The pipe is equipped with various instruments to measure important parameters. Stationary or
movable thermocouples are used for wall temperature measurements. The stationary
thermocouples can measure only temperatures, i.e., burnout/dryout at specific locations in the
pipe. Therefore, these may not capture initial CHF value and location. Moreover, no
information on subsequent CHF and dry-patch spreading can be obtained with these
thermocouples. Moveable thermocouples provide coverage almost over the entire heated
length because they are moved during the experiment along the internal wall of the heated
pipe. They provide measurements of initial CHF, subsequent CHF occurrences, and dry-patch
spreading. However, they are more conservative, expensive, and time-consuming (when
scanning the entire pipe surface area).
More complex and sophisticated CHF measurements are performed in CANDU fuel bundle
simulators. One such simulator is shown in Figure 118 [LEU2004]. This fuel bundle simulator
consists of a 37-element full-scale bundle 6 m long. The bundle is provided with a simulator of
junctions and appendages and fully resembles a CANDU fuel channel. The wall heating
distribution takes into account bundle endplates and fuel distribution along the string of fuel
bundles. This device can simulate non-uniform axial and radial power distribution. Sliding
thermocouples are inserted into a number of fuel element simulators that are indirectly heated
by coils connected to the electric current source. Figure 119 shows an axial power distribution
in such a fuel bundle simulator, and as is evident in the figure, any axial variation in heat flux
can easily be simulated. Figure 120 shows a close-up view of one part of a fuel bundle and a
sliding thermocouple moving inside along the fuel element, with a sample of a measured fuel
sheath temperature [LEU2004].
Figure 118 CANDU bundle experimental simulator
Figure 119 Axial flux distribution in full-scale Figure 120 CHF variation along a heated
bundle test surface
One of the most important aspects and of CHF in fuel bundles is the impact of spacing devices,
which enhances CHF as the spacing value increases. In addition, the impact of radial power
distribution is important because it can result in asymmetric CHF occurrence in the bundle.
However, the 37-element and 43-element CANDU fuel bundle designs have enhanced features
that improve CHF design.
Figure 121 shows various spacer devices used in fuel bundle design for various reactor types
[IAEA2001]. These devices have a dual purpose: to keep the bundle tight, i.e., to minimize
radial vibration, and to enhance turbulence and thus improve CHF robustness (i.e., increase the
CHF value). In general, turbulence in pipes and fuel bundles creates better fluid mixing and
therefore easier removal of bubbles generated in sub-cooled boiling on the walls. This in turn
enables a fuel element to achieve higher heat fluxes without causing CHF by bubble crowding
on the fuel sheath surface.
Figure 121 Different types of spacer devices for CHF robustness

Figure 122 presents an illustrative example of the impact of spacer devices on CHF [IAEA2001].
On the y-axis, local critical heat flux is shown. The flow along this CANDU channel is from left
to right. Because sub-cooling is higher at the inlet, CHF is also higher. As fluid moves along the
channel, the pressure decreases, quality increases, and bulk boiling starts. The net effect of
these changes is a reduction in the CHF along the channel. However, as fluid approaches a
bundle spacer, fluid turbulence increases, causing a sharp CHF increase. If the fuel rods were
smooth along the length of the fuel channel and there were no spacers or other appendages,
the CHF would have been reduced much faster.
The benefits of more turbulence as a method of enhancing CHF robustness must be weighed
against the negative impact of higher turbulence and local pressure losses on the primary
pump head. Therefore, having too much turbulence and local pressure loss causes the primary
pump head, pump design, and particularly pump operation to become prohibitively expensive,
which may outweigh the benefit with regard to CHF.
Figure 122 CHF behaviour around fuel bundle spacer devices

CHF data have a wide range of application, and this is one of the most important R&D areas for
reactor design, safety analysis, and licensing. CHF margins are the central issue in the licensing
submissions for nuclear reactors, and this application is a strong driver for full-scale fuel bundle
data, for quantification of the impact of separate effects, and for assessment of CHF
uncertainties.
CHF R&D work is important to reach an adequate understanding of the phenomena involved
using fundamental and bundle-specific studies and parametric and separate-effects tests.
Certainly, CHF R&D is essential to develop prediction methods, i.e., CHF correlations and
models, and to validate these correlations, models, and reactor safety codes. The CHF tests
used for this purpose are a combination of bundle and separate-effects tests. The key
applications of CHF prediction methods are in ROP (or NOP) and reactor safety codes, fuel
channel behaviour codes, and fuel behaviour codes.
Experimentally measured CHF values are based on tests with defined inlet conditions. CHF
values based on inlet-flow conditions are limited to a specific channel geometry and heated
length. However, reactor safety analyzes use the local-conditions approach based on local
pressure, mass flux, and quality to predict local CHF. For given inlet conditions, the local CHF
conditions may vary in terms of thermodynamic quality (or enthalpy). In such cases, low
quality (or enthalpy) is appropriate for short heated lengths and high quality (or enthalpy) for
long heated lengths. It is important to evaluate local thermodynamic quality using inlet flow
conditions and power.
When using CHF data, it is important to mention that all CHF data are obtained under steady-
state conditions in terms of power, mass flux, and sub-cooling. In safety analysis, CHF
prediction tools developed for steady-state conditions are used to simulate reactor transients.
If the transient is slow, i.e., longer than 10 s, no significant impact on the models is expected.
However, for fast transients (shorter than 1 s), a significant effect on CHF is likely.
Another important CHF application is related to thermal-hydraulic and neutronic responses to
CHF occurrence (partial bundle dryout). This application is strongly affected by the level of
success in predicting the spread of dryout in a fuel bundle. It has been experimentally
observed that the average CHF value that occurs over a large portion of the fuel sheath surface
can be much higher than the initial CHF at some cross-sectional locations. To capture such
important effects in dryout spreading, sub-channel codes are the most promising tool.
7.6.5 Transition and film boiling heat transfer

Reactor fuel channel operation beyond dryout (beyond CHF) is possible in certain transient and
accident regimes. This section provides information on post-dryout heat transfer phenomena
and on the correlations used for these heat transfer regimes.
Post-CHF (or post-dryout) heat transfer occurs when the surface temperature becomes too
high to maintain continuous liquid contact and the surface becomes covered by a continuous or
intermittent vapour blanket. Post-CHF heat transfer includes transition boiling, illustrated in
Figure 29 and Figure 123, where intermittent wetting of the heated surface takes place, and
film boiling, where the heated surface is too hot to permit liquid contact. The boundary
between these post-CHF heat transfer modes is the minimum film boiling temperature, or TMFB.
Due to the poor heat transport properties of vapour, high heated surface temperatures are
often encountered during film boiling.
7.6.5.1 Post-dryout phenomena
Several figures presented in previous sub-sections of this chapter are relevant to post-dryout
heat transfer. Figure 26 shows heat transfer regimes in vertical fuel channels, with clearly
shown areas of post-dryout heat transfer regimes in the upper part of the channel. Figure 28
shows a 3-D representation of the boiling surface for water, with transition and film boiling
indicated. Figure 29 shows a conceptual illustration of the boiling curve, again showing post-
dryout heat transfer beyond the CHF point. Figure 33 shows temperature profiles along a
vertical fuel channel experiencing various flow and heat transfer regimes, including post-dryout
heat transfer regimes in the upper part of the figure.
Post-CHF heat transfer is initiated as soon as the CHF condition is exceeded and continues until
quenching (rewetting) of the surface is achieved. Depending on the particular scenario and
flow conditions present, various heat transfer modes of the boiling curve may be distributed
along a heated surface, or a series of heat transfer modes can succeed each other over time at
the same location, as is the case during transients.
During sustained post-dryout heat transfer, the heated surface is cooled mainly by vapour flow.
Because of the relatively low heat-transfer coefficient for this heat transfer regime, the fuel
sheath is exposed to high temperature that may lead to damage. In this respect, it is important
to understand and determine the surface conditions and the coolant flow parameters at the
time CHF is reached, because these determine the impact of the post-dryout condition on the
fuel sheath.
Figure 123 shows various post-dryout flow types and heat transfer regimes on a vertical surface
[LEU2004, POP2014]. These are highly similar to those that occur in a horizontal fuel channel.
Type I post-dryout identifies a situation in which the heated surface has lost contact with the
liquid, but at one location, contact has been re-established, and this point acts as a heat sink.
In this case, heat transfer from the internal parts of the wall and from surrounding areas occurs
towards this heat sink, thus cooling the wall near the liquid-wall contact and preparing it for
quenching.
Type II represents a situation in which a heated surface in post-dryout regime is flooded by
liquid water from the top. This figure clearly shows the quench front moving downward. The
upper part of the surface is well cooled, and obviously heat transfer occurs from the lower,
hotter portions of the wall upwards towards the cooler area, which therefore acts as a local
heat sink, preparing the lower part of the surface for quenching. Type III represents a situation
similar to Type II, but in this case, quenching proceeds in an upward direction by flooding from
the bottom. Type II and Type III post-dryout regimes combine into one regime for a horizontal
surface.
Figure 123 Types of transition boiling and film

Type IV mode occurs when the heated surface in post-dryout regime is cooled by impingement
of droplets flowing past the heated wall. Each droplet impingement on the heated surface
provides a quick intermittent miniature local heat sink. However, although local surface
cooling by one droplet impingement is infinitesimally small, as many droplets impinge on the
surface, the integral effect of heat removal is significant, thus preparing the surface for
rewetting.
Types V and VI apply to horizontal surfaces in pool boiling. Type V shows droplets impinging on
the surface, with an effect similar to Type IV, whereas Type VI represents a regime in which a
water layer floats on top of a vapour film, a situation similar to that described for Type I.
Transition Boiling
Berenson [IAEA2001] provides one of the best definitions of transition boiling: a combination of
unstable film boiling and unstable nucleate boiling alternately existing at any given location on
a heating surface. The variation in heat transfer rate with temperature is primarily the result of
a change in the fraction of time that each boiling regime exists at a given location.
Figure 124 conceptually shows transition boiling in a pool boiling regime. Pockets of vapour are
visible on the heated surface and may move intermittently back and forth along the surface, as
is typical for this pool boiling situation. At any surface temperature in excess of the CHF
temperature, the heated surface is partially covered with unstable vapour patches varying over
space and time, with frequent replacement of vapour patches by fluid. Although this may
seem similar to transition pool boiling as described above, the introduction of the convective
component improves the film boiling component by reducing the vapour film thickness and
changing the heat transfer mode, whether dry or wet, from free convection to forced
convection.
Figure 124 Concept of transition boiling in pool boiling

As shown in Figure 29, the slope of the curve for transition boiling is negative, which means
that this is an unstable heat transfer regime. This heat transfer regime can occur on a
temperature-controlled surface at conditions beyond CHF. Examples of temperature-
controlled surfaces are surfaces with high thermal inertia, surfaces heated by condensing
steam (which controls temperature at the saturation point), and fuel element surfaces during
fast transients.
A heat-flux-controlled surface at conditions beyond CHF can experience transition boiling only
under conditions under which the slope of the transition boiling curve is positive (dq/dt > 0),
which is possible for mass quality above 50%, as shown in Figure 31. Note that as temperature
increases gradually with increasing heat flux, transition boiling intermittently wets and dries
the surface. In the case of high quality, even when the surface is continuously dry, the
temperature increase is not drastic due to effective convective heat transfer. Examples of
heat-flux-controlled surfaces are electric heaters and fuel elements during fast transients.
If the slope of the transition boiling curve is negative (dq/dt < 0), a rapid transition from
nucleate boiling to film boiling usually occurs, and therefore transition boiling is not
encountered (or occurs very briefly).
The periodic contacts between liquid and heated surface in the transition boiling region of the
boiling curve result in formation of large amounts of vapour, which force liquid away from the
surface and create an unstable vapour film or blanket. Because of this, the surface heat flux
and the surface temperature can vary both with time and with position on the heater. The
average heat transfer coefficient decreases as the temperature increases because the time of
contact between the liquid and the heater surface is decreased.
Heat transfer in transition boiling occurs primarily by liquid-solid contact. At the critical heat
flux point, the contact-area (or time) fraction is generally 100%, i.e., most of the surface is in
good contact with the liquid. Post-CHF conditions result in a sudden large decrease of the
contact fraction between the liquid and the heated wall surface. Thereafter, most of the heat
transferred during transition boiling will be due to droplet-wall interaction. Initially, at surface
temperatures just in excess of the boiling crisis temperature, a significant fraction of the
droplets deposit on the heated surface, but at higher wall superheats, vapour repulsion forces
become significant in repelling most of the droplets before they can contact the heated
surface. The repelled droplets contribute to heat transfer by inducing turbulence in the
boundary layer to enhance convective heat transfer to the vapour.
Figure 29 shows that transition boiling ends when the heated wall surface temperature drops
to a minimum value, called the minimum film boiling temperature, TMFB. Several names are
given to the minimum film boiling temperature. They include rewetting temperature, quench
temperature, Leidenfrost temperature, film boiling collapse temperature, and others.
During dryout and wall temperature increase, the minimum film boiling temperature separates
the high-temperature region, where film boiling or vapour cooling takes place, from the lower-
temperature region, where much more efficient heat transfer in transition boiling occurs.
Knowledge of the minimum film boiling temperature is particularly important in reactor safety
assessments.
During surface quenching (such as emergency core cooling), rewetting commences at the
minimum film boiling temperature and proceeds until nucleate boiling is established at a much
lower wall temperature. For this reason, predicting the minimum film boiling temperature as a
function of the system parameters is very important because heat transfer coefficients on
either side of the minimum film boiling temperature can differ by orders of magnitude.
Generally, TMFB is defined as the temperature at the minimum heat flux.
In summary, no liquid/sheath contact is possible for TW > TMFB, but possible liquid/fuel sheath
contact exists for TW < TMFB:
qMIN
TW ,MIN  . (279)
hFB
Film boiling
Two theories have been proposed for analytical prediction of the minimum film boiling
temperature. One theory says that the minimum temperature is a thermodynamic property of
the fluid (i.e., the maximum liquid temperature) and therefore is primarily a function of
pressure. According to this theory, the minimum film boiling temperature is associated with
the maximum liquid superheat (beyond which the nucleation rate is infinite and liquid cannot
exist). This theory comes from the equation of state and from homogeneous nucleation
theory.
Figure 125 Minimum film pool boiling – Taylor instability
The other minimum film boiling theory suggests that rewetting commences due to
hydrodynamic instabilities that depend on the velocities, densities, and viscosities of both
phases as well as the surface tension at the liquid-vapour interface. By this theory, the
minimum heat flux associated with the minimum film boiling temperature can be predicted
from Taylor’s instability criterion. Figure 125 illustrates the theory of Taylor’s instability
[LEU2004, POP2014, DEL1981, BER1981]. This figure shows a chunk of liquid that floats on top
of a vapour film above a heated surface. The instability results from gravity acting on the liquid
as a heavier phase on the top, and on the vapour as a much lighter phase at the bottom.
During fast transitions, where insufficient time is available to develop hydrodynamic forces
fully, rewetting can be expected to be thermodynamically controlled, whereas for low flows
and low pressures, where sufficient time is available and the volumetric expansion of the fluid
near the wall is large, rewetting is more likely to be hydrodynamically controlled. Once
rewetting has occurred locally, the rewetting front can then propagate at a rate controlled
primarily by axial heat conduction in the heated wall.
Figure 126 [LEU2004, POP2014] shows various film boiling types in pool boiling and forced
convection flow situations.
Figure 126 Film boiling types

Film boiling is generally defined as that mode of boiling heat transfer in which only the vapour
phase is in contact with the heated surface (see Figure 125 and Figure 126). The term is used in
forced convective boiling to refer to conditions under which the liquid does not contact the
heated surface, but is usually in one of the following forms (shown on Figure 126 and Figure 32)
a) a dispersed spray of droplets, normally encountered at void fractions in excess of
80% (liquid-deficient or dispersed-flow film boiling regime);
b) a continuous liquid core (surrounded by a vapour annulus which may contain

entrained droplets), usually encountered at void fractions below 40% (inverted
annular film boiling or inverted annular-flow boiling regime (IAFB)); and
c) a transition between these two cases, which can take the form of an inverted slug
flow for low to medium flow.
Figure 127 Pool film boiling mechanisms with horizontal heated wall
Figure 128 Pool film boiling patterns

The flow and heat transfer regimes types described above are illustrated in Figure 126. This
figure also shows an example of stratified flow film boiling that is relevant to CANDU reactor
channels. Figure 128 shows pool film boiling at a vertical wall and at a horizontal cylinder in
stagnant water.
The dispersed flow film boiling regime, shown in Figure 129 at the bottom, is the most
commonly encountered and has been well studied. Its heated surface temperature is
moderate. In the inverted annular and inverted slug flow regimes, shown at the top of Figure
129, excessive surface temperatures are frequently encountered (see Figure 104 and Figure
105). Radiation heat transfer, which is unimportant in transition boiling, becomes important in
film boiling, particularly at low flows, low void fractions, and surface temperatures in excess of
700°C.
Figure 129 Forced convection film boiling patterns

When the flow and heat transfer regime is film boiling with the presence of liquid, several heat
transfer regimes occur, as shown in Figure 130. On the left side of Figure 130, a film boiling
regime with droplets in the middle of the pipe is illustrated. On the right side of Figure 130, a
film boiling regime with a liquid core in the middle of the pipe is illustrated [POP2014,
LEU2004]. These two flow regimes correspond to the regimes shown in Figure 129. The y-axis
represents temperature or velocity, whereas the x-axis represents distance from the heated
wall, along with heat flux direction.
In both cases, close to the wall, a gradual change of temperature or velocity is indicated, as
appropriate for a boundary layer in a laminar flow regime. On the left side, the liquid droplets
are assumed to be small, as is appropriate for dispersed flow. In such a case, the droplets are
assumed to be at saturation. Hence, the temperature profile in the channel shows a
superheated wall (temperature above the saturation temperature for a given pressure) and
superheated steam in the pipe core.
Figure 130 Film boiling heat transfer temperature and velocity profiles
Away from the wall, temperature drops quickly to a lower value in the boundary layer, but still
remains above saturation. At the steam-droplet interface, temperature is close to saturation
and continues to be at saturation inside the droplet. The velocity is near zero along the wall
and rises in the boundary layer to a value representing steam bulk velocity. The steam velocity
is higher than the liquid velocity, and this velocity difference, and the associated interfacial
friction between the phases, carries the droplets upward. Because water is a much heavier
fluid than steam, a large velocity difference between water and steam is needed to produce
sufficient interfacial friction to carry the droplets in steam flow.
The heat flux direction on the left side of Figure 130 is from the wall to the vapour, and then
from the vapour to the liquid-droplet interface. Because the liquid droplets are saturated, any
heat added to the droplets results in a certain mass of liquid evaporating and joining the
vapour flow, and therefore liquid droplets continue to lose mass as they flow along the
superheated vapour flow. Heat radiation from the wall to the vapour and the droplets is
important only if the wall temperature is substantially above the saturation and steam
temperatures.
The right side of Figure 130 shows a film boiling regime with a liquid core in the pipe. In this
case, the velocity profile is similar to the one on the left, with the vapour velocity being higher
than the liquid velocity because the vapour is a much lighter fluid. The temperature profile
from the heated wall to the vapour is also similar. However, because the liquid core contains a
large volume of liquid, its internal temperature is assumed to be sub-cooled. Hence, the
vapour-liquid interface must be at saturation to maintain a continuous temperature profile. A
thin thermal boundary layer around the vapour-liquid interface with a steep temperature
change is evident.
In the inverted annular flow regime, few entrained droplets can be present in the steam film,
whereas the bulk of the liquid is in the form of a continuous liquid core that may contain
entrained bubbles. In dryout, the continuous liquid core becomes separated from the wall by a
steam layer that can accommodate steep velocity gradients. As a stable steam blanket is
formed, heat is transferred from the wall to the steam and subsequently from the steam to the
liquid core. For very thin steam films, heat transfer from the wall to the liquid is primarily by
conduction across a laminar steam film. For thicker steam films, turbulent flow occurs in the
film, and the liquid-steam interface becomes agitated. Heat transfer across the wavy steam-
liquid interface takes place by forced convection. This mode of heat transfer is much more
efficient than single-phase convective heat transfer between the smooth wall and the steam.
Hence, it is assumed that the bulk of the steam is at or close to the liquid core interface
temperature (i.e., the saturation temperature).
The heat flux direction on the right side of Figure 130 is from the wall to the steam, from the
steam to the liquid interface, and from the liquid interface to the liquid internal body. Note
that the heat flux from the steam to the liquid interface is partitioned into a part that is used to
evaporate a portion of the liquid interface and a part that is directed to the internal liquid body
as convective heating of the liquid. As mentioned above, heat radiation from the wall to the
steam and the liquid core is important only if the wall temperature is substantially above the
saturation and steam temperatures.
Heat transfer from the wall to the vapour and from the vapour to the liquid will depend on the
conditions of the surfaces, the flow patterns, and other parameters. This topic and the specific
heat transfer correlations are described in the section below.
Figure 131 shows various heat transfer regimes in film boiling of a dispersed droplet regime, as
described above [POP2014, LEU2004]. Solid lines in this figure indicate heat fluxes and their
directions, whereas the dashed line indicates mass transfer by evaporation.
Figure 131 Film boiling heat transfer mechanisms
7.6.5.2 Post-dryout prediction methods and heat transfer correlations
As mentioned before for other heat transfer regimes, many heat transfer correlations have
been developed for post-dryout heat transfer regimes. The key correlations and those that are
most widely in use are covered in this section.
Accurate prediction of wall temperature in the film boiling regime is of vital importance in
accident safety analysis of the core and steam generators of advanced water-cooled reactors.
The following four methods for estimating film boiling heat transfer are commonly used:
i. Semi-theoretical equations for pool film boiling;
ii. Semi-theoretical models to predict flow film boiling. They are based on the appropriate
constitutive equations, some of which are empirical in nature;
iii. Purely empirical correlations for flow film boiling, which do not account for any of the
physics, but instead assume a forced convective-type correlation; and
iv. Phenomenological equations for flow film boiling, which account for thermal
disequilibrium and attempt to predict the “true” vapour quality and vapour
temperature.
Transition boiling
The wall temperature at the initiation of transition boiling can be calculated using the CHF
correlation (the lower limit temperature) [IAEA2001]:
qCHF
TW ,CHF   TSAT , (280)
hNB
where qCHF is the heat flux at CHF and hNB is the nucleate boiling heat transfer coefficient.
The upper wall temperature limit that occurs in transition boiling is the minimum wall film
boiling temperature:
qMFB
TMFB   TSAT , (281)
hMFB
where qMFB is the minimum film-boiling heat flux and hFB is the film-boiling heat transfer
coefficient. Note that in general, heat transfer coefficients decrease from nucleate boiling to
transition boiling and then to film boiling, i.e., hNB > hTB > hFB.
Transition boiling is encountered only within a relatively narrow temperature range. For
example, for water at 10 MPa, the transition boiling wall temperature range is 300oC < TTB <
373oC.
Tong [IAEA2001, TON1996] has developed a transition boiling correlation that is frequently
used:
kg
hTB  39.75  e 0.144 T  2.3 105 e 105/ T Rev0.8 Prv0.4 . (282)
De
Very often, correlations containing boiling and convective components (valid for transition and
film boiling) are used, such as [IAEA2001]:
k 
hTB  A  e BTSAT   v   a  Rebv  Prvc . (283)
 D
An example of a phenomenological correlation (valid for transition and film boiling) is the one
proposed by Tong & Young and by Iloeje [IAEA2001]:
q  qdc  qndc  qconv . (284)

An example of an empirical correlation (valid for transition boiling only) independent of CHF
and minimum film boiling has been proposed by Ellion [IAEA2001]:
qTB  4.56 1011   TSAT 

2.4
. (285)
Film boiling
In addition to the above correlations, which are applicable to both transition and film boiling,
the minimum film boiling temperature can be given by the following correlation [IAEA2001,
DEL1981, COL1972]:
h 104
TW , MIN  284.7  0.0441 P  3.72 106 P 2  , Pressure  9000 kPa
(2.82  0.00122 P)h fg
. (286)
PCRIT  P
TW , MIN  TSAT  TW , MIN , 9000 kPa , Pressure  9000 kPa
PCRIT  9000
There is general agreement that the modified Bromley equation [IAEA2001] for film boiling can
be used for horizontal surfaces in pool boiling:
1/ 4
  3      v   r  g 1 g     v  
hFB  0.62   v v    . (287)
  Ts   v 2  
This correlation is valid for pressures of 100–9000 kPa, mass fluxes of 50–4500 kg.m-2s-1,
qualities of 0.15–0.40, sub-cooling of 0oC–50oC, and tube diameters of 9–12 mm.
A general correlation for post-CHF conditions in tubes has been proposed by Groeneveld (1973)
[IAEA2001]:
b
a  k g    g   
  Reg  x    1  x )    Prg Y
c d
h pcr ,ld
De    l   
. (288)
0.4
  G  De
Y  1  0.1  l  1 
  1  x  , Reg 
 g  g
For tubes, a=1.09ˑ10-3, b=0.989, c=1.41, d=-1.15. This correlation is valid for equivalent
diameters of 2.5–25 mm, pressures of 6.8–21.5 MPa, mass fluxes of 700–5300 kg/m2s, heat
fluxes of 120–2100 kW/m2, and qualities of 0.1–0.9.
Post-dryout look-up table
The post-dryout bundle look-up table method appears to be a suitable approach for the
following reasons:
i. Relatively simple to use;
ii. Provides correct asymptotic and parametric trends;
iii. Uses the most universal method with the best overall fit to the fully developed film
boiling data base; and
iv. With embedded modifications for complex geometries, it can be used to account for
CANDU fuel bundle effects such as geometry and spacer devices.
Table 12 Sample of PDO look-up table [IAEA2001]
Xe ΔTW = ΔTW = ΔTW = ΔTW = ΔTW = ΔTW = ΔTW = ΔTW = ΔTW = ΔTW = ΔTW =
50 K 100 K 200 K 300 K 400 K 500 K 600 K 750 K 900 K 1050 K 1200 K
-0.10 1.5664 1.504 1.474 1.358 1.177 1.134 1.280 1.395 1.532 1.705 1.833
-0.05 1.414 1.311 1.275 1.194 1.026 0.981 1.131 1.210 1.356 1.499 1.625
0.00 1.257 1.164 1.106 1.071 1.006 1.048 1.120 1.198 1.283 1.369 1.456
0.05 1.513 1.393 1.223 1.126 1.141 1.270 1.273 1.352 1.371 1.397 1.430
0.10 1.654 1.547 1.318 1.264 1.361 1.476 1.459 1.492 1.483 1.483 1.472
0.20 1.700 1.617 1.427 1.394 1.504 1.655 1.643 1.704 1.704 1.712 1.688
0.40 2.629 2.582 2.547 2.368 2.334 2.393 2.401 2.447 2.456 2.470 2.479
0.60 4.229 4.008 3.750 3.483 3.324 3.285 3.344 3.046 3.119 3.188 3.249
0.80 5.507 5.203 4.488 4.088 3.983 3.936 3.963 3.906 3.955 4.012 4.088
1.00 8.103 7.061 5.500 4.692 4.446 4.605 4.796 5.061 5.264 5.457 5.627
1.20 10.27 8.893 6.784 5.176 5.028 5.471 5.639 6.038 6.296 6.549 6.775
1.40 10.80 9.683 7.906 6.412 5.854 6.206 5.983 6.632 6.846 7.065 7.275
1.60 8.327 8.087 7.856 7.222 6.694 6.780 6.763 6.936 7.107 7.277 7.477
(2001 Film Boiling Table, P=9000 kPa, G=1500kg.m-2, H in kW.m.K-1)

A sample extract from a post-dryout look-up table is provided in Table 12. This table provides
the heat transfer coefficient for given values of pressure, mass flux, quality, and wall superheat.
Current look-up tables do not yet properly account for developing flow effects, in particular
those that usually occur during accident scenarios. This effect is accounted for using specific
correction factors. The film boiling look-up table method has been used for the following
applications:
i. As a normalized database for validation of film boiling models; and
ii. As an alternative to certain film boiling models which cover only limited ranges of flow
conditions.
7.6.5.3 Post-dryout heat transfer R&D in CANDU fuel bundles
Prediction of post-dryout heat transfer in fuel bundles, particularly film boiling heat transfer, is
much more complex than that of CHF. Aside from requiring a fourth parameter in the look-up
table (“heat flux”), non-equilibrium effects must also be considered, especially in the region
just downstream of the quench, near flow obstructions, and at low flows. Fuel bundles are
equipped with bundle appendages (as in CANDUs) or grid spacers. These appendages have a
CHF- and heat transfer-enhancing effect and also a de-superheating effect, thus reducing the
degree of disequilibrium. They can also create multiple quench fronts.
Many experimental tests have been conducted with regard to post-dryout heat transfer to
investigate phenomena, develop correlations, and test prediction methods. For such
experiments, well-instrumented fuel bundles are required. Figure 132 show a cross section of a
fuel bundle with 37 elements at the dryout plane and the appendages plane. A number of fuel
element simulators were equipped with movable temperature thermocouples. For each of the
instrumented fuel elements, using different colour codes, dry-patch spreading is indicated by a
circular line around the element for each overpower test. Different behaviour is evident for
different fuel elements, depending on the flow and heat transfer regimes that occur. These
experiments provide very useful information about post-dryout heat transfer in a fuel bundle.
The objective of these experiments is to measure temperature profiles in various flow regimes,
such as dispersed flow film boiling, and to measure dryout or dry-patch spreading around the
bundle. The sheath temperature in such a heat transfer regime rises gradually with power or
decreasing flow, i.e., it is controllable with flow condition variation. Maximum sheath
temperature is predictable and occurs at locations just upstream of appendage planes. Dry
patches are stable and propagate gradually with flow condition variations, which makes these
experiments easily controllable and measurable. In Figure 132, it can be readily seen that at
different overpower levels (i.e., power above nominal power), a different extent of wall
circumference was in dryout for different fuel pins [LEU2004]. On the left side, a test fuel
bundle is shown, with instrumented fuel pins shaded. The right side shows, using different
colours, the portions of the circumference of a fuel pin that were measured to be in dryout at
different overpower levels. Generally, the higher the power above nominal power, the larger
was the portion of the circumference that was in dryout.
Figure 132 Bundle dry patches in the dryout plane

Two types of tests were conducted, with water and with Freon. The water experiments
provided direct confirmation of post-dryout behaviour and closely resembled the phenomena
that occur in a CANDU fuel channel. However, these experiments are rather expensive. Freon
experiments use a lighter element, Freon R134a, which has a much lower saturation pressure,
and therefore these experiments are easier to perform, induce less burden on the survivability
of the materials at post-dryout conditions, and enable reuse of the experimental test
equipment. However, for these experiments, scaling techniques are needed to convert the
collected information to conditions applicable to CANDU.
Many post-dryout water experiments with fuel bundles have used fixed thermocouples
attached to the sheath. In these experiments, surface coverage was limited; details of sheath-
temperature variation downstream of appendages and spacer grids were not available; and the
extent of dry patches, in both axial and circumferential directions, could not be quantified.
Experiments have also been performed with CANDU 37-element bundles in high-pressure
steam-water flow using movable thermocouples. In this case, increased surface coverage was
achieved, i.e., more temperature measurements were collected at different places along the
fuel element. However, a limited range of powers was applied to avoid damaging the fuel
simulator.
A CANDU fuel bundle simulator for post-dryout water experiments is shown in Figure 118. It
consists of a 6 m-long full-scale bundle string of fuel elements with junctions and appendages.
In these experiments, a uniform axial power profile and a non-uniform radial power profile
were used, simulating natural uranium fuel. Sliding thermocouples were installed inside rods
at several downstream bundles in the string.
CANDU bundle post-dryout Freon experiments were also performed with movable
thermocouples inside the elements of a CANDU 37-element bundle simulator. In these
experiments, because of the lower power, temperature, and pressure conditions, fine axial and
radial movements were achievable, providing detailed temperature profiles. Freon-134a was
used as a coolant, enabling low operating power and sheath temperature and high overpower
ratios (i.e., local-to-critical power ratios). With Freon, a relatively wide range of test conditions
could be investigated because these experiments were not as expensive as the water
experiments.
Figure 133 shows a circumferential distribution of sheath temperatures in a Freon-cooled post-
dryout bundle experiment as a function of overpower percentage [LEU2004]. Nominal power
means 0% overpower. This figure shows the results of tests on various overpower situations,
up to 60% higher than nominal power.
Figure 133 Bundle circumferential temperature profiles

Virtually all film boiling prediction methods are based on correlations derived for tubes. Using
them to predict fuel-bundle sheath temperatures is common practice, but the following
bundle-specific factors and impacts must be considered:
i. Bundle enthalpy and flow imbalance;
ii. Heat transfer enhancement downstream of grids or spacers;
iii. Adjacent surface wetness or cooling near the pressure tube wall;
iv. Narrow gaps between elements

v. Lower wall friction in dry portions of bundles (resulting in higher flow in dry sub-
channels);
vi. Non-circular sub-channel cross-sectional shape;
vii. Presence of axial dry streaks in partially dry fuel bundles.
Reactor safety codes, particularly sub-channel codes that can model variation of flow
parameters across a fuel bundle, may account for some, but not all of these effects. When the
heat flux of a rod surface facing a given sub-channel exceeds the local CHF, both the wall-fluid
heat transfer coefficient and the wall friction factor are reduced drastically. By keeping track of
the circumferential dry-patch fraction (CDF) and the axial dry-patch length (ADL) for each rod
facing each sub-channel, the flow and enthalpy distribution as well as that of the film-boiling
heat transfer coefficient can be evaluated. This enables evaluation of the fuel temperature
distribution and prediction of the extent of fuel melting. This approach is now being
incorporated in some sub-channel codes to enable detailed prediction of the cladding
temperature distribution.
Aside from cross-sectional differences, the global and local effects of the grid or spacers on wall
heat transfer, quench behaviour, and interface mass and energy transport are usually not well
known, or at best are included by means of an empirical fix for each grid spacer configuration.
In general, (grid) spacers can have the following effects:
i. Promote rewetting downstream of the grid due to the higher turbulence level (i.e., can
induce multiple quench fronts);
ii. Act as cooling fins;
iii. Cause de-superheating of vapour;
iv. Cause an increase in the liquid-vapour interfacial area by breaking up droplets or liquid
core;
v. Homogenize flow by promoting turbulence.
7.6.5.4 Post-dryout applications and analyzes
The application of post-dryout heat transfer is focussed on the safety analysis of postulated
events in power reactors, including CANDU. More details about safety analysis of power
reactors are provided in Chapter 16, and in this section, these topics are considered only from
the perspective of heat transfer.
7.6.6 Single-phase heat transfer to vapour

Single-phase heat transfer to superheated steam is important because it provides an
asymptotic value for film boiling heat transfer for cases when the actual quality approached
100%. A number of tube-based correlations have been proposed; all are of the Dittus-Boelter
type and give similar predictions. The following equation is frequently used.
Miropolskiy (1975) equation [IAEA2001, DEL1981, BER1981]:
Nu g  h g D /  g  0.028  Re g0.8 Prg0.4   w /  g  , (289)

1.15
where Reg=G⋅X⋅D/μg and ρw is vapour density at wall surface temperature. This correlation is
valid for pressures of 4–22 MPa, mass fluxes up to 0.42 Mgˑm-2s-1, ρl/ρg=0.5–0.9, and Re=105–
106.
7.6.7 Problems
1. Explain the most important aspects of the three approaches of calculating reactor
dryout margins: by CHF ratio, by critical CHF power ratio, and by critical power ratio.
Provide comments on the applicability of these three approaches.
2. Describe the three main mechanisms that lead to dryout in forced flow in horizontal
heated pipes. With respect to these mechanisms, provide explanation on the difference
between heated tubes and fuel channels with fuel bundles.
3. Provide explanation about the difference in the following terms boiling crises CHF,
dryout, burnout, departure from nucleate boiling, and post-dryout in fuel bundles.
Provide examples of conditions where these terms are applicable.
4. Explain the most important parameters that affect CHF, and explain the parametric
trends that are used to gain appropriate understanding of the dryout in heated pipes
and fuel bundles.
5. Explain the main transition boiling and film boiling types in vertical and horizontal pipes
and channels, list the heat transfer modes and provide explanation on the importance
of these modes.
6. A PWR reactor uses UO2 cylindrical fuel pellets, 12.7-mm diameter, with a helium-filled
gap of 0.075 mm, and Zr cladding of 0.762 mm. The fuel rods are arranged in a square
lattice, with a pitch of 1.8 cm. At a particular section, the bulk water temperature and
velocity are 270°C and 4.5 m/s respectively. The volumetric heat generation at this
location is 5 x 104 kW/m3. Calculate (a) the convective heat transfer coefficient using
the Dittus-Boelter equation, and (b) the minimum system pressure so that no boiling
occurs in the film.
7. Calculate the local CHF for vertical upward flow of water inside a uniformly heated tube
(0.012 m ID) at the following conditions: pressure 8200 kPa, mass flux 5000 kg.m-2.s-1,
and quality 0.175. Use the most appropriate correlation for these flow conditions.
8. Using the most appropriate correlation calculate the post-dryout wall temperature for a
vertical water-cooled tube of 0.015 m inside diameter. The following parameters are
given: inlet temperature 250oC, inlet pressure 9 500 kPa, mass flux 3961 kg.m-2.s-1, local
pressure 9400 kPa, local heat flux 1500 kW/m2, local quality 0.225.
9. Evaluate the dryout power for a 3658-m-long tube of 0.0095 m ID, heated with a
symmetric-cosine heat flux. The flux profile has been discretized as
Length 0.05 0.20 0.40 0.60 0.81 1.01 1.21 1.42 1.62 1.82
0.0
(m) 1 3 6 7 3 6 9 2 6 9
qloc/qavg 0.3 0.37 0.45 0.66 0.86 1.03 1.17 1.27 1.33 1.38 1.35
Length 2.03 2.23 2.43 2.64 2.84 3.04 3.25 3.45 3.60 3.65
(m) 2 5 8 2 5 8 1 4 7 8
qloc/qavg 1.38 1.33 1.27 1.17 1.03 0.86 0.66 0.45 0.37 0.30
More nodes can be introduced to improve the accuracy of the predictions of dryout
location and power. Theoretically, an infinite number of nodes should be used for a
smooth profile. For a stepped profile, the number of nodes corresponds to the number
of steps since dryout generally occurs at the end of a step.
The flow conditions are: pressure 6930 kPa, mass flux 2050 kg.m-2.s-1, and inlet
subcooling 280.75 kJ/kg.
10. A pressure tube, with inside diameter of 103.86 mm, contains twelve 37-element
bundles of the following dimension: bundle length: 495.3 mm, and fuel element outer
diameter 13.06 mm.
Total power generated over the bundle string is 6.3 MW, and the axial power profile is
uniform. The channel is cooled with a flow of light water at the following conditions:
outlet pressure 9 MPa, mass flow rate: 17 kg/s, inlet fluid temperature 265C.
Evaluate dryout power ratio for this channel, and surface temperature distribution
along the bundle string at an overpower of 10% (overpower is defined as (String-
Power/Dryout-Power – 1)*100%).
Previous research findings appropriate to this problem:
 Heat-transfer coefficient at nucleate boiling for the 37-element bundle is 120
kW/(m2oK) at the pressure of 9 MPa, mass flux of 5 Mg/(m2 s) and quality of
20%.
 Heat-transfer coefficient at single-phase liquid flow is about 50% of that at
nucleate boiling.
 CHF in the 37-element bundle is 20% lower than that in tubes based on cross-
sectional average flow conditions.
 Heat-transfer coefficient at fully developed film boiling for the 37-element
bundle is the same as that for tubes.
 Overall loss coefficient is 1.7 for a 37-element bundle inside a pressure tube of
103.86 mm.
 The homogeneous equation is the most appropriate for two-phase multiplier.
Basic assumptions:
 Nucleate boiling heat transfer coefficient does not vary with flow conditions.
 Fully developed film-boiling heat transfer is reached once local heat flux exceeds
CHF.
 Equilibrium flow in the channel.
 Pressure drop due to acceleration is negligible.
11. In the middle of a reactor channel of a PWR reactor the bulk coolant temperature of the
turbulent coolant flow is 310oC at pressure of 14 MPa. The heat transfer coefficient in
nuclear boiling is 4.44 MW/m2 oK. Calculate the minimum heat flux at that location to
initiate nucleate boiling.
12. Water enters a heated tube of 2.5 cm diameter at 12 MPa and velocity of 2.5 m/s. The
linear heat load in the tube is constant at 100 kW/m. Calculate the distance in the tube
at which nucleate boiling starts, and distance at which it converts to bulk boiling.
Compare the heat transfer coefficients for forced convective heat transfer with these
for nucleate boiling and bulk boiling.
13. A pipe of 15 mm diameter and 3.5 m length is heated with a constant heat flux, and
total heat load of 200 kW (appropriately insulated, i.e., no heat losses). The pipe is
cooled by water at 150oC and pressure of 2 MPa, which flows through the pipe with
enters into the pipe with a velocity of 3 m/s. Calculate the mass flux of the water, and
the minimum ratio of the heat flux in the pipe and the critical heat flux.
14. Boiling water reactor operates at average pressure in the core of 8 MPa, and the
recirculation ratio of 7:1. The average velocity of the two-phase mixture in the core is 8
m/s. Calculate the critical heat flux at the exit of the reactor core channel.
15. List and explain the most important geometrical factors that have significant impact on
CHF in horizontal fuel channels.
7.7 Critical flow

Consider a situation in which fluid discharges from a container at high pressure to a container
at lower pressure, or to ambient at lower pressure (back-pressure). When the back-pressure is
reduced below a constant upstream pressure in a flow system, flow begins because of the
pressure gradient established in the connecting channel between the upstream container and
downstream towards the channel exit. The flow increases as the back-pressure is reduced
further. When the back-pressure is sufficiently reduced to cause the flow velocity at the
channel exit to be equal to the speed of sound at the channel exit temperature and pressure,
the mass flow at the exit attains a maximum value. Further reduction in the back-pressure
results in no further increase in mass flow rate; the flow remains at the maximum value and is
referred to as “critical”. The channel, and the flow, are also called “restricted” or “chocked”.
Because the back-pressure is less than the pressure resulting in critical flow, free fluid
expansion can occur between the exit (at exit pressure) and the ambient (at back-pressure)
outside the channel, and the flow takes on a paraboloid shape in this region. This situation can
be observed by many flow and measurement systems.
The critical flow phenomenon occurs in both single- and two- phase flow when the velocity of
the two-phase mixture is controlled by its upper limit, i.e., the sonic velocity of the mixture.
That velocity is a very strong function of quality, varying from ~300 m/s for single-phase steam
to ~1400 m/s for single-phase liquid water.
In nuclear reactors, this phenomenon is important for safety analysis of both boiling and
pressurized systems. A break in a primary coolant pipe causes two-phase critical flow in either
system because even in a pressurized reactor, the reduction in hot coolant pressure from about
10 MPa to near-atmospheric causes flashing and two-phase flow. This kind of break results in a
rapid loss of coolant and is considered to be a credible design basis accident in power reactors
built to date. This is further explained in Chapter 16.
An evaluation of the flow rate in critical two-phase systems is therefore important for designing
emergency cooling systems and for determining the extent and causes of damage in accidents.
Further details on modelling critical flow in nuclear reactor primary heat transport systems are
provided in Chapter 7.
7.8 Water hammer

Water hammer is an important phenomenon that requires attention by thermal-hydraulic
designers and plant operators. This phenomenon can result in equipment damage and can
affect the performance or functionality of certain equipment. Therefore, it is important to
have a good understanding of this phenomenon so that appropriate prevention, mitigation,
and accommodation measures can be undertaken to reduce the risk of its occurrence in
nuclear power plant cooling systems.
Detailed coverage of the water hammer phenomenon and its prevention, mitigation, and
accommodation are beyond the scope of this textbook, but appropriate information can be
found in the literature [EPRI1996] and in the suggestions for further reading at the end of this
textbook. This section reviews basic information about water hammer and provides
elementary guidance to thermal-hydraulic engineers and designers.
7.8.1 Types of water hammer

Water hammer events have been observed and reported widely in PWR, BWR, CANDU, and
other nuclear power plants in various coolant or support systems. These events have provided
valuable information on the types and strength of the water hammer effect, which has been
used to develop defensive strategies.
Water hammer events can be classified into two broad groups: water hammer events in single-
phase liquid or gas, and two-phase water hammer events.
Most water hammer events in nuclear power plants fall into the first group and are caused by
rapid acceleration or deceleration of single-phase liquid flowing in a piping system. Rapid
acceleration can be created by rapid valve closure, which results in pressure waves that
propagate in the piping system at sonic velocity. Figure 134 shows two examples of single-
phase water hammer; Case A shows rapid valve closure causing water hammer. Case B shows
check-valve action downstream of a pump triggering water hammer [EPRI1996].
a) Fast valve closure b) Pump check valve closure

Figure 134 Rapid valve operation
The necessary physical and mechanical conditions required for water hammer to occur exist all
the time in nuclear power plants. Most nuclear power plants have design and operational
provisions to prevent severe water hammer from occurring, but nevertheless, in unanticipated
situations, water hammer events do occur. Usually water hammer transients are typically
caused by pump start, pump trip, control or isolation valve operation, check valve closure,
safety or relief valve operation, main steam turbine trip, and filling of empty systems. All these
normally occur in single-phase flow conditions and can be easily predicted by conventional
water hammer analysis techniques.
Two-phase water hammer events can be of several types. These events are more complex to
analyze, have a strong impact on pump operation, and can cause significant damage to pipe
components and pipe supports. Six transient scenarios lead to water hammer events with
various probabilities and strengths, with most of them related to steam condensation-induced
water hammer.
1. Flow of sub-cooled water with condensing steam in a vertical pipe (Figure 135)
This type of water hammer is a typical BWR case in which, during a large LOCA event,
steam lines discharge water into the suppression pool. If the valve is quickly closed or
reduced significantly, a steam pocket remaining in the steam line can condense rapidly,
which pulls water from the suppression pool towards the valve, thus creating water
hammer [EPRI1996, POP2014].
Figure 135 Flow of sub-cooled water with steam condensing in vertical pipe (water cannon)
2. Counter-current flow of steam and water in a horizontal pipe (Figure 136)

This situation occurs in PWR feedwater lines and in CANDU feeders [POP2014]. In
horizontal stratified flow, a large interface area exists between sub-cooled water and
steam. Therefore, the steam flow is subjected to fast condensation, which induces an
increase in steam flow from the upper pipe. Under certain conditions, the counter-
current flow can transition to slug flow and create a steam pocket. When the steam
pocket collapses as a result of condensation, severe water hammer can occur. Long
feeder sections are particularly vulnerable to steam trapping and this type of water
hammer. The advanced CANDU design addresses this problem by having feeders
slightly tilted, so that an escape route for steam is provided.
Figure 136 Counter-current flow of steam and water in a horizontal pipe

3. Pressurized water entering a vertical pipe filled with steam (Figure 137)
This type of water hammer can occur in any of the PWR and BWR reactors with refilling
of a steam-filled pipe either from the bottom or from the top [EPRI1996]. If pressurized
sub-cooled water enters a pipe filled with steam in a piston flow regime, it can create
steam trapping. The filling rate for this situation is determined by the inertia of the
liquid and the pressure induced by the pump or other refilling device. Water hammer
occurs in this situation if the refilling rate is higher than the bubble rise velocity for the
top refilling case. This type of water hammer can also occur in CANDU reactors.
Figure 137 Pressurized water flowing in a vertical pipe filled with steam
4. Hot water entering a low-pressure line (Figure 138)
This type of water hammer can occur in PWRs, BWRs, and CANDUs. In this case, hot
water flowing down the pipe will flash and create a pressure wave flowing upstream,
thus creating water hammer on the upstream valve. The cold part of the sub-cooled
water that was initially outside and below the pressure vessel can create water hammer
on the downstream valve as it enters the steam-filled part of the pipe [EPRI1996].
Figure 138 Hot water entering a lower-pressure line

5. Water slug driven by steam (Figure 139)
This type of water hammer can occur in a piping system that collects condensate
upstream of a closed valve or can collect condensate in certain portions of the piping
systems that are normally filled with steam. Water hammer can occur when the valve
opens and this steam drives the water slug at high speed. The slug can be accelerated
significantly and can create a water hammer effect in a restricted part or at the closed
end of the pipe [EPRI1996].
Figure 139 Steam-driven water slug in piping system
6. Water filling a voided pipe line (Figure 140)

In this situation, as a pump trips, water from the vertical section recedes downwards
and creates an empty vacuum-filled space above. When the pump restarts, it will push
water to refill the pipe (and perhaps condense some residual steam), thus creating a
significant water hammer [EPRI1996].
Figure 140 Filling of voided piping system
7.8.2 Analytical models and computer codes for water hammer analysis
Water hammer assessment is performed using analytical models and computer programs
designed for this purpose. A detailed discussion of this topic is beyond the scope of this
textbook, and hence this chapter provides only a brief summary.
The analytical models consist of governing differential equations for transient single-phase or
two-phase flow. Usually, the flow is assumed to be one-dimensional because it describes flow
in piping systems. The governing differential equations are solved by appropriate numerical
techniques. Then a network model is created in which physical representations of various
system components are introduced as boundary conditions. Usually, a number of constitutive
correlations are used to cover system-specific phenomena.
Usually, analytical models for liquid, gas, or two-phase mixtures are based on rigid and elastic
column theories of the piping system. It is important to recognize that the severity of water
hammer is proportional to the level of rigidity of the piping system. Therefore, an increased
level of piping elasticity reduces the consequences of water hammer because elastic pipe
systems can absorb the energy of the pressure waves traveling through the system.
Software packages vary in complexity depending on the processes modelled. The more
sophisticated packages may have one or more of the following features:
 Multiphase flow capabilities;
 An algorithm for cavitation growth and collapse;
 Unsteady-state friction: the pressure waves dampen as turbulence is generated and due
to variations in the flow velocity distribution;
 Varying bulk modulus for higher pressures (water becomes less compressible);
 Fluid-structure interaction: the pipeline reacts on the varying pressures and causes
pressure waves itself.
Most computer programs that can model water hammer are the same as those used to model
general piping networks. However, computer codes may not be completely successful in
modelling all types of water hammer, particularly condensation-induced water hammer, which
is quite challenging. The RELAP5 code [REL2010] is an example of such a network code that can
model certain types of water hammer for PWRs and BWRs. For CANDU reactors, the CATHENA
computer code [CAT1998, CAT2000, CAT2005] has been validated for use in analysis of certain
types of water hammer.
7.8.3 Diagnostics and assessment of water hammer

Diagnostic and assessment techniques for water hammer consist of systematic review of a
thermal-hydraulic network in terms of vulnerabilities to and consequences of water hammer.
The assessment principles are based mainly on lessons learned and root cause analysis.
Nuclear power plant design features and operating characteristics require specific water
hammer inspection processes.
The general diagnostic analysis consists of three major steps: study of deformation evidence,
determination of water hammer source, and determination of root causes. Each of these steps
consists of several sub-steps as follows:
a) Collect evidence of water hammer impact by field inspections, interviews, and review of
operating conditions;
b) Identify pipe network characteristics that may lead to water hammer;
c) Determine water hammer consequences by identifying pipe movement, pipe
deformations, and pipe support deformations;
d) Determine the piping or other thermal-hydraulic equipment initial conditions (for
example, water-filled or steam-filled piping);
e) Determine the source of water hammer by identifying the origin point of the pressure
wave or water slug;
f) Compare the operating conditions and consequences of the various possible water
hammer mechanisms;
g) Determine the root causes; and
h) Define possible corrective action in design space and/or procedural space.
The review of systems and equipment susceptibility to water hammer usually consists of the
following activities:
i. Collection of plant data to establish plant-specific thermal-hydraulic network
configurations and identify component features and functions;
ii. Determine industry practices with respect to thermal-hydraulic network testing,
maintenance, and operating procedures that result in water hammer vulnerabilities;
iii. Perform plant assessment with respect to prevention, diagnosis, and assessment of
water hammer;
iv. Review plant-specific thermal-hydraulic system operating procedures and designs with
respect to water hammer occurrence;
v. Document the findings of the above assessment; and
vi. Develop recommendations for improving plant-specific design features or operating
procedures to reduce the probability of water hammer.
7.8.4 Prevention, mitigation, and accommodation of water hammer

This section provides general high-level information on measures taken to prevent, mitigate,
and/or accommodate certain types of water hammer events in nuclear power plants. The
measures that can be taken can be broadly grouped into design measures and operational
(procedural) measures.
Prevention of water hammer is the best option when and where possible. It includes design
and operational measures aimed at preventing or significantly reducing the water hammer
hazard. However, this option may not be available in many situations, especially in thermal-
hydraulic equipment operating in a transient following an accident.
Mitigation of water hammer includes design and operational measures to reduce the impact of
water hammer in cases where it cannot be prevented. This technique is mostly used in
situations in which prevention is not possible.
If a water hammer hazard cannot be eliminated by design or operational procedure, the next
step is to accommodate the residual impact that cannot be mitigated. A number of techniques
can be used to achieve this, such as strengthening and reinforcement of pipe support systems,
or designing piping systems that take water hammer loads into account.
General recommendations for reducing water hammer hazard are the following:
 Avoid direct contact of steam with sub-cooled water in a horizontal pipe (this is
particularly important for CANDU feeder horizontal runs);
 Avoid opening a valve adjacent to a steam pocket (identified or suspected), or ensure
that the valve opens slowly to provide an escape path for the steam pocket;
 Open isolation valves slowly where stagnant hot water coexists with hot water in the
same line upstream and downstream of the valve;
 Inspect condensate removal systems to avoid accumulation of a water slug in the steam
line;
 Inspect check valves routinely, especially those located on a pump discharge line, to
ensure that the valve is not stuck open;
 Avoid starting a pump to fill a long line at significantly higher elevation, which could be
voided for various reasons (such as leaking valves).
Specific recommendations for reducing each type of water hammer are provided in the
following points [EPRI1996].
1. Rapid valve operation
Rapid valve actuation is defined as rapid closure of a stuck-open check valve or
malfunctioning of a valve actuator. Hence, any system with this type of valve is at risk
of experiencing this type of water hammer.
Prevention or mitigation by design involves valve modifications to prevent this type of
valve operation. Mitigation or prevention by procedure aims to avoid valve cycling in
systems with significant flow, or perhaps uses a special monitoring capability to sense a
pressure difference across a valve and to use it as a signal for actuating the valve.
2. Flow of sub-cooled water with condensing steam in a vertical pipe
This water hammer phenomenon is unique to steam turbine exhaust piping that
discharges into a body of water. This happens when reduced steam supply to the
turbine traps steam in the exhaust piping.
The only way to prevent or mitigate this type of water hammer is to install a vacuum
breaker with sufficient capacity to replace the condensing steam in the exhaust piping
from a tripped turbine.
3. Counter-current flow of steam and water in a horizontal pipe
This type of water hammer happens in steam generators when a horizontal feedwater
pipe is slowly filled in counter-current flow. It can also happen in horizontal sections of
CANDU feeders when injecting cold ECC water.
One design option for mitigation is to shorten the horizontal portions of piping systems,
or to have the horizontal piping slightly inclined. A procedural modification may involve
better control of the water flow to affected areas.
4. Pressurized water entering a vertical pipe filled with steam
This type of water hammer can happen unexpectedly due to valve leakage, which can
lead to accumulating hot water pockets (which could quickly flash or condense) in the
closed parts of piping systems. This kind of water hammer can also occur if high-
pressure water is pumped into a system filled with steam (such as injection of ECC into
steam-filled CANDU headers or feeders).
One mitigation design change involves keeping the system full of water and preventing
valve leakage. In addition, installation of void detection devices at places where valve
back-leakage may occur can be a solution. A procedural modification may involve
venting steam pockets before injecting water or starting pumps.
5. Hot water entering a low-pressure line
This type of water hammer is typical of heater drain systems in cases where a reservoir
of hot water is being discharged into low-pressure piping.
A possible preventative design change is to modify the control system to avoid cyclic
dumping of hot water. Installing a flow control valve with slowly increasing flow can
also address this problem. A procedural change may involve placing a high-pressure
feedwater heater back in operation without inadvertently actuating the emergency
dump system.
6. Water slug driven by steam

This type of water hammer can happen in pipes in which condensate accumulation
occurs upstream or downstream of a closed valve.
A design fix for this water hammer is to ensure proper inclination and condensate
drainage from piping systems. Pipes subject to sagging could also be better supported
to avoid sags that can collect condensate. A procedural change could involve slowly
initiating steam flow after opening a closed valve to ensure gradual acceleration of the
condensate pocket.
7. Water filling a voided pipe line
This type of water hammer happens typically in long piping systems with an elevation
change of 10 m or more. In this situation, stopping the water flow can result in
formation of a vacuum in the upper parts of the pipe. Restarting the flow leads to
refilling the pipe and to interaction of water and steam surfaces.
A design fix for this kind of water hammer is to keep the system filled with water.
Installing a vacuum breaker can also help. A procedural improvement could involve
keeping the pump discharge valve closed and then opening it slowly as the pump is
restarted.
7.8.5 Problems
1. Explain the reasons and consequences of water hummer in thermal-hydraulic equipment.
Provide examples of possible water hummer situations in the CANDU heat transport
systems, and indicate the strategies for mitigating the effects.
2. List the types of water hummer in CANDU heat transport system, and for each type list
and explain the strategy for avoiding or mitigating it.
7.9 Natural circulation

New generations of nuclear reactor designs under development rely on passive systems,
including use of natural circulation. All Generation IV reactors are required to make use of
natural circulation as a heat transfer mechanism for removing heat from the reactor core, for
some designs even under normal operating conditions. However, currently operating reactors
are also required to use natural circulation to remove decay heat in accident situations when
forced circulation in the core is lost.
Natural circulation occurs in many different situations, such as a heater immersed into a fluid;
an open flame in air; a chimney-driven fire; a surge of hot water into a pool of cold water; a
heat source and heat sink connected by piping and located at different elevations; hot steam
and cold air mixing in a containment building. In all these situations, natural circulation plays
an essential role in transferring heat from one part of a system to another, or between
systems.
7.9.1 Natural circulation phenomenon

Natural circulation will occur in a reactor primary loop in the absence of pumped flow
whenever buoyant forces caused by differences in loop fluid densities are sufficient to
overcome the flow resistance of loop components (steam generators, primary coolant pumps,
etc.). Fluid density differences occur as a result of fluid heating in the core region (causing the
liquid to become less dense) and fluid cooling in the steam generators (causing the fluid to
become denser). Figure 141 shows the principles of natural circulation in a closed loop with a
heat source located at a low elevation and a heat sink at a high elevation connected by piping.
On the right side of this figure, a similar loop is shown in two-phase operation with a steam
drum extracting steam and adding feedwater to the loop. This figure is similar to the situation
in a nuclear power plant. Note that in this figure, two-phase flow is indicated out of the
reactor, which is typical of BWR and CANDU reactors.
Figure 141 Principles of natural circulation

The buoyancy forces resulting from these density differences cause fluid to circulate through
the primary loops, providing a means of removing core decay heat. Depending on the primary
loop fluid inventory, natural circulation consists of three distinct cooling modes: single-phase,
two-phase (liquid continuous), and reflux condensation (or boiler-condenser mode for once-
through steam generators). Progression from the single-phase mode through the two-phase
and reflux condensation modes occurs as primary system liquid inventory decreases, as shown
in Figure 142 [IAEA2005, IAEA2012, POP2012a]. Natural circulation flow in a sub-cooled
primary heat transport loop (such as in PWRs) is driven by temperature-induced density
gradients, enhanced by a thermal centre elevation difference between the hot (core) and cold
(steam generator) regions in the primary loop. This density gradient produces a buoyancy
force that drives the natural circulation flow. Hence, single-phase natural circulation is the flow
of an essentially sub-cooled primary liquid driven by liquid density differences within the
primary loop.
Figure 142 Various types of natural circulation

In a reactor where boiling occurs in the primary heat transport system, vapour generated in the
core enters the hot leg and flows along with the saturated liquid to the steam generator, where
at least some of the vapour is condensed. Hence, density gradients are affected in two-phase
mode, not only by temperature differences, but also by voids in the primary loop. In both
single-phase and two-phase natural circulation, the mass flow rate is the most important heat
removal parameter.
In reflux condensation, the vapour generated in the core flows through the hot leg, is
condensed in the steam generator, and flows back to the core as a liquid. In this mode, the
loop mass flow rate has a negligible effect because the primary heat removal mechanism is
vapour condensation.
The three modes of natural circulation are distinguishable based upon characteristic mass flow
rates, loop temperature differences, and basic phenomenological differences. The dominant
heat transfer mechanism in single-phase natural circulation cooling is convection, making the
loop flow rate the most important parameter governing heat removal. Heat generated by the
core is transported away from the reactor vessel through the hot leg to steam generators (heat
sink) by means of the sub-cooled primary liquid.
Advantages of natural circulation

a) Elimination of pumps. Elimination of the primary circulating pumps in some Generation
4 designs not only reduces capital, operating, and maintenance costs, but also
eliminates all safety issues associated with circulating pump failure.
b) Better flow distribution. Use of pumps can cause misdistributions of pressure in the
headers, leading to misdistribution of flow in the parallel channels. Moreover, using
pumps requires orificing to provide flow in the fuel channels corresponding to each
channel’s power. Natural circulation provides a natural channel flow distribution.
c) Flow characteristics. In a natural circulation system, flow increases with power,
whereas in a forced-circulation two-phase system, flow decreases with increasing
power. This has specific advantages in steam generating power plants.
d) Safety aspects. Because natural circulation is based on a natural physical law, it is not
expected to fail, unlike fluid-moving machines such as pumps. This aspect of natural
circulation has enabled its application in many safety systems. In all current designs of
nuclear power reactors, natural circulation is a backup for decay heat removal in the
event of a pumping power failure.
e) Simplicity. Because of the need to minimize pressure losses to enhance flow rates,
designers of natural circulation systems tend to eliminate all unnecessary pipe bends,
elbows, etc. The result is a system with a simple piping layout that can be fabricated in
the factory.
Challenges of natural circulation
a) Low driving force. One of the drawbacks of natural circulation systems is that their
driving force is low. The most straightforward way to increase the driving force is to
increase the loop height, which may not be economical. In addition, use of tall risers
can make natural circulation systems slender in structure and may raise seismic
concerns. Due to these reasons, the incremental height of natural circulation systems
compared to the corresponding forced circulation systems is often limited to less than
10 m.
This challenge has a specific impact on channel-type reactors such as CANDU. As
explained in this section, at low driving force in a CANDU, certain channels in the core
may not sustain forward flow, i.e., channel flow direction cannot be accurately
estimated.
b) Low system pressure losses. With low driving force, the only way to obtain reasonably
large flow rates is to design for low pressure losses. There are several measures to
achieve low system pressure losses:
i. Use of large-diameter components and pipes. The most straightforward way to
reduce pressure losses is to use large-diameter components. However, this also
results in increased cost and enhanced system volume, both of which have
economic and safety implications.
ii. Simplified system. Generally refers to the simplified piping and equipment
layout of the system, such as minimization of U-bends, elbows, loop seals, etc.
This not only results in a simplified system, but also results in low pressure losses
and prevents phase separation from promoting natural circulation flow.
iii. Elimination of components. An example in this case is the possible elimination
of mechanical separators.
c) Low mass flux. Low driving force and the consequent use of large-diameter
components result in low mass flux in natural circulation systems compared to forced
circulation systems. With low mass flux, the allowable maximum channel power is
lower, leading to larger core volume than in a forced circulation system of the same
rating.
d) Instability effects. Although instability is common to both forced and natural circulation
systems, the latter are inherently less stable than forced circulation systems. This is
attributable to the nonlinear nature of the natural circulation phenomenon, where any
change in the driving force affects the flow, which in turn affects the driving force,
possibly leading to oscillatory behaviour.
e) Low pressure and low flow regime. In natural circulation systems, the flow rate is a
strong function of power and system pressure. Moreover, the flow is stagnant when
the reactor power is zero during initial start-up. The operating conditions of natural
circulation systems can fall into the low-power, low-flow regime, where validated
thermal-hydraulic relationships are not readily available.
f) Specification of a start-up and operating procedure. It is well known that most boiling
systems exhibit instabilities at low pressures and low qualities. Therefore, natural
circulation reactors must be started up from a stagnant low-pressure, low-temperature
condition. During the pressure and power raising process, passing through an unstable
zone shall be avoided because instability can cause premature CHF occurrence.
g) Low CHF. The basis for thermal margin is the CHF, which depends on geometric and
operating parameters. Because flow in natural circulation reactors is less, designers of
these reactors tend to use the maximum allowable exit quality to minimize their size.
This means that the CHF value of the reactor tends to be significantly lower than in the
forced circulation case. This calls for several measures to increase CHF.
As shown on the left side of Figure 141, due to the difference in densities between the vertical
legs, a pressure difference is created between stations ‘a’ and ‘b’, which is the cause of the
flow. At steady state, the driving buoyancy force is balanced by the retarding frictional force,
thus providing a basis for flow rate estimation using the following equation [IAEA2005,
IAEA2012]:
R W 2
gH  c  h   , (290)
2
where g (m/s2) is acceleration due to gravity, R (m-4) is hydraulic resistance, and W (kg/s) is
mass flow rate. Hence, the mass flow rate induced by natural circulation is given by:
1
 2  A 2   c   h  gH  2
W   . (291)
 R 
Natural circulation flow is enhanced by increasing the loop height H and the density difference
(ρc-ρh), as well as by decreasing the hydraulic resistance R.
By introducing loop resistance models, the above equation can be further expanded into the
following form:
1
 2
 
2  A2  c   h  gH  .
W  N (292)
  fL  
   i i  Ki  
i
 i 1  Di  
The natural circulation phenomenon can be classified in terms of the state of the working fluid,
the interactions with the surroundings, loop geometry, body force field, system inventory, and
number of heated channels. Figure 142 shows certain types of natural circulation in a
rectangular geometry.
The state of the working fluid can be single-phase, two-phase, or supercritical. The single-
phase natural circulation is important as a decay power removal mechanism in many operating
and new designs of nuclear power reactors. Two-phase natural circulation normally occurs in
BWRs and CANDUs under certain operating regimes.
In terms of interactions with the surroundings, closed-loop and open-loop natural circulation
can be distinguished. The closed loop involves only energy exchange with the surroundings,
whereas the open loop involves both energy and mass exchange.
In terms of loop geometry, natural circulation can occur in rectangular, U-tube, toroidal, and
figure-of-eight forms (the last being specific to the CANDU design). The body force field has an
important impact and can be gravitation (in most cases), centrifugal force, or both for certain
specific applications.
In terms of system inventory, natural circulation can occur in full systems (in single-phase
mode), in partially filled systems (in two-phase mode), or with boiling in reflux condensation
mode. In any of these types, natural circulation can occur in single-channel or parallel-channel
loops, as with the CANDU design.
Natural circulation in a containment building is a specific form of single-phase multi-component
natural circulation.
7.9.2 CANDU natural circulation

CANDU reactors have a very different design than LWR reactors, and therefore the
effectiveness of natural circulation must be carefully assessed. The performance of natural
circulation in a CANDU reactor is an important part of reactor safety under certain operating or
accident conditions. Therefore, a natural circulation test is performed in each CANDU plant
during commissioning to demonstrate that natural circulation works adequately and that the
computer codes predict the observed behaviour sufficiently well.

Figure 143 shows a CANDU primary heat transport system on the left side and a view of the
CANDU header on the right side [POP2012a, POP2014]. The CANDU header plays an important
role in natural circulation, particularly the outlet header because it contains flows of about 2%–
4% quality.
A CANDU reactor cooling system is arranged in a figure-of-eight loop configuration, typically
with two loops (i.e., two figure-of-eight loops). Figure 143 shows on the left side an elevation
indication for various components. The height difference between the bottom channel and the
top of the steam generators is about 22 metres. The height difference between the reactor top
channels and the bottom of the steam generators is about 3 metres. Due to the difference in
densities between the vertical branches, a pressure difference is created, which creates the
potential for natural circulation flow around the loop.
From Figure 143, it is clear that the height difference in CANDU is greater than in PWRs, and
according to Eq. (291), this enhances natural circulation. However, flow resistance in the
CANDU primary loop is greater than in LWRs because of the relatively small-diameter feeders,
channel end-fittings, and shorter fuel bundles, which according to Eq. (291), reduces the
natural circulation driving force. In addition, the CANDU core consists of many fuel channels in
parallel that have similar geometry, but are at different elevations. This will create differences
in the natural circulation driving force.
Figure 143 Natural circulation in CANDU

A CANDU header, shown on the right side of Figure 143, has multiple feeder connections at
different elevations and angles. When the header is partly voided, this geometry can result in
interesting phenomena. One of these is vapour pool-through, in which steam is sucked into the
feeder even if the feeder connection to the header is below the collapsed liquid level in the
header. Because the scope of this textbook is limited, details of this phenomenon cannot be
covered in this section.
At very low pressure drop between inlet and outlet headers, the feeder flow is low and could
possibly reverse direction in some feeders and channels. However, flow direction in the
channel is not as important as the mass flow rate. Hence, as long as there is sufficient flow in
the feeders and the fuel channel, decay heat generated in the core can be removed effectively,
regardless of the flow direction. Experiments have provided information on flow direction in
CANDU feeders and core channels under natural circulation and have demonstrated that decay
heat removal can be adequate during natural circulation for certain combinations of decay
power levels and loop inventory.
The key parameters for CANDU natural circulation are the primary coolant inventory in the
core, the secondary pressure, and the primary power. At very low powers in CANDU, two
natural circulation loops are created: one between channels and headers, and the other one
from headers to steam generators (see Figure 143). This is important because it demonstrates
that channel flow direction is not an essential parameter in determining the effectiveness of
natural circulation for decay heat removal.
Figure 144 [ING1992, POP2012b]] shows the experimentally observed onset of bi-directional
flow in a CANDU core under natural circulation. It is evident that the heat transport system
inventory plays an important role, along with channel power and pressure. Figure 145
[POP2012a, POP2012b] shows the criteria for flow reversal in a CANDU core. The upper and
lower boundaries are shown in terms of secondary-side pressure and loop-integrated void. It is
evident that at higher loop void and lower pressure, the flow reversal band is significantly
expanded.
One very specific application of natural circulation in CANDU is heat rejection by the feeders to
the containment environment. At very low powers, decay heat can be passively rejected by the
feeder pipes directly to the containment atmosphere.
The Canadian Nuclear Safety Commission issued Generic Action Item GAI 90G02 in 1990,
requesting utilities to demonstrate reactor cooling adequacy in natural circulation and address
uncertainty in the RD-14M experiments with respect to core cooling with reduced loop water
inventory in a CANDU loop (LOCA combined with loss of forced circulation). This GAI was
closed for all CANDU stations in Canada, thus demonstrating the cooling adequacy of natural
circulation in CANDU.
CANDU core cooling in the absence of forced flow (CCAFF), i.e., natural circulation, has been
extensively studied over the past 50 years. The CANDU CCAFF terminology defines the
following natural circulation modes: a) single-phase thermo-siphoning; b) two-phase thermo-
siphoning, c) intermittent buoyancy-induced flow (IBIF), and continuous steam venting.
Figure 144 Onset of bi-directional Figure 145 Criteria for flow reversal in a CANDU core
flow in CANDU
Single-phase thermo-siphoning
In this CCAFF mode, continuous heat input to the coolant raises the temperature in the fuel
channel (as shown in Figure 143). This induces a buoyancy force whereby the hot liquid rises to
the steam generator and deposits its heat. From the steam generator, a lower-temperature
(denser) liquid continues to flow down back to the core. In this case, the buoyancy forces are
sufficient to overcome the static pressure caused by the elevation of the boilers with respect to
the fuel channels. This is a continuous process and is capable of removing a substantial amount
of decay heat indefinitely as long as the loop does not drain beyond about 70%, and as long as
the heat is removed from the steam generators.
Two-phase thermo-siphoning
This is a similar phenomenon to liquid thermo-siphoning, but with continuous boiling of the
liquid, and thus steam content in the flow provides an additional buoyancy force for heat
transfer to the steam generators (resulting from the liquid-vapour density difference). It has
been experimentally demonstrated that flow is still continuous in a given direction, with a large
quantity of vapour at the channel outlet. For this thermo-siphoning mode, it is important to
maintain a certain flow rate rather than a certain flow direction. The flow rate in this type of
natural circulation depends on the amount of void in the system and the channel power.
Intermittent buoyancy-induced flow (IBIF)
If the flow rate is reduced to a very small value, or if the fluid stagnates in the channel, the IBIF
phenomenon occurs. Numerous analytical and experimental studies of IBIF have been
conducted in recent decades to address the Canadian Nuclear Safety Commission-issued
Generic Action Item GAI 90G02 on CCAFF. The closure of this GAI provided answers to many
issues related to fuel and fuel channel behaviour in a CANDU channel when the flow stops,
such as:
 First channel in which boiling occurs;
 Role of static pressure (i.e., channel elevation);

 Location in the channel where boiling occurs;
 Vapour accumulation and outflow from the channel;
 Heat transfer from uncovered fuel elements;
 Impact of void rate generation;
 Fuel sheath temperature for uncovered fuel elements.
Figure 146 shows the four steps in the IBIF cycle, which are explained below [POP2012a,
POP2014]. The IBIF cycle repeats itself, providing channel cooling when the flow is stopped.
Figure 146a shows a CANDU fuel channel in which flow has stopped and the power profile has
a certain cosine shape. As forced circulation is lost, flow almost stagnates in the channel, and
therefore heat transfer and removal from the fuel are lost. As a result, because of decay power
generation, fuel temperature rises, along with fluid temperature. After some time, the coolant
in the channel reaches saturation and thermal stratification. Steam starts to develop at the top
of the channel.
Figure 146 Intermittent buoyancy-induced flow in CANDU

Figure 146b shows a large bubble being generated because of liquid evaporation. The
uncovered fuel elements begin to heat up faster (and local heating of the pressure tube
occurs). The large vapour bubble grows outward from the centre of the channel towards the
end-fittings, and uncovered fuel-element sections stop producing vapour.
Figure 146c shows further progression of an IBIF cycle, as steam reaches and heats up the end-
fittings. Because the end-fittings have a very large mass of relatively cold material (during a
reactor shutdown, everything approaches 60°C), and it takes a while to heat it up before the
bubble can “push” itself by. This delays venting because the longer the fuel is uncovered, the
hotter it gets. Before steam can push through, several sub-cycles may occur, with steam
condensation at the end-fitting edge. The process continues until the steam creates a flow
path to the feeders.
Figure 146d shows the last part of an IBIF cycle, when the steam pushes through to the feeder
and the header. Usually steam push-through happens on one side or the other first. It does
not necessarily follow the usual flow direction. Once one side starts to “vent” (vapour rushes
out of that side of the channel), this creates a pressure force, which sucks in cold liquid from
the opposite header.
After venting is completed, the fuel channel refills with sub-cooled or saturated water from the
other header. At this point, the previous IBIF cycle is completed, and a new cycle begins.
Figure 147 shows an envelope of CANDU fuel sheath temperature histories during an IBIF cycle
[POP2012a, POP2014]. While the channel is full and sub-cooled, the sheath temperatures are
below saturation (left part of the diagram). As the channel saturates and remains saturated for
a certain time, sheath temperatures remain near saturation. Then, as bubbles are generated in
the upper part of the channel and fuel elements are exposed, the fuel sheath temperatures of
these fuel elements rise to above 400°C (during the heating period, and depending on fuel
element power). When steam venting occurs at the end of an IBIF cycle and the channel is
refilled with water from the other header, sheath temperatures fall to and below saturation
(the last part of the diagram).
Figure 147 Typical temperature histories in a CANDU channel in natural circulation
7.9.3 CANDU natural circulation experimental studies

In support of natural circulation knowledge base development, numerous experimental tests
and assessments have been conducted in Canada on CANDU-typical experimental facilities.
The most important experimental facilities in this area are:
 RD-14 natural circulation experiments. RD-14 is a full-scale, full-elevation CANDU-
typical experimental facility at AECL with two figure-of-eight passes containing one fuel
channel in each loop.
 RD-14M natural circulation experiments. RD-14M is a full-scale, full-elevation CANDU-
typical experimental facility at AECL with two figure-of-eight passes containing five fuel
channels in each loop.
 CWIT standing-start experiments. CWIT is a full-scale experimental facility containing
two full-scale fuel channels connected to inlet and outlet header at the Stern
Laboratories.
More details of these experimental facilities, experimental results, and conclusions are
provided in [IAEA2005, IAEA2012]. The experimental work in this area was performed to
improve the knowledge base and to provide data for development and validation of computer
programs and models.
7.9.4 Natural circulation calculations and predictions

Mathematical modelling of natural circulation is based on several approaches, including
empirical models based on experimental data; analytical mechanistic models based on
fundamental principles approaches that use conveniently established simplifying assumptions;
and general-purpose computer programs based on detailed application of field conservation
equations.
CATHENA [CAT2005], TUF [TUF2001], and RELAP [2010] are examples of large one-dimensional
two-fluid thermal-hydraulic computer codes used to simulate natural circulation. Recently,
more attention has been devoted to use of CFD computer models [HAS2012] for natural
circulation analysis in cases where more detailed information is needed. More details about
the use of thermal-hydraulic computer programs are provided in Chapter 7.
Various mechanistic models have been proposed in the literature to analyze natural circulation
in the CANDU reactor core and feeder system. These models use mass, momentum, and
energy conservation accompanied by drift-flux momentum equations. Often, a macroscopic
mass, momentum, and energy balance in a six-equation configuration is also used. Details of
this topic are beyond the scope of this textbook, but information can be found in the section
for suggested further reading. One example of one mechanistic model is provided below.
Algebraic equations or field equations properly solved by simplified numerical models or by
complex computer codes can be used to predict natural circulation performance under a wide
variety of geometrical conditions and technological systems. The fundamental principles
approach is based on the following assumptions: constant core inlet enthalpy; cross-section
averaged fluid properties; homogeneous two-phase flow; thermal equilibrium; form losses
dominating loop resistance. These assumptions are applied to the mass, momentum, and
energy equations for each loop component to obtain the conservation equations. The
equations are then integrated over their respective single-phase and two-phase regions to
obtain the loop balance equations.
Assuming the natural circulation loop shown in Figure 141, the transient momentum equation
can be written as:
N
 Li  dW
 A 
i 1  i  dt
 g  l  TP  H nc
. (293)
W2   1  fL   Ac   l
2
 1  fL   Ac   
2
    K        K     
l Ac2  SP  2  Dh i  Ai   TP TP  2  Dh i  Ai   

The energy balance equation for transient natural circulation can be written as:
d  eM  el 
Mloop  W  hTP  hl   qSG  qloss . (294)
dt
The equilibrium two-phase quality in the core exit can be defined as:
hTP  hl
xe  . (295)
hlg
The homogeneous mixture density is defined using the following relation:
l
TP  , (296)
 l   g 
1  xe 
  
 g 
where:
Li is the length of a loop component; K is the coefficient of local resistance;
Ai is the area of a loop component; eM is the mixture internal energy;
Ac is the area of a loop restriction; el is the liquid internal energy;
Hnc is the elevation difference driving the ̇qSG is the steam generator heat sink [kJ/s];
natural circulation ̇ loss
q is the loop heat losses;
Di is the loop element diameter; hTP is two-phase enthalpy;
f is the friction factor; hl is liquid enthalpy,
These equations can be manipulated to yield the following equations for steady-state mass
flow rate in a two-phase natural circulation loop with uniform diameter:
 2 g   r  TP H nc Qh Drb Ar2  b  l 
W ss   . (297)
p  rb N G 
The geometric contribution to the friction factor is defined as:
Lt  NSP  leff  NB
 leff  NTP
 leff 
NG    1b 2b   LO   1b 2b   LO   1b 2 b   . (298)
2 2
Dr  i 1  Di Ai  i  N SP 1  Di Ai i i  N B 1  Di Ai i 
The Reynolds and Grashoff mixture numbers are defined as:

DW Dr3 l3 h gQh z
Rem  r ss Grm  . (299)
Ar l Ar l3
In these equations:
βh is the thermal expansion coefficient leff is the effective length of a component;
Qh (°C) A is the flow area (m2);
Δz is the total heat input rate (W); µ is the dynamic viscosity (Ns/m2);
is the centreline elevation difference in r is a subscript indicating a reference
the loop between cooler and heater value.
7.9.5 Problems
1. Explain the importance of natural circulation in the CANDU primary heat transport
system, and the design decisions that help to enhance it.
2. Explain the key aspects of the Intermittent Buoyancy Induced Flow (IBIF) in CANDU
channels, list and explain the key IBIF parameters.
3. Calculate natural circulation flow in a PWR reactor primary cooling loop assuming
typical geometrical and thermal-hydraulic parameters for a PWR reactor (primary loop is
in subcooled boiling flow in normal operation, i.e., single phase liquid flow).
8 References
[AECL1978] G.A. Pon, “Evolution of CANDU Reactor Design”, AECL-6351, August 1978.
[AECL1981] “CANDU Nuclear Power System”, AECL-TDSI-105, January 1981.
[AECL1997] “Canada Enters the Nuclear Age – A Technical History of Atomic Energy of
Canada Limited”, AECL, McGill-Queen’s University Press, Toronto, 1997.
[AECL2005] “CANDU 6 Technical Outline”, AECL, September 2005.
[AECL2009a] “Enhanced CANDU 6 Technical Summary”, AECL, 2009.
[AECL2009b] “ACR-1000 Technical Description”, AECL 10820-01371-TED-001, 2009.
[AECL2010] “Enhanced CANDU 6 (EC6) Design Description”, IAEA Advanced Water-Cooled
Reactors Web-Based Report, 2010.
[AHM2009] R. Ahmadi, A. Nouri-Borujardi, J. Jafari, I. Tabatabari, “Experimental Study of
Onset of Subcooled Annular Flow Boiling”, Progress in Nuclear Energy Journal,
Vol 51, pp.361-365, 2009.
[APR2011] “Advanced Pressurized Reactor (AP1000)”, Design Control Document, Rev. 19,
Westinghouse, USNRC Website, 2011.
[BER1981] A.E. Bergles, et al., Two-Phase Flow and Heat Transfer in the Power and
Process Industries, Hemisphere Publishing, 1981.
[BIR1960] R.B. Bird, W.E. Stewart, and E.N. Lightfoot, Transport Phenomena, Wiley, New
York, 1960.
[BPR2000] “Bruce: A Safety Report”, Brice Power, 2000.
[CAR2015] “Carnot Principle”, Wikipedia, Internet, 2015.
[CAT1998] B.N. Hanna, et al., “CATHENA: A Thermal-Hydraulic Code for CANDU
Analysis”, Nuclear Engineering and Design 180 (1998) 113-131, and AECL-
1194.
[CAT2000] “CATHENA Mod 3.5c/Mod o Theory Manual”, COG-00-008, Nov. 2000.
[CAT2005] “CATHENA Mod 3.5d Theory Manual”, AECL, 153-112020-STM-001, Dec.
2005.
[CLA2015] “Clausius Statement”, Wikipedia, Internet, 2015.
[COL1972] J.C. Collier, Convective Boiling and Condensation, McGraw-Hill, 1972.
[CRA1957] Flow of Fluids Through Valves, Fittings, and Pipes, Crane, 1957.
[DEL1981] J.M. Delhaye, et al., Thermo-Hydraulics of Two-Phase Systems for Industrial
Design and Nuclear Engineering, Hemisphere Publishing, 1981.
[ELW1978] M.M. El-Wakil, Nuclear Heat Transport, American Nuclear Society, 1978.
[ELW1990] M.M. El-Wakil, Nuclear Energy Conversion, American Nuclear Society, 1990.
[EPRI1999] Nuclear Reactor Generations, EPRI document EPRI NP-6780-L, Volume 1,
1999.
[EPRI1996] Water Hammer Handbook for Nuclear Plant Engineers and Operators, EPRI
Technical Report, Palo Alto, California, 1996.
[GAR1999] B. Garland, “Reactor Thermal-Hydraulic Design”, Graduate course EP716

lectures, Engineering Physics, McMaster University, 1999.
[GLA1967] S. Glasstone, A. Sesonske, Nuclear Reactor Engineering, Van Nostrand
Reinhold, 1967.
[HAS2012] Y.A. Hassan, Computational Fluid Dynamics for Natural Circulation, Texas
A&M University, Texas, USA, for the N.IN.E Nuclear and Industrial Engineering
Natural Circulation Course, Beijing, China, 27 July–31 Aug. 2012.
[HET1982] G. Hetsroni, Handbook of Multiphase Systems, Hemisphere and McGraw-Hill,
Washington, 1992.
[IAEA2001] Thermal-Hydraulic Relationships for Advanced Water-Cooled Reactors, IAEA
TECDOC-1203, Vienna, 2001.
[IAEA2005] Natural Circulation in Water-Cooled Nuclear Power Plants, IAEA TECDOC-
1474, Vienna, 2005.
[IAEA2012] Natural Circulation Phenomena and Modelling for Advanced Water-Cooled
Reactors, IAEA TECDOC-1677, Vienna, 2012.
[IDE1996] I.E. Idelchik, Handbook of Hydraulic Resistance, 3rd Edition, Begell House, New
York, 1996.
[ING1992] P.J. Ingham, Natural Circulation Experiments in the RD_14M Multiple Channel
Thermal-Hydraulic Loop, AERCL, Pinawa, Manitoba, Canada, 1992.
[KEV2015] “Kelvin Statement”, Wikipedia, Internet, 2015.
[KNI2002] R.A. Knief, Nuclear Engineering—Theory and Technology of Commercial
Power, American Nuclear Society, 2008.
[KRI1987] V.S. Krishnan, P. Gulshani, “Stability of Natural-Circulation Flow in a CANDU-
Type Fuel Channel”, AECL, Pinawa, Manitoba, Canada, in Journal of Nuclear
Engineering and Design, Vol. 99, pp. 403–412, 1987.
[KPL2015] Kelvin-Planck Statement, Wikipedia, Internet, 2015.
[KOK2009] K.D. Kok (Ed.), Nuclear Engineering Textbook, CRC Press, 2009.
[LAH1977] R.T. Lahey, Jr., and F.J. Moody, The Thermal Hydraulics of a Boiling Water
Nuclear Reactor, ANS&AEC Monograph Series on Nuclear Science and
Technology, ANS, 1977.
[LEU2004] L. Leung, Reactor Thermal-Hydraulic Design: CHF, Post-Dryout, and Pressure
Drop Lectures, Graduate course UN804 lectures, UNENE graduate courses,
Chalk River Laboratories, AECL, 2004.
[MOO1944] L.F. Moody, “Friction in Pipes”, Transactions of the ASME, pp. 666-671, 1944.
[OPG2002] Darlington Safety Report, Ontario Power Generation, 2002.
[OPG2000] Pickering: A Safety Report, Ontario Power Generation, 2000.
[PLA2015] Planck Principle, Wikipedia, Internet, 2015.
[POP2000] N.K. Popov, U.S. Rohatgi, H.E.C. Rummens, R.B. Duffy, “Two-Phase Flow
Stability in Low-Pressure Parallel Vertical Heated Annuli”, ICONE8, Baltimore,
MD, USA, April 2–6, 2000.
[POP2012a] N.K. Popov, Natural Circulation Behaviour in PHWRs, Engineering Physics,

McMaster University, Hamilton, Canada, for the N.IN.E Nuclear and Industrial
Engineering, Natural Circulation Course, Beijing, China, 27 July–31 August
2012.
[POP2012b] N.K. Popov, PHWR Natural Circulation Experimental Program in Canada,
Engineering Physics, McMaster University, Hamilton, Canada, for the N.IN.E
Nuclear and Industrial Engineering, Natural Circulation Course, Beijing, China,
27 July–31 August 2012.
[POP2013] N. Popov, Reactor Thermal-Hydraulic Design, Graduate course UN804
lectures, UNENE Graduate Courses, 2013.
[POP2014] N. Popov, Reactor Thermal-Hydraulic Design, Graduate course EP716 lectures,
Engineering Physics, McMaster University, 2014.
[REL2005] “RELAP5 - A Computer Program for Transient Thermal-Hydraulic Analysis of
Nuclear Reactors and Related Systems”, INEEL-EXT-98-00834, September
1996.
[REL2010] “RELAP5/MOD 3.3 Code Manual”, USNRC, NUREG/CR-5535/Rev P4, October
2010.
[ROH1983] W.M. Rohsenov, J.P. Hartnett, E.N. Ganic, Handbook of Heat Transfer
Fundamentals, McGraw-Hill, New York, 1985.
[SOP1975] Y.F. Chang, “The SOPHT Program and its Applications”, Proceedings,
Simulation Symposium on Reactor Dynamics and Plant Control, May 26–27,
CRNL, published as CRNL 1369, 1975.
[SAA1966] M.A. Saad, Thermodynamics for Engineers, Prentice-Hall, 1996.
[SOP1977] Y.F. Chang, A Thermal-Hydraulic System Simulation Model of the Reactor,
Boiler, and Heat Transport System (SOPHT), CNS-37-2, Ontario Hydro,
September 1977.
[TOD1990a] N.E. Todreas, M.S. Kazimi, Nuclear Systems I—Thermal Hydraulic
Fundamentals, MIT, CRC Press, Taylor & Francis, 2011.
[TOD1990b] N.E. Todreas, M.S. Kazimi, Nuclear Systems II—Elements of Thermal Hydraulic
Design, MIT, CRC Press, Taylor & Francis, 1990.
[TON1965] L.S. Tong, Boiling Heat Transfer and Two-Phase Flow, Wiley, 1965.
[TON1975] L.S. Tong, Boiling Heat Transfer and Two-Phase Flow, Krieger, 1975.
[TON1996] L.S. Tong, J. Weisman, Thermal Analysis of Pressurized Water Reactors,
American Nuclear Society, 1996.
[TUF2010] TUF Thermal-Hydraulics Computer Program, N-REP-06631.02-10010 R00,
Ontario Power Generation, October 2001.
[WAL1969] G.B. Wallis, One-Dimensional Two-Phase Flow, McGraw-Hill, 1969.
[WEI1977] J. Weisman, Elements of Nuclear Reactor Design, Elsevier, 1983.
[WIK2015] “Steam-Water Diagrams”, Internet, Wikipedia, 2015.
[WNA2009] “Nuclear Reactor Generations”, WNA Web site, 2009.
9 Further Reading
[DUD1976] J.J. Duderstadt, L.J. Hamilton, Nuclear Reactor Analysis, Wiley, New York, ISBN
0-471-22363-8, 1976.
[GIN1981] J.J. Ginoux ,Two-Phase Flows and Heat Transfer with Application to Nuclear
Reactor Design Problems, Hemisphere, 1981.
[GOV1972] G.W. Govier, K. Aziz, The Flow of Complex Mixtures in Pipes, Van Nostrand
Reinhold, 1972.
[HSU1976] Y. Hsu, R. Graham, Transport Processes in Boiling and Two-Phase Systems,
Hemisphere, 1976.
[ISH1975] M. Ishii, T. Hibiki, Thermo-Fluid Dynamic Theory of Two-Phase Flow, Springer,
2010.
[LAM1977] J.R. Lamarsh, Introduction to Nuclear Engineering, Addison-Wesley, 2001.
[LEW1977] E.E. Lewis, Nuclear Reactor Safety, Wiley, 1977.
[ROA1976] P.J. Roache, Computational Fluid Dynamics, Hermosa, 1976.
10 Glossary
ACR Advanced CANDU Reactor MCHFR Minimum CHF Ratio
ADL Axial Dry-Patch Fraction MCHFPR Minimum CHF Power Ratio
AECL Atomic Energy of Canada Limited MCPR Minimum Power Ration
AFD Axial Flux Distribution NPSH Net Positive Suction Head
AGR Advanced Gas-Cooled Reactor NPSHA Net Positive Suction Head Available
BP Bruce Power NPSHR Net Positive Suction Head Required
BWR Boiling Water Reactor NOP Nominal Operating Power
CANDU CANada Deuterium Uranium NOP Neutron Overpower Protection
CANFLEX CANDU Flexible Fuelling NPD Nuclear Demonstration Plant
CCAFF Core Cooling in Absence of Forced Flow NPP Nuclear Power Plant
CDF Circumferential Drypatch Fraction OCR Organic Coolant Reactor
CHF Critical Heat Flux OH Ontario Hydro
CHFR CHF Ratio OPG Ontario Power generation
CHFPR CHF Power OSV Onset of Significant Void
CP Critical Power PHTS Primary Heat Transport System
CPR Critical Power Ratio PHWR Pressurized Heavy Water Reactor
DNB Departure from Nucleate Boiling PWR Pressurized Water Reactor
ECC Emergency Core Cooling RBMK High Power Channel Type Reactor
ECI Emergency Coolant Injection RCS Reactor Cooling System
EPRI Electric Power Research Institute RFD Radial Flux Distribution
GAI Generic Action Item RIH Reactor Inlet Header
HTGCR High Temperature Gas Cooled Reactor ROH Reactor Outlet Header
HTS Heat Transport System ROP Reactor Overpower Protection
IAEA International Atomic Energy Agency R&D Research & Development
IBIF Intermittent Buoyancy-Induced Flow R&M Reliability and Maintainability
LMFBR Liquid Metal Breeder Reactor SGHWR Steam Generating Heavy Water
LOCA Loss of Coolant Accident QA Reactor
LOFA Loss of Flow Accident WNA Quality Assurance
LORA Loss of Reactivity Accident WWER World Nuclear Association
LWR Light Water Reactors Water Cooled and Moderated Power
Reactor (from Russian VVER)
11 Nomenclature
A cross section area, [m2] ν fluid velocity, [m/s]
CHF critical heat flux, [kW/m2] W mass flow rate, [kg/s], work, [kJ]
cp specific heat at constant pressure, [J/(kg K)] w specific work, [kJ/kg]
cv specific heat at constant volume, [J/(kg K)] We Weber number
DH hydraulic diameter, [m] x mass quality
d diameter, [m] xth thermodynamic quality
dh hydraulic diameter, [m] Z height, [m]
dx axial derivative
Greek Symbols
EK kinetic energy, [kJ]
EP potential energy, [kJ] α void fraction, thermal diffusivity, [m2/s]
EU internal energy, [kJ] γ adiabatic exponent, SG pinch fraction
e specific energy, [kJ/kg] ΔP pressure drop, [kPa/m2]σ surface tension,
f friction factor [N/m]
G mass flux, [kg/s-m2] ηt thermodynamic efficiency
g gravitational constant, [m/s2] µ dynamic viscosity of the fluid [Pa·s],
H total enthalpy, [kJ] [N·s/m²], [kg/(m·s)]
H pump head, [m] ρ density [kg/m3]
h specific enthalpy [kJ/kg] σ surface tension, [N/m]
h convective heat transfer coefficient, τ pump torque, [Nm], friction stress, [N/m2]
[kW/m2 °K] υ specific volume, [m3/kg], kinematic viscosity,
hfg latent heat of vapourization, [kJ/(kg °K)] [m2/s]
j volumetric flux (superficial velocity) [m/s] Φ neutron flux, [neutron/(m2-s)]
K concentrated friction factor φ two-phase friction multiplier
k thermal conductivity, [kW/m °K] ω pump rotational speed
L length, characteristic linear dimension, [m]
Subscripts
l characteristic length,
typically the droplet diameter, [m] a acceleration
M, m mass, [kg] C cold, cladding, fuel sheath
ṁ mass flow, [kg/s] CL centreline
NOP neutron over power cold cold side
Nu Nusselt number CRIT critical
P perimeter, [m] DNB departure of nuclear boiling
p fluid pressure, [kPa] e equivalent
Pr Prandtl number G gap
Q Core power, [kW], volumetric flow, [m3/s] g gas, gap
Q″′ Core power density, [kW/m3] F fuel
q″′ volumetric heat generation rate, [kW/m3] f fuel, liquid
q″ surface heat flux, [kW/m2] fg latent heat of vapourization
q′ linear power, [kW/m] FL fluid
q energy generation per fuel element, [kW] FW feed water
R gas constant, [Nˑm/(kg °K)], [J/(kg °K)] H hot, homogenious
r radius, [m] Hot hot side
Re Reynolds number h hydraulic
ROP reactor over power MFB minimum film boiling
S slip ratio, S total entropy, [kJ] m mixture
s specific entropy, [kJ/kg] n vector perpendicular to surface
T temperature, [°K] P, p primary
t time, [s] S steam, saturation
t thickness, [m] SG steam generator
U steam generator heat transfer coefficient, SH sheath
[kJ/(m2-s)] TB transition boiling
U internal energy, [kJ] w wall
u specific internal energy, [kJ/kg]
V volume, [m3]
12 Acknowledgements
I gratefully thank the following reviewers for their hard work and excellent comments during the
development of this Chapter. Their feedback has much improved it. Of course, the responsibility for
any errors or omissions is entirely the author’s.
Bill Garland
Lawrence Leung
Dave Novog
Terry Rogers
Victor Snell
Dave Wright
I also thank Diana Bouchard for expertly editing and assembling the final copy.
13 APPENDIX A – Reactor Types

13.1 General evolution of reactor designs
The categorization of reactor designs into generations was originally suggested by the U.S. DOE
and only reflected the technology evolution timeline. The attributes were assigned later, based
on observed characteristics.
The nuclear industry typically categorizes nuclear reactors as follows (see Figure 148):
 Generation I – early prototype reactor designs
 Generation II – commercial power reactor designs
 Generation III/III+ – designs with significant improvements in safety and economics
 Generation IV – future designs.
Note that most currently operating reactors are of Gen II design, and reactors currently under
construction are of Gen III/III+ design.
As reactor safety and economics became an important benchmark, the Gen III/III+ reactors were
given certain attributes to characterize them more fully. Two separate sets of attributes
(requirements or guidelines) were developed by EPRI in 1999 and by WNA in 2009.
Figure 148 Reactor generations
In the EPRI documentation [EPRI1999], the following attributes are shown to be typically
possessed by a Gen III/III+ design:
1. design simplification; 6. regulatory stabilization; 12. adequate quality
2. design margins; 7. technology standardization; assurance;
3. human factors; 8. proven nature of design; 13. improved security and
4. reactor safety; 9. enhanced maintainability; sabotage protection; and
5. design basis and safety 10. enhanced constructability; 14. reduced environmental
margin; 11. improved economics; impact.
The WNA definition of the third generation [WNA2009] suggested the following attributes for
Gen III/III+ designs:
1. standard design to 4. higher availability and 9. less waste;
expedite licensing; longer operating life; 10. use of burnable
2. reduced cost and 5. reduced possibility of core absorbers;
construction time; meltdown; 11. more passive
3. simpler and more rugged 6. resistance to damage; systems;
design for easier 7. reduced radiation release; 12. load following.
operation and 8. higher burn-up;
maintenance;
Because the above two sets of guidelines overlap somewhat and expectations are not quantified,
it is difficult to compare designs in quantifiable terms. More importantly, there are conflicting
expectations, such as expecting changes to improve safety and at the same time expecting the
design to be well proven. A detailed discussion of implementation and compliance with these
objectives of current advanced design is beyond the objectives of this section, but note that
none of the modern advanced designs claims to achieve all the expectations described above.
13.2 Pressurized water reactors

13.2.1 General information
Pressurized water reactors (PWRs) are currently one of the most commercially successful reactor
types in operation; they originate from the U.S. nuclear industry. PWRs have been built in large
numbers around the world, with many different versions and design variations. However, they
all share the common characteristics of a light water coolant and moderator and enriched
uranium fuel. This section provides an overview of the PWR reactor type, particularly from the
perspective of thermal-hydraulic design of the reactor cooling system (RCS).
The PWR has, to date, been the world’s most widely accepted power reactor type [KNI2002,
KOK2009, WEI1997, TON1996, APR2011, LAH1977]. It got its start in the development of PWR
propulsion reactors for United States (U.S.) nuclear submarines. The first PWR power reactor
was designed and built by Westinghouse Electric Company at Shippingport in the United States,
followed by a number of designs and units built in the United States by a number of vendors.
In France, during the early stage of development, Framatome designed and built a number of
units as a licensee to Westinghouse until 1984. Later, Framatome, presently renamed Areva,
continued the development of PWRs in France. In this way, France became a world leader in
PWR development and implementation.
In Japan, Mitsubishi continued development of PWR reactors. In Russia, following submarine
and icebreaker design and technology, a similar PWR design, the water-cooled, water-
moderated power reactor (WWER), was developed.
Typical PWR reactor thermal-hydraulic parameters are provided in Table 13 [APR2011,
POP2014]. To ensure efficient energy conversion from the reactor core to the steam turbine, the
primary coolant pressure is typically 10–15 MPa and the primary coolant temperature 290°C–
330°C, whereas the secondary loop pressure is typically 5–7 MPa.
In the PWR reactor design, ordinary light water is used both as coolant and moderator. A
simplified reactor flow diagram is shown in Figure 149. The reactor coolant loops include
primary, secondary, and tertiary loops. In the primary coolant loop, the coolant picks up the
heat produced in the core and leaves the reactor at a temperature that is sufficient to generate
steam in the steam generator. Then the coolant is returned to the reactor core by means of
pumps. In the secondary loop, the steam generated in the steam generator expands in the
steam turbine, where its energy is transferred into mechanical energy. In the tertiary loop, the
exhaust steam from the turbine is condensed in the condenser, thus transferring heat to the
outside heat sink.
Table 13 Typical PWR reactor parameters
Reactor Fuel
Thermal output, MWth 3800 Fuel pellet material UO2
Electrical output, MWe 1300 Pellet outer diameter, mm 8.19
Thermal efficiency, % 34 Rod outer diameter, mm 9.5
Specific power, kW/kg(U) 33 Zircaloy cladding thickness, mm 0.57
Power density, kW/L 102 Rods per bundle (17 x 17) 264
Ave. linear heat flux, kW/m 17.5 Bundles in core 193
Rod heat flux ave./max, MW/m2 0.584/1.46
Vessel Core
Outer diameter, m 4.4 Length, m 4.17
Height, m 13.6 Outer diameter, m 3.37
Wall thickness, m 0.22 Pressure, MPa 15.5
Steam generator Inlet temperature, °C 292
No. 4 Outlet temperature, °C 329
Outlet pressure, MPa 6.9 Mass flow rate, kg/s 531
Outlet temperature, ⁰C 284
Mass flow rate, kg/s 528
A typical PWR plant and a typical single primary coolant system loop are shown schematically in
Figure 149a, and an isometric view with four primary coolant loops is shown in Figure 149b. The
number of loops, usually two to four, is determined by the reactor power so that the size of the
primary pumps and steam generators is reasonable and obtainable from a commercial vendor.
Each primary coolant loop has one primary pump and one steam generator. All loops are
connected through the reactor vessel

vessel; one of the loops has a pressurizer connected.
13.2.2 Reactor vessel

Several typical PWR reactor vessel designs are illustrated in Figure 150 [KNI2002, KOK2009,
WEI1997, TON1996, APR2011].. Figure 150a shows a quarter section view of the vessel internal
i
structures, in which the reactor core and the fuel elements are also shown. Figure 150b shows a
cross section of the reactor vessel body with the internal structures of the core and the control
guide tubes above the core. Figure 150cc shows an external view of the reactor vessel with a
number of inlet and outlet openings for connection to the inlet and outlet piping.
piping The reactor
pressure vessel height and diameter depend on the design and the thermal power produced,
with a typical height of about 12
12–14 m and a diameter of 5–7 7 m. The vessel is constructed of
forged low-alloy
alloy carbon steel with a wall thickness of 20–25 25 cm (depending on coolant
conditions) and an approximately
ely 3-mm stainless-steel cladding on the inner surface to increase
robustness to water corrosion and erosion.
a) PWR four-loop
loop primary system b) PWR primary coolant loop cross section
Figure 149 PWR primary reactor cooling system
As shown in Figure 149a,a, the nuclear reactor core and fuel assemblies are located in the
reactor pressure vessel. The water coolant at high p pressure
ressure (~14 MPa) is circulated by
external pumps into the reactor vessel, flows upwards through the fuel assemblies, out of
the vessel to heat exchangers, and from the heat exchangers back to the pumps. The
water temperature at the exit of the reactor co
core
re is lower than the saturation
temperature, i.e., the water is sub
sub-cooled,
cooled, which has an important impact on reactor
thermal-hydraulic
hydraulic margins, as will be explained in later sections of this chapter.
a) PWR vessel internal systems b) PWR vessel cross section
Figure 150 Typical PWR reactor vessel
Figure 150a and Figure 150b show the internal structures of a typical PWR vessel. The reactor
core is located in the lower part of the reactor vessel and is surrounded by a cylindrical inner
vessel called the core barrel. On the inner side of the core barrel, the fuel bundles are
surrounded by a multi-faced shroud that constrains flow in the radial direction. The bottom of
the core barrel has radial sliding pins that give the barrel freedom to accommodate thermal
expansion, but do not allow barrel rotation. Around the outside of the core barrel is a metal
jacket called the thermal shield (approximately one-third of the core length, situated around the
core centre). The role of these thermal shrouds is to absorb core neutron radiation in the radial
direction, preventing excessive radiation impact on the reactor vessel and thus reducing
radiation aging effects on the vessel.
The top and bottom of the reactor core are supported by the upper core plate and the lower
support plate. Both plates support the fuel assemblies in appropriate positions in the core. The
reactor control elements and safety shutdown rods pass through the guide tubes that penetrate
the upper core plate. The reactor instrumentation typically (in older designs) passes through the
lower part of the reactor vessel and penetrates the core through the lower support plate.
The reactor coolant enters the vessel through the inlet pipes (the so-called “cold legs” shown in
Figure 149a), then turn downwards in the so-called reactor downcomer and along the outside of
the reactor barrel and inside the reactor vessel, thus coming down to the reactor lower plenum.
There, the coolant turns upwards, flows through the lower support plate, and enters the core
from the bottom. The coolant flows through the core upward and exits in the reactor upper
plenum. From there, the coolant flow exits the vessel through the exit pipes (the “hot legs”).
The upward flow in the core provides good conditions for thermo-siphoning should primary
pump flow be lost in certain design basis accidents.

Steam generators transfer the heat generated by the reactor in the primary loop to the
secondary loop, where the steam turbine transfers the thermal energy into mechanical energy,
which in the electric generator is further transferred into electrical energy. Therefore, steam
generators are an important part of the heat transfer path connecting the two loops.
Figure 7 shows a typical vertical U-tube steam generator (some PWRs use once-through vertical
steam generators, and the Russian WWER design uses horizontal double U-tube steam
generators). A typical vertical steam generator is typically 21 m high and 4.5 m wide in the upper
portion. Primary coolant enters the steam generator at the bottom inlet chamber and flows
upward through the support plate into a U-shaped bundle of small-diameter tubes. Then it flows
back down to the outlet chamber at the bottom of the steam generator. Thus, the outside
surface of the U-tubes in the steam generators transfers heat from the primary to the secondary
side.
On the secondary side of the steam generators, water boils, producing saturated steam that
drives the turbine. This steam is generated in the steam generator at ~5–7 MPa depending on
the station design. The feedwater coming into the steam generator secondary loop enters at the
side, at the bottom above the support plate. It passes through the pre-heater region, where it is
heated to saturation temperature and then boils, passes through two sets of steam dryers, and
exits the upper plenum, which is located at the top of the steam generator.
The design of the steam generators is similar between the CANDU and the PWR designs.
However, some PWR steam generator designs do not have a pre-heater designed at the
feedwater entrance to the steam generator secondary side.
13.2.4 Primary pumps

Figure 9 presents a view of a typical PWR primary pump [KNI2002, KOK2009, WEI1997,
TON1996, APR2011]. It contains a single-stage impeller and an impeller casing. This is the most
important part of the pump because it provides the driving force for the primary flow. Above
the impeller casing, the rest of the pump consists of a number of pump seals to insulate the
primary flow, a flywheel set to provide inertia for the pump to continue operating for some time
after power is lost, and an electric pump motor.
The primary pumps are a vital component in the reactor heat transport system because they
drive the coolant flow by forced circulation through the primary heat transport system. Hence,
they are essential to fulfill the primary function of the reactor cooling system, to provide
continuous cooling of the reactor core in normal operation, during transients, and during reactor
shutdown. The pump start-up and shut-down sequences are routine operations of the reactor
heat transport system.
Pump failures, whether due to loss of power or inadvertent operator action, are events that
must be considered explicitly in the reactor design. The reactor primary cooling system is
designed to allow and enhance natural circulation in the absence of pump flow. This is achieved
by placing the reactor core (heat source) at the lowest point in the primary loop and the pumps
and steam generators (heat sinks) at the highest point in the primary loop.
The design of the primary pump is similar between CANDU and PWR reactors.
13.2.5 Pressurizer
The pressure in the reactor primary coolant system is maintained at a certain level by a
pressurizer. Figure 8 shows a view of a typical pressurizer.
13.2.6 Reactor refuelling and reactivity control

The fuel is in the form of assemblies (or bundles) of enriched UO2 rods, clad in zirconium alloy (or
in some earlier designs, in austenitic stainless steel). As shown in Figure 151, these fuel
assemblies are square or hexagonal in shape (depending on the reactor type) and contain an
array of fuel rods within each assembly. The fuel assemblies are arranged in a tightly packed
lattice configuration, with either square or triangular pitch (depending on the reactor type).
Typically, reactors designed by GE-Westinghouse use a square fuel assembly and square lattice
pitch, whereas the Russian-designed WWERs use a hexagonal fuel assembly arranged in a
triangular pitch.
The fuel assemblies have an open lattice, which enables radial cross flow of the cooling water,
i.e., mixing of coolant between adjacent fuel assemblies. The coolant does not boil in the core
because its temperature at the core exit is still below saturation.
To refuel a PWR reactor, it must be shut down and the primary loop cooled and depressurized.
Several days after reactor shutdown, the decay heat load of the primary loop decreases
sufficiently to enable reactor depressurization and the start of refuelling operations. Following
cooling and depressurization, the reactor pit (or reactor bunker) is filled with water to at least a
10-m height above the reactor head to provide adequate radiation shielding for the refuelling
personnel working at the top of the reactor pit. Then the reactor head is opened and removed,
enabling access to the reactor core. The fuel assemblies are individually manipulated
underwater, repositioned in the core, fresh fuel assemblies brought into the core, and some
irradiated assemblies removed from the core. Typically, the fresh fuel assemblies and irradiated
fuel assemblies are positioned and reshuffled in the core in a predetermined sequence. The
refuelling crew typically uses computer programs to ensure optimal fuel bundle distribution and
to monitor fuel bundle histories in the core.
The irradiated fuel bundles that are removed from the core are transported underwater to the
irradiated fuel bay for storage and cooling for a number of years. Removing the fuel from the
irradiated fuel bays depends on the national policies of specific jurisdictions for handling highly
radioactive irradiated fuel or subjecting it to reprocessing.
a) Typical fuel assembly structure b) Typical fuel assembly view
Figure 151 Typical PWR square fuel assembly design
Refuelling is normally done on an 18–24 month basis and typically takes 4–5 days. However,
refuelling outages in PWRs can last up to one month or sometimes longer when the
preparations, cooldown, maintenance activities, and return to full power are taken into account.
To operate for long periods without refuelling, the new fuel is enriched in 235U (up to a maximum
of 5%). When the fuel is fresh, the excess reactivity in the reactor core is compensated for by a
neutron poison dissolved in the coolant/moderator water and by other devices. As the fuel
burns up, the dissolved poison in the coolant/moderator is gradually removed by ion-exchange
columns.
Reactivity control in a PWR reactor is performed by full-length control rods. Levelling neutron
flux and heat generation in the core is performed by various control rod banks of full or partial
length, and by other means in newer PWR designs (i.e., burnable absorber material in fuel and in
“grey” control rods). Intermediate- and long-term reactivity control in the core is achieved by
soluble poison (called “chemical shim”); typically, boric acid is used for this purpose. The critical
soluble boron concentration is reduced in the core with core burn-up, i.e., with buildup of 135Xe
and 149Sm. Therefore, during the initial burn-up of a few hundred MWd/t, the boron
concentration is reduced sharply by up to 40%, followed later by a more gradual reduction to
compensate for fuel burn-up and generation of fission products in the fuel.
13.2.7 Fuel assembly and fuel pin design

Figure 151 shows a typical PWR fuel assembly design [KNI2002, KOK2009, WEI1997, TON1996,
APR2011]. As shown in Figure 151a, the fuel assembly typically consists of fuel pins (also called
fuel rods or fuel elements) arranged in a square pitch with a certain number of grid plates
(spacers) that hold the fuel pins in position and constrain them in the radial direction. Other
than by the support plates, the PWR fuel bundle does not restrict the coolant flow in the radial
direction because there is no flow tube around the bundle. Therefore, all fuel pins in the core
create a continuous matrix of fuel pins throughout the core.
a) Typical fuel rod design b) Typical fuel rod grid design
Figure 152 Typical PWR fuel design
The guide tubes shown in Figure 151a are provided for each fuel assembly. They are intended to
receive the control rods (shown in Figure 151a) from the top as a cluster control assembly with a
certain number of control rods. Fuel assemblies that do not have control assemblies are plugged
from the top by a special plugging device to prevent coolant bypass though them (they have a
larger hydraulic diameter than the rest of the fuel matrix).
The top and bottom nozzles are welded with the guide tubes and create a rigid fuel assembly
structure. The fuel pins pass though the grid plates, but are not welded to the grid plates to
allow for unrestricted thermal expansion. The bottom nozzle contains four openings at the four
support legs, which fit into the reactor core bottom plate, and which deliver coolant to the fuel
assembly
Figure 152 shows the fuel pin design [POP2014]. The fuel pin design is similar for the LWR and
CANDU designs. In fact, all LWRs and CANDUs, regardless of the manufacturer and country of
origin, have a similar fuel pin design.
The fuel pin consists of a Zircaloy tube into which fuel pellets are inserted. The tube is then filled
with gas and sealed. The fuel pellets have a height-to-diameter ratio of about 1.5 and have
dished surfaces on both ends, which help to align them properly in the vertical direction,
accommodate fuel swelling, and collect gases generated during fission. Numerous fuel pellets
are located in each fuel pin and form a vertical fuel pellet stack within the pin. At the top end of
the fuel pin, the fuel pellets are held tightly in a vertical direction by a spring that is installed and
sealed in the tube.
Figure 153 Typical PWR fuel assembly support Figure 154 Typical PWR reactor control sites
spacer plates
Figure 153 shows the design of the fuel bundle support plates [POP2014]. Each spacer plate has
dimples on each side of the square frame, which hold the fuel pin in the radial direction. The
fuel pin can slide along the spacer plates in the vertical direction, thus accommodating thermal
expansion. The spacer plates also have mixing vanes in the radial direction that help to enhance
turbulence along the fuel pin, resulting in higher thermal margins.
Figure 154 shows cross sections of the fuel bundles and the positions of the 21 control-rod guide
tubes [POP2014]. Typically, a number of positions in the fuel assembly are not filled with fuel,
but instead with guide tubes that are used as sites for insertion of control rods. Although not all
bundles will have control rods inserted in them, they are all made with guide tubes so that they
can be installed at any core position regardless of whether they will receive control rods or not.
Those assemblies that do not receive control rods will receive plug assemblies, which will insert
short plug rods to block coolant bypass through the empty guide tubes (the plug assembly is
similar to the assembly shown in Figure 155b, except that its plug pins are shorter).
Figure 155a shows a typical PWR control assembly design. Figure 155b shows one immovable
control assembly for long-term reactivity control in certain fixed positions in the core [POP2014].
The movable control-rod assembly consists of 21 control rods held by radial arms, which are
connected to the central holding shaft, which in turn is connected to a control drive assembly.
The movable control assembly moves up and down vertically as driven by the reactor control
system. The immovable control bundle also has 21 control rods, which are connected at the top
plate. The top plate has a shaft with a spring to lock the control bundle into position so that it
will remain stationary between two fuel replacement cycles.
a) Movable control assembly (vertical) b) Immovable control assembly (vertical)
Figure 155 Typical PWR control assembly (cluster) design
Figure 156 shows a typical PWR control rod design [POP2014]. The control rods are of various
design and material composition depending on their role in the reactor core. The control rods
that are part of the reactor control system are filled with neutron-absorbing material along their
full length in the core (usually a mix of 80% silver, 15% indium, and 5% cadmium, which is the
material with the largest neutron-absorption capability used in modern reactors). The
immovable control rods (Figure 156a) can be filled over their full length or partially filled at the
top, bottom, or middle. They can also be filled with neutron-absorbing material of various
thicknesses in the annulus (Figure 156a). The immovable control rods are filled with a mix of
neutron-absorbing material that reduces absorption capability by neutron capture (diminishing
absorption potential per unit volume with radiation exposure), or by materials that retain
absorption capability over several control isotope generations.
a) Solid control rod (vertical) b) Annular control rod (vertical)
Figure 156 Typical PWR control-rod design
PWR control-rod guide tubes penetrate through the reactor pressure vessel head and hold each
control assembly cluster in place. The control assemblies are driven in and out of the core by
control-rod drive devices that are located outside the reactor vessel. The control-rod shaft is
located in the guide tubes that extend through the reactor cap, in the space above the core, and
holds the control-rod top. Special attention is devoted to the seals of the control-rod guide
tubes and assemblies because they are located in the vessel at a pressure of 15 MPa, whereas
their shafts and guide tubes are at atmospheric pressure. Because of the large pressure
difference between the inside of the vessel and the atmosphere outside of the reactor vessel,
special attention is devoted to preventing control-rod ejection from the vessel, which can create
a severe large positive reactivity insertion.
13.2.8 Reactor core design

The fuel assemblies in the reactor core are typically 3.6–4.2 m long and have from 14 x 14 to 19 x
19 fuel rod positions. There are typically 120–170 fuel assemblies in the core, depending on the
reactor thermal power and reactor type, arranged in a square or hexagonal lattice in the core.
The fuel assemblies are enveloped by a metal shroud which follows the core shape at the sides
of the peripheral assemblies (see Figure 150a). The metal shroud does not permit fluid
movement in the radial direction, but fluid movement in the radial direction is not restricted by
the fuel assemblies themselves.
The core flow is in the upward direction, which in the absence of forced circulation (when the
primary pumps are off) enhances the potential for natural circulation, thus removing core decay
heat.
13.2.9 WWER reactor design

The WWER reactors are of Russian design, and their design follows in principle the original U.S.
PWR design [KNI2002, KOK2009,. Figure 157a shows an isometric view of the primary cooling
loop and the reactor vessel. The primary cooling system is similar to the PWR primary cooling
system, with three or four primary loops depending on reactor power. There are some
differences in the fuel assembly design, the most significant being that it is not open in the radial
direction because it is surrounded by guide tubes. Figure 157b shows a reactor vessel similar to
the American PWR vessel design. Figure 157c shows a WWER steam generator of horizontal
design and containing double horizontal U-tube assemblies on the primary side connected at the
middle of the steam generator. In addition, steam is collected at the top of the steam generator
in a separate steam collector tube.
a) Primary loop b) Reactor vessel
Figure 157 Russian WWWR reactor design
13.3 Boiling water reactors

Boiling water reactors (BWRs) also originate from the U.S. nuclear industry, but unlike PWRs,
they are designed to allow coolant boiling in the reactor vessel, and hence they do not have a
separate heat exchanger (steam generator) for steam production. There are numerous BWRs in
operation around the world, although fewer than the PWR designs. BWRs feature a lower
primary heat transport system pressure than PWRs for the same power output, which results in
decreased thickness of the reactor pressure vessel (which amounts to an economic benefit).
Figure 158 shows two BWR reactor versions, BWR/4 and BWR/6 [KNI2002, KOK2009, LAH1977,
POP2014].
Figure 158 BWR reactor design

BWR plants have been designed in two main design streams: direct-cycle and dual-cycle. In a
direct-cycle plant, which is the most common in operation, steam is produced in the steam
generator and is moved to the turbine generator as dry steam. In a dual-cycle plant, in addition
to steam production in the reactor, steam is also produced in a secondary steam generator fed
by saturated water from the moisture separator.
Figure 159 shows a simplified flow diagram of a direct-cycle BWR system [LAH1977]. The direct
cycle has the advantages of lower capital cost and simplicity, which increases system reliability.
There is no separate steam generator in the cooling system because the reactor generates
steam. Older BWR designs could not follow the load, but are to a certain extent self-controlling
due to the negative void reactivity. Modern BWR designs are capable of load following by
coolant recirculation control.
Figure 159 Typical BWR reactor vessel coolant circulation

With this arrangement, the turbine plant is “active” due to activity induced in the reactor coolant
(primarily N-16). As a result, the turbine generator plant is less accessible during operation;
fortunately, however, this activity decays quickly following shutdown, permitting normal direct
access for maintenance.
Figure 160 BWR reactor fuel assemblies in the core

The typical BWR reactor vessel, shown in Figure 158, has the steam production and collection
part in the upper part of the vessel, the reactor core in the middle of the vessel, and the control
assembly and instrumentation access through the reactor bottom and the lower plenum. The
reactor core has an active length of typically 3.5–4 m, similar to PWRs. The core contains
between 500 and 800 fuel assemblies, each containing arrays of fuel rods with a similar design to
those in PWRs.
Figure 159 shows the coolant circulation in the BWR core. The recirculation system consists of
recirculation pump loops located on the sides of the reactor vessel. In the BWR design
illustrated, the reactor vessel downcomer jet internal pumps are located on the outer core
periphery, arranged in pairs, with each having a common inlet riser. The driving flow from the
external pumps passes through the jet pump nozzles and acquires high velocity and momentum.
The momentum exchange entrains the recirculation flow, usually called suction flow. The
combination of the external pump-driven flow and the suction flow enters the throat of the jet
pump, thus ensuring adequate flow through the core. The jet pump design has the advantages
of fewer moving parts, higher reliability, and robustness in accident situations. The feedwater
inlet is in the upper part of the core, and the coolant falls down by gravity towards the inlet of
the jet pumps.
BWR fuel assemblies are typically of square design (see the left part of Figure 160), but the fuel
rods are situated in a square flow tube which does not allow flow mixing in the radial direction
(see the right part of Figure 160) [KNI2002, KOK2009]. Because the BWR reactor is designed to
operate with saturated water that boils in the core, a good level of turbulence is already
achieved, and therefore there is no need to promote radial mixing to enhance turbulence, as in
the PWR design. The left side of Figure 160 shows a BWR fuel assembly with a square tube and
the spacer grids inside holding the fuel pins.
Figure 161 shows a cross section of four fuel assemblies, on the left 10 x 10, and on the right 7 x
7 fuel rod arrays [KNI2002, KOK2009, LAH1977, POP2014]. The cruciform control assembly is
located between the fuel bundles. Clearly, between the fuel assemblies, a gap is maintained for
the control assembly to move in a vertical direction through the core.
Figure 161 BWR fuel and control assemblies
The control assembly is moved through the core from the bottom of the reactor.
reactor Therefore, the
control assembly holding the plate and drive mechanisms is at the bottom of the core and needs
need
to work against gravity.
13.4 Gas-cooled reactors

Gas-cooled
cooled reactors (AGRs) have been in operation for many years and were promoted by the
U.K. nuclear industry due to their potential for higher thermal efficiency and relatively low
stored energy in the primary coolant
coolant, which reduced containment requirements. AGRs were
primarily designed and built in the United Kingdom. They use gas as the reactor coolant and
natural or enriched uranium as fuel. In recent years
years, some of these reactors have been shut
down, and others are planned for shutdown soon. The U.K. nuclear industry has abandoned this
design and does not plan to use it for future reactors. This section provides a brief overview of
the AGR reactor type and a brief history of its development in recent decades.
13.4.1 Magnox reactors

Magnox reactors are graphite-moderated,
moderated, CO2 gas-cooled
cooled reactors fuelled with natural uranium
metal clad with a magnesium alloy called Magnox. They derived their generic name from this
last feature. Figure 162 shows a schematic arrangement of one version of this reactor type
[POP2014].
This type of reactor was pioneered by the British and French and was a natural outgrowth of
earlier air-cooled, graphite-moderated
moderated research and plutonium production reactors. r A
significant number were built in Britain and France
France, with a few exported to other countries.
Early versions used steel reactor pressure vessels with external heat exchangers (boilers) and gas
circulating blowers. Later versions, as shown in Figure 163, used pre-stressed
stressed concrete pressure
vessels incorporating the reactor core, heat exchangers
exchangers, and coolant circulation blowers. This
was primarily
arily a cost reduction measure, although safety advantages in terms of low risk of
coolant system rupture were also claimed.
Figure 162 Typical Magnox reactor design

Primarily because of coolant temperature limitations imposed by the uranium metal fuel and the
Magnox cladding, only relatively modest turbine steam conditions are achievable, limiting the
station overall efficiency to ~30%.
As is typical of all natural uranium
anium power reactors, the Magnox reactors are fuelled on-load.
on This
is necessary because large quantities of excess reactivity, in the form of additional U-235,
U are not
“built into” the new fuel.
The in-service
service availability of the Magnox reactors has prov
proven
en to be relatively good. On-load
On
refuelling helps in this regard. Nevertheless, their relatively high capital cost and relatively
modest achievable fuel utilization has led to the abandonment of plans for constructing further
reactors of this type. Presently,
sently, in the United Kingdom, all Magnox reactors are shut down and
undergoing decommissioning.
13.4.2 AGR reactors

The AGR (advanced gas-cooled
cooled reacto
reactor) was developed in the United Kingdom as a successor to
the Magnox line of reactors [POP2014]
[POP2014]. A number of these ese reactors were successfully operating
in the United Kingdom,, but presently they are agaging and require replacement.. They differ from
the latest Magnox reactors primarily in the fuel used. The fuel is UO2 clad in stainless steel. This
design permits higher
her fuel temperatures and coolant temperatures to be achieved, leading to
conditions similar to those in conventio
conventional fossil-fuel steam plants (16.5 MPa, 550°C).
The fuel is designed as a cluster of small
small-diameter rods, which permits relatively high power
p
levels. This reduces the size of the reactor core relative to the Magnox reactors,
reactors where the fuel
is in the form of large single elements. However, because of these fuel changes, the AGR
requires some fuel enrichment. Figure 163 shows a view of an AGR reactor
ctor type diagram.
diagram
The British have decided that the AGR is not fully competitive with other types of power
reactors. Hence, this design, like the Magnox type, appears to be “dead-ended
ended” and is being
replaced by other reactor types, such as advanced PWRs and advanced BWRs.
Figure 163 Typical AGR reactor design
13.4.3 HTGCR Reactors

The high temperature gas cooled reactor (HTGCR) differs from the AGR in two major respects
[POP2014].. The first is the use of helium as the coolant in place of CO2. This permits even higher
coolant temperatures without inducing a chemical reaction with the graphite moderator. The
second relates to the fuel. In very early designs, tthe he fuel was supposed to be fully enriched
(93%) U-235
235 mixed with thorium. Thorium absorbs neutrons and is converted, after a
radioactive decay chain, to U-233
233, which is fissile. As a result, the reactivity of the fuel remains
high even after very long irradiation, the UU-233 replacing the U-235 235 as the latter is burned up.
The fuel is in the form of mixed carbides. It is manufactured in very small spheres which are
coated with pyrolytic graphite, the latter providing the cladding. These spheres are compacted
into holes in large graphite assemblies, forming an integral fuel and moderator assembly.
assembly
Figure 164 shows an advanced version of the HTGCR reactor design currently being developed as
part of the Generation IV development efforts. The very high coolant temperatures achievable
lead to high steam cycle efficiencies,
iciencies, or alternatively make possible the ultimate use of gas
turbines directly driven by the coolant. The development of the direct
direct-cycle
cycle gas turbine version
would be particularly attractive in this regard.
Figure 164 Typical HTGCR reactor design
13.5 Channel-type
type reactors
13.5.1 SGHWR reactor
The steam generating heavy water reactor (SGHWR) has been developed and operated in the
United Kingdom,, Japan, and Italy for a number of years. The best example of this reactor type
that operated successfully in Japan for more than 20 years is the Fugen reactor (currently
decommissioned) [POP2014].
Figure 165 SGHWR reactor

Figure 165 shows a schematic of this reactor type. The heat transport system contains vertical
pressure tubes, light water coolant, heavy water moderator in a calandria vessel, and
recirculating pumps used to return water from steam drums to the lower end of pressure tubes.
The fuel was low-enriched uranium with on-power refuelling. The reactor heat transport system
was located in a containment building, thus providing protection during accidents.
Another type of SGHWR design was the CANDU Gentilly-1 reactor, which was shut down and is
now decommissioned. It also used vertical pressure tubes with natural uranium, heavy water
moderator, and light water coolant.
13.5.2 CANDU reactor

The CANDU fuel channel type reactors have been developed in Canada, have achieved great
commercial success around the world, and are being still developed and prepared for new builds
in Canada and internationally. Figure 2 shows the CANDU reactor cooling systems and Figure 3
the reactor heat transport system and moderator cooling system.
13.5.3 RBMK reactor

The Russian version of the channel reactor is the RBMK reactor (an acronym for High Power
Channel Type Reactor in Russian) [POP2014]. This reactor design was built in Russia and in
neighbouring states in the 1980s and 1990s. The RBMK reactors had two basic power levels, one
around 1000 MWe, the other around 1400 MWe.
Following the Chernobyl accident, many of these reactors were shut down, and they continue to
be shut down, with the exception of a few, such as the Kursk and Leningrad reactors that will
continue to operate until the early 2020s. The RBMK design was abandoned by the Russian
nuclear industry after the Chernobyl accident. However, many design changes have been
implemented in the operating units to improve their safety during their remaining operational
life.
Figure 166 shows the reactor cooling systems (primary, secondary, and tertiary), a view of the
reactor core, and a cross section of a RBMK plant. The reactor core is very large (21.6 x 21.6 x
25.5 m). The vessel contains the graphite stack filled with a helium-nitrogen mixture to ensure
an inert atmosphere. The reactor fuel is UO2 with a fuel rod design similar to the LWR design,
with 2%–4% enrichment in U-235. Refuelling is performed on-power to achieve a better balance
of neutron flux and core utilization and to optimize production of Pu-239.
Figure 166 RBMK reactor design

The reactor has two independent cooling circuits, each with four primary circulation pumps
(including one stand-by). The cooling water is fed to the reactor through lower water lines to a
common header, which is split into 22 distribution headers, each feeding 38–41 fuel channels, in
which water boils. The mixture of water and steam is conveyed by the upper steam lines, one
for each fuel channel, from the reactor top to the steam separators. Steam with 15% quality is
taken from the separators by two steam collectors per separator, combined, and then conveyed
to the turbine hall. The nominal total flow through the core is ~46,000–48,000 m3/h. The
nominal temperature of the water at the inlet of the reactor is about 270°C and the outlet
temperature 284°C, at a pressure of 6.9 MPa.
The fuel channels consist of Zircaloy pressure tubes that pass through the channels in the centre
of the graphite blocks. There are 1661 fuel channels in the core and 211 control channels. There
is only a small clearance between each pressure channel and a graphite block, which makes the
design vulnerable to damage. If a pressure tube channel deforms due to high internal pressure
in an accident situation, the deformation can cause significant pressure straining of the graphite
blocks and lead to damage, which can then possibly propagate to neighbouring channels.
Light water was used as coolant and graphite as moderator. This combination of core materials
leads to a highly positive void coefficient, which results in high power excursions in the case of a
loss of coolant accident.
Because of certain similarities to the CANDU design, and because of the unpopular experience
with the RBMK reactors, CANDU has often been criticized as potentially vulnerable in the same
way as the RBMK reactors. As described in Chapter 16, many studies have been performed since
the Chernobyl accident and have demonstrated that CANDU has different physical
characteristics that make the CANDU design behave quite differently from the RBMK reactors,
and that therefore CANDU reactors are not as vulnerable in accidents as RBMK reactors.
13.6 Fast breeder reactors

Liquid metal fast breeder reactors (LMFBRs) were designed many years ago, and some were built
as prototype versions in the United States, France, and Russia, and later by other countries as
part of pilot programs. These pilot-prototype versions had varying degrees of operational
success, and therefore many were shut down or restarted over the past several decades. These
reactors use a fast neutron spectrum to maintain the fission chain reaction and are cooled by
molten metal coolants. Figure 167 shows the design sketch of a typical pool-type (left) and a
loop-type fast breeder reactor [POP2014].
Whereas all the reactors previously described are of the thermal type, i.e., the fissions in the fuel
are primarily induced by thermal neutrons, it is also possible to sustain a chain reaction with
high-energy neutrons, i.e., fission neutrons, provided that the fuel is highly enriched with fissile
material such as U-235 or Pu-239. Furthermore, on average, more than two neutrons are born
per fission. One of these neutrons is required to induce the next fission, leaving a surplus of
somewhat more than one neutron which can be absorbed by a “fertile” material such as U-238,
producing fissile Pu-239. It is therefore possible to produce fissile material as rapidly as it is used
up; this is called “breeding”. In fact, it is possible to produce more fissile material than is used
because the average number of neutrons produced per fission is greater than 2. The excess is
referred to as the “breeding gain”. This can be achieved only if the neutron economy is high, i.e.,
if relatively few neutrons are wasted.
This possibility of breeding is a very attractive means of extending the power available from
uranium because less than 1% of natural uranium is fissile. If a substantial part of the other
~99% could be converted to fissile Pu-239 as a by-product of reactor operation, then the world’s
uranium reserves could be stretched enormously. The fast breeder reactor is one way of doing
this, which explains the widespread interest in this concept.
Figure 167 Fast reactor design

The basic arrangement of a liquid metal-cooled fast breeder reactor is shown in Figure 167. The
reactor core consists of a closely packed array of highly enriched (U-235 or Pu-239) oxide rods
clad in a high temperature-resistant metal. This core is surrounded on all sides by a “breeder
blanket” of fertile U-238 rods (also in clad oxide form). The excess fission neutrons produced in
the core “leak” out of the core and are absorbed in the blanket rods. Both the core and blanket
are cooled by a flow of liquid sodium. This sodium is in turn cooled in heat exchangers and
returned to the reactor by more or less conventional centrifugal pumps. The heat exchangers
are cooled by a second flow of sodium, which in turn is cooled in a second set of heat
exchangers, which produce steam for the turbine. The purpose of this intermediate sodium loop
is to provide isolation between the sodium cooling the reactor and the steam and water in the
turbine cycle, thereby practically eliminating the in-leakage of water and preventing it from
contaminating the reactor coolant.
Despite this intermediate loop, the reactor operating temperature is sufficiently high to permit
steam to be produced under modern fossil-fuelled plant conditions (~16.5 MPa with a single
reheat to 540°C). There are a number of new Generation IV designs for fast breeder reactors,
but none of them has been built yet, even as a prototype.
1
CHAPTER 7
Thermalhydraulic Analysis
prepared by:
Dr. Wm. J. Garland, Professor Emeritus,
Department of Engineering Physics,
McMaster University, Hamilton, Ontario, Canada
Summary:
This chapter is concerned with thermalhydraulic analysis of the process systems that are
required to transport heat energy away from the nuclear reactor source and transform this heat
energy into useful work (generally electrical energy. Thermal hydraulic system behaviour is
largely determined by the simultaneous solution of the equations that govern the four variables
(flow, pressure, density and enthalpy). The general mass, energy and momentum conservation
equations are presented in general terms and are simplified to the common approximate forms
used in systems modelling. The equation of state that is required for closure is explored with
particular emphasis on implementation. Process system solution algorithms are investigated.
Table of Contents
1 Introduction ............................................................................................................................ 4
1.1 Learning Outcomes ......................................................................................................... 5
1.2 The Chapter Layout......................................................................................................... 6
2 Basic Equations for Thermalhydraulic Systems Analysis......................................................... 7
2.1 Introduction .................................................................................................................... 7
2.2 Conservation ................................................................................................................... 7
2.3 Conservation of Mass ................................................................................................... 10
2.4 Conservation of Momentum......................................................................................... 12
2.5 Conservation of Energy................................................................................................. 15
2.6 The Equation of State.................................................................................................... 20
2.7 Empirical Correlations................................................................................................... 20
2.8 Solution Overview......................................................................................................... 21
2.9 Problems ....................................................................................................................... 23
3 Nodalization .......................................................................................................................... 24
3.1 Introduction .................................................................................................................. 24
3.2 The Node-Link Concept................................................................................................. 24
3.3 Nodal Diffusion ............................................................................................................. 28
3.4 Examples ....................................................................................................................... 32
3.5 Matrix Notation............................................................................................................. 34
©UNENE, all rights reserved. For educational use only, no assumed liability. Thermalhydraulic Analysis – December 2016
3.6 Exercises........................................................................................................................ 35
4 Equation of State................................................................................................................... 36
4.1 Introduction .................................................................................................................. 36
4.2 Thermodynamic Properties .......................................................................................... 36
4.3 The Iterative Method .................................................................................................... 38
4.4 The Rate Method .......................................................................................................... 40
4.5 Water Property Fits....................................................................................................... 46
4.6 Problems ....................................................................................................................... 48
5 The Rate Form of the Equation of State ............................................................................... 49
5.1 Introduction .................................................................................................................. 49
5.2 The Rate Form............................................................................................................... 49
5.3 Numerical Investigations: a Simple Case ................................................................... 50
5.4 Numerical Investigations: a Practical Case.................................................................... 61
5.5 Discussion and Conclusion............................................................................................ 66
5.6 Problems ....................................................................................................................... 67
6 Thermalhydraulic Network Simulation ................................................................................. 68
6.1 Introduction .................................................................................................................. 68
6.2 Porsching’s Method ...................................................................................................... 68
6.3 Derivation of FIBS.......................................................................................................... 69
6.4 Special Cases ................................................................................................................. 74
6.5 Programming Notes ...................................................................................................... 75
6.6 Conclusion..................................................................................................................... 77
6.7 Problems ....................................................................................................................... 77
7 Case Study: Heat Transport System Stability ........................................................................ 78
8 References............................................................................................................................. 78
9 Nomenclature ....................................................................................................................... 82
List of Figures
Figure 1 The four cournerstone single phase flow equations and the flow of information
between them....................................................................................................................... 21
Figure 2 The four cornerstone equations for the two-fluid model. .......................................... 22
Figure 3 The four cornerstone equations for the full two-fluid model with equal pressure of
the two phases...................................................................................................................... 23
Figure 4 A general node and connecting links........................................................................... 24
Figure 5 Two connected nodes.................................................................................................. 25
Figure 6 Node-link setup for a simple pipe................................................................................ 26
Figure 7 Node-link setup for an area change in a pipe.............................................................. 27
Figure 8 Illustration of convection and diffusion....................................................................... 29
Figure 9 Transmission of a step change using the plug flow model and the mixing tank model
(1 to 50 tanks). ...................................................................................................................... 30
Figure 10 Transmission of a step change using the plug flow model and a feeder model with
Thermalhydraulic Analysis 3
skewing due to differences in transit times. ......................................................................... 31

Figure 11 Simple Tee junction.................................................................................................... 32
Figure 12 Simple Y junction. ...................................................................................................... 32
Figure 13 Node-link diagram: ¼ circuit CANDU. ........................................................................ 33
Figure 14 Sample node-link connection for a header................................................................ 33
Figure 15 4 Node - 5 link diagram.............................................................................................. 35
Figure 16 P-v-T surface for water .............................................................................................. 37
Figure 17 Numerical search for P given ρ and h for a two-phase mixture. ............................... 38
Figure 18 Error correction scheme for pressure in two-phase. .................................................... 39
Figure 19 Density vs. pressure at various temperatures in sub-cooled water........................... 47
Figure 20 Basis for curve fitting in the subcooled region. ......................................................... 47
Figure 21 Simple 2-node, 1-link system..................................................................................... 51
Figure 22 Program flow diagram for the normal method. ........................................................ 52
Figure 23 Program flow diagram for the rate method. ............................................................. 54
Figure 24 Number of iterations per pressure routine call for the normal method with a time
step of 0.01 seconds and a pressure tolerance of 0.001 of full scale (10 MPa). .................. 56
Figure 25 Integrated flow error for the rate method and the normal method for various fixed
time steps, convergence tolerances and adjustment factors. .............................................. 56
Figure 26 Flow vs. time for the implicit forms of the normal and rate methods. ..................... 61
Figure 27 Schematic of control volumes in the pressurizer....................................................... 62
Figure 28 Pressurizer's pressure transient for the rate method................................................ 65
Figure 29 Pressurizer's pressure transient for the normal method with error tolerance of 0.2%.
............................................................................................................................................... 65
Figure 30 Average number of iterations per pressure routine call for the normal method in
simulating the pressurizer problem. ..................................................................................... 66
List of Tables
Table 1 Figure of merit comparisons of the normal and rate forms of the equation of state for
various convergence criteria (simple case)........................................................................... 58
1 Introduction
This chapter is concerned with thermal hydraulic analysis of the process systems that are
required to transport heat energy away from the nuclear reactor source and transform this heat
energy into useful work (generally electrical energy). Thermal hydraulic design of the process
systems is covered in the previous chapter. Design and analysis are, of course, tightly coupled.
Nuclear systems design is guided by analysis results. Analysis, in turn, is performed on a
specific design to determine its performance. It is a complex, iterative dance. Design and
analysis of the reactor process involves a number of interrelated systems:
 reactor core
 heat transport system
 steam generators
 turbines
 pressure control system
 coolant inventory control systems
 power control systems;
a number of system components:
 valves
 pumps
 pipes
 vessels
 heat exchangers;
and a number of engineering and science disciplines:
 reactor physics
 heat transfer
 fluid mechanics
 thermodynamics
 chemistry
 metallurgy
 control
 stress analysis.
The heat transport system (HTS) is of central importance since it is the interface between the
heat source and the heat sink. Good HTS performance is essential to reactor integrity, plant
performance and safety. Herein, the scope is limited to the modelling tools used in thermal
hydraulic analysis of the HTS. This chapter is a systems level chapter, not a components level
one. Component modelling is limited to approximate models that are appropriate for systems
analysis. Detailed multidimensional modelling of complex components such as steam
generators, pumps, calandria vessels, headers, etc., are not attempted.
This chapter is primarily about the interplay the two main actors in hydraulic systems: flow and
pressure. But because we are dealing with systems involving the transfer of heat, local density
and enthalpy determine the pressure. Hence, thermal hydraulic system behaviour is largely
determined by the simultaneous solution of the equations that govern these four variables
(flow, pressure, density and enthalpy).

The overall objectives for this chapter are as follows:
 The student should be able to explain the roles played by mass, flow, energy and
pressure in thermalhydraulic simulation.
 The student should be able to derive appropriate forms of the governing equations, and
develop a flow diagram and pseudo-code for a thermalhydraulic system simulator from
first principles.
 The student should be able to build a thermalhydraulic system simulator from first
principles.
 The student should be able to identify the terms and symbols used in thermalhydraulics.
 The student should be able to distinguish between the differential and integral form and
be able to choose, with justification, the correct form to use in various situations.
 The student should be able to recall typical values and units of parameters.
 The student should be able to recognize key physical phenomena.
 The student should be able to recognize the coupling between mass, momentum,
energy and pressure in thermalhydraulic systems.
 The student should be able to choose approximations as appropriate (# of dimesions,
transient or steady state, averaging, spatial resolution, etc.) with justification.
 The student should be able to develop, with justification, a node-link diagram given a
thermalhydraulic system.
 The student should be able to construct the matrix form of the conservation equations
for a given node-link structure.
 The student should be able to calculate any dependent thermodynamic property given
any two independent state variables using (a) the steam tables, (b) supplied codes, (c)
supplied curve fits to the steam tables.
 The student should be able to develop a flow diagram and pseudo-code for the
calculation of P and T given density and enthalpy.
 The student should be able to explain the pressure and temperature response of a
volume of fluid to perturbations given the F and G functions.
 The student should be able to develop a flow diagram and pseudo-code for the rate
method of the equation of state.
 The student should be able to develop a computer code implementing the rate method
of the equation of state.
 The student should be able to model a simple thermalhydraulic network using the
integral form of the conservation equations and the rate form of the equation of state.
The student should be able to check for reasonableness of the answers.
 The student should be able to apply the various numerical methodologies (fully explicit
to fully implicit) to special cases of the thermalhydraulic system equations.
 The student should be able to produce a general node-link code based on the
cumulative concepts presented in this course.
 The student should be able to evaluate the efficacy of the various numerical algorithms.
1.2 The Chapter Layout

To lay the groundwork for thermal hydraulic systems analysis, Section 2 presents the general
mass, energy and momentum conservation equations in very general terms and proceeds to
derive the common approximate forms used in systems modelling. Section 3 shows how to
model hydraulic piping networks as a system of nodes connected by links and elaborates on the
appropriate equation forms for these node-link approximations. The conservation equations
requires a relationship between pressure, temperature, density and energy (the equation of
state) for closure. In Section 4, the equation of state is explored with particular emphasis on
implementation. Sections 5 and 6 cover numerical considerations. Explicit, semi-implicit and
implicit techniques are presented. At this point the reader is almost ready to conduct thermal
hydraulic simulations. Chapter 6 on Thermalhydraulic design completes the picture by
providing heat transfer and hydraulic correlations that are needed for the simulations. A
survey of industrial strength tools is outside the scope of this chapter as is a discussion of
phenomena identification and evaluation, and code verification and validation.
As with design, there is no one best model for a given analysis task, nor is there even one best
solution procedure. Good simulation is evolutionary; we learn from past successes and
failures, incorporate the latest experimental, theoretical and numerical results, employ sound
engineering principles and a solid understanding of the basics to engineer each and every new
simulation tool.
2 Basic Equations for Thermalhydraulic Systems Analysis

2.1 Introduction
This section presents the basic mass, momentum and energy equations used in typical
computer codes for thermalhydraulic simulation. The equations are derived from first
principles and the necessary approximations lead to the requirements for empirical
correlations. Closure is obtained by the equation of state.
The known territory of the basic mass, energy and momentum conservation equations (Bird et
al [BIR60]) is explored, herein, from the perspective of thermalhydraulic systems analysis for
nuclear reactors. See also [CAR95] for seminal coverage of this topic.
Invariably in the modelling of fluids, the conservation equations are cast in one of two main
forms (Currie [CUR74]): integral or distributed approach. The distributed (differential) form
sees infrequent use in the analysis of thermalhydraulic systems since the cost and complexity of
such a detailed analysis on even a single complex component of a system is enormous, which
makes this route to the analysis of systems of such complex components unrealizable.
Recourse is generally made to the integral or lumped form so that inter-relationships of various
components comprising a system can be simulated. Necessarily, the models used for the
individual components are much simpler than that of the detailed models based on the
distributed approach. Great care must be taken to ensure that the simpler models of the
integral approach are properly formulated and not misused.
It behooves us, then, to develop the models used in thermalhydraulic systems analysis from first
principles. This will provide a traceable and verifiable methodology to aid development and
validation of system codes, to elucidate the necessary assumptions made, to show pitfalls, to
show the common roots and genealogy of specific tools like FLASH [POR69], SOPHT [CHA75a,
CHA75b, CHA77a, CHA77b, SKE75, SKE80], RETRAN [AGE82], FIREBIRD [LIN79], CATHENA
[HAN95], etc., and to help guide future development.
The exploration proceeds by first establishing and discussing the general principle of
conservation. Next, this general principle is applied in turn to mass, momentum and energy to
arrive at the specific forms commonly seen in the systems approach. Closure is then given via
the equation of state and by supporting empirical correlations. Finally, the ideas developed
are codified in a diagrammatical representation to aid in the physical interpretation of these
systems of equations and to provide a summary of the main characteristics of fluid systems.
2.2 Conservation
We start, both historically and pedagogically, with a basic experimental observation:
"CONSERVATION".
This was, and is, most easily understood in terms of mass:
"WHAT GOES IN MUST COME OUT UNLESS IT STAYS THERE
OR IS GENERATED OR LOST SOMEHOW".
Although this should be self-evident, it is important to realize that this is an experimental

observation.
If we further assume that we have a continuum, we can mathematically recast our basic
experimental observation for any field variable, ψ:
  V ψdV =   ΓdV +  S  n ds
D
Dt (1)
V S
where
DDt = substantial derivative1 = change due to time variations plus change due to
movement in space at the velocity of the field variable, ψ,
V = arbitrary fluid volume,
Γ = net sum of local sources and local sinks of the field variable, ψ, within the
volume V,
ψ = field variable such as mass, momentum, energy, etc.,
t = time,
s = surface bounding the volume, V,
n = unit vector normal to the surface, and
S = net sum of local sources and local sinks of the fluid variable, ψ, on the surface s.
We can now use Reynold’s Transport Theorem [CUR74]:
ψ dV =  dV +  ψv  n ds
ψ
D
Dt V V t S
(2)
where
t = local time derivative, and
v = velocity of the field variable,
to give
dV = -  ψ v  n ds +  Γ dV +  S  n ds .
ψ
V t S V S
(3)
In words, this states that the change in the conserved field variable ψ in the volume V is due to
surface flux plus sources minus sinks. We can use another mathematical identity (Gauss’
1  D d
For a lucid discussion of the three time derivatives, , , , see [BIR60, pp 73-74].
t t dt
Divergence Theorem):
S A  n ds =    A dV , (4)

V
where
A = any vector, such as velocity, and
 = Del operator (eg.  = x i + y j + ...) .
Thus equation 3 can be rewritten:
dV = -    ψ v dV +  Γ dV +    S dV .

ψ
V t S V V
(5)
If we assume that this statement is universally true, i.e. for any volume within the system under
consideration, then the following identity must hold at each point in space:
ψ
= -  ψ v + Γ +   S . (6)
t
This is the distributed or microscopic form. Equation 3 is the lumped or macroscopic form.
They are equivalent and one can move freely back and forth between the two forms as long as
the field variables are continuous.
The above derivation path is not unique. One could start with an incremental volume and
derive (1) via (6). It is largely a question of personal choice and the end use. One school of
thought, attended by most scientists, applied mathematicians and academics, since they usually
deal with the local or microscopic approach, focuses on the conversion of the surface integrals
to volume integrals using Gauss’ Theorem. The volume integrals are then dropped giving the
partial differential or microscopic form. This path works well when a detailed analysis is
desired, such as subchannel flow in fuel bundles, moderator circulation in the calandria, etc.
The second school, which sees more favour among engineers, particularly in the chemical
industry, evaluates the surface integrals as they stand without converting to volume integrals.
This leads to a lumped or macroscopic approach useful for network analysis, distillation towers,
etc.
There exists a very large number of possible derivations, each with its own advantages and
disadvantages. As more and more detail is picked up in each class of models, numerical means
have to be used. In the limit of large numbers of nodes or mesh points, etc., both methods
converge to the same solution.
Since the above equations are basic to all subsequent modelling of thermalhydraulic systems,
one should keep in mind the basis for these equations:
1) Conservation as an experimental observation.
This is usually taken for granted. However, when the conservation equations for separate
phases in a mixture are under consideration, the various sinks and sources of mass, momentum
and energy are not entirely known and the interpretation of experimental data can be difficult
because of the complexity. It helps to keep in mind the distinctly different roles that we have
historically assigned to the players in the conservation process:
a) the local time derivative, ψt,
b) the advection term (flux), ψv,
c) the local sinks and sources, Γ, within a volume and
d) the local sinks and sources, S, on the surface of a volume.
If a clarity of form is adopted by establishing and maintaining a one-to-one correspondence
between the form and the physical processes, then a substantial pedagogical tool will have been
achieved. This proves invaluable in experimental design (to zero in on a particular process or
parameter), model formulation and interpretation, data analysis and presentation, correlation
development, etc. A model could lose its generality because, for instance, fluxes across
interfaces are written as a term in Γ, thus making the interfacial flux a local phenomena rather
than a boundary phenomena. This may be acceptable for a single geometry but causes the
model to break down when applied to diverse geometries.
2) The field variables are continuous within the volume V.
This is also usually taken for granted. But care must be exercised in multiphase flow where
discontinuities abound. A common approach, taken to simplify the complexity of multiphase
flow, is to average the terms in the conservation equations across the cross-sectional area of the
flow path. One could speculate that the error introduced in this manner could separate the
model from reality enough to make the solutions be "unreal", i.e. complex numbers,
singularities, etc. Further, fluctuating parameters are often smoothed by averaging over an
appropriate Δt. These averaged parameters and products of parameters are used in models
and compared to experiments. But there is no guarantee that, for instance,
1  1
ψ v dt = Δt  
ψ dt
1
  
v dt .
Δt Δt  Δt  Δt Δt 
Thus the use of time averaged parameters can lead to additional errors. Indeed, because of
the possibility of error due to space and time discontinuities, several investigators have offered
rigorous treatments for the distributed approach (see, for example, Delhaye [DEL81]). There is
no reason why these treatments could not be applied to the lumped approach, as well. But, at
this time, there is little incentive to do so since grid coarseness and experimental data are larger
sources of error. As always, the operative rule is - BUYER BEWARE.
We now proceed to treat the mass, momentum and energy equations in turn.
2.3 Conservation of Mass

Historically, mass was the first variable observed to be conserved:
V (γ ρ ) dV = -  γk ρk vk  n ds +  Γk dV +  Sk n  ds

(7)
t k k S V s
where
ρk = density of phase k (1 = liquid, 2 = vapour),
γk = volume fraction of phase, k, in volume V, and
Γk, Sk = phase sinks and sources, including chemical and nuclear effects.
The average density is defined as:
ρ = γ1 ρ1 + γ2 ρ2 = (1 - α) ρ1 + α ρ2 , (8)
whereρ = average density, and
α = void fraction.
Adding both phases together, equation 7 becomes:
V [(1 - α) ρ1 + α ρ2] dV = -  [(1 - α) ρ1 v1 + α ρ2 v2]  n ds


t S
(9)
= +  (Γ1 + Γ2) dV +  (S1 + S2)  n ds.
V S
In our case, Γ1 = -Γ2 (liquid boils or vapour condenses) and Sk = 0 (no mass sources or sinks at
surfaces). Therefore:
dV = -  ρ v  n ds
ρ
V t S
(10)
where
ρ v = (1 - α) ρ1 v1 + α ρ2 v2 . (11)
If we apply Gauss’ Theorem and drop the integrals we have:
ρ
+ ρv = 0 (12)
t
or

[(1 - α) ρ1 + α ρ2] +   [(1 - α) ρ1 v1 + α ρ2 v2] = 0 . (13)
t
This is the distributed form useful for modelling detailed flow patterns such as in the calandria,
vessels, steam generators and headers. Component codes such as THIRST [CAR81a] and
COBRA [BNW76] use this approach.
In contrast, system codes such as SOPHT [CHA77a], based on Porsching’s work [POR71], use the
lumped equations. These codes represent a hydraulic network of pipes by nodes joined by
links, discussed in detail in section 3. Mass, pressure and energy changes occur at the nodes.
Momentum changes occur in the links. Thus the network is treated on a macroscopic scale
requiring an integral approach to the fundamental equations. Flow details in pipes are not
considered. That is, diffusion, dispersion, advection, flow regimes, flow profiles, etc. are not
fundamentally accounted for but are covered by empirical correlations. Averaging techniques,
commonly used in the distributed approach are not used in the lumped approach mainly
because there is little incentive to do so. The main sources of error lie elsewhere, mainly in
the coarseness of the discretization in the direction of flow (i.e. node size) and in friction factors
and heat transfer coefficients.
Now,  ρdV is the mass, Mi, in the volume, Vi, of the ith node. Also, for our case, the surface
integral can be written as surface integrals over the individual flow paths into and out of the
volume or node. That is,
-  ρ v  n ds = j ρj vj Aj , (14)
S
where j represents inflow and outflow links with vj > 0 for inflow and <0 for outflow. Inherent
in equation 11 is the assumption that the integral, S ρ v  n ds can be replaced by the simple
product ρj vj Aj. This implies known or assumed (usually uniform) velocity and density profiles
across the face of the link (or pipe).
Thus we now have:
Mi
= j ρj vj Aj   Wj , (15)
t
where Wj is the mass flow. This is the typical representation in system codes. Thus for the
node-link type equations, we must add two more assumptions:
i) nodalization, and
ii) assumed velocity and density profile across the cross-section of a pipe.
These assumptions have far reaching ramifications that may not be immediately obvious. This
is discussed in more detail in section 3.
To conclude our progressive simplification, we note the steady state form of equation 15:
 ρ v A   W = 0. (16)
j j j j j
For a simple circular flow loop, the mass flow rate at steady state is a constant at any point in
the loop. Local area and density variations thus give rise to velocity variations around the
loop.
Local velocity then is:
W
v = . (17)
ρA
2.4 Conservation of Momentum

Newton observed that momentum is conserved, i.e. a body moves in a straight line unless
forced to do otherwise. This is equivalent to a force balance if the inertial force (a momentum
sink of sorts) is recognized. In the integral sense, the rate of change of momentum is equal to
the forces acting on the fluid. Thus:
D
Dt V γk ρk vk dV = S σk  n ds +  γk ρk fk dV +  Mk dV ,
V V
(18)
where
σ is the stress tensor (i.e., short range or surface effects including pressure, viscosity, etc.),
f is the long range or body force (i.e., gravity),
and M is the momentum interchange function accounting for phase change effects.
Using Reynold’s Transport Theorem, we get:
V (γ ρ v ) dV +  (γk ρk vk) (vk  n) ds


t k k k S
(19)
=  σk  n ds +  γk ρk fk dV +  Mk dV .
S V V
Adding both phases together as per the mass equation, we find:
V ρ v dV +  ρ v (vn) ds =  σ  n ds +  ρ f dV .


(20)
t S S V
To get the microscopic form we use Gauss’s theorem and drop the volume integral as before to
leave:

(ρ v) +   ρ v v =   σ + ρ f . (21)
t
The stress tensor, σ, can be split into the normal and shear components:
σ = - PI + τ , (22)
where P is the pressure, I is the unity tensor and τ is the shear stress tensor. This enables the
explicit use of pressure and helps maintain our tenuous link with reality. Of course, it can
equally be introduced in the integral form, equation 20, or as a separate pressure for each
phase in equation 19. At any rate, equation 21 becomes:

(ρ v) +   ρ v v = -P +   τ + ρ f . (23)
t
This is the form commonly seen in the literature, useful for distributed modelling as per the
mass conservation equation. The term, τ, is usually replaced by an empirical relation. For
the system codes using the node-link structure, we switch back to the macroscopic form,
Equation 20.
If the surface integral for the advective term is performed over the inlet and outlet areas of the
pipe (link) in question, then:
 v( v  n)ds   v( v  n)ds   v( v  n)ds

S Ain Aout
(24)
where AIN is the flow inlet area and AOUT is the flow outlet area. If we assume the properties
are constant over the areas, then:
v
V  AININvINvIN  A OUT OUT v OUT vOUT   σk n  ds   fdV (25)
t S V
Alternatively we could perform a cross-sectional average of each term, usually denoted by < >,
where <( )> = 1A  ( ) ds. If we assume the properties, V, ρ and A are constant along
S
the length of the pipe, then the second and third terms cancel.
Equation 25 can be rewritten as:
v  
V
t

   PI  nds        f dV 
(26)
V  fL v v g
 A OUTPOUT  AINPIN    k   LA sin     
L D  2g c  gc 
where gc is the gravitational constant, g is the acceleration due to gravity and where τ and ρf
evaluated by empirical correlations (the standard friction factor) plus an elevation change term
(θ is the angle w.r.t. the horizontal). Note that is AOUT  AIN then, even for constant pressure,
there is a net force on the volume causing it to accelerate if it were not restrained. In a
restrained system such as HTS piping, the piping supports exert an equal and opposite force on
the volume. Thus when the area differences are explicitly modelled, the appropriate body
forces must be included. Generally, it is simpler to use an average or representative area for
the IN and OUT surfaces and to add entrance and exit frictional losses explicitly in the (fl/D+k)
term.
Assuming one dimensional flow and defining the mass flow as W  ρVA, and L as the pipe
length, equation 26 becomes:
W A   fL  W 
2
g
 PIN  POUT     k  2
 A sin() (27)
t L  D  2g cA  gc
which is the form typically used in system codes.

If circumstances require, extra terms can be added. For instance, if a pump is present this can
be considered to be an external force acting through head, ΔPpump. Equation 27 would then
become:
W
L   A OUTPOUT  AINPIN  APPUMP (28)
t
The momentum flux terms (Aρv2) in equation 25 could also be added if large area or property
changes were present or the effect could be included in the friction term.
In the steady state, for a constant area pipe with no pump and no elevation change:
2 2
 fL  V  fL  W
PIN  POUT    k    k 2  PPUMP  ... (29)
D  2g c  D  2A gc
As a final note, the assumptions made for the mixture momentum equation are thus similar to
those made for the mixture mass equation and the same comments apply. One cannot hope
to accurately model such phenomena as void propagation and other two phase transient flow
effects using lumped single phase equations unless a large number of nodes and links are used.
2.5 Conservation of Energy

By the early 1800’s, philosophical jumps were made in recognizing that heat was not a
substance and in the emergence of electromagnetic theory. The concept of energy as we now
think of it was formulated and it was found that energy, too, was conserved, as long as we
carefully identify all the different forms of energy (kinetic, chemical, potential, nuclear, internal,
electromagnetic, ...).
The mathematical statement of the conservation of energy is:
γk ρk ek + 2 vk  dV = -  qk  n ds +  Ek dV
1 2
D
Dt V 
  S V
(30)
+  γk ρk fk  vk dV +  (σk  n)  vk ds ,
V S
where
ek = internal energy of phase k,
qk = surface heat flux for phase k, and
Ek = internal heat sources and sinks of phase k.
The left hand side is the substantial derivative of the internal plus kinetic energy. The right
hand side terms are, respectively:
1) surface heat flux,
2) internal sources and sinks,
3) work due to long range body forces (gravity, etc.),
4) work due to short range forces (surface tension, pressure, etc.).
Using Reynold’s Transport Theorem again:
V    1 2
γk ρk ek + 2 vk  dV +
t   
S  1 2
γk ρk ek + 2 vk  vk  n ds
 
(31)
= -  qk  n ds +  Ek dV +  γk ρk fk  vk dV +  (σk  n)  vk ds .
S V V S
Summing over k, the mixture equation becomes:
 ρ e + 2 ρv2  dV +   ρe + 2 ρ v2  v  n ds
1 1
V  
t 



S 


(32)
= -  q  n ds +  E dV +  ρ f  v dV +  (σ  v)  v ds ,
S V V S
where
ρe = γ1ρ1e1 + γ2ρ2e2 E = E1+E2, etc.
Using Gauss’ Theorem to change some of the surface integrals to volume integrals:
 ρ e + 2 ρv2  dV +  ρe v  n ds +    2 ρ v2 v  dV
1 1
V  
t 

 S V 


(33)
= -  q  n ds +  E dV +  ρ f  v dV +    (σ  v) dV .
S V V V
Since
σ = -PI + τ ,
V   (σ  v) dV =  [  (τ  v) -   (P v) ] dV .
V
This is the total energy equation, composed of thermal terms and mechanical terms. We can
separate the two by first generating the mechanical terms from the momentum equation
(equation 20). Forming the dot product with velocity we get:
V (ρ v)  v dV +  v  (  ρ v v) dV =  v  (  τ) dV


t V V
(34)
-  v   P dV +  ρ f  v dV .
V V
Now
v  (  τ) =  (τ  v) - τ : v , (35)
v   P =   (P v) - P   v , (36)
  1   1 
v (ρ v) =  ρ v  v = 2 ρ v 2  (37)
t t 2  t  
and
1 
v  (  ρ v v) =   2 ρ v2 v . (38)
 
Using these identities and subtracting equation 34 from equation 33, we get:
V (ρ e) dV +  ρe v  n ds = -  q  n ds

t s S
(39)
+  E dV +  τ:v dV -  P  v dV .
V V V
This is the thermal form of the energy equation. This form of the energy equation can be used
to generate the thermal conductance equation for solids. By setting fluid velocity to zero and
converting surface integrals to volume integrals we get the distributed form:

(ρe) = -   q + E , (40)
t
where E is the internal energy generation rate term.
From thermodynamics, for solids, we have:
 e T
(ρe)  ρ  ρ Cv , (41)
t t t
and using Fourier’s law for heat conduction:
q = - kT , (42)
we have the classical form of the heat conduction equation:
T
ρ Cv = kT + E
t (43)
= k 2 T + E space independent k .
This is useful for determining the temperature distributions in boiler tube walls, piping walls and
reactor fuel pencils. To generate the node-link forms we now turn back to the integral form of
equation 39. If we assume that the density and enthalpy are uniform over the node (the
volume in question), then
U
V 
t
(ρ e) dV =
t
, (44)
where
U  Vρe = LAρe. (45)
The integral of the transport term can be written over the flow surfaces:
S ρ e v  n ds =   ρ e v  n ds +   ρ e v  n ds + .... , (46)

A1 A2
where A1, A2, etc., are the pipe flow cross-sectional areas. For inflow, vn is negative. For
outflow, vn is positive. Assuming uniform velocity, enthalpy and density across the link (pipe)
cross-section gives:
 ev  nds    evA  

INFLOW
i
OUTFLOW
evA   WINeIN   WOUT eOUT (47)
The heat flux and generation terms of the thermal energy equation can be lumped into a
loosely defined heat source for the volume.
-  q  n ds +  E dV  Q . (48)
S V
Therefore, the thermal energy equation becomes:

U
  WINeIN   WOUT eOUT  Q   τ : vdV   P  vdV (49)
t
The last two terms are the irreversible and reversible internal energy conversion, respectively.
Some system codes track enthalpy rather than internal energy. Defining:
P
h  enthalpy  e  and H  Vh. (50)

we can rewrite equation 39 as follows:
dV +  (ρ h - P) v  n ds = -  q  n ds
 (ρ h - P)
V t s S
(51)
+  E dV +  τ:v dV -  P  v dV .
V V V
Collecting the pressure terms and simplifying yields:
V (ρ h) dV +  ρ h v  n ds = -  q  n ds +  E dV

t S S V
(52)
+  τ:v +  dV +  P v  n ds -  P  v dV .
P
V V t S V
The surface integral over P can be transformed into a volume integral using Gauss’ theorem and
combined with the last term to give:
S P v  n ds -  P   v dV =    (P v) dV -  P   v dV

V V V
(53)
=  v   PdV .
V
The enthalpy flux terms can be evaluated in the same manner that the energy flux terms were
in equations 46-47. Thus,
 v  nds   W h   W
S
IN IN h
OUT OUT . (54)
Finally, using equations 48, 50, 53-54, equation 52 becomes:

H  P 
   WINhIN   WOUThOUT  Q   τ : vdV     v P  dV. (55)
t  t 
The integral term involving pressure is often neglected since it is usually negligible compared to
the other terms. For instance, the typical CANDU Heat Transport System operates at a
pressure of 10 MPa, a fluid velocity of ~10 ms, and a pressure gradient of less than 70 kPam.
This translates into roughly 10 kJkg while e is approximately 1000 kJkg.
The turbulent heating term is usually approximated by adding pump heat as a specific form of
Q.
equation 55 in the steady state, neglecting turbulent heating and the pressure terms, is the
familiar:
Q    WINhIN   WOUThOUT . (56)
For a reactor or a boiler (one flow in, one flow out):

Q  W hOUT  hIN   WCp  TOUT  TIN  in single phase. (57)
Another special case of equation 55 is obtained by expanding the term Q as per equation 48:
-  q  n ds +  E dV  Q . (48)
S V
Using Newton’s Law of cooling for convection:

q  n = hN (T - Ts) , (58)
where
qn = heat flux normal to surface, s,
T = Temperature of fluid
Ts = Temperature of surface (wall), and
hN = heat transfer coefficient,
Equation 55, neglecting the pressure terms, becomes:
h P  e T 
V V V  V C v    WINhIN  WOUTh OUT
t t  t t  (59)
 AhN  T  Ts   VE   τ : vdV
which is useful for accounting for heat transfer between the fluid and the pipe or tube walls (eg:
boiler heat transfer).
The heat transfer coefficient, hN, is supplied through empirical relations. The turbulent heating
term V τ:  v dV generally can be neglected or added as a pump heat term.
2.6 The Equation of State

From the conservation equations, we have three equations for each phase (mass, momentum
and energy conservation) and four unknowns:
1) density, ρ or mass, Vρ,
2) velocity, v, or mass flow, W, or momentum, ρv,
3) energy, e, or enthalpy, h = e + Pρ, or temperature, T = fn(e) or fn(h), and
4) pressure, P.
The fourth equation required for closure is the equation of state:
P = fn(h,ρ) ρ = fn(P,T) , etc. (60)
Thermodynamic equilibrium is usually assumed. For water, H2O or D2O, tables of properties
give the required functional relationship. Often, a curve fit of the tables is used. This data is
input to the computer codes and utilized in table lookup schemes or directly via the parametric
curve fits.
The equation of state is discussed in detail in Section 4.
2.7 Empirical Correlations

As previously discussed, supporting relations are required to provide the necessary information
for the conservation and state equations. The primary areas where support is needed are:
1) relationship between quality and void fractions, i.e., slip velocities in two phase flow (to
link the mass and enthalpy via the state equation);
2) the stress tensor, τ (effects of wall shear, turbulence, flow regime and fluid properties on
momentum or, in a word: friction);
3) heat transfer coefficients (to give the heat energy transfer for a given temperature
distribution in heat exchangers, including steam generators and reactors);
4) thermodynamic properties for the equation of state;
5) flow regime maps to guide the selection of empirical correlations appropriate to the
flow regime in question;
6) special component data for pumps, valves, steam drums, pressurizers, bleed or degasser
condensers, etc; and
7) critical heat flux information (this is not needed for the solution of the process equations
but a measure of engineering limits is needed to guide the use of the solutions of the
process equations as applied to process design;
The above list of correlations, large enough in its own right, is but a subset of the full list that
would be required were it not for a number of key simplifying assumptions made in the
derivation of the basic equations. The three major assumptions made for the primary heat
transport system are:
1) one dimensional flow;
2) thermal equilibrium (except for the pressurizer under insurge); and
3) one fluid model (i.e. mixture equations).
These are required because of state of the art limitations (however, two fluid models are being
used increasingly in recent years.). Empirical correlations were discussed in more detail in
Chapter 6.
2.8 Solution Overview

Because of the complexity of solving the mass, momentum and energy equations plus
supporting equations of state and empirical correlations all subject to initial and boundary
conditions, it is quite easy to "not see the forest for the trees". A skeleton overview may help
in this regard. Figure 1 illustrates the equations and the information links between them. In
words, the momentum equation gives the flows or velocities from one node to another, or from
one grid point to another, based on a given pressure, flow, mass and energy distribution. The
updated flows are used by the mass and energy equations to update the mass and energy
contents at each location. The new mass and energy are given to the equation of state to
update the pressure distribution. The new pressure, along with the new densities and
energies are used by the momentum equation, and so on. In this manner, a time history of
the fluid evolution is obtained. Of course, only the main variables are noted. The numerous
Figure 1 The four cournerstone single phase flow equations and the flow of
information between them.
and diverse empirical correlations require updates on the main variables and many secondary
variables. This information also "flows" around the calculation.
A further point to note on the solution overview is that each phase in a multiphase flow has a
main information flow path as shown in figure 2. In the full UVUEUP (unequal velocity, energy
and pressure) model, there are two distinct phases: one for the vapour phase and one for the
liquid phase. If a simplified model was imposed, this essentially means that the planes would
touch at some point. For instance, if equal pressure in both phases was assumed, then figure
3 would result. Here, the equation of state is common to both planes.
The HEM (homogeneous equilibrium model) is the fully collapsed case where both planes
collapse into one (figure 1). You may find these images to be useful in conceptualizing the
basic equations and how they fit together.
The precise solution procedure that you might employ is case dependent. At present, no
general solution scheme exists because the nuances of specific problems are subtle and
because one cannot usually afford to ignore the efficiency and cost savings gained by tuning a
method to a particular case. The economics of using a case specific code are changing,
however, with developments in the microcomputer field and with the realization that total
design and analysis time can often be reduced by using a less efficient but more robust code.
Codes such as SOPHT and CATHENA [HAN95] are a direct result of this realization. The near
term evolution will likely be affected mostly by microcomputer developments.
Figure 2 The four cornerstone equations for the two-fluid model.
Figure 3 The four cornerstone equations for the full two-fluid model with
equal pressure of the two phases.
2.9 Problems
1. Referring to figure 1:
a. Explain the inter-relationship between the mass, momentum and energy equations
and the equation of state.
b. For the integral form, devise a simple solution scheme for the transient equations.
Show what equations are being solved and how they are being solved. Flow chart
your scheme.
3 Nodalization
3.1 Introduction
This section focuses
ses on establishing a rationale for, and the setting up of, the geometric
representation of thermalhydraulic systems. The hydraulic network is represented by a series
of interconnected nodes to form a node-link diagram.
The exploration proceeds by first establishing and discussing the governing rationale. Next,
limitations of the approximation are presented and examples are given. Finally, the matrix
approach is used to capture the system geometry in a succinct form.
3.2 The Node-Link

Link Concept
From section 2 we have the integral mass, momentum and energy equations for an arbitrary
volume, i, with material flow through various surfaces, designated by the subs
subscript
cript j (see figure
4):
M
Mi
= j ρj vj Aj   Wj , (61)
tt
W A   fL  W 
2
 PIN  POUT     k  2
 Ag / g c sin() (62)
t L  D  2g cA 
H
  WINhIN   WOUThOUT  Q (63)
t
Figure 4 A general node and connecting links.
These mass and energy equations are averaged over the volume in question, hence they do not
capture any detail within the volume. KnowKnowing
ing the mass and energy of a volume, the
equation of state gives the pressure. Flow, however, is driven by pressure differences. Hence
it naturally follows that the momentum equation should be applied between the points of
known pressure, ie, between vol
volumes.
umes. In the distributed approach, this is called the staggered
grid method. In the lumped approach, it is called the node
node-link method and is illustrated in
figure 5.. Volumes are represented by nodes, flow paths are represented by links.
Figure 5 Two connected nodes.

To assign nodes and links to a given piping configuration, say the simple pipe of figure 6, it is
best to first focus on the flow modelling. The key question to ask is: Where should the link
endpoints (ie the node centres) be placed? The node centre locations define the positions at
which the pressure will be evaluated and this is important for correct flow calculations. For
constant area pipes, the placement is not critical but at junctions and area changes,
change modelling is
simplified if node centres are placed at junctions and pipe area changes. For the case shown
in figure 7 (a), if the flow were going from left to right and the junction resistance were included
in link 1, then the pressure at node 2 would be the pressure just downstream of the junction.
Carefully plan the node-link link configuration to match the problem at hand. If possible, avoid
running links across area changes since this inserts ambiguity into the flow area of the link
model, as illustrated in figure 7 (b).
Figure 6 Node-link setup for a simple pipe.
Figure 7 Node-link setup for an area change in a pipe.
Hydraulic friction can be affected by flow direction. Figure 7 illustrates a simple pipe flow
situation wherein there is a step change in area. Flow from left to right experiences a different
junction resistance than flow going from right to left. Direction dependent resistances are
usually modelled explicitly in the system codes. The momentum flux terms, Aρv2, can either
be modelled explicitly or through the resistance coefficient, k. Note that a simple force
balance around the junction would show that there is a net lateral force on the pipe. This
force imbalance would have to be accounted for by a body force if different inlet and outlet pipe
areas were used. This is another reason that links are chosen to coincide with constant pipe
length sections.
The properties of the fluid within the link are a result of the properties of the upstream node.
As the fluid is transported along a flow path (ie along the link), the link properties will change
over time. Naturally there will be a transport delay but given that the nodal properties are
themselves average values that change relatively slowly over time, system simulation codes
typically assume that the link properties are just the same as the upstream nodal properties.
For most purposes this is an acceptable assumption that can be lessened by using more nodes
and links in the model. One has to be careful, however, of flow reversal situations that involve
two-phase flow since this can lead to rapid and large density changes in the link.
The node volume is usually assigned as the fluid surrounding the node centre as shown in figure
7(a). But this is not a critical assignment; the node “centres” can actually be at the edge of the
volume if that proves convenient. From a numerical point of view, it is beneficial to divide the
hydraulic network up into volumes of roughly equal size since the properties in small volumes
can change very rapidly and thus force the use of correspondingly small time steps. This
rationalization of the volume assignments may force the user to take some liberties with the
notion of a node “centre”.
To recap, the momentum equation is used to solve for Wj in all links, driven by upstream and
downstream pressure differences and retarded / accelerated by friction, elevation change,
pumps, etc. that appear in the links. This flow transports mass and energy to and from the
nodes. Local heat sources and sinks, such as surface heat fluxes, are modelled at the nodes.
3.3 Nodal Diffusion

In the node - link representation of flow in a pipe, no flow detail is considered as the fluid
moves along the pipe. Therefore, no diffusion, dispersion, advection, flow profiles or flow
regimes are explicitly permitted. This is not too crude an approximation for the calculation of
pressure drops and flows but for modelling the propagation of disturbances, this approach is
inadequate as it stands unless a large number of nodes and links are used.
To show this, consider a homogeneous or bubbly flow through a pipe, as in the two-phase
regions of typical heat transport systems in nuclear reactors, modelled in system codes as nodes
connected by links. Perfect mixing at the nodes is assumed. Flow in a pipe, however, has
aspects of plug flow. That is, flow is transmitted along the pipe relatively undisturbed. If no
diffusion or turbulent dispersion existed, a sharp discontinuity in a property would propagate
undisturbed. Figure 8 shows how the discontinuity would move in time and space. The left
to right movement is due to the velocity, v, while the spreading out is due to diffusion. If a
single mixing tank (node) represented a section of pipe of volume, V m3, and volumetric flow, f
m3/s, then a step change to zero in a field property, C, (which could be concentration or density)
entering the node would be an exponential by the time it left the node, that is:
t

COUT  CIN  e 
(64)
where τ = V/f ; τ is also the transmission time for the plug flow model. If the pipe were
modelled by two nodes in series,
2t
 2t  
COUT NODE2  CIN NODE1  1   e  (65)
 
and in general, for n nodes:
nt N k 1
  nt  1
COUT NODE2  CIN NODE1e 
  
k 1    n  1 !
(66)
Figure 8 Illustration of convection and diffusion.
Figure 9 compares the transmission of a step change for various numbers of nodes and the plug
flow model. It is easy to see whwhyy the codes model void propagation poorly. A very large
number of nodes are needed to transmit a disturbance without appreciable distortion. The
phase relationships or timing, of the propagation is very important in determining the stability
of a thermal hydraulic system. A pocket of void reaching a given destination at an earlier or
later time may enhance or cancel the phenomenon in question. The smearing of a wave front
Figure 9 Transmission of a step change using the plug flow model

and the mixing tank model (1 to 50 tanks).
alters the timing and gain and hence affects stability. The slow convergence of the
t mixing
model to the plug flow model explains the typically slow convergence of such system codes as
more nodes are added to increase accuracy.
Thus, nodalization creates a form of diffusion in much the same manner as finite difference
schemes create numerical diffusion (see, for instance, Roache [10]). Attaining convergence in
nodalization is, in essence, converging tthe
he model to plug flow behaviour. But is the flow in
typical heat transport systems plug flow?
Flow in the CANDU feeders (38 to 76 mm) at 15 M Msec
sec may indeed be plug flow. But some
turbulent mixing does take place. More importantly, the feeders are of va
varying
rying length and the
flow has a spectrum of qualities. This gives quite a spectrum in transit time. This will skew
the propagation of a disturbance as illustrated in figure 10.. Thus, depending on the transit
time spectrum, a 5 node approximation (say) m may
ay be quite a good representation. The risers
and headers may also give more diffusion than plug flow. These pipes are large diameter and
the flow is turbulent. Very little is know of flow regimes and propagation properties in these
situations.
In short, careful attention should be given to nodalization for meaningful simulation, quite apart
from the normal numerical concerns such as the Courant limit, etc.
Figure 10 Transmission of a step change using the plug flow model

mod and
a feeder model with skewing due to differences in transit times.
3.4 Examples
In figures 11 to 13,, some common piping situations are depicted. In figure 11, a simple Tee
junction, note that each link has a unique junction resistance associated with the
th flow path of
that link. Note also that a link has a unique upstream node and a unique downstream node.
Links are always terminated by nodes at either end; in effect, the nodes provide boundary
conditions for the links. There are 2 nodes per link, no mo
more,
re, no less. A node, on the other
hand, can have many links connected to it.
The Y junction of figure 12 has a node link structure that is identical to the Tee junction. The
differences in the two types of junctions are captured in the details of the correlations
co for
friction, flow regimes, etc.
Figure 13 shows a CANDU HTS header and connecting piping. Note that there is no best or
unique node-link
link representation. The requirements of the problem at hand dictates the
number of nodes and links and the la layout
yout of the representation. For instance, it is useful to
place a node centre at the point of a pressure measuring device so that experimental data can
be more readily compared to the simulation results.
Figure 14 shows a typical nod-link
link diagram for a CANDU Heat Transport System simulation.
Figure 11 Simple Tee junction.
Figure 12 Simple Y junction.
Figure 14 Sample node

node-link connection for a header.
Figure 13 Node-link diagram: ¼ circuit CANDU.
3.5 Matrix Notation

As we shall see, it is sometimes expedient to cast the governing equations in matrix form. To
illustrate, consider the node-link network of figure 15. Nominal flow directions are assigned to
be positive in the normal flow direction. The mass balance equations for the 4 nodes are:
dM1
dt = -W1+W4
dM2
dt = +W1-W2+W5
(67)
dM3
dt = +W2-W3
dM4
dt = +W3-W4-W5
 i  dMi then the mass balance equations can be written

If we define M
dt
M 1   1 0 0 1 0   W1 
 
     W2 
 M2    1 1 0 0 1    W   m
  AMW  w (68)
M 3   0 1 1 0 0   3 
     W4 
  4   0 0 1 1 1   W 
M 
 5
where AMW is a 4x5 matrix (number of rows = number of nodes N=4, and number of columns =
number of links L=5) and w is the flow vector. Generally, upper case bold will be used to
indicate a matrix and lower case bold will be used to indicate a vector. The superscript MW
denotes that the matrix relates to the mass equation and to the flow vector. It also indicates
the size of the matrix (nodes x links)
There can be up to L entries in a row but only 2 non-zero entries in any column - no more, no
less. The AMW matrix uniquely defines the geometry. The matrix is most easily constructed
on a column by column basis, ie on a link by link basis: for each link (column vector) place a -1 in
the location of the upstream node and a +1 in the location of the downstream node. As we
shall see, all other matrices that arise in the solution to the mass, momentum and energy
equations can be derived from the structure of AMW. This is very handy for computer coding
purposes.
Figure 15 4 Node - 5 link diagram.
3.6 Exercises
1. For the 4 node, 5 link example of section 3.5, write out the flow and enthalpy equations as
individual equations and in matrix form. Compare the structure of AWP and AHW to AMW,
where the superscript WP denotes that the matrix relates to the flow equation and the
pressure vector, and the superscript HW denotes that the matrix relates to the enthalpy
equation and the flow vector.
2. For the case of 2 connected, open tanks of water with surfaces at different elevations, set up
the node-link diagram and the mass, momentum and enthalpy equations.
3. What would be different if the tanks in question 2 were closed?
4 Equation of State
4.1 Introduction
As discussed in the previous sections, the momentum equation gives an update on the flows or
velocities from one node to another, or from one grid point to another, based on a given
pressure, flow, mass and enthalpy distribution. The updated flows are used by the mass and
enthalpy equations to update the mass and enthalpy contents at each location. This
information is given to the equation of state to update the pressure distribution which, along
with the new densities and enthalpies is used by the momentum equation, and so on. In this
manner, a time history of the fluid evolution is obtained. Of course, only the main variables
are noted. The numerous and diverse empirical correlations require updates on the main
variables and many secondary variables. This information also "flows" around the calculation.
The exploration of the appropriate forms of the equation of state to use for systems analysis
begins by reflecting on the thermodynamics and the iterative method of finding pressure.
Next a non-iterative method is offered as an improvement. This leads naturally to the water
property evaluation. Fast, accurate curve fits are presented.
4.2 Thermodynamic Properties

From a thermodynamics viewpoint (see, for instance Sears [SEA75], the equation of state of a
substance is a relationship between any four thermodynamic properties of the substance, three
of which are independent. An example of the equation of state involves pressure P, volume V,
temperature T and mass of system:
π (P, V, T, M) = 0 (69)
If any three of the four properties are fixed, the fourth is determined.
The equation of state can also be written in a form which depends only on the nature of the
system and not on how much of the substance is present, hence all extensive properties are
replaced by their corresponding specific values. Thus
π (P, v, T) = 0 (70)
is the specific value form of the above equation of state, where v is the specific volume. If any
two of the thermodynamic properties are fixed, the third is determined.
From a thermodynamic point of view, the appropriate way to present water properties is by
tables or formula for each property expressed as a function of the independent parameters P
and T, as per Meyer [MEY67 or Haar [HAR84] (figure 16). Thus given values of pressure and
temperature, the calculation of other thermodynamic properties is usually straightforward.
On the other hand, the determination of pressure from known values of other thermodynamic
properties is not direct since interpolation and iteration is required. Unfortunately, T and P are
rarely the independent parameters in system dynamics since the numerical solution of the
conservation equations yield mass and energy as a function of time. Hence, from the point of
view of the equation of state, it is mass and energy which are the independent parameters.
Consequently, system codes are hampered by the form of water property data commonly
available.
Figure 16 P-v-T surface for water .
A key point to note is that the conservation equations are all cast as rate equations whereas the
equation of state is typically written as an algebraic equation. This arises from the basic
assumption that, although the properties of mass, momentum and energy must be traced or
solved as a function of time and space, the corresponding local pressure is a pure function of
the local state of the fluid. Process dynamics are not considered. This is the essence of the
equilibrium assumption (in a like manner, of course, we invariably use steady state heat transfer
coefficients, etc. in dynamic processes). Historically, this mixture of form arose because
thermodynamics endeavours were concerned with equilibrium states and not with system
processes. System modellers, on the other hand, emphasized system dynamics and used what
was available for constitutive relations. System modellers are more concerned with numerical
problems.
But the decisive role of the equation of state in determining system dynamics was recognized
early. Paynter [PAY60] identifies the power throughput as being the most important
parameter for system dynamics. Power is composed of the product of effort (i.e. force or
pressure) and flow. Porsching [POR71] correctly identifies the important role of flow in his work
and by keying the formulation of node-link networks to flow, stable, efficient and accurate
solution schemes result. However, the role of pressure has not received the equivalent
acknowledgement. Although the system dynamics are captured in Porsching’s Jacobian, the
essence of the system dynamics is not apparent. Nahavandi [NAH70] comes much closer to
recognizing the role of pressure
ure and explicitly casts the equation of state in rate form.
Unfortunately, the system essence is again not apparent because Nahavandi’s form is very case
specific.
Most other popular schemes, for instance, Agee [AGE83], use the algebraic form of the
equation
tion of state. This treatment puts the pressure determination on the same level as heat
transfer coefficients. Thus, although numerical solution of the resulting equation sets give
correct answers (to within the accuracy of the assumption), intuition is not generated and
time consuming iterations must be performed to get a pressure consistent with the local state
parameters.
We look first at such an iterative scheme and then consider a more efficient alternative (the rate
method).
4.3 The Iterative Method

Given the density and enthalpy of a volume of water, the task at hand is to find the associated
values of pressure and temperature. Figure 17 shows qualitatively the relation between
density, ρ, and enthalpy, h, for a given P. At low enthalpy, the fluid is single phase liquid and
the density is high. As heat is added and the fluid reaches saturation temperature, vapour is
generated to form a two-phase
phase mixture and the density approaches the vapour density. The
curve is well behaved and continuous making it a suitable candidate for numerical search
routines.
Figure 17 Numerical search for P given ρ and h for a two-phase

phase mixture.
We start the iteration procedure by guessing a pressure. Usually in system transient t

simulation codes, the value of P at a previous time step is a good choice. Given P we calculate
hfsat and hgsat, the saturation enthalpies for the liquid and vapour phases, respectively. If h <
hfsat then the fluid is single phase liquid. If h > hgsat then the fluid is single phase vapour.
phase mixture with a quality, x  [0,1].
Otherwise the fluid is a two-phase
The case of two-phase
phase equilibrium is considered first. Subsequently, the equations are
extended to cover single phase and two
two-phase non-equilibrium fluid.
4.3.1 Two-Phase
Phase Equilibrium Fluid
For two-phase
phase fluid, the density and enthalpy are functions of the pressure and quality. Since
we know the density, ρ, we can estimate the quality (xest) for the guessed P (assuming a
homogeneous mixture) since:
1
v = ρ = vf(P) + xest vfg(P) (71)
and thus calculate the enthalpy based on the guessed P:

hest = hf(P) + xest hfg(P) (72)
This estimated value of h will differ from the known value of h. This d difference
ifference is used to drive
the iteration, ie, to update the guessed pressure as illustrated in figure 18:
Δh
ΔP = (73)
(hP)ρ
The denominator in equation 5 must be evaluated numerically if analytical expressions are not
available. The pressure is updated via:
P = P + ΔP (74)
and the iteration is repeated until the pressure has converged to some tolerance. The
temperature is just the temperature of saturated fluid at that pressure.
Figure 18 Error correction scheme for pressure in two

two-phase.
phase.
4.3.2 Single-Phase Sub-cooled and Superheated Fluid

For single phase fluid, the density and enthalpy are functions of P and T, ie:
ρ = ρ(P, T) h = h(P, T) (75)
For a guessed P and T, ρ and h can be found directly from the water property tables. But this is
just an estimate since P and T are guessed. The true values of ρ and h lie some distance away
and, to a first approximation, the true values and the estimated values are related by a Taylor’s
series expansion:
ρ ρ
ρ = ρest +  ΔT +  ΔP (76)
TP PT
h h
h = hest + PΔT + PTΔP (77)
T 
Defining Δρ = ρ - ρest and Δh = h - hest, we solve for ΔP and ΔT:
h ρ
 Δρ -  Δh
TP TP
ΔP = (78)
ρ dh ρ h
  -  
PT dT P TP PT
h ρ
 Δρ -  Δh
PT TT
ΔT = (79)
ρ dh ρ h
  -  
TP dPT PT TP
or, more compactly,
ΔP = G1P Δρ + G2P Δh (80)
ΔT = G1T Δρ + G2T Δh (81)
The derivatives must be evaluated numerically if analytical expressions are not available.
The pressure and temperature are updated via:
P = P + ΔP T = T + ΔT (82)
and the iteration is repeated until the pressure and temperature have converged to some
tolerance.
4.4 The Rate Method

We next consider a scheme (called the Rate Method) that eliminates the need for iteration with
no loss in accuracy. The case of two-phase equilibrium is considered first in order to illustrate
the method. Subsequently, the equations are extended to cover single phase and two-phase
non-equilibrium fluid.
4.4.1 Two-Phase Equilibrium

For a two-phase homogeneous mixture we have:
v = vf + xvfg (83)
h = hf + xhfg (84)
where vfg  vg-vf and hfg  hg-hf.

We wish to relate rates of change of pressure to rates of change in ρ and h. Specifically, we
desire:
dP dρ dh
dP = G1 dρ + G2 dh dt = G1 dt + G2 dt (85)
since dρdt and dhdt (or equivalently, dMdt and dHdt) are available from the mass and
enthalpy conservation equations. First concentrating on the case of constant ρ (or v), to
obtain G2, we differentiate equation (16) to gives:
dh h dP hf x hfg dP

=   =  + h + x  . (86)
dt Pρ dt  P fg P P  dt
Using equation (83), holding v constant (i.e., ρ = constant):
v - vf
  v 
dx  fg  1 vf vfg
= = -  + x . (87)
dP P vfg  P P 
Substituting this into equation (86) gives:
dh  hf hfg hfg vf vfg  dP

=  + x -  + x  (88)
dt  P P vfg  P P   dt
or equally:
dP vfg dh
dt =  hf hfg vf vfg  dt
 vfg  +x  - hfg  +x 
  P P   P P   (89)
vfgdh dh
= dt = G2 dt .
{DENOMINATOR}
This gives the pressure rate response due to an enthalpy rate change, holding ρ constant.
If we repeat the above but holding h constant we find:
dP - hfg dv hfg v2 dρ dρ
dt = {DENOMINATOR} dt = {DENOMINATOR} dt = G1 dt . (90)
Note that G1 and G2 are functions that depend only on the local saturation fluid properties and
their slopes at the local pressure.
Combining equations 89 and 90 to get the total pressure rate response when both h and ρ are
varying:
dP dρ dh
= G (P, x) + G (P, x) dt . (91)
dt 1 dt 2
This is the rate form of the equation of state for two-phase equilibrium fluid in terms of the
intensive rate properties, dρdt and dhdt, which are obtained from the continuity equations.
Equation 91 can be cast in the extensive form by noting that, since ρ = MV and h = HM,
dρ 1 dM M dV
= V dt - 2 dt (92)
dt V
and
dh 1 dH H dM
dt = M dt - M2 dt . (93)
Substituting into equation 91 and collecting terms:

dP  G1 G2 H dM G2 dH G1 M dV
=  -  +
dt V M2  dt M dt - V2 dt . (94)
After some simplification and rearrangement we find:

dM dH dV
F1 dt + F2 dt + F3 dt
dP
dt = Mg F4 + Mf F5 (95)
where:
F1 = hg vf - hf vg
F2 = vg - vf
F3 = hf - hg
hg vg
F4 = (vg - vf) - (hg - hf)
P P (96)
hf vf
F5 = (vg - vf) - (h - hf)
P P g
Mg  x M
Mf  (1 x) M .
The F functions are smooth, slowly varying functions of pressure provided good curve fits are
used. The latest steam tables [HAA84] were used to fit saturated properties to less than 14%
accuracy using low order polynomials and exponentials [GAR88]. Considerable effort was
spent on obtaining accuracy and continuous derivatives over the full pressure range. The fact
that good fits are available means that the F functions are well behaved which in turn makes the
rate form of the equation of state extremely well behaved, as shown later.
The G functions are also well behaved for the same reasons.
The F and G functions have direct physical interpretations which aid in generating intuition.
The F functions relate changes in the extensive properties, M, H and V, to changes in pressure.
The G functions related changes in the intensive properties, ρ and h, to changes in pressure.
Often, a simple numerical evaluation of these functions during a simulation aids in developing
an appreciation of the changing roles of the key actors in a dynamic simulation.
For instance, because F1 is negative, we immediately see that adding mass to a fixed volume of
liquid with fixed total enthalpy will cause a depressurization (because the specific enthalpy,
h = HM, is decreased). But, since G1 is positive, an increase in density in a fluid of fixed
specific enthalpy causes a pressurization.
4.4.2 Single-Phase Sub-cooled and Superheated Fluid

For the single-phase sub-cooled or superheated case, we do not have to account for the sorting
out between phases as we did for the two phase case. thus the derivation is more direct and
less complex. We could simply use:
P = π (ρ, h)(97)
to give:
dP P dρ P dh
dt = ρh dt + hρ dt (98)
but, since the steam tables are given as a function of P and T, the slopes in equation (98) are not
easily obtained. To cast the pressure rate equation in terms of the independent variables, P
and T, consider:
ρ = ρ (P, T) (99)
and
h = h (P, T) (100)
Note that the non-equilibrium case requires the explicit tracking of the temperature in addition
to pressure. Taking derivatives of Equations (99) and (100):
dρ ρ  dP ρ  dT
= T dt +  (101)
dt P  T P dt
and
dh h  dP h  dT
dt = P T dt + T P dt . (102)
But we desire:
dP dρ dh
dt = G1P dt + G2P dt (103)
and
dT dρ dh
dt = G1T dt + G2T dt . (104)
This is easily obtained by solving equations (101) and (102) for dPdt and dTdt to yield:
h dρ ρ dh
 - 
dP T P dt TP dt
dt = ρ dh (105)
ρ h
T dTP -  
P  TP PT
and
h dρ ρ dh
 - 
dT PT dt TT dt
dt = ρ dh (106)
ρ h
  -  
TP dPT PT T P
which is the intensive form we desire.
The extensive form is obtained as for the two-phase equilibrium case. Equations (92) and (93)
are substituted into equations (105) and (105) and after rearrangement we find:
dM dH dV
F1P dt + F2P dt + F3P dt
dP
dt = Mv F4P + Ml F5P (107)
and
dM dH dV
F1T dt + F2T dt + F3T dt
dT
dt = Mv F4T + Ml F5T (108)
where
(109)
4.4.3 Two-Phase Non-Equilibrium

The rate form for the equation of state for the two-phase non-equilibrium case is a simple
extension of the single-phase non-equilibrium case. The liquid and vapour phases are treated
independently to give:
dPk dρk k
dhk
+ G (110)
k
dt = G 1P dt 2P dt
dTk dρk k
dhk
+ G (111)
k
dt = G 1T dt 2T dt
where k represents either l or v for the liquid or vapour phases respectively. In general, the 6
equation model (3 continuity equations for each phase) would be used for the general unequal
temperature, unequal velocity, unequal pressure situation. Thus dρkdt and dhkdt are
available to the rate form of the equation of state.
4.5 Water Property Fits

To facilitate the calculation of light water properties, the 1984 standard tables were accurately
curve fitted as discussed in detail at www.nuceng.ca/water/h2ohome.htm. These fitted
functions are supplied in the files H2OPROP.FOR and H2OPROP.C for user convenience. These
FORTRAN and C functions cover a wide range of pressures and temperatures and should be
sufficient for most nuclear reactor simulations,
with the exception of severe accidents that
generate extreme conditions. These functions Heavy water properties have of course
are fast and more than accurate enough given the been developed in-house at AECL and
other errors in system simulation [GAR88, GAR89, Ontario Power Generation for use in
GAR92]. their industry computer tool-set. Some
The basic overall approach taken in the curve tools are available in the public domain;
fitting task was that, since the more difficult region see www.nuceng.ca/d2o/d2ohome.htm.
to fit was the transition from single to two-phase
and since most power plants operate at or near this region, careful attention would be paid the
phase transition region at the expense of accuracy away from the saturation line, if necessary.
Thus, the first major step was to accurately fit the saturation lines. Then, since density,
enthalpy and other properties vary more strongly with T than with P (as shown in figure 19), the
property in question, say density, would be calculated based on the deviation from the
saturation value at the given T, ie:
ρ
PT(
ρ(P,T) = ρsat(T) +  P - Psat(T)) (112)
Figure 20 illustrates the strategy. It should be obvious by now that not only the properties
need to be fitted but the slopes are needed as well. Both the properties and the slopes of the
properties must be free of discontinuities if numerical searches are to converge.
Having derived the desired rate forms for the equation of state, we proceed to section 5 to
illustrate the utility of the approach.
Figure 19 Density vs. pressure at various temperatures in sub

sub-cooled
cooled water.
Figure 20 Basis for curve fitting in the subcooled region.

4.6 Problems
4. Using the spread sheet macros for Microsoft Excel supplied by G.R. McGee (as per
http://www.nuceng.ca/water/h2ohome.htm), calculate and plot the density, enthalpy,
quality and void fraction for a range of pressures ( 1 to 100 atmospheres) and
temperatures(50 C to 350 C). Make sure you cover the subcooled, saturated and
superheated ranges.
5. Using the code, WATERA.EXE (see http://www.nuceng.ca/water/h2ohome.htm):
a. Calculate ρ and h for P=10 MPa and T=300 C. Increase the temperature in steps
to see the approach to two-phase.
b. Using ρ and h slightly different than that found in (a), calculate P and T.
c. Practice calculating ρ given h and P.
5 The Rate Form of the Equation of State

5.1 Introduction
In conjunction with the usual rate forms of the conservation equations, the time derivative form
of the Equation of State is investigated from a numerical consideration point of view. By
recasting the equation of state in a form that is on equal footing with the system conservation
equations, several advantages are found. The rate method is found to be more intuitive for
system analysis, more appropriate for eigenvalues extraction, as well as easier to program and
to implement. Numerically, the rate method is found [GAR87a] to be more efficient and as
accurate than the traditional iterative method.
First, the derivation of the rate form of the Equation of State is presented. Systematic
comparison between the new method and the traditional iterative method is made by applying
the methods to a simple flow problem. The comparison is then extended to a practical
engineering problem requiring accurate prediction of pressure.
5.2 The Rate Form

Presently, the conservation equations are all cast as rate equations whereas the equation of
state is typically written as an algebraic equation [AGE83]. This arises from the basic
assumption that, although the properties of mass, momentum and energy must be traced or
solved as a function of time and space, the corresponding local pressure is a pure function of
the local state of the fluid. Hence the equation of state is considered only as a constitutive
equation. This treatment puts the pressure determinations on the same level as heat transfer
coefficients. Although numerical solution of the resulting equation sets give correct answers
(to within the accuracy of the assumption), intuition is not generated and time-consuming
iterations must be performed to get a pressure consistent with the local state parameters.
The time derivative form of the Equation of State is investigated, herein, in conjunction with the
usual rate forms of the conservation equations. This gives an equation set with two distinct
advantages over the use of algebraic form of the Equation of State normally used.
The first advantage is that the equation set used consists of four equations for each node or
point in space, characterizing the four main actors: mass, flow, energy and pressure. This
consistent formulation permits the straight-forward extraction of the system eigenvalues (or
characteristics) without having to solve the equations numerically.
The second advantage is that the rate form of the Equation of State permits the numerical
calculation of the pressure without iteration. The calculation time for the pressure was found
to be reduced by a factor of more than 20 in some cases (where the flow was rapidly varying)
and, at worst, the rate form was no slower than the algebraic form. In addition, because the
pressure can be explicitly expressed in terms of slowly varying system parameters and flow, an
implicit numeric scheme is easily formulated and coded.
This section will concentrate on this numerical aspect of the equation of state.
The equation of state has been discussed in section 4 where we saw that the determination of
pressure from known values of other thermodynamic properties is not direct. Interpolation and
iteration is required because the independent (known) parameters are temperature, T, and
pressure, P. Unfortunately, T and P are rarely the independent parameters in system dynamics
since the numerical solution of the conservation equations yield mass and energy as a function
of time. Hence, from the point of view of the equation of state, it is mass and energy which
are the independent parameters. Consequently, system codes are hampered by the form of
water property data.
Having derived the desired rate forms for the equation of state in section 4, we proceed to
illustrate the utility of the approach.
5.3 Numerical Investigations: a Simple Case

The simple two-node, one-link system is (Figure 21) chosen to illustrate the effectiveness of the
rate form of the equation of state in eliminating the inner iteration loop in thermalhydraulic
simulations. In general, the task is to solve the matrix equation,
u
= Au + b (113)
t
over the time domain of interest. The key point that we wish to discuss is the difference in the
normal method (where u = {M1, H1, W, M2, H2}) and the rate method (where u = {M1, H1, P1, W,
M2, H2, P2}). For simplicity and clarity, we first summarize work for a fixed time step Euler
integration:
ut+Δt = ut + Δt[Au + b] (114)
As we shall see, this is sufficient to generate some observations on the utility of the rate
method. These observations then guide us in the use of more complicated and efficient
algorithms.
5.3.1 Normal Method

The normal method obtains the value of pressure at time, t+Δt, from an iteration (as discussed
previously) on the equation of state using the values of mass and enthalpy at time, t+Δt, i.e. the
new pressure must satisfy:
Pt+Δt = fn(ρt+Δt, ht+Δt) (115)

where both ρ and h are pressure dependent functions. Any iteration requires a starting guess
and a feedback mechanism. Here, the starting guess for pressure is the value at time, t: Pt.
Feedback in the Newton-Raphson scheme is generated by using an older value of pressure, Pt-Δt,
to estimate slopes. Since the slope, hP, was readily available from the rate method, we
chose to use this slope to guide feedback. Thus, in the comparison of methods, we have
borrowed from the rate method to enhance the normal method. This provides a stronger test
of the rate method.
Thus we can now generate our next pressure guess from:

h-hest
Pnew = Pguess + *ADJ (116)
hP
where h is the known value of h at t+Δt and hest is the estimated h based on the guessed
pressure as discussed in detail in section 4. ADJ is an adjustment factor [0, 1], to allow
experimentation with the amount of feedback. This iteration on pressure continues until a
convergence criteria, Perr, is satisfied. The converged pressure is used in the outer loop in the
momentum equation and the time can be advanced one time step. Figure 22 summarizes the
logic flow.
Figure 21 Simple 2-node, 1-link system.
Figure 22 Program flow diagram for the normal method.
5.3.2 Rate Method

The rate method obtains the value of pressure at time, t+Δt, directly from the rate equation as
is done for the conservation equations. Equation 95, gives the rate of change of pressure
which can be solved simultaneously with the conservation equations if substitutions for dMdt
and dHdt are made, leading to:
u
= Au + b (117)
t
where u = {M1 H1, P1, W, M2, H2, P2} .
Thus:
= Pi + Δt[Au + b]i (118)
t+Δt t
Pi
No inner iteration is required, as shown in Figure 23.

One problem with this approach is that the pressure may drift away from a value consistent
with the mass and energy. This problem does not arise with the conservation equations
because the equations are conservative in form, by design. It is not possible to cast the rate
form of the equation of state in conservative form since pressure is simply not a conserved
property. We can surmount the drift problem by using the feedback philosophy of the normal
method. Thus the new pressure is given by:
h-hest
= Pi + Δt[Au + b]i + *ADJ (119)
t+Δt t
Pi
hP
This correction term uses only readily available information in a non-iterative manner.
In essence, the main effective difference between the normal and rate method is that during
the time step between t and t+Δt the normal method employs parameters such as density,
quality etc. derived from the pressure at time, t+Δt, whereas the rate form employs parameters
derived from the pressure and rate of change of pressure at time, t. The normal method is not
necessarily more accurate, it is simply forcibly implicit in its treatment of pressure. The rate
method can be implicit (as we shall see) but it need not be. Without experimentation it is not
evident whether the necessity of iteration in the normal method is outweighed by the possible
advantages of the implicit pressure treatment.
The next sections test these issues with numerical experiments.
Figure 23 Program flow diagram for the rate method.
5.3.3 Comparison
The two node, one link numerical case under consideration is summarized in figure 21. Perhaps
the most startling difference between the normal and rate methods is the difference in
programming effort. The rate form was found to be extremely easy to implement since the
equation form is the same as the continuity equations. The normal method took roughly
twice the time to implement since separate control of the pressure logic is required. This
arises directly from the treatment of pressure in the normal method: it is the odd man out.
The second startling difference was ease of execution of the rate form compared to the normal
form. The normal form required experimentation with both the pressure convergence
tolerance, Perr, and the adjustment factor, ADJ, since the solution was sensitive to both
parameters. The rate method contains only the adjustment factor ADJ. The first few runs of
the rate method showed that since the correction term for drift (h-hest)(hp) is always several
orders of magnitude below the primary update term, Δt{A u + b}, the solution was not at all
sensitive to the value of ADJ. Thus the rate method proved easier to program and easier to
run than the normal method.
We look at the number of iterations required for pressure convergence as a function of Perr and
ADJ for the normal method without regard to accuracy. For a Δt of 0.01sec, Pert = 10-3 (fraction
of the full scale pressure of 10 MPa), the effect of ADJ is seen in figure 24. This result is typical:
an adjustment factor of 1 gives rapid convergence (one or two iterations) except where very
large pressure changes occur. For the case of very rapid changes, the full feedback (ADJ = 1)
causes overshoot. Overall, however, the time spent for pressure calculation is about the same,
independent of ADJ.
Allowing a larger pressure error had the expected result of reducing the number of iterations
needed per routine call. But choosing a smaller time step (say .001) did not have a drastic
effect on the peak interations required. The rate method, of course, always used 1 iteration
per routine call and the adjustment factor ADJ was found to be unimportant since the drift
correction factor amounted to no more than 1% of the total pressure update term.
The integrated error for both methods is shown in figure 25. Both methods converge rapidly
to the benchmark. The value of Perr is not overcritical. A value of Perr consistent with
tolerances set for other simulation variables is recommended. The time spent per each
iteration is roughly comparable for both methods. The main difference is that the rate method
requires the evaluation of the F functions over and above the property calls common to both
methods. This minor penalty is insignificant in all cases studied since the number of iterations
 call dominated the calculation time.
In summary, to this point, the rate method is easier to implement, more robust and is equal to
the normal method at worst, more than 20 times faster under certain conditions. We now
look at incorporating a variable time step to see how each method compares.
Figure 24 Number of iterations per pressure routine call for the normal method with a
time step of 0.01 seconds and a pressure tolerance of 0.001 of full scale (10 MPa).
Figure 25 Integrated flow error for the rate method and the normal method for
various fixed time steps, convergence tolerances and adjustment factors.
Typical variable time step algorithms require some measure of the rate of change of the main
variables to guide the Δt choice. The matrix equation, equation 113, provides the rates that
we need. Since the rate method incorporated the pressure into the u vector, the rate of
change of pressure is immediately available. For the normal method, the rate of change of
pressure has to be estimated from previous history (which is no good for predicting the onset of
rapid changes) or by trial and error. The trial and error method employed here is to calculate
the Δt as the minimum of the time steps calculated from:
(fractional tolerance)x(scale factor for ui)
Δti = (120)
uit
This restricts Δt so that no parameter changes more than the prescribed fraction for that
parameter. This can be implemented in a non-iterative manner for the rate method.
However, for the normal method, the above minimum Δt based on u is used as the test Δt for
the pressure routine and the rate of change of pressure is estimated as:
P Pt+Δt - Pt
= Δt (121)
t
The Δt is then scaled down if the pressure change is too large for that iteration. Then the new
Δt is tested to ensure that it indeed satisfies the pressure change limit. This iteration loop has
within it the old inner loop.
It is expected then, that the normal method will not perform as well as the rate method
primarily because of the "loop within a loop" inherent in the normal method as applied to
typical system simulation codes.
A number of cases were studied and the results of the normal method were compared to the
rate method. The figure of merit was chosen as
10,000
F.O.M.= (122)
(integrated error)x(total pressure routine time)x(No. of adjustable parameters)
Thus, an accurate, fast and robust method achieves a high figure of merit. Some results are
listed in table 1. Derating a method with more adjustable parameters is deemed appropriate
because of the figure of merit should reflect the effort involved in using that method. On
average, about 6 runs of the normal method, with various Perr and ADJ were needed to scope
out the solution field compared to 1 run for the rate method. Thus a derating of 2 is not an
inappropriate measure of robustness or effort required.
The results indicate that the rate method is a consistently better method than the normal
method in terms of numerical performance. We see no reason why this improvement would
not exist for any thermal hydraulic system in which pressure field determination is required.
Table 1 Figure of merit comparisons of the normal and rate forms of the
equation of state for various convergence criteria (simple case).
Next we briefly discuss implicit numerical schemes.

The nodal equations are:
dM1 dM2
dt = -W dt = +W (123)
dH1 dH2
dt = -h1W dt = +h1W (124)
dMi dHi
dPi F1 dt + F2 dt
dt = MgF4 + MfF5 , i = 1,2 (125)
Considering just the flow and pressure rate equations, we have (after substituting in for dMdt
and dHdt):
dW A A
dt = L(P1-P2) - LK|W|W (126)
and
dP1 dP2
dt = -χ1W dt = +χ2W (127)
where χ1 and χ2 are > 0 and are given by:

F1 + hF2
χ = M F +MF (128)
g 4 f 5
evaluated at the local property values of nodes 1 and 2.

Employing a scheme that is implicit in flow, the difference equations are cast
Wt+Δt-Wt A t+Δt t+Δt A
= L (P1 -P2 ) - L K|Wt|Wt+Δt (129)
Δt
t+Δt t
P -Pi t+Δt t
i
= ±χiWt+Δt impliesP -P = ±χiWt+ΔtΔt (130)
Δt i i
Collecting terms and solving for the new flow:
 A A   A t t 
Wt+Δt = 1+LK|Wt|Δt + L (χ1+χ2)Δt2-1Wt + L P1-P2 Δt
    ( ) (131)
This is the implicit time advancement algorithm employing the rate form of the equation of
state. For the normal method, the pressure rate equation in terms of flow (i.e., equation 130)
is not available to allow an implicit formulation of the pressure. Consequently, the implicit
time advancement algorithm for the normal method is:
 A t+Δt t+Δt 
Wt+Δt = 1+LK|Wt|Δt-1Wt + L P1 -P2 Δt ( )
A
(132)
   
To appreciate the difference between equations 131 and 132, consider the eigenvalues and
vectors of
u(t)
= A(u,t)u(t) (133)
t
If we assume, over the time step under consideration, that A = constant and has distinct
eigenvalues, then the solution to equation 133 can be written as:
N
u(t) =  uleαlt (134)
l=1
where ul = eigenvectors
αl = eigenvalues.
It can be shown that for the explicit formalism, the numerical solution is equivalent to:
N
ut+Δt =  (1+αlΔt)ul (135)
l=1
while the implicit form is:

Nul
ut+Δt =  (1-α Δt) (136)
l=1 l
The eigenvalues can often be large and negative. Thus, at some Δt, the factor (1+αlΔt) can go
negative in the explicit solution causing each subsequent evaluation of u to oscillate in sign and
go unstable. For the implicit method, the contributions due to large negative eignevalues
decays away as Δt  . Thus the implict formalism tend to be very well behaved at large time
steps. Positive eigenvalues, by a similar argument pose a threat to the implicit form.
However, this is not a practical problem because αlΔt is kept <<1 for accuracy reasons. Thus,
as long as the solution algorithm contains a check on the rate of growth or decay (effectively the
dominant eigenvalues) then the implicit form is well behaved.
With this digression in mind, we see that the implicit rate formalism (equation 131) has more of
the system behaviour represented implicitly than the normal method (equation 132). Thus,
we might expect the rate from to be more stable than the normal form. Indeed, this was
found to be the case as shown in figure 26. For a fixed and large time step (0.1sec.) the
normal method showed the classic numerical instability due to the explicit pressure treatment.
The rate form is well damped and very stable, showing that this method should permit the user
to "calculate through" pressure spikes if they are not of interest.
Figure 26 Flow vs. time for the implicit forms of the

normal and rate methods.
5.4 Numerical Investigations: a Practical Case

The comparison between the normal and rate methods is extended to a practical application
where a two node homogeneous model is used to simulate a transient of a small pressurizer
operating at near-atmospheric pressure. The procedure is briefly described in the following
[SOL85].
Figure 27 illustrates the problem. Steam and stratified liquid water in the pressurizer are
schematically shown as two control volumes (nodes). The nodal fluids are assumed to be at
saturated two-phase conditions corresponding to the pressure at their respective control
volumes. The overall boundary conditions to the system are the steam bleed flow at the top
of the pressurizer, the flow into and out of the pressurizer through the surge line, heat input
from heaters at the bottom of the pressurizer and heat loss to pressurizer wall.
The rate of change of mass, MS in the steam control volume and ML in the liquid control volume,
can be expressed by the following:
dMs
dt = -WSTB-WCD-WCI+WEI+WBR (137)
dML
dt = WSRL-WEI-WBR+WCD+WCI (138)
where WSTB is the steam bleed flow, WSRL is the surge line inflow, WCI is the interface
condensation rate at the liquid surface separating the steam control volume from the liquid
control volume, WEI is the interface evaporation rate at the same liquid surface, WCD is the flow
of condensate droplets (liquid phase) from the bulk of the steam control volume toward the
liquid control volume, and WBR is the rising flow of bubbles (gas phase) from the bulk of liquid
volume toward the steam volume.
Figure 27 Schematic of control volumes in the pressurizer.
The rate of change of energy in the two control volumes can be expressed by the rate of change
in the total enthalpy, HS and HL, in the steam and liquid control volumes respectively:
dHS
dt = -WSTBhgST-WCDhfST-WCIhgST+WEIhsLQ+WBRhgLQ-QWS+QTR-(1-β)[(1-δ)QCOND+QEVPR]
(139)
and
dHL
dt = WSRLhSRL-WEIhfLQ-WBRhgLQ+WCIhfST+WCDhfST-QWL+QPWR-QTR-β[(1-δ)QCOND+QEVPR]
(140)
where hSRL is the specific enthalpy of the fluid in the surge line, hgST and hfST are respectively the
saturated gas phase specific enthalpy and the saturated liquid phase specific enthalpy in the
steam control volume, hgLQ and hfLQ are respectively the saturated gas phase specific enthalpy
and the saturated liquid phase specific enthalpy in the liquid control volume, QWS and QWL are
the rate of heat loss to the wall in the steam control volume and in the liquid control volume
respectively, QTR is the heat transfer rate from the liquid control volume to the steam control
volume due to any temperature gradient, excluding those due to interface evaporation and
condensation; QCOND is the rate of energy released by the condensing steam to both the steam
and liquid control volumes during the interface condensation process and QEVPR is rate of energy
absorbed by the evaporating liquid from both the steam and liquid control volumes during the
interface evaporation process. The constant, β, represents the fraction of these energies
distributed to or contributed by the liquid control volume. The ratio δ represents the portion
of energy released during the interface condensation that is lost to the wall.
The calculation of swelling and shrinking of control volumes is only done for the liquid control
volume and the volume in the steam control volumes will be related to the volume in the liquid
control volume, VL, as:
dVS dVL
dt = - dt (141)
The swelling and shrinking of the liquid control volume as well as values of WSTB, WSRL, WCI, WEI,
WCD, WBR, QWS, QWL, QTR, QPWR, β and δ are calculated using analytical or empirical constitutive
equations. The majority of these parameters depend directly or indirectly on pressure. Any
inaccurate prediction of pressure during a numerical simulation will result in severe numerical
instability. Hence the above problem is a good testing ground for comparing the
performances of the two methods.
During the test simulation, the pressurizer is initially at a quasi-steady state. The steam
pressure is at 96.3 kPa. The steam bleed flow, WSTB, heater power QPWR and heat losses QWL
and QWS are at their quasi-steady values, maintaining the saturation condition of the pressurizer.
At time = 11 sec., the steam bleed valve is closed and WSTB drops to zero while QPWR is increased
to a fixed value of 300 Watts. At time = 16 sec., the steam bleed valve is reopened and its set
point set at 80 kPa.

Since the thermodynamic properties in the steam control volume and the liquid control volume
are functions of PS and PL (pressures of the respective control volumes), there are seven
unknowns, namely: MS, ML, HS, HL, VS (or VL), PS and PL. Adding two equations of state, one
for each control volume, will complete the equation set:
MS HS 
PS = fn(ρS,hS) = fn V ,M  (142)
 S S
ML HL 
PL = fn(ρL,hL) = fn V ,M  (143)
 L L
Both the normal iterative method and the rate method are tested to solve. The following
observations are made:
1. Using the normal method, the choice of adjusting P to converge on h given ρ or
converging on ρ given h is found to be very important in providing a stable numerical
result. At time step = 10 msec, no complete simulation result can be generated when ρ
was the adjusted variable. An explanation of this can be given by referring to G1(P,x),
or Pρ, This factor is proportional to the square of [x vg(P) + (1-x)vf(P)]. However, the
direction of change in the saturated gas phase specific volume with pressure is opposite
to that of saturated liquid phase specific volume:
2. dvfdP > 0
3. dvgdP < 0
4. Therefore, a fluctuation in the value of pressure during an iteration process will amplify
the fluctuation in the value of predicted density when that method is used;
5. Using enthalpy as the adjusted variable to converge on P, simulation results can be
generated if an error tolerance E of less than 0.2% is used. The error tolerance is
defined as:
ABS(h-hestimate)
6. E = h x100%
7. Figure 28 shows the transient of PL and PS for E = 0.2%. Unstable solutions result for E
higher than 0.2%. The average number of iteration is found to depend on the error
tolerance as shown in figure 30.
8. On the other hand, the performance of the rate method is much more convincing in
both accuracy and efficiency. The transient of PL and PS predicted using the rate
method is shown in Figure 29.
Figure 29 Pressurizer's pressure transient for the normal method with

error tolerance of 0.2%.
Figure 28 Pressurizer's pressure transient for the rate method.
Figure 30 Average number of iterations per pressure routine call for the
normal method in simulating the pressurizer problem.
5.5 Discussion and Conclusion

The rate form is a cogent expression of the equation of state that is distinct from the normal
algebraic form. The essential difference is that the rate form expresses the relationship
between the rates of change of the state variables, while the normal form relates the static
values of the state variables. Although this is stating the obvious, the change in viewpoint is
revealing.
No barrier is perceived to applying the rate form to the multi-nodelink case, to the distributed
form of the basic equations, and to eigenvalue extraction (numerical or analytical).
Although we have not made use of it in this work, the non-equilibrium form (equations 110 and
111) is provocative. It entices one to view the non-equilibrium situation as the essentially
dynamic situation that it is and helps to focus our attention on the thermal relaxation. Given
the temperature rate equations, the non-equilibrium situation should be easy to incorporate
without a major code rewrite.
We conclude by restating our major findings. The rate method offers many advantages:
1. It is more intuitive for system work. It permits a proper focus on the two main
actors, flow and pressure.
2. The same form is appropriate for eigenvalue extraction as well as numerical
simulation. This extends the usefulness of coding.
3. Programs are easier to implement.
4. Programs are more robust and require less hand holding.
5. Time step control and detection of rapid changes (like phase changes) is
improved.
Overall the method is usually faster and more accurate. Time savings peaked at a ratio of 26
for the cases considered.
5.6 Problems
6. Consider 2 connected volumes of water with conditions as shown in figure 21. Model this
with 2 nodes and 1 link.
a. Solve for the pressure and flow histories using the normal iterative method for the
equation of state,
b. Solve for the pressure and flow histories using the non-iterative rate method.
c. Compare the two solutions and comment.
7. Vary the initial conditions of question 1 so as to cause void collapse in volume 2 during the
transient. What problems can you anticipate? Solve this case by both methods.
6 Thermalhydraulic Network Simulation

6.1 Introduction
This section introduces some more advanced numerical algorithms for solving systems of
ordinary differential equations such as found in the modelling of thermalhydraulic networks.
Explicit algorithms are simple to devise and program but they are restricted in time step so as to
ensure stability. The more implicit the formulation, the more stable the solution in most
instances. Larger time steps can be used for implicit algorithms but the accompanying matrix
manipulation is computationally costly. Herein, we explore the tradeoffs.
Porcsching’s method is explored to show the methodology and its limitations. Then the rate
form of the equation of state is used with the conservation equations to develop a generalized
fully implicit (at least in terms of the main variables) formalism. Porsching’s method is a
special case of the general method. The section concludes with some programming notes.
6.2 Porsching’s Method

One of the more successful algorithms for thermalhydraulic simulation is based on the work of
Porsching [POR69, POR71]. This algorithm, involving the Jacobian (derivative of the system state
matrix), is used originally in the computer program FLASH-4 [POR69] and subsequently in the
Ontario Hydro program SOPHT [CHA77] and evolved into forms used in RETRAN [AGE82].
The strength of Porsching's approach lies in its recognition of flow as the most important
dependent parameter and, hence, its fully implicit treatment of flow. This leads to excellent
numerically stability, consistency and convergence. Further, the Jacobian permits a generalized
approach to the linearization of nonlinear systems. This allows the development of a system
state matrix which contains all the system dynamics in terms of the dependent parameters of
mass, energy and flow. Back substitution finally gives a matrix rate equation in terms of the
system flow (the unknown) and the system derivatives. While this approach is certainly a proven
and successful one, it has some disadvantages. The matrix rate equation involving the Jacobian
is as complicated as it is general. The resulting expressions are somewhat obtuse and it is
difficult to obtain an intuitive feel for the system. This complexity also hinders implementation
in a simulation code and makes error tracking a tedious process. The pervasiveness and
obtuseness of the algorithm begs a revisit so as to distil the salient features, leaving them
exposed for pedagogy and further scrutiny.
Section 5 discussed the use of the Rate Form of the equation of state. This work showed that by
casting the equation of state in the form of a rate equation rather than the normal algebraic
form, the system state matrix can be more logically formed from the normal conservation rate
equations for mass, energy and momentum plus the pressure rate equation. These form the
four cornerstone equations in thermalhydraulic systems analysis (figure 1). Numerical
implementation of the rate form proved to be very successful, leading to roughly a factor of 10
improvement over the algebraic form of the equation of state, largely due to the iterative
nature of the algebraic form. Incorporating the implicit pressure dependency in the numerical
method also drastically improved the numerical stability.

Since Porsching's method also carried the pressure dependency implicitly (via the Jacobian), the
question arises as to how the Rate form compares the Porsching's method. This section is
devoted to an explanatory derivation of the fully-implicit back-substituted form (FIBS), which is
a more general than the Rate form. It is shown that the Porsching form is identical to the Rate
form and is a subset of the fully-implicit back-substituted form and is easily derived from it
[GAR87b]. The FIBS form thus offers an alternative to Porsching, is found to be of some
pedagogical usefulness and is far more intuitive and easier to code.
6.3 Derivation of FIBS

Following Porsching [POR71], the general form of system equations can be written
u = f(t,u) (144)
where u is the vector of dependent mass, total enthalpy and flow variables {Mi, Hi, Wj} for all
nodes i=1..N and all links, j=1..L. Equation 1 is linearized, assuming no explicit t dependence to
give:
u = ft + Δt J u (145)
or
Δu = Δt ft + Δt J Δu (146)
to give
[I-Δt J]Δu=Δt ft (147)
where J is the system Jacobian, composed of elements fk /ul.
For typical thermalhydraulic systems using the node-link notation2:
dWj Aj
dt = LJ (Pu + SWPΔPu - Pd - SWPΔPd) + kj (Wj + SWWΔWj)2 + bwj
(148)
ΔWj
= Δt
2
Porsching actually uses U, total energy rather than H, total enthalpy in a hybrid form:
 (H M ) W -  (H M ) W + Q
Ui = jd j j j ju j j j i
There is no advantage to tracking both H and U in a simulation; thus in this course, H is used
throughout.
Typically bwj = (Aj/Lj) (hjρjg + ΔPpump) where hj = height.
dMi   (W + S ΔMi
= (W
dt jd j + S MW ΔW j ) - ju j MW ΔW j )  Δt (149)
dHi  (Hj + SHH ΔHj)  (Hj + SHH ΔHj)

= (W + S ΔW ) - (W + S ΔW j (Mj + SHM ΔMj) + Qi
)
dt jd j HW j (Mj + SHM ΔMj) ju j HW
 WjHj SHW Hj SHH Wj SHM Wj Hj

= + ΔW + ΔH - ΔMj
jd  Mj Mj j Mj j 2 
 M
j 
 WjHj SHW Hj SHH Wj SHM Wj Hj
- + ΔW + ΔH - ΔMj+ Qi
ju  Mj Mj j Mj j 2 
 M
j 
ΔHi
 Δt
(150)
Pi Pi Pi
ΔPi = ΔMi + ΔHi + ΔVi
Mi Hi Vi
(151)
ΔPi ΔMi ΔHi
Δt = Cli Δt + C2i Δt for constant volume.
where j indicates a sum over all links for which the node i is a downstream (d) or upstream (u)
node.
Switches, S, are used to provide user control over the degree of implicitness:
0 = explicit
1 = implicit.
The system unknowns to be solved for are ΔW, ΔM, ΔH and ΔP using equations 5, 6, 7 and 8.
The general strategy is to reduce the number of unknowns so that the size of the matrices to be
inverted in the simultaneous solution of these equations is reduced. The mass equation 6 is
simple and is used to eliminate ΔM in terms of ΔW. Flow is chosen as the prime variable since it
is the main actor in thermalhydraulic systems. The enthalpy equation poses a problem as it is
too complex to permit a simple substitution. Porsching surmounts this by setting SHH = SHM = 0,
ie making the solution explicit in specific enthalpy. However, we need not make this
assumption; by casting the equations in matrix notation, the full implicitness can be retained
while still allowing the back substitutions to be made.
Proceeding then, using matrix notation:
ΔM = Δt AMW[Wt+SMW ΔW] (152)
where, for a 4 node - 5 link example (Figure 15):

links
 1 0 0 1 0 
 
1 1 0 0 1 
A MW 
 0 1 1 0 0 
 
 0 0 1 1 1  nodes (153)
This matrix contains the total system geometry. It is constructed by the following procedure:
For each column (link), insert -1 for the upstream node and +1 for the downstream node for
that link since the link supplies (adds) flow to the downstream node and takes it away from the
upstream node. Flow reversal is handled automatically since the sign of W will take care of
mass accounting properly.
The form of other matrices in the following are derivable from AMW. This can be used to
advantage in coding. The input data for each link need only contain pointers to the upstream
node and the downstream node for that link. This allows AMW to be created. In short, the
upstream node and downstream node for each link completely defines the geometry and this
can be used to programming advantage.
The flow equation is:
{
ΔW = Δt AWP [Pt+SWPΔP] + AWW[Wt+2SWWΔW]+BW } (154)
Where:
 -k1|W1| 0 
 
 -k 2 |W2 | 0 
A mWW =  0 0  (155)
 
 0 0 
 0 -k 5 |W5 | 

 A1 /L1 A1 /L1 0 0 
 
 0 -A 2 /L 2 -A 2 /L 2 0 
Am =  0
WP
0 A 3 /L3 -A 3 /L3  (156)
 
 -A 4 /L 4 0 0 A 4 /L 4 
 0 A 5 /L5 
 -A 5 /L5 0
note that AWP is formed easily from AMW by the following procedure:
First multiply AMW by {-A1/L1, -A2/L2, ... -A5/L5}-1

Then transpose the resulting matrix to give AWP.
A1L1(h1ρ1g + ΔPpump1)
BW =  A2L2(h2ρ1g + ΔPpump2)  (157)
  
  
Finally:
(
ΔH = Δt AHW[Wt+SHWΔW]+SHHAHH ΔH* - SHMAHM ΔM*+BH )
* *
(158)
where ΔH* and ΔM* refer to the enthalpy and mass associated with upstream properties of the
links (ie the transported properties). Thus
ΔH ΔM
ΔH21 ΔM12
 
ΔH* = ΔH3 ,
 
ΔM* = ΔM3 (159)
ΔH4 ΔM4
ΔH4 ΔM4
 -H1 /M1 0 0 H 4 /M 4 0 
 
HW 
H1 /M1 -H 2 /M 2 0 0 H 4 /M 4 
Am = (160)
 0 H 2 /M 2 -H3 /M 3 0 0 
 
 0 0 H 3 /M 3 -H 4 /M 4 -H 4 /M 4 
For each link, the elements of the column are formed from the link flow, Wj and the upstream
properties (H and M). Each link has a sink and source node.
Similarly
 -W1 /M1 0 0 W4 /M 4 0 
 
HH* 
W1 /M1 -W2 /M 2 0 0 W5 /M 4 
Am = (161)
 0 W2 /M 2 -W3 /M 3 0 0 
 
 0 0 W3 /M 3 -W4 /M 4 -W5 /M 4 
 -W1H1 /M12 0 0 W4 H 4 /M 42 0 
 2 2 2 
HM  W1H1 /M1 -W2 H 2 /M 0 0 W5 H 4 /M 
Am = 2 4
(162)
 0 W2 H 2 /M 2
-W3 H 3 /M 32 0 0 
 2

 0 0 W3 H 3 /M 32 -W4 H 4 /M 42 -W5 H 4 /M 42 
We wish to write the matrix equations eliminating the * parameters, ie convert ΔH* to ΔH, ΔM*
to ΔM. To do this we introduce a transfer matrix, ILN so that
ΔH* = ILNΔH (163)

where
nodes
1 0 0 0
0 1 0 0 

I mLN  0 0 1 0
 
0 0 0 1
0 0 0 1 
 links (164)
where ILN is formed by entering 1 for the node that is the upstream or source node for each link.
Now, we can define:
AHH* ΔH* = AHH* ILN ΔH (165)

 AHH ΔH
and AHM* ΔM* = AHM* ILN ΔM (166)
 AHM ΔM.
Thus
ΔH = Δt {AHW (Wt + SHW ΔW) + SHH AHH ΔH - SHM AHM ΔM + BH} (167)
Substituting in the mass equation 9:
ΔH = Δt { AHW (W + SHW ΔW) + SHH AHH ΔH - Δt SHM AHM AMW (Wt + WMW ΔW) + BH } (168)
Solving for ΔH:
ΔH = Δt[I - Δt SHH AHH]-1 {AHW (Wt + SHW ΔW) - Δt SHM AHM AMW(Wt + SMW ΔW) + BH} (169)
So now we have ΔM and ΔH in terms of ΔW. Recalling equation 8, in matrix notation, we
have:
ΔP = C1 ΔM + C2 ΔH, (170)
where
 C11 0 0 0 
 
0 C12 0 0 
Cm1   (171)
 0 0 C13 0 
 
 0 0 0 C14 
Similarly for C2.

We can back-substitute ΔM and ΔH into equation 8 and the result into the flow equation to
leave a matrix equation in ΔW only, which can be solved by traditional numeric means. Hence,
ΔP = Δt C1 AMW (Wt + SMW ΔW) + Δt C2 [I - Δt SHH AHH]-1 [AHW (Wt + SHW ΔW)
- Δt SHM AHM AMW (Wt + SMW ΔW) + BH]
 Δt APW1 Wt + Δt APW2 ΔW + Δt BP (172)
where : APW1 = C1 AMW + C2 [I - Δt SHH AHH]-1 [AHW - Δt SHM AHM AMW] (173)
PW2 MW HH -1 HW HM MW
A = SMW C1 A + C2 [I - Δt SHH A ] [SHW A -Δt SHM SMW A A ] (174)
BP = C2 [I - Δt SHH AHH]-1 BH (175)
Thus:
ΔW = Δt {AWP [Pt + Δt SWP (APW1 Wt + APW2 ΔW + BP)] + AWW [Wt + 2SWW AWW ΔW] + BW } (176)
Collecting terms in ΔW:
[I - Δt(2 SWW AWW + Δt SWP AWP APW2)] ΔW
= Δt {[AWW + Δt SWP AWP APW1] Wt + BW + AWP [Pt + Δt SWP BP]} (177)
which is of the form
A ΔW = B
which can be solved by conventional means to yield ΔW. Then we can directly calculate ΔM,
ΔH and ΔP using equations 152, 158 (or 167), and 170. Associated changes in temperature
can be obtained as for pressure, using the appropriate equation of state coefficients.
6.4 Special Cases

To summarize, the general solution is given by the following equations:
APW1 = C1 AMW + C2 [I - Δt SHH AHH]-1 [AHW - Δt SHM AHM AMW] (178)
APW2 = SMW C1 AMW + C2 [I - Δt SHH AHH]-1 [SHW AHW -Δt SHM SMW AHM AMW] (179)
BP = C2 [I - Δt SHH AHH]-1 BH (180)
[I - Δt(2 SWW AWW + Δt SWP AWP APW2)] ΔW
= Δt { [AWW + Δt SWP AWP APW1] Wt + BW + AWP [Pt + Δt SWP BP] } (181)
ΔM = Δt AMW [Wt + SMW ΔW] (182)
ΔH = Δt { AHW (Wt + SHW ΔW) + SHH AHH ΔH - SHM AHM ΔM + BH } (183)
ΔP = C1 ΔM + C2 ΔH (184)
Special cases of this general algorithm are as follows:
Fully explicit: all S's = 0
APW1 = C1 AMW + C2 AHW (185)
APW2 = 0 (186)
BP = C2 BH (187)
 ΔW = Δt { AWW Wt + BW + AWP Pt } (188)
ΔM = Δt AMW Wt (189)
HW t H
ΔH = Δt { A W +B } (190)
ΔP = C1 ΔM + C2 ΔH, (191)
as expected.
Porsching's semi-implicit (SHH = 0 and SHM = 0, all other S's = 1)
APW1 = C1 AMW + C2 AHW (192)
APW2 = C1 AMW + C2 AHW (193)
P H
B = C2 B (194)
[I - Δt(2 AWW + Δt AWP APW2)] ΔW
= Δt { [AWW + Δt AWP APW1] Wt + BW + AWP [Pt + Δt BP] } (195)
ΔM = Δt AMW [Wt + ΔW] (196)
ΔH = Δt { AHW (Wt + ΔW) + BH } (197)
ΔP = C1 ΔM + C2 ΔH (198)
Fully Implicit: All S’s = 1
APW1 = C1 AMW + C2 [I - Δt AHH]-1 [AHW - Δt AHM AMW] (199)
APW2 = C1 AMW + C2 [I - Δt AHH]-1 [AHW - Δt AHM AMW] (200)
BP = C2 [I - Δt AHH]-1 BH (201)
[I - Δt(2 AWW + Δt AWP APW2)] ΔW
= Δt { [AWW + Δt AWP APW1] Wt + BW + AWP [Pt + Δt BP] } (202)
MW t
ΔM = Δt A [W + ΔW] (203)
ΔH = Δt { AHW (Wt + ΔW) + AHH ΔH - AHMΔM + BH } (204)
ΔP = C1 ΔM + C2 ΔH (205)
6.5 Programming Notes

It should be noted that the full system geometry is contained in AMW. All other matrices are
derived from this matrix and node/link properties. Programming is thus very straightforward.
In addition, the switches, S, can be varied at will to control the degree of implications of the
system variables, W, M, H and P.
The fully-implicit method is more complicated than the semi-implicit method in that it requires
the addition and multiplication of more matrices as well as a matrix inversion. The effect of
these additional operations is quite costly, especially when a large number of nodes is needed.
In one case study [HOS89], for 9 nodes and links, the cost is a 50% increase in iteration time. But
this becomes a 250% increase as one approaches the 36 node/link case. By handling the matrix
operations as efficiently as possible, some increase in speed should be attainable for both
models. Using efficient assembly routines (rather than FORTRAN) for the matrix operations
yielded a 10 to 20% reduction (increasing from 9 nodes to 36 nodes) in the time per iteration for
the semi-implicit method and a 15 to 25% reduction in the fully-implicit case.
Usually the matrices contain mostly zeros and, in the case of a circular loop, may be diagonally
dominant in nature (i.e. non-zero elements occupy one, two or three stripes through the
matrix). By writing routines specific to the nodal layout for handling the matrix operations,
significant gains in speed may be possible. However, the simulator will no longer be general in
nature and the routines may have to be changed if the nodal layout is altered.
If the multiplication of two large matrices is desired, say NxN in dimension, the time to carry out
the operation (N3 multiplications and N3 additions) can be very significant. However, it is
possible to reduce the number of individual operations without losing the generality of the
method. Take, for example, the multiplication of AWP and APW. The rows in the former term
pertain to links and the columns to nodes. Each row will only contain two terms located in the
columns corresponding to the upstream and downstream nodes of that particular link. Thus,
knowing which are the upstream and downstream nodes for every link, it is only necessary to
do two multiplications and one addition to obtain each element of the product matrix (2N2
multiplications and N2 additions). By taking advantage of having only two elements in each row
of the former term or only two elements in each column of the latter term wherever possible,
significant savings in time may be observed. With this improvement in the code, a cut in time
by a factor of two for 18 nodes and by a factor of three for 36 nodes, regardless of the method
(semi- or fully-implicit) was obtained. The cost of the fully-implicit method is reduced slightly to
a 32% increase in iteration time over the semi-implicit method when 9 nodes and 9 links are
used. This becomes a 214% increase as one approaches the 36 node case.
Since the focus of this section is to provide a less obtuse and more general derivation of
thermalhydraulic system equations than Porsching's method, a full comparison of the
performance of the fully- and semi-implicit methods will not be made. Suffice it to say that, in
general, the semi-implicit method has a Courant limit on the maximum time step that can be
taken in order to ensure stability. The fully-implicit method does not have this limitation. As the
Courant time step limit is determined by the nodal residence time, the time step limit is
dependant on the node sizes and the flows through the nodes. Practical simulations have a
further time step constraints such as: the tracking of movement of valves, the maintenance of
accuracy, synchronizing of report times, etc. Thus, the choice between the semi- or fully-implicit
method depends on the time per iteration multiplied by the number of iterations required to
reach the largest time step permitted by the simulation problem. For example, for a 9 node
case, the semi-implicit method required 0.10 seconds per iteration and required 2 iterations to
meet the report time of 1.0 seconds. The fully-implicit method meet the report time in one
iteration which took 0.14 seconds. At 36 nodes however, the semi-implicit method took 2 x 0.71
seconds while the fully-implicit method took 2.12 seconds. Clearly, one method is not superior
to the other in all cases.
Pressure determination involves the use of property derivatives. To avoid the numerical
problems associated with discontinuities, smooth functions for properties must be used, such
as those derived by [GAR88, GAR89 and GAR92]. These functions and routines permit the
quick and fast evaluation of ΔP and ΔT given ΔM and ΔH for all water phases. Automatic
adjustment is provided to prevent P and T drift from values consistent with current M and H
values. These routines are non-iterative, essential for real-time simulation.
6.6 Conclusion
The FIBS approach for thermalhydraulic system simulation has been compared to the classic
work of Porsching. Porsching's algorithm is derived as a subset of the fully implicit approach.
Focusing on the system Jacobian, as Porsching did, focuses on the perturbation of the system as
a whole. Although general, it tends to obscure the interaction of the main players in typical
thermalhydraulic systems: flow and pressure. The FIBS form is shown to be more general
than Porsching's method, yet less obtuse. The interplay of flow and pressure is clarified and
coding is simplified.
6.7 Problems
1. Rewrite the conservation equations for the 4 node, 5 link case with various explicit /
implicit switches set for the following cases:
a. fully explicit
b. diagonally implicit
c. semi-implicit solution scheme (implicit in flow and pressure, explicit in mass and
enthalpy)
d. fully-implicit solution scheme (implicit in flow and pressure, mass and enthalpy).
2. Build a simulation code that solves the thermalhydraulic equations for a general
node-link network for the explicit case using the supplied skeleton code as a starting
point. Use the node-link diagrams and equations as developed in section 3, the
water property routines as developed in section 4, the rate form of the equation of
state as developed in section 5 and the explicit solution as developed in this section.
3. Improve upon your solution to question 2 by implementing a diagonally implicit

solution procedure. Is the solution more stable? Is there a cost penalty?
4. Implement a semi-implicit solution scheme (implicit in flow and pressure, explicit in

mass and enthalpy). Is the solution more stable? Is there a cost penalty?
5. Implement a fully-implicit solution scheme (implicit in flow and pressure, mass and
enthalpy). Is the solution more stable? Is there a cost penalty?
7 Case Study: Heat Transport System Stability

As mentioned in section 6.2, Porsching’s Method is one of the more successful algorithms for
thermalhydraulic simulation. This algorithm was used in the Ontario Hydro program SOPHT
to investigate Heat Transport System stability for the CANDU reactors with a figure-of-eight heat
transport loop configuration, such as the CANDU 6 and Darlington, since each loop potentially
has two-phase water at the outlet headers at high power. Since there are 2 outlet headers per
loop (one at each end of the reactor) separated by 2 single phase regions, we have a coupled
spring-mass system that, under the right conditions, could give undesirable flow and pressure
oscillations. To enhance heat transport system stability, a reactor outlet interconnect was
provided in each loop. This is discussed in detail in [GAR84, GAR86] but, of relevance to this
chapter, the code predictions based on the best estimate of the heat transport conditions
compared well to the plant tests, confirming both the mathematical methodology and the
efficacy of the one dimensional homogeneous thermalhydraulic model for overall system
simulation.
8 References
AGE82 L.J. Agee, “RETRAN Thermalhydraulic Analyses: Theory and Applications”, Progress in
Nuclear Energy, Vol. 10, No. 1, pp. 19-67, 1982.
AGE83 L.J. Agee, M.P Paulsen and E.D Hughes, “Equations of State for Nonequilibrium
Two-Phase Flow Models”, Transient Two-Phase Flow Proceedings of the Third CSNI
Specialist Meeting, Hemisphere Publishing Corporation, 1983.
BAN80 S. Banerjee, “Two-Phase Hydrodynamics: Models and Mechanisms”, Keynote Paper,
ANSASME Topical Meeting, Nuclear Reactor Thermal-Hydraulics, Saratoga, New York,
October 1980.
BER81 A.E. Bergles et al., Two-Phase Flow and Heat Transfer in the Power and Process
Industries, Hemisphere Publishing, 1981.
BIR60 R.B. Bird, W.E. Stewart and E.N. Lightfoot, Transport Phenomena, John Wiley and Sons,
Inc., New York, 1960.
BLO71 Benjamin S. Bloom, J. Thomas Hastings, George F. Madaus, “Handbook on Formative
and Summative Evaluation of Student Learning”, McGraw-Hill, Library of Congress
75-129488, 1971.
BNW76 COBRA-IV: An Interim Version of COBRA for Thermalhydraulic Analysis of Rod
Bundle Nuclear Fuel Elements and Cores, Battelle Pacific Northwest Laboratories,
BNWL-1962, March 1976.
CAR81a M.B. Carver, L.N. Carlucci and W.W.R. Inch, Thermalhydraulics in Recirculating Steam
Generators, THIRST Code User's Manual, AECL-7254, April 1981.
CAR81b M.B. Carver, “Numerical Simulation Involving Large Systems of Equations”, Keynote
Lecture, United Kingdom Simulation Council, 1981 Conference on Computer Simulation,
Harrogate, May 1981.

CAR95 M.B. Carver, “Numerical Solution of the Thermalhydraulic Conservation Equations from
Fundamental Concepts to Multidimensional Two-Fluid Analysis”, Atomic Energy of
Canada report AECL-11387, ARD-TD-550, August 1995, amended January 2017,
https://canteach.candu.org.
CHA75a Y.F. Chang, “The SOPHT Program and its Applications”, Proceedings of the 1975
Simulation Symposium on Reactor Dynamics and Plant Control, May 26-27, CRNL,
published as CRNL 1369, 1975.
CHA75b Y.F. Chang, SOPHT-B Bruce GS A Plant Control Simulation Engineers' Manual and
Program Specifications, Ontario Hydro CNS-37.3, June 1975.
CHA77a C.Y.F. Chang and J. Skears, "SOPHT - A Computer Model for CANDU-PHWR Heat
Transport Networks and Their Control", Nucl. Technol. 35, 591, 1977.
CHA77b Y.F. Chang, A Thermalhydraulic System Simulation Model for the Reactor, Boiler and
Heat Transport System (SOPHT), Ontario Hydro CNS 37-2, Sept. 1977.
CHE77 Bindi Chexal, NUCIRC - A Computer Code for Nuclear Heat Transport Circuit
Thermohydraulic Analysis, User's Instruction Manual, TDAI-116, AECL, Jan. 1977.
CHO74 W.G. Choe, J. Weisman, Flow Patterns and Pressure Drop in Concurrent Vapour-Liquid
Flow, A State-of-the-Art Report, AECL Lib 146648, Sept. 1974.
COL72 J.C. Collier, Convective Boiling and Condensation, McGraw-Hill, 1972.
CRA57 Crane, Flow of Fluids Through Valves, Fittings and Pipe, 410-C (Engineering Units),
410-m (metric), 1957.
CUR74 I.G. Currie, Fundamental Mechanics of Fluids, McGraw-Hill, Inc., 1974.
DEL81 J.M. Delhaye et al., Thermohydraulics of Two-Phase Systems for Industrial Design and
Nuclear Engineering, Hemisphere Publishing, 1981.
ELH80 M.A. El-Hawary, Verification of HYDNA-2 and HYDNA-3 Codes against Pertinent
Experimental Data, AECL TDAI-231, December 1980.
FIR84 A.P. Firla, “Approximate Computational Formulas for Fast Calculation of Heavy Water
Thermodynamic Properties”, 10th Annual Simulation Symposium on Reactor Dynamics
and Plant Control, Saint John, New Brunswick, 1984.
GAR79 W. Garland, SOPHT-D User's Guide, Revision 3, Ontario Hydro, Nuclear Systems
Department, 1979-03-01.
[GAR84] Wm. J. Garland, W.J.G. Brimley, "CANDU 600 Heat Transport System Flow Stability",
l0th Simulation Symposium on Reactor Dynamics and Plant Control, Saint John, N.B., 34
pages, April 9-10, 1984. [www.nuceng.ca/papers/conf.htm, paper C8]
[GAR86] Wm. J. Garland and S. H. Pang, "CANDU 600 Heat Transport Flow Stability", Nuclear
Technology, Vol. 75, #, pp. 239-260, December 1986.
[www.nuceng.ca/papers/journal.htm, paper J4]

GAR87a W.J. Garland and R. Sollychin, "The Rate Form of State for Thermalhydraulic Systems:
Numerical Considerations", Engineering Computations, Vol. 4, December 1987.
GAR87b Wm. J. Garland, "A Comparison of the Rate Form of the Equation of State to the
Jacobian Form", 13th Symposium on Simulation of Reactor Dynamics and Plant Control,
Chalk River Nuclear Laboratories, Chalk River, Ont., 28 pages, April 27-28, 1987.
[www.nuceng.ca/papers/conf.htm, C21]
GAR88 W.J. Garland and J.D. Hoskins, "Approximate Functions for the Fast Calculation of Light
Water Properties at Saturation", Int'l J. of Multiphase Flow, Vol. 14, No. 3, pp 333-348,
1988. [www.nuceng.ca/papers/journal.htm, paper J6]
GAR89 W.J. Garland and B.J. Hand, "Simple Functions for the Fast Approximation of Light
Water Thermodynamic Properties", Nuclear Engineering & Design, Vol. 113, pp.
21-34,1989. [www.nuceng.ca/papers/journal.htm, paper J8]
GAR92 Wm. J. Garland, R. J. Wilson, J. Bartak, J. Cizek, M. Stasny and I. Zentrich, "Extensions
to the Approximation Functions for the Fast Calculation of Saturated Water Properties",
Nuclear Engineering and Design, # 136, pp. 381-388, 1992.
GIN81 Ginoux, Two-Phase Flows and Heat Transfer with Application to Nuclear Reactor Design
Problems, Hemisphere Publishing, 1981.
HAA84 L. Haar, J.S. Gallagher, and G.S Knell, NBSNRC STEAM TABLES: Thermodynamic and
Transport Properties and Computer Programs for Vapor and Liquid States of Water in SI
Units, Hemisphere Publishing Corporation, 1984.
HAN95 B.N. Hanna and M.E. Laveck, ed., CATHENA Input Reference, Atomic Energy of Canada
Ltd., RC-982-4 / COG-93-140 (Vol. 4), Rev. 0.4, 1995.
HOS89 J. D. Hoskins, A Heat Transport System Operator Companion for a CANDU Nuclear
Reactor, Masters Thesis, McMaster University, Hamilton, Ontario, Canada, July, 1989.
HSU76 Y. Hsu and R. Graham, Transport Processes in Boiling and Two-Phase Systems,
Hemisphere Publishing, 1976.
IDE60 Idel’chik, Handbook of Hydraulic Resistance - Coefficients of Local Resistance and of
Friction, National Science Foundation, Washington, D.C. (Translation),1960.
ITT73 ITT Grinnel Corp., Piping Design and Engineering, Providence, Rhode Island, 1973.
KAY79 R.A. Kay, AESOP, Atomic Energy Station Optimization Program), TDAI-114/Revision 1,
AECL-EC, December 1976, revised by M. Gold March 1979.
KIN73 F.K. King, Steam Generators for CANDU Nuclear Power Stations, Sizing and Cost Analysis
Program, User's Instruction Manual, no file: see N. Subash, Process Systems
Development Department, AECL-EC, May 1973, revision 1 by N. Subash, July 1976.
LAH77 R.T. Lahey, Jr., and F.J. Moody, The Thermal Hydraulics of a Boiling Water Nuclear
Reactor, ANSAEC Monograph Series on Nuclear Science and Technology, ANS, 1977.
LIN79 M.R. Lin et al., FIREBIRD-III Program Description, AECL-7533, September 1979.
MER80 E.E. Merlo, et al, HYDNA-3 Program Description, AECL TDAI-205, April 1980.
MEY67 Meyer, C.A. et al., Thermodynamic and Transport Properties of Steam, ASME, New
York, 1967.
MIL71 D.S. Millar, Internal Flow, A Guide to Losses in Pipe and Duct Systems, British
Hydromechanics Research Association, 1971.
MUR68 J.H. Murphy and J.A. Redfield, WASP - A Program to Generate Water and Steam
Thermodynamic Properties, WAPD-TM-839, 1968.
NAH70 A.N. Nahavandi, and S. Makkenchery, “An Improved Pressurizer Model with Bubble
Rise and Condensate Drop Dynamics”, Nuclear Engineering and Design 12, p135-147,
1970.
PAY60 H.M. Paynter, Analysis and Design of Engineering Systems, MIT Press, 1960.
POR69 T.A. Porsching, J.H. Murphy, J.A. Redfield, and V.C. Davis, FLASH-4: A Fully Implicit
FORTRAN IV Program for the Digital Simulation of Transients in a Reactor Plant,
WAPD-TM-840, Mar. 1969.
POR71 T.A. Porsching, J.H. Murphy, J.A. Redfield, "Stable Numerical Integration of Conservation
Equations for Hydraulic Networks", Nuc. Sci. and Eng. 43, p 218-225, 1971.
REL76 RELAP4/MOD5 - A Computer Program for Transient Thermalhydraulic Analysis of
Nuclear Reactors and Related Systems, User's Manual, Vol. I: RELAP4/MOD5 Description,
Vol II: Program Implementation, Vol. III: Checkout Applications, ANCR-NUREG-1335,
Sept. 1976.
ROA76 Patrick J. Roache, Computational Fluid Dynamics, Hermosa Publishers, P.O. Box 8172,
Albuquerque, NM, 87108, 1976.
SEA75 F.W Sears and G.L. Salinger, Thermodynamics, Kinetic Theory, and Statistical
Thermodynamics, Addison-Wesley Publishing Company, 1975.
SKE75 J. Skears and T. Toong, SOPHT Programmers Manual, Revision 0.0, Ontario Hydro,
Engineering Systems Department, May 1975.
SKE80 J. Skears and T. Toong, SOPHT User's Manual, Version 2.0, Ontario Hydro, Engineering
Systems Department, January 1980.
SOL85 R. Sollychin,, S.A. Adebiyi, and W.J. Garland, “The Development of a Non-Iterative
Equation of State for Two-phase Flow Systems”, 11th Annual Symposium on Simulation
of Reactor Dynamics and Plant Control, Kingston, Ontario, 1985.
STE48 A.J. Stepanoff, Centrifugal and Axial Flow Pumps, John Wiley & Sons, Inc. 1948.
TON65 L.S. Tong, Boiling Heat Transfer and Two-Phase Flow, John Wiley and Sons, Inc., 1965.
YAN78 A. Yang, 600MW-SOPHT Control Simulation Manual, AECL file XX-63330-225-010 and
CENI-92, August 1978.
9 Nomenclature
A area γ phase volume fraction

A arbitrary vector Γ local sink or source
C concentration ψ field variable
Cp heat capacity at constant pressure ρ density
Cv heat capacity at constant volume σ stress tensor
e specific internal energy θ angle with respect to horizontal
E internal heat source or sink τ shear stress tensor
f friction factor
f long range or body force Operators
gc gravitational constant
g acceleration due to gravity 
partial time derivative
h specific enthalpy t
hN heat transfer coefficient d
H total enthalpy in volume, V dt total time derivative
I unity tensor D
k head loss coefficient Dt substantial time derivative
L length  Del operator
M mass in volume, V
M momentum interchange vector  ( )dV volume integral
V
n unit vector normal to the surface
P pressure  ( )ds
S
surface integral
q heat flux 1
A 
Q lumped heat source or sink  ( )  ( )ds cross sectional average
s surface bounding volume, V S
S surface sink or source

t time Subscripts
T temperature f liquid (fluid) phase
U total internal energy in volume, V g vapour (gaseous) phase
V arbitrary fluid volume i summation index for nodes
v velocity vector j summation index for links
W mass flow k 1, 2 (1 = liquid, 2 = vapour)
x quality (weight fraction) S surface
SAT saturated
Greek IN ingoing
α void fraction OUT outgoing
10 Acknowledgements
This chapter is based on a graduate level course given by the author at McMaster University and
on research funded by McMaster University and the Natural Sciences and Engineering Research
Council of Canada. The author is indebted to Atomic Energy of Canada Ltd. and to Ontario
Power Generation since this work is inextricably linked to past involvements with these
companies. Special thanks go to students, John Hoskins and Brian Hand, for water property
development, and to Raymond Sollychin for his central role in the development of the rate
method for the equation of state. A general thanks to all students over the years; their
freshness is a constant delight. Their sharp eyes caught many a mistake; of course the
responsibility for any errors or omissions lies entirely with the author.
1
CHAPTER 8
Nuclear Plant Systems
prepared by
Dr. Robin Chaplin
Summary:
This chapter deals with the main components of a CANDU nuclear power plant. It follows the
general path of energy production from the nuclear reactor to the electrical generator with
descriptions of the systems and equipment essential for the purpose of generating thermal
energy, transferring heat and converting heat into mechanical and then electrical energy. Some
key auxiliary systems are included where these are an essential part of the energy processes as
are the basic control systems which maintain stable operational conditions and ensure safety of
the plant. This chapter leads into Chapter 9 which describes how a typical CANDU plant oper-
ates.
Table of Contents
1 Introduction ............................................................................................................................ 5
1.1 General Configuration..................................................................................................... 5
2 Nuclear Reactor....................................................................................................................... 5
2.1 General Arrangement ..................................................................................................... 5
2.2 Calandria Arrangement................................................................................................... 7
2.3 Moderator Temperature Control .................................................................................... 7
2.4 Moderator Reactivity Control ......................................................................................... 8
2.5 Core Shielding ................................................................................................................. 9
2.6 Reactor Control Systems and Devices............................................................................. 9
2.7 Fuel Configuration......................................................................................................... 17
2.8 Fuel Channel Conditions ............................................................................................... 18
2.9 Fuel Handling ................................................................................................................ 19
3 Heat Transport System.......................................................................................................... 21
3.1 General Arrangement ................................................................................................... 21
3.2 Heat Transport Pumps .................................................................................................. 24
3.3 Pressurizer..................................................................................................................... 24
3.4 Inventory Control System.............................................................................................. 25
3.5 Shutdown Cooling System ............................................................................................ 25
3.6 Partial Boiling in CANDU Reactors ................................................................................ 26
3.7 Pressure and Level Effects............................................................................................. 27
3.8 Temperature and Nuclear Considerations .................................................................... 27
©UNENE, all rights reserved. For educational use only, no assumed liability. Nuclear Plant Systems – July 2016
4 Steam Generators ................................................................................................................. 28

4.1 Steam Generator Function............................................................................................ 28
4.2 Steam Generator Conditions ........................................................................................ 28
4.3 Pressure Control............................................................................................................ 30
4.4 Level Control ................................................................................................................. 31
4.5 Comparison of CANDU and PWR Systems .................................................................... 31
5 Steam System........................................................................................................................ 33
5.2 Steam Bypass System.................................................................................................... 34
5.3 Condenser Steam Discharge Valves .............................................................................. 36
5.4 Turbine Control Valves .................................................................................................. 36
6 Steam Turbine ....................................................................................................................... 36
6.2 Turbine Cylinders .......................................................................................................... 37
6.3 Turbine Rotors............................................................................................................... 39
6.4 Turbine Blading ............................................................................................................. 40
6.5 Turbine Sealing Principles ............................................................................................. 42
6.6 Turbine Seals ................................................................................................................. 43
6.7 Steam Sealing System ................................................................................................... 43
6.8 Turbine Bearings ........................................................................................................... 44
6.9 Bearing Lubrication ....................................................................................................... 45
7 Reheating and Feed Heating................................................................................................. 47
7.1 Steam Separators and Reheaters.................................................................................. 47
7.2 Deaerator ...................................................................................................................... 47
7.3 Feedwater Heaters........................................................................................................ 48
8 Condenser ............................................................................................................................. 49
9 Electrical Generator .............................................................................................................. 51
9.1 Generator Rotor ............................................................................................................ 51
9.2 Generator Stator ........................................................................................................... 52
9.3 Excitation System .......................................................................................................... 52
9.4 Generator Systems........................................................................................................ 52
10 Power Transmission .............................................................................................................. 54
10.1 Generator Main Connections........................................................................................ 54
10.2 Generator Circuit Breakers............................................................................................ 54
10.3 Generator Transformers................................................................................................ 54
10.4 Transformer Connections.............................................................................................. 54
11 Problems ............................................................................................................................... 55
Nuclear Plant Systems 3
List of Figures
Figure 1 Reactor vault and assembly .............................................................................................. 6
Figure 2 Feeder tube assembly on reactor face.............................................................................. 7
Figure 3 Moderator cooling and purification system...................................................................... 8
Figure 4 CANDU 6 reactor assembly ............................................................................................. 11
Figure 5 Reactor cross section ...................................................................................................... 12
Figure 6 Reactor longitudinal section ........................................................................................... 13
Figure 7 Reactor plan view............................................................................................................ 14
Figure 8 Shutdown System 2 (SDS2) ............................................................................................. 17
Figure 9 37-element fuel bundle .................................................................................................. 18
Figure 10 Reactor face with fuelling machine............................................................................... 20
Figure 11 Typical fuelling machine and bridge ............................................................................. 20
Figure 12 Orientation of fuelling machine with respect to reactor.............................................. 21
Figure 13 Simplified heat transport and moderator circuits ........................................................ 22
Figure 14 Diagrammatic arrangement of figure-of-eight heat transport system......................... 23
Figure 15 Heat transport purification and feed system................................................................ 25
Figure 16 Shutdown cooling system ............................................................................................. 26
Figure 18 Temperature profiles in steam generator ..................................................................... 30
Figure 19 Steam generator for PWR system ................................................................................. 32
Figure 21 Steam and feedwater system for a CANDU reactor...................................................... 34
Figure 22 Main steam system (Point Lepreau) (courtesy of NB Power) ....................................... 35
Figure 23 High pressure turbine cross section for nuclear unit (courtesy of Eskom)................... 38
Figure 24 Low pressure turbine cross section for nuclear unit (courtesy of Eskom).................... 38
Figure 25 Electrical generator for a large turbine (courtesy of Siemens) ..................................... 40
Figure 26 Construction of a 900 MW steam turbine for a nuclear unit (courtesy of Eskom)....... 41
Figure 27 Spring mounted labyrinth seal to accommodate shaft deflection ............................... 42
Figure 28 Principle of operation of LP turbine shaft seal (courtesy of NB Power) ....................... 43
Figure 29 Turbine steam sealing system ....................................................................................... 44
Figure 30 Fixing points for turbine rotors and casings (courtesy of NB Power) ........................... 45
Figure 31 Turbine generator shaft catenary ................................................................................. 45
Figure 32 Turbine lubricating oil system ....................................................................................... 46
Figure 33 Deaerator and deaerator storage tank (courtesy of NB Power) ................................... 48
Figure 34 Feedwater heater with integral drain cooler (courtesy of NB Power).......................... 49
Figure 35 Condenser longitudinal section (courtesy of NB Power) .............................................. 50
Figure 36 Condenser cross section (courtesy of NB Power) ......................................................... 50
Figure 37 Insertion of rotor into stator of generator (courtesy of NB Power).............................. 53
Figure 38 Arrangement of three single phase generator transformers........................................ 55
List of Tables
Table 1 CANDU fuel characteristics............................................................................................... 18
Table 2 Fuel channel parameters (CANDU 6)................................................................................ 19
Table 3 Steam generator conditions (CANDU 6)........................................................................... 30
Table 4 Turbine generator conditions (CANDU 6)......................................................................... 33
1 Introduction
1.1 General Configuration
Specific units in a typical CANDU nuclear plant include a nuclear reactor in which heat is generated
by nuclear fission, a heat transport system to convey this heat from the reactor to the steam cycle,
a steam generator in which steam is generated, a steam system to convey steam from the steam
generator to the steam turbine, a steam turbine where available heat energy is converted into
mechanical energy, a condenser where unavailable heat energy is rejected to the cooling water
system, a steam reheating and feedwater heating system to improve cycle efficiency, an electrical
generator to convert mechanical energy into electrical energy, and an electrical system to generate
electrical energy for distribution to consumers. To balance the flow of energy through these
components and to regulate plant output, a control system is required. Each of these major
systems will be described in this chapter.
All thermal reactors have four important components: fuel, moderator, coolant, and control
elements. The moderator is that part of the nuclear reactor where neutrons are reduced in energy
to increase the probability of fission when they re-enter the fuel. Spacing of the fuel assemblies
within the moderator is therefore of key importance. Heat must be removed by the coolant from
the fuel assemblies at a high rate without making the fuel sustain excessive temperatures. This is
achieved by having multiple small diameter fuel rods (elements) in one fuel assembly. These two
requirements dictate the general reactor configuration. A unique aspect of the CANDU reactor is
the horizontal arrangement of fuel channels within the reactor calandria. These channels contain
the fuel bundles and are surrounded by the heavy water moderator, whereas heavy water coolant
flows through the channels.
Heat is generated in the nuclear reactor, and electricity is produced by the electrical generator. The
production and discharge of this energy must be perfectly balanced. The steam generator is the
key component for maintaining this balance, that is, the rate of steam generation must match the
rate of steam flow to the steam turbine, or vice versa. Any change in these two energy flows will
immediately be reflected as a change in steam pressure in the steam generator. Therefore, steam
pressure is a key control parameter for the entire plant under all operating conditions.
2 Nuclear Reactor
2.1 General Arrangement
For all CANDU reactors, the horizontal fuel channels are arranged in a square array across the
reactor face, as shown in Figure 1. The calandria, which contains the moderator at essentially
atmospheric pressure, contains axial tubes in a similar square array to accommodate the
pressurized fuel channels or pressure tubes, which in turn contain the fuel bundles and the
primary system coolant. Figure 2 shows a close-up view of the feeder pipes conveying coolant
to and from the pressure tubes at one end of the reactor.
Figure 1 Reactor vault and assembly
Figure 2 Feeder tube assembly on reactor face
2.2 Calandria Arrangement

The calandria contains the moderator for the reactor. It consists essentially of a horizontally
mounted cylindrical tank containing many horizontal tubes in a square array. Pressure tubes
containing fuel and coolant are fitted inside the calandria tubes, with a small annular gas filled
space to provide heat insulation. This annular space is maintained by garter springs which resist
sagging of the pressure tubes and thus minimize the risk of contact with the calandria tubes.
The gas used is carbon dioxide because it has low thermal conductivity, shows little tendency to
promote corrosion, and does not produce significant activation products. The carbon dioxide is
circulated in the annular space to cool it and to monitor it for any leakage from the calandria
and pressure tubes. The calandria contains heavy water as moderator, which fills the space
between the calandria tubes. The spacing of the calandria tubes and hence the fuel channels is
greater than optimum with respect to neutron moderation, with the result that the reactor is
over-moderated.
2.3 Moderator Temperature Control

Heat is generated in the heavy water moderator by the neutron moderating process, where the
kinetic energy of the neutrons is converted into heat energy, and by absorption of gamma
radiation. Both these processes intensify with a rise in reactor power. Heat is also transferred
to the moderator through the carbon dioxide in the annular space between the pressure tubes
and calandria tubes. This process is largely independent of reactor power because the fuel
channel temperature is fairly constant at all power levels. The total amount of heat generated
in and transferred to the moderator is approximately 5% of total reactor heat production. This
heat must be removed to maintain the desired temperature in the moderator.
The moderator temperature is maintained between 60°C and 65°C by an external cooling
system, as illustrated in Figure 3. This system includes circulating pumps and heat exchangers,
which reject heat to the service water system. Although the relatively large temperature
difference between the reactor coolant and the moderator induces some heat loss, the rela-
tively cool moderator can provide some measure of cooling to the fuel in the event of a loss of
coolant accident and the coincident failure of the emergency coolant injection system. Because
the reactor’s configuration makes it over-moderated, any void formation within the moderator
would cause an increase in reactivity, leading to unstable reactivity conditions. Hence, it is
desirable to maintain an adequate margin below boiling conditions.
Figure 3 Moderator cooling and purification system

The moderator temperature has an effect on the reactivity of the reactor as its temperature
rises from ambient conditions to operating conditions during reactor warm-up. For equilibrium
fuel conditions, that is, a mix of used and new fuel that arises after several months of continu-
ous refuelling, the temperature coefficient of reactivity is positive, meaning that the reactivity
increases by about 4 mk as the temperature goes from 30°C to 70°C. This phenomenon is due
mainly to two factors. One is the lower density of the moderator at higher temperatures,
resulting in less neutron absorption. The other is the increased fission cross section of pluto-
nium as the increase in temperature takes the neutron energy into a resonance peak of Pu-239.
2.4 Moderator Reactivity Control

To control moderator reactivity, soluble poisons are added to the heavy water at appropriate
concentrations under certain operating conditions through a moderator liquid poison addition
system. This system consists of mixing tanks for boron and gadolinium, also shown in Figure 3,
where the solutions are formulated, as well as flow transmitters at the circulating pump inlet to
measure the amount entering the moderator system.
Both gadolinium and boron have strong neutron absorbing isotopes, namely Gd-155, Gd-157,
and B-10. They gradually lose their ability to absorb neutrons as neutrons are absorbed to form
a heavier isotope. The burnout rate for boron is slower than that of gadolinium, making it more
suitable for longer term operations such as compensation for reactivity loss due to fuel burnup.
When a reactor is charged with new fuel, the excess reactivity from the fuel can be balanced by
a certain concentration of boron. As the fuel burns up and neutron absorbing fission products
are produced, the effectiveness of the boron decreases, making it possible to maintain reasona-
bly constant reactivity with naturally small adjustments. The burnout rate for gadolinium is
faster than that of boron, and therefore it is more suitable for short term reactivity adjustments
such as those necessary during xenon transients. The xenon concentration in the fuel is lower
during periods of low load and builds up following a large power increase. Gadolinium can
compensate for this and burns out at a rate roughly corresponding to xenon buildup. Gadolin-
ium can also be used to compensate for a lack of xenon in the reactor during an outage to
ensure that the reactor will remain subcritical.
The shutdown system, which will be described later, injects liquid poison directly into the
moderator in the event of certain reactor trips. Gadolinium is used at sufficient concentration
to ensure an immediate reduction in reactivity to well below critical conditions. Upon reactor
restart, it must be removed by the moderator purification system. Gadolinium is easily removed
by ion exchange resins, whereas boron, being a weaker ion, is not easily removed. This makes
gadolinium better suited for use in the shutdown system, but it still takes several hours to
remove it and almost as long as the ensuing outage due to xenon poisoning.
A moderator purification system, also shown in Figure 3, is connected to the moderator cooling
system to maintain heavy water purity and to remove or adjust the concentration of soluble
poisons such as gadolinium or boron which may have been introduced for shutdown purposes
or to assist in reactivity control. The non-active soluble poison (after neutron absorption) must
be removed to maintain the general purity of the moderator. The purification system has
strainers to remove particulate material such as resin fines from the system, followed by mixed
bed ion exchange columns to remove dissolved material in the form of positive and negative
ions.
2.5 Core Shielding

The reactor core is shielded at the ends by end shields and around the periphery by a thermal
shield to absorb neutron and gamma radiation which would otherwise create excessive radia-
tion fields at the reactor faces and high temperatures in the calandria structure. The shields
contain water which is circulated and cooled. The end shields contain steel balls as well to
ensure sufficiently low radiation fields during routine maintenance shutdowns. The combina-
tion of water and steel balls is very effective in attenuating both neutron and gamma radiation.
2.6 Reactor Control Systems and Devices

Heavy water is a very weak neutron absorber, which is advantageous with regard to fuel utiliza-
tion. To compensate for reactivity changes in the fuel, the absorption properties of the modera-
tor can be modified by adding a soluble neutron absorber in small concentrations. Gadolinium
and boron both have strong neutron absorbing isotopes, namely Gd-155, Gd-157, and B-10.
These can be added through the liquid poison system and removed through the moderator
purification system using ion exchangers. Light water is a stronger neutron absorber than heavy
water, and therefore the concentration of light water in the heavy water moderator must be
limited to avoid excessive neutron absorption. The normal isotopic concentration of D2O in the
moderator is 99.5%. Light water, however, can be conveniently used as a control device within
the reactor, where its volume in vertical tubes can be varied.
Reactor power is controlled and maintained at the required level by several devices which
absorb neutrons and thereby control the nuclear fission process. These systems and devices
are:
 Liquid poison in moderator
 Liquid zone control absorbers
 Mechanical rod control absorbers
 Mechanical rod adjusters.
In addition, a complete reactor shutdown can be achieved by operating one or both of the
following devices in addition to those listed above:
 Mechanical rod shut-off absorbers (Shutdown System 1)
 Liquid poison injection system (Shutdown System 2).
Appropriate signals to operate the control systems are obtained from neutron flux detectors
within the core and ion chambers alongside the core as well as from other sensors in the steam
system and turbine generator. The main control devices are shown as part of the reactor
assembly in Figure 4.
The control devices described below are generally applicable to CANDU 6 reactors. The same
design philosophy applies to other CANDU reactors, but the number of devices may be different
if the reactor core is of different size.
In a CANDU 6 reactor core, there are 26 in-core vertical flux detectors and 7 in-core horizontal
flux detectors. Twelve vertically installed detectors provide flux information to Shutdown
System 1, while three horizontal flux detectors supply flux data to Shutdown System 2. The
remaining vertical flux detectors send information to the liquid zone control absorbers and the
solid control rod absorbers. All these in-core flux detectors can provide data for three
dimensional flux mapping of the reactor core.
Figure 5, Figure 6, and Figure 7 show three orthogonal views of a CANDU 6 reactor. These show
the relative positions and arrangement of various components and devices.
Figure 4 CANDU 6 reactor assembly
Figure 5 Reactor cross section
Figure 6 Reactor longitudinal section
Figure 7 Reactor plan view
2.6.1 Moderator Poison Addition System

Liquid poison such as boron or gadolinium can be added to the moderator to adjust reactor
reactivity. Because these are both neutron absorbers, they reduce neutron efficiency and hence
adversely affect fuel utilization. However, as described earlier, their use is necessary under
certain transient conditions to maintain reactivity within regular control limits. Such conditions
arise when a reactor is started with a new charge of fresh fuel, when a significant amount of
fresh fuel is added, and during large increases in load when xenon transients need to be
accommodated.
2.6.2 Liquid zone control absorbers

Critical reactivity balance is mostly and predominantly maintained by the six liquid zone control
absorbers located vertically within the reactor core interspersed among the mechanical control
absorbers, adjuster rods, and shut-off rods. These liquid zone absorbers contain light water in
two or three separate sections, forming a total of fourteen compartments. These separate and
independent absorbers enable spatial control of neutron flux. Light water absorbs neutrons
more strongly than heavy water and acts like a control rod as light water is added or removed
through inlet and outlet pipes, where a difference in flow rate will vary the level. Helium gas is
maintained above the light water, and the difference in pressure between the inlet and outlet,
at the bottom and top of the water column respectively, indicates the level of light water. The
quantity of water in each compartment is controlled by automated valves receiving signals from
the flux detectors through the control computers of the reactor regulating system.
2.6.3 Mechanical rod control absorbers

Besides the liquid zone control absorbers, there are also four mechanical control absorbers
made from a variety of alloys, predominantly cadmium encased in stainless steel. These
absorbers adjust the flux level when a greater reactivity rate or depth than can be provided with
the liquid zone control absorbers is required. They can bring about a fast reactor power
reduction such as a reactor stepback. They may be driven in or out of the core at variable speed
or dropped into the core by the release of a clutch. The design of these solid control absorbers
is virtually identical to that of the shut-off rods described below.
2.6.4 Mechanical rod adjusters

Adjuster rods are used in the reactor to optimize the neutron flux profile for reactor power,
compensate for fuel burnup, and provide excess reactivity to overcome Xenon-135 surges
following a power reduction. CANDU reactors are designed to develop full power when these
adjuster rods are fully inserted into the reactor core. There are 21 adjuster rods arranged in
three rows vertically between the other control devices. The adjuster elements are raised or
lowered individually by winching drives similar to those described for the shut-off rods, except
that these adjusters do not fall into the reactor in the event of a shutdown signal from the flux
detectors because they would normally be fully inserted during normal operation. When the
adjusters are fully withdrawn, the reactivity gain is about 16 mk, which is sufficient to overcome
minor xenon transients during power manoeuvering.
2.6.5 Mechanical rod shut-off absorbers

The signals received from the flux detectors or other plant indicators, which are routed to
Shutdown System 1 (SDS1), control a system of 28 vertical shut-off rods. These shut-off rods are
composed of cadmium and stainless steel absorber elements and are used to shut the reactor
down under normal and emergency conditions. When these rods are inserted into the core,
they are aided by perforated guide tubes made of zirconium alloy that are open to the pressure
of the moderator. These shut-off rods are inserted into or withdrawn from the core by means
of a stainless steel cable which is wound on a drum in the rod drive mechanism. This drum is
driven by an electric motor with an electromagnetic friction clutch. When a shut-off signal is
received, the clutch de-energizes, and the shut-off rods fall into the reactor core inside the
zirconium alloy guide tubes under the influence of gravity. While the clutch is energized, the
shut-off rods can be driven into or out of the reactor core by the electric drive motor.
2.6.6 Liquid poison injection system

This system, unlike the Moderator Poison Addition System, is a safety system as opposed to a
control system. It injects a large quantity of soluble poison into the moderator to reduce
reactivity drastically rather than adding a controlled amount to maintain reactivity at the
required level.
Shutdown System 2 (SDS2), as illustrated diagrammatically in Figure 8, is activated based on
information received from the three horizontal in-core flux detectors or other plant indicators.
The poison that is injected into the core is gadolinium nitrate dissolved in heavy water. This
poison is contained in six tanks outside the reactor vault in an accessible area for routine
maintenance. Each of the six poison tanks is connected to a nozzle pipe in the calandria by
stainless steel pipes. These nozzle pipes are oriented horizontally in the reactor core, each with
several orifice holes situated to provide maximum negative reactivity changes when the poison
is injected. The nozzle pipes themselves are made of zirconium alloy.
The gadolinium nitrate poison is injected into the core by helium under pressure. Pressurized
helium in this tank is isolated from the poison tanks by triplicated quick-acting valves. When
these valves are opened, the pressure forces the poison into the moderator very quickly. A ball
in the poison tanks prevents blowthrough of helium gas and ensures that calandria pressure
does not exceed its design value.
SDS2 is typically capable of supplying at least 50 mk of negative reactivity worth, although a
negative reactivity of 30 mk is considered adequate to bring about complete and rapid
shutdown. This system acts as a backup to SDS1 and should be activated only in the event of a
failure of SDS1 or if the initiating event is so severe or so fast that SDS1 has not arrested the
transient. Because liquid poison is injected into the moderator, quick recovery from such a trip
is not possible.
In the original CANDU design concept, it was decided that there should be two independent and
different shutdown systems operated from different sources and configured differently. To
achieve this objective, the SDS1 and SDS2 systems inject negative reactivity in different ways
and even in different orientations within the reactor, that is, one vertically and one horizontally.
Figure 8 Shutdown System 2 (SDS2)
2.7 Fuel Configuration

In most CANDU reactors, the individual fuel bundles have 37 fuel elements, each containing a
stack of natural uranium dioxide fuel pellets. Within each fuel element are approximately 30
fuel pellets, giving an effective stack length of 480 mm. The fuel elements are held in place by
end plates which are perforated to enable coolant to flow axially and are prevented from
contacting one another or the pressure tube by spacers. The element sheaths are made of
zircaloy to minimize neutron absorption. There are 12 fuel bundles in each fuel channel of the
CANDU 6 reactor. The configuration of the reactor with horizontal fuel channels enables the
reactor to be refuelled while on load and also facilitates partial refuelling of a single channel
because the fuel bundles do not have to be linked together as in a vertical channel. A 37-
element fuel bundle is shown in Figure 9, while Table 1 gives details of a 37-element fuel
bundle.
Figure 9 37-element fuel bundle
Table 1 CANDU fuel characteristics
Fuel Type Sintered natural uranium dioxide

Pellets Cylindrical with concave dished ends
Number in Elements 30
Elements per Bundle 37
Bundles per Fuel Channel 12
Pellet Diameter 12.16 mm
Pellet Stack Length 480 mm
Cladding Zircaloy-4
Cladding Outside Diameter 13.08 mm
Bundle Outside Diameter 102.4 mm
Bundle Overall Length 495 mm
2.8 Fuel Channel Conditions

A temperature rise takes place along the fuel channel from the inlet to the outlet until satura-
tion conditions are reached, after which the temperature remains approximately constant. This
temperature rise is not linear because the rate of temperature rise depends upon the local rate
of heat release and hence on the neutron flux, which drops off towards both the inlet and
outlet of the channel. The most severe fuel conditions occur where both the heat flux and
coolant temperature are high. This occurs normally somewhat downstream of the channel
midpoint. Typical fuel channel conditions are given in Table 2.
Table 2 Fuel channel parameters (CANDU 6)
Fuel Channel Material Zirconium-Niobium Alloy

Inside Diameter 103.38 mm
Wall Thickness 4.19 mm
Inlet Pressure 11.04 MPa (a)
Outlet Pressure 10.30 MPa (a)
Inlet Temperature 267°C
Outlet Temperature 312°C
Maximum Coolant Flow 24 kg/s
Average Coolant Flow 20 kg/s
2.9 Fuel Handling

Short term reactivity changes occur continuously during normal operation due to load changes
or to minor transient conditions. Short term reactivity control is maintained in the reactor core
by liquid zone control absorbers (for load changes) and adjuster rods (for transients). Long term
reactivity changes occur slowly as fuel is used up and fission products accumulate in the fuel.
Long term reactivity control is achieved through on-line refuelling to maintain constant reactiv-
ity while fuel is burned. By varying the frequency of refuelling in different areas of the reactor
core, the flux and hence the power distribution can be very effectively and accurately con-
trolled.
The CANDU reactor with its horizontal fuel channels has a reactor face at each end, onto which
two fuelling machines must attach while maintaining pressure and flow in the channel. Figure
10 shows a typical reactor face with a fuelling machine. This is a larger reactor with 480 fuel
channels (Darlington) with a later design of fuelling machine than the one illustrated in Figure
11 and Figure 12.
Fuel is inserted into and removed from the fuel channels by two remotely controlled fuelling
machines which attach to the ends of a selected fuel channel. A general view of a typical
fuelling machine (Pickering) as seen from the reactor side is shown in Figure 11. These fuelling
machines are fully automatic and remotely controlled. The equipment itself consists of a
fuelling machine mounted on a suspended carriage. The carriage travels on a track which can
be raised or lowered by a gantry so that the fuelling machines can reach any fuel channel.
Heavy water and electricity supplies are provided by a flexible catenary which connects the
machine to the stationary auxiliary systems.
Figure 10 Reactor face with fuelling machine
Figure 11 Typical fuelling machine and bridge

The arrangement with respect to the reactor face is shown in Figure 12. By raising or lowering
the gantry and traversing the carriage, the fuelling machine can reach any channel. When the
gantry is at its lowest point, the machine can turn 90° in the horizontal plane, as shown by the
dotted image, to receive new fuel or to discharge spent fuel in the service area. After
attachment to the reactor, the fuelling machines open the fuel channel and remove the shield
plug. One fuelling machine carries the new fuel in its magazine, while the other machine at the
opposite end of the fuel channel receives and stores spent fuel in its own magazine. Each
magazine can revolve and has several slots, each of which can store two fuel bundles. Two new
fuel bundles from the fuelling machine are inserted at the coolant inlet end, while two used
bundles are removed at the coolant outlet end. Normally, depending on the particular CANDU
design, eight bundles are changed during one channel refuelling. In this way, the two bundles at
each end where the neutron flux is low spend twice as long in the reactor than those in the high
flux zone in the middle. Flow through the channel is maintained, and the pressure difference so
derived is used to move the bundles into the channel and to provide a driving force to facilitate
their removal. Spent fuel is later transferred to the irradiated fuel bay, where it is stored under
water. The average time taken to refuel one fuel channel is about two and one half hours,
meaning that the refuelling process is almost continuous during normal working hours.
Figure 12 Orientation of fuelling machine with respect to reactor

New fuel is loaded directly into the magazine of the fuelling machine, but irradiated fuel must
be remotely handled and transferred under shielded conditions to the irradiated fuel bay for
interim storage during its initial cooldown period.
3 Heat Transport System

The fundamental function of the primary coolant circuit or heat transport system is to circulate the
heavy water coolant and transport heat from the fuel elements to the steam system. The entire
circuit consists of pumps, headers, feeder pipes, pressure tubes, and steam generators, as shown in
Figure 13. In addition, there is a single pressurizer in the circuit. The coolant used in CANDU
reactors is heavy water to minimize neutron absorption. In the pressure tubes, the coolant flows
are in opposite directions in adjacent tubes. Thus the coolant passes from the steam generators at
one end of the reactor through half the tubes to the steam generators at the other end of the
reactor and back again through the adjacent tubes.
There are two steam generators and two heat transport pumps at each end of the reactor, and
therefore the entire system is divided into two independent loops, with each serving half the
reactor. They are linked by a balance pipe with restricting orifices so that the same pressure is
maintained in each loop and only one pressurizer is required. This arrangement is illustrated
diagrammatically in Figure 14. The advantage of this arrangement is that, in the event of a serious
loss of coolant accident, only half the reactor will be affected initially, with minimum subsequent
loss of coolant from the other half.
Figure 13 Simplified heat transport and moderator circuits
Figure 14 Diagrammatic arrangement of figure-of-eight heat transport system

As coolant leaves the reactor, it passes through the outlet feeder tubes on each reactor face and
then to the respective headers. From here, it enters the U-tube steam generators situated at
opposite ends of the reactor. In the steam generator, the pressurized heavy water exchanges
heat with the light water on the steam side of the steam generator. The now cooled heavy
water passes to the heat transport pumps, one for each steam generator, and returns to the
reactor core through headers and inlet feeder tubes at each end of the reactor.
The heat transport system serves to remove heat from the reactor under both power producing
and shutdown conditions. The flow is maintained by the heat transport pumps under most
conditions, but with the steam generators located above the reactor, natural circulation is
possible under shutdown conditions.
Pressure in the fuel channels is maintained at about 10 MPa at the outlet end, and the coolant
is generally subcooled, although above about 90% full load power, some boiling occurs in the
most highly rated fuel channels. At full power, the steam exit quality is typically 3% to 5%,
which gives a voidage of 26% to 37% at 10 MPa. This steam is subsequently mixed with sub-
cooled coolant from the lower rated channels to give conditions very close to saturation at the
steam generator inlet though the flow to the lower rated channels is somewhat reduced to
minimize the difference in exit quality.
The heat transport pumps are driven by electrical motors. Should power supply to the motors
fail, forcing a reactor trip, the pumps have sufficient built-in inertia to run down over a period of
two to three minutes. This allows sufficient time for natural circulation or thermosiphoning to
be established.
Pressure in the heat transport circuit is maintained by a pressurizer in most CANDU reactors.
This operates at saturation conditions and is approximately half filled with liquid, with the
remaining space containing vapour. It is connected to the heat transport system by a pipe
which is long enough to maintain its temperature independently of the heat transport system.
This enables the reactor outlet header pressure to be controlled regardless of the heat trans-
port system temperature.

There is a feed-and-bleed system to inject or extract heavy water from the heat transport
system. This enables the correct inventory to be maintained under transient conditions such as
start-up and shutdown and to compensate for leakage. It also enables external clean-up and
chemical treatment of coolant under lower pressure and temperature conditions, as well as
recovery of leakage and upgrading of heavy water.
3.2 Heat Transport Pumps

Four heat transport pumps circulate heavy water coolant around the system. They are vertical
single stage centrifugal pumps driven by induction motors. Mechanical seals prevent leakage of
reactor coolant. These seals are supplied with cooled heavy water, some of which leaks out-
wards and some into the pump, thus maintaining the correct operating temperature at the
pump seals and preventing loss of reactor coolant. In the CANDU 6, the pumps each draw
approximately 5 MW of electrical power, giving a total of 17 MW of energy input to the heat
transport system. This input is used to raise the temperature and pressure of the reactor
coolant system during warm-up from cold shutdown conditions.
3.3 Pressurizer
The pressurizer maintains pressure in the heat transport system by maintaining a cushion of
vapour above the heavy water in the lower half. Pressure can be reduced by venting vapour at
the top or by spraying water into the vapour space and can be increased by heating the water
with electric heaters at the bottom. During normal operation, the heaters generally operate
periodically because the pressurizer loses heat naturally to its surroundings. Some plants rely
on venting vapour only to reduce pressure and do not have water sprays. Level in the pressur-
izer is ramped up as the temperature of the heat transport system increases to accommodate
partially the thermal expansion of the reactor coolant during power raising. Under cold shut-
down conditions, the pressurizer is normally isolated from the heat transport system to main-
tain an elevated temperature in it.
During warm-up, the pressurizer is isolated from the heat transport system, meaning that the
system is in the so-called solid mode of operation as opposed to the normal mode of operation
where the pressurizer is connected to the system. In this solid mode, pressure control is by feed
and bleed action where the flow rate of heavy water in and out of the system is adjusted to
maintain the correct pressure. In the normal mode, the feed and bleed action of the inventory
control system adjusts the coolant inventory by maintaining the pressurizer level at its set-point.
To avoid over-pressure conditions in the heat transport system, pressure relief valves on the
reactor outlet headers discharge excess coolant to the degasser condenser of the inventory
control system. If this action is insufficient to alleviate the over-pressure condition, the control
system will step back the reactor to a lower power level, and if this still does not reduce pres-
sure sufficiently, it will trip the reactor.
Note that the steam generators serve as a heat sink for the heat transport system, and therefore
reducing pressure and hence temperature in the steam generators will have the effect of
reducing temperature in the heat transport system, with a resulting reduction in pressure as
well.
3.4 Inventory Control System

The inventory control system serves to maintain pressure in the heat transport system when the
pressurizer is isolated and to maintain level in the pressurizer under normal operating condi-
tions. It also serves as a reserve source of heavy water to compensate for minor leakage and to
maintain heavy water purity by means of the purification system. This requires continuous
bleeding and feeding of heavy water so that it can be treated appropriately. Figure 15 shows a
diagrammatic arrangement of a typical inventory control system for purification and feeding of
coolant to the heat transport system. The heavy water is at high temperature and pressure
upon leaving the system. Therefore, it must be treated under either high pressure or low
pressure conditions. Figure 15 shows a system with high pressure purification to minimize
pressure reduction and pumping back to pressure, but the feed and bleed operation and heavy
water storage are at lower pressure, with the coolant initially passing through a degasser before
being pumped back into the system. Furthermore, the ion exchange resins cannot withstand
high temperatures, and therefore cooling to about 50°C is required before purification. It is
advantageous to use the heat removed from the bleed flow to heat the feed flow before it is
returned to the heat transport system. Therefore, the system includes interchangers as well as
service water coolers.
Figure 15 Heat transport purification and feed system
3.5 Shutdown Cooling System

When the reactor is shut down, decay heat is removed through the steam generators and the
steam reject system. If the reactor must be cooled down, the steam reject system can function
well until the steam generator pressure drops so low that the volume flow rate through the
steam reject or condenser steam discharge valves is insufficient for effective cooling. This
occurs at about 150°C to 165°C. Operation of the shutdown cooling system is initiated before
this stage is reached and removes heat directly from the heat transport system. A typical
system is shown in Figure 16. It can handle about 10% to 15% of the main system flow and is
capable of removing 1% to 3% of reactor full power heat.
Figure 16 Shutdown cooling system
3.6 Partial Boiling in CANDU Reactors

Although CANDU reactors have pressurized primary coolant systems where boiling is
suppressed, local boiling does occur in the highest rated fuel channels. The steam produced is
subsequently partially condensed when it mixes with coolant from lower rated fuel channels.
This has some definite advantages. Slight boiling enhances the heat transfer coefficient on the
fuel element surfaces, thus promoting heat transfer. It also increases the heat removal capacity
of the coolant because the enthalpy of the mixture leaving the fuel channel is increased.
Thirdly, after mixing with coolant from nonboiling fuel channels, the average coolant exit
temperature is at saturated conditions with some vapour present, or at least closer to saturated
conditions than it would be if there were no boiling. This means that the coolant entering the
steam generators is at as high a temperature as possible considering the prevailing pressure in
the system. The degree of boiling is limited to about 3% by weight to avoid flow instability
problems and to maintain a sufficient margin to avoid dryout of the fuel elements. This small
amount of steam is nevertheless about 25% by volume at prevailing pressures and is not
insignificant with regard to reactivity effects because it does provide slight positive feedback.
This occurs because the fuel channel spacing is greater than optimum, resulting in over-
moderation of neutrons. Voidage reduces the moderating effect and reduces neutron
absorption, creating conditions closer to the optimum.
3.7 Pressure and Level Effects

Pressure in the heat transport system must be maintained within certain limits to avoid over-
pressure conditions if it is too high or excessive vapour production and pump cavitation if it is
too low. Although these problems can be easily controlled by venting some vapour (if the
pressure is too high) or heating the liquid (if the pressure is too low), the required liquid level in
the pressurizer must also be maintained, and therefore level and pressure are interdependent,
although independently controlled. Level changes occur during warm-up and shutdown and
also during power increases and decreases due to changes in average coolant temperature. As
the average coolant temperature rises and increases the coolant volume, the excess is pushed
into the pressurizer, causing an increase in level. During warm-up from the cold shutdown
condition to the hot standby condition, as much as 60 m3 may be displaced due to swelling. The
resulting rise in pressurizer level is accommodated by the feed and bleed system, which
temporarily increases the extraction rate. With an increase in power, there is additional
swelling due to a further increase in average temperature and the onset of limited vapour
generation. Note that the steam generator steam pressure and hence steam temperature are
kept constant, and therefore, to achieve an increased heat transfer rate, the temperature
difference between the primary side reactor coolant and the secondary side water and steam
must increase. Hence, the average reactor coolant temperature in the steam generator must
increase, which causes an increase in its specific volume. This last effect accounts for about 10
m3 to 20 m3 displacement of reactor coolant. This is accommodated by ramping up the set-
point of the pressurizer level control system. Power transients therefore do not force a change
in inventory and have minimal impact on the feed and bleed system.
3.8 Temperature and Nuclear Considerations

Temperature and voidage both affect reactor reactivity. Generally, temperature causes a change
in reactivity as the coolant goes from cold shutdown conditions to hot shutdown conditions.
This leads to a negative coefficient of reactivity with new fuel and a slight positive coefficient
with equilibrium fuel. During normal operation, when some boiling occurs in the fuel channels,
the effect of voidage becomes significant. Because so little moderator is present between the
closest fuel rods, there is little initial moderating of neutrons, and the fast fission factor in-
creases. Moreover, with higher neutron energies, there is less probability of resonance absorp-
tion. Both these effects contribute to an increase in reactivity, giving a positive coefficient of
reactivity as voidage increases. In the overall design, these effects are compensated for by
other negative coefficients and by the reactor regulating system.
During operation of the primary system, the heavy water isotopic concentration is slowly
downgraded due to leakage and chemical treatment. The increase in light water increases
neutron absorption. This reduces neutron economy and results in an increase in fuel usage.
Moreover, with reduced heavy water isotopic concentration, additional reactivity worth must be
added to maintain criticality. In the event of a loss of coolant accident and channel void forma-
tion, the increase in positive reactivity will be greater. To ensure that this will be controllable,
there is a lower limit on heavy water isotopic concentration. This is generally in the range of
97.5% to 98.5%, while the normal value is around 99.5%.
There is also an upper limit on fuel channel heavy water isotopic concentration. It should
always be less than the calandria heavy water isotopic concentration by a small margin. This is
necessary because, in the event of an in-core loss of coolant accident, the possible penetration
of channel coolant into the calandria may displace moderator containing neutron poison. The
concentration of light water must therefore be sufficient to absorb at least as many neutrons as
the poison and light water in the moderator would have absorbed before the accident.
4 Steam Generators
4.1 Steam Generator Function
The steam generator serves the primary function of supplying energy to the turbine for power
production. It also serves the secondary function, but one most important from the safety
point of view, of removing energy from the reactor. Immediately after a reactor trip, the decay
heat of the fission products amounts to some 7% of full power heat production. The most
practical method of removing this heat is through the steam generator. Hence, it is necessary to
maintain a certain minimum water inventory in the steam generator and to ensure that
adequate reserve supplies of feed water are available.
The steam generator also serves as the main controlling link between the nuclear reactor and
the turbine generator. The power flows of the primary coolant system and the secondary steam
system must be balanced under all conditions. Any imbalance will cause an accumulation or a
depletion of the total heat content of the steam generator inventory, which in turn will result in
a rise or fall of steam pressure. Steam pressure is therefore a key parameter in plant control.
The steam generator separates the primary coolant circuit from the secondary steam circuit and
hence prevents radioactive products in the reactor coolant from entering the steam turbine and
the feedwater heating system. This barrier is an additional safety feature and also minimizes
the area over which radiation monitoring is required.
As an example of the importance of the steam generator in performing an essential safety
function, consider a nuclear plant with an electrical output of 600 MW. The reactor produces
approximately 1800 MW of heat, and the decay heat amounts to about 120 MW immediately
after a reactor trip from full power. This is a substantial amount, the removal of which must be
guaranteed under all conditions, including for a certain period after shutdown, because it takes
almost 10 hours for this heat to fall to a tenth of the initial value, that is, about 12 MW for the
example given. Separate smaller capacity heat removal systems can handle these lesser heat
flows on their own after the reactor has been shut down for some time, although they are often
put into service while the steam generators are still handling the major portion of the heat flow.
4.2 Steam Generator Conditions

The primary coolant enters the steam generators at nearly saturated conditions or saturated
with some steam present. As shown in Figure 17, it passes through the inverted U-tubes at high
velocity, resulting in high heat transfer coefficients on the inside of the tubes. Water on the
outside of the tubes is at lower pressure and therefore boils under the prevailing conditions.
This boiling results in high heat transfer coefficients on the outside of the tubes and induces a
strong upward convection current which promotes recirculation of the water in the steam
generator, while the steam produced is separated for use in the steam system. A feature of the
steam generator is the high heat transfer rate across the tubes compared with other types of
heat exchangers. This high heat transfer rate leads to a large temperature difference across the
tubes themselves, as well as in the boundary layer on each side of the tubes. Typically, the
temperature differences in the two boundary layers and across the tube itself are all
approximately the same. This means that the temperature drop across the tubes is about one
third of the total temperature difference and cannot be disregarded as in some other heat
exchangers such as the main condenser.
Ideally, heat exchangers should be counter-flow to minimize the temperature difference
between the hot and cold fluids. This is not possible with a phase change on one side of the
heat exchanger. Steam is therefore generated on the secondary side at a somewhat lower
temperature than the primary coolant inlet temperature. To maintain good thermodynamic
efficiency, a partial counter-flow arrangement is achieved by the subcooled feed water, which
enters the steam generator being passed through a preheater section before mixing with the
saturated recirculating water. The preheater receives heat from the primary coolant just before
its exit to minimize large temperature differences across which heat is transferred. The
temperature profile along the length of the tubes is shown in Figure 18. Table 3 gives typical
operating parameters of a CANDU 6 steam generator.
Figure 17 Steam generator for CANDU system
Figure 18 Temperature profiles in steam generator

Table 3 Steam generator conditions (CANDU 6)
Type Vertical U-Tube with Preheater

Number of Steam Generators 4
2
Heat Transfer Area (each) 3127 m
Heat Transferred (total) 2064 MW
Coolant Flow Rate (total) 7600 kg/s
Coolant Inlet Temperature 312°C
Coolant Outlet Temperature 266°C
Steam Flow Rate (total) 1047 kg/s
Feedwater Flow Rate 959 kg/s
Reheater Return Flow Rate 89 kg/s
Blowdown Flow Rate 1 kg/s
Steam Outlet Pressure 4.69 MPa (a)
Steam Outlet Temperature 260°C
Feedwater Inlet Temperature 187°C
Reheater Drain Inlet Temperature 258°C
4.3 Pressure Control

Any mismatch in the rate of heat flow to the steam generator on the primary side or from it on
the secondary side will result in a change in steam pressure. Control of steam generator pres-
sure is therefore achieved by reactor heat input or steam turbine output. It therefore becomes
part of overall plant control, which is covered in Chapter 10.
A change in steam pressure in the steam generator affects steam generator water level. A
reduction in pressure causes the expansion of existing vapour bubbles in the water and addi-
tional evolution of vapour due to the lower saturation conditions. The result is a rise in water
level or swelling of the inventory. An increase in pressure has the opposite effect, shrinking of
the inventory. These are transient conditions, and the original conditions are restored when
steam pressure returns to normal.
A change in load also affects the level. A simultaneous increase in reactor output and turbine
generator load will not affect steam pressure, but will result in increased evolution of vapour
within the water space. This will in turn cause a rise in water level or swelling of the inventory.
The opposite will cause shrinking of the inventory. These are steady state effects, and hence
there is a relationship between load and level. The level control system takes this into account
by ramping up the level set-point as load increases.
4.4 Level Control

Considering the effects described above, it is evident that level control is a critical aspect of
steam generator performance. A change in pressure is usually induced by a change in load. An
increase in turbine load, for example, will cause a drop in steam pressure as more steam is
drawn from the system. This will cause transient swelling, and the rise in water level would
indicate that less feed water is required, when in fact more will be required to meet the in-
creased steam demand. Inventory control by level only will result in excessive level fluctuations.
To solve this problem, the control system compares steam flow and feedwater flow and com-
bines the result with the level indication. Such a three element level control system gives better
control in normal operation, while single element level control using level measurement only is
satisfactory for low power operation.
4.5 Comparison of CANDU and PWR Systems

A comparison of steam generators as used in the PWR and CANDU systems is instructive in
comparing their relative merits. Both are very similar in construction. Natural circulation drives
the flow upwards over the tube bundle and downwards through a peripheral annulus after
separation of the steam, which occurs in cyclonic separators. Above the cyclones are steam
dryers, so that the steam leaving the steam generator has a moisture content of only about
0.25%. The fundamental difference between the PWR and CANDU steam generators is shown
in Figure 19 and Figure 20 respectively. The feed water enters the PWR steam generator at the
top near the normal water level and mixes with the recirculating water before passing down the
annulus. The recirculating water is therefore somewhat subcooled when it re-enters the tube
bundle. In the CANDU steam generator, the feed water enters at the bottom and passes over
the outlet side of the tube bundle before mixing with the recirculating water. The feed water is
thus preheated by the exiting coolant in a counter-flow arrangement, while the recirculating
water within the steam generator is not subject to subcooling. This is an advantage from a
thermodynamic point of view because hot and cold fluids are not mixed, but limits must be
placed on the feedwater temperature range to avoid thermal shock to the steam generator
tubes, and a deaerator with hot storage capabilities is essential. In the PWR, this is not a
problem because thorough mixing occurs before the water enters the tube bundle. This means
that the PWR steam generator can receive feed water over a much wider temperature range
and that hot storage of feed water is not essential.
Figure 19 Steam generator for PWR system
Figure 20 Steam generator for CANDU system
5 Steam System
From the steam generators, steam is piped to the high pressure turbine at a temperature of
about 260°C, as shown in Figure 21, where it expands partially in the high pressure turbine,
yielding some useful work. The steam is then put through a separator and reheater, where
moisture in the steam is removed in a cyclonic separator and the steam brought up to higher
temperature in a heat exchanger using steam directly from the steam generator. The super-
heated steam is then passed in parallel through two or three low pressure turbines where the
remaining energy is removed from the expanding steam. The steam is then passed to a con-
denser where it is condensed. This condensate is pumped through low pressure feedwater
heaters where it is preheated. It is then passed through a deaerator and pumped through high
pressure feedwater heaters for further preheating. It re-enters the steam generator at a tem-
perature of about 180°C. Table 4 gives typical technical and operating parameters for a CANDU
6 steam turbine.
Table 4 Turbine generator conditions (CANDU 6)
Number of HP Cylinders 1 (double flow)

Number of LP Cylinders 3 (double flow)
Number of Moisture Separators 4
Number of Reheaters 2
Steam Flow Rate (total) 1047 kg/s
Steam Flow Rate (turbine) 957 kg/s
HP Steam Inlet Pressure 4.55 MPa (a)
HP Steam Inlet Temperature 258°C
LP Steam Inlet Pressure 0.588 MPa (a)
LP Steam Inlet Temperature 242°C
Steam Exhaust Pressure 0.0042 MPa (a) (typical)
Steam Exhaust Temperature 30°C (typical)
Figure 21 Steam and feedwater system for a CANDU reactor

Control of steam flow is important to protect the steam generators as well as the steam turbine.
Provision is therefore made to release excess steam to avoid over-pressure in the steam genera-
tors and to stop steam flow to the turbine if electrical output is terminated or if a component or
system fails on the turbine generator.
5.2 Steam Bypass System

The main steam system is somewhat more complex than that shown diagrammatically in Figure
21 because provision must be made to discharge excess steam when it is not required or cannot
be used. It is advantageous to recover this steam, condense it, and return it to the feedwater
system, but if this is not possible, it must be discharged to the atmosphere. Figure 22 shows the
steam system for a CANDU 6 unit with the various valves to control steam flow under various
conditions.
Figure 22 Main steam system (Point Lepreau) (courtesy of NB Power)

As mentioned above, the steam generator has the important secondary function of removing
heat from the reactor after a turbine trip. Under such circumstances, the excess steam must be
diverted elsewhere. Excess steam may be vented to the atmosphere through atmospheric
discharge valves. This, however, is wasteful, and it is better to recover the steam in the con-
denser and to reuse the condensate. All plants therefore have a steam bypass system to enable
excess steam to be dumped to the condenser through the condenser steam discharge valves
and recovered by the condensate system. The condenser bypass system is usually designed to
accept the maximum quantity of steam that can be handled by the condenser to enable excess
steam to be bypassed under a range of reactor operating conditions. In the event that the
atmospheric discharge valves and condenser steam discharge valves are both inoperable, there
are conventional boiler safety valves on the steam generators to enable discharge of excess
steam.
During plant start-up, the reactor load is increased progressively through a series of checks on
its operation. During this time, it might not be desirable to run the turbine, and hence excess
steam would be discharged to the condenser, bypassing the turbine.
Another situation is a turbine trip or grid system disruption requiring a temporary rejection of
load. Such a large load reduction would result in a buildup of neutron absorbing xenon in the
reactor to the extent that it would force a reactor shutdown. To avoid this situation, the reactor
load must be maintained at about 60%. By dumping the equivalent quantity of steam into the
condenser, reactor load can be maintained at a sufficiently high level to avoid xenon poisoning
and a long term forced shutdown. Because the steam cycle has a thermal efficiency of about
30%, almost all the remaining 70% of heat input can be rejected to the environment through
the condenser. The maximum heat rejection capabilities of the condenser therefore roughly
match the minimum load required by the reactor to avoid excessive xenon buildup and reactor
poisoning.
5.3 Condenser Steam Discharge Valves

The condenser steam discharge valves open progressively to control steam generator pressure
by discharging an appropriate amount of surplus steam. Under certain circumstances, when the
turbine is unavailable and it is desirable to maintain load on a nuclear reactor to prevent poison-
ing out with xenon buildup, substantial quantities of steam may have to be discharged for an
extended period. This steam enters the condenser space at high velocity and in a hot super-
heated condition. To avoid thermal damage to the condenser structure and tubes and to the
condenser neck expansion joint, cool condensate is sprayed into the steam to cool it to normal
condenser temperatures. Provision must also be made to avoid flow induced vibration due to
the high velocities generated. A typical location of the steam discharge headers is shown in
Figure 35 and Figure 36 in the section on the condenser later in this chapter.
Certain limits are imposed on the operation of the condenser steam discharge valves. Under
particular circumstances, the condenser steam discharge valves are prevented from opening
and discharging steam. Circumstances leading to this are reduced condenser vacuum or high
turbine load when the ability of the condenser to absorb additional thermal load is limited.
5.4 Turbine Control Valves

The steam turbine has stop valves and governor valves on all high pressure turbine inlet pipes.
The governor valves control the steam flow to the turbine and hence the turbine output, while
the stop valves close in an emergency to stop all flow of steam to the turbine. In the event of a
turbine trip, they operate together as a back-up to one another, but during a load rejection, the
stop valves stay open. In addition, most turbines have stop valves and all have intercept valves
on all the low pressure turbine inlet pipes to serve the same purpose with regard to the steam
trapped in the moisture separator and reheater system. Expansion of this steam could acceler-
ate the turbine generator to overspeed conditions if it were allowed to enter the low pressure
turbine. Separate steam release valves enable this trapped steam to be released to the con-
denser.
6 Steam Turbine
Steam turbines consist essentially of a casing to which stationary blades are fixed on the inside
and a rotor carrying moving blades on its periphery. The rotor is fitted inside the casing, with
the rows of moving blades penetrating between the rows of fixed blades. Therefore, steam
flowing through the turbine passes alternately through fixed and moving blades, with the fixed
blades directing the steam at the correct angle for entry into the moving blades. Both casings
and rotors must be constructed to minimize damaging thermal stresses, and the moving blades
must be fitted securely to the rotor to withstand high centrifugal forces.
Where the rotor shaft passes through the ends of the casing, a seal is required to prevent steam
leakage and air ingress. In addition, within the casing, seals are required to prevent steam from
leaking around the blades rather than passing through them. Turbine seals are of the labyrinth
type where there is no mechanical contact between fixed and rotating parts. Leakage is there-
fore not really eliminated, but merely controlled to minimal amounts.
The rotor shafts are carried on bearings and are linked together and to the electrical generator.
Bearings must be properly aligned to accommodate the natural gravitational bending of the
shaft. Allowance must also be made for differential expansion between the rotors and the
casings during thermal transients. Both must be free to expand without upsetting the align-
ment, while allowing the rotors to expand more quickly and to a greater degree than the casing.
Lubrication is required for the bearings. Multiple pumps driven by alternative power sources
ensure adequate lubrication under all operating circumstances.
6.2 Turbine Cylinders

Typical units applicable to CANDU plants have a high pressure cylinder and three low pressure
cylinders. These are designed to accommodate the increasing specific volume of the steam as it
expands down to subatmospheric pressures. Provision is also made for steam quality improve-
ment by reheating and for the extraction of partially expanded steam for feedwater heating.
To withstand pressure, thick cylinder walls are required, but to minimize thermal stress, abrupt
changes in thickness and asymmetrical arrangements should be avoided. This leads inevitably
to smooth rounded profiles of the stress bearing components. In addition, uniform heating of
components is desirable to avoid differential expansion and undue thermal stress. This requires
sections of uniform thickness and provision for steam circulation within the casing, as shown in
Figure 23, to promote uniform temperature changes, particularly during unit start-up. Steam
access into and out of the cylinder must also be accommodated, which requires special nozzles
and casing reinforcement in these areas. Furthermore, overall expansion of the components
must be accommodated, and therefore the cylinders must be mounted on sliding pads with
keys to maintain alignment.
To assemble the turbine and to disassemble it for maintenance, the casing must be split in some
way. The joint is normally horizontal so that the upper half can be removed, leaving the lower
half in position along with the rotating parts.
The high pressure turbine is fairly compact due to the low specific volume of steam at this point
in the steam cycle. Provision is made to extract steam for feedwater heating and for steam
circulation within the casing to promote uniform heating and cooling. Figure 23 shows the high
pressure cylinder of a 900 MW turbine for a PWR, which has high pressure steam conditions
fairly close to those of a CANDU reactor.
Figure 23 High pressure turbine cross section for nuclear unit (courtesy of Eskom)
Low pressure turbines are very much larger because they must accommodate the increase in
specific volume as the steam expands. Low pressure turbines are therefore of different con-
struction. An inner casing supports the fixed blading and has annular channels through which
steam is extracted for feedwater heating. Surrounding each entire low pressure turbine is an
exhaust hood into which the exhaust steam flows before passing into the condenser below the
turbine. Figure 24 shows the low pressure cylinder of a 900 MW turbine for a PWR which has
low pressure steam conditions fairly close to those of a CANDU reactor.
Figure 24 Low pressure turbine cross section for nuclear unit (courtesy of Eskom)
Diaphragms carrying the fixed blades are fixed inside the turbine cylinders. Like the cylinder
casings, they are split horizontally to enable assembly around the turbine rotor. Because there
is a pressure drop across the fixed blades in both impulse and reaction turbines, sealing must be
provided between the diaphragms and the rotor to minimize steam leakage.
In turbines with reaction blading, significant axial thrust is developed by the steam flow. Double
flow turbines are normally used so that this axial thrust is balanced by opposing forces. Tur-
bines with impulse blading do not develop this large thrust, but the low pressure blading near
the turbine exhaust must have a reaction component due to the length of the blades. There-
fore, all large low pressure turbines are double flow.
6.3 Turbine Rotors

The moving blades are mounted on and transmit power to the turbine rotors. Their shafts are
coupled together and drive the generator rotor. Rotors must be able to transmit the applied
torque and to withstand the gravitational force due to their mass. As with casings, rotors are
subject to high temperatures and must be built to minimize thermal stress during temperature
transients. A certain degree of rigidity is important to minimize vibration during full speed
operation as well as during run-up and shutdown. Rotors of small turbines are usually ma-
chined from solid forgings, but large turbines and especially low pressure turbines are usually
built up from discs shrunk onto a stepped shaft.
The blades are attached to the turbine rotors in various ways and in such a pattern as to mini-
mize unbalanced forces. Even with close manufacturing tolerances, there is a slight mass
variation between blades. The moment of each blade with respect to the shaft centre is meas-
ured and the blades arranged around the circumference to minimize out of balance forces and
to ensure uniform stress on the turbine discs. Even so, some imbalance is inevitable, and the
turbine rotor must be dynamically balanced, usually at full speed, in a specially designed tunnel
or pit.
The generator rotor is invariably the longest of the rotors making up the entire turbine genera-
tor. A typical generator is illustrated in Figure 25, which shows the rotor. The rotor carries the
field coils which excite the stator coils to produce electric power. Because the field coils carry
heavy currents, they must be sufficiently robust and well cooled. The individual current carrying
bars must be well insulated and securely mounted to withstand centrifugal forces. Cooling is by
hydrogen under pressure circulating through the bars. The hydrogen is driven by fans mounted
on the rotor, which contribute to the full speed no-load frictional resistance felt by the machine.
More frictional resistance arises from the shaft bearings of all rotors. The stator is also cooled
by hydrogen and often by water passing through the windings as well. Electric power is pro-
duced at high voltage, and insulation must be designed accordingly.
Figure 25 Electrical generator for a large turbine (courtesy of Siemens)
6.4 Turbine Blading

Turbine blades are either fixed or moving and are shaped so that energy transfer takes place by
the impulse or reaction principle. The blade design is governed by the steam conditions in the
turbine, the desired steam velocities and directions, and the steam forces on the blades. In
addition, consideration must be given to geometrical limitations and dynamic forces arising
from high speed turbine rotation. The moving blades in particular are subject to very high
centrifugal forces and are sensitive to vibration induced by turbine rotation and steam flow.
Figure 26 shows a large nuclear steam turbine under construction. This turbine receives satu-
rated steam at 5.5 MPa from a pressurized water reactor, operates at 1500 revolutions per
minute, and has a nominal rating of 900 MW. In the figure, the lower half of a diaphragm
containing one row of fixed blades for one low pressure cylinder is seen being lowered into
position. Also shown are the moving blades attached to the rotor of another low pressure
cylinder. In the background is one complete rotor for a low pressure turbine. It is evident from
this photograph that the fixed blades are securely fixed at both ends in a robust diaphragm. The
moving blades, however, are attached only to the rotor discs and may or may not have tip
shrouding or lacing wires to increase rigidity. Both tip shrouding and lacing wires are used in
this machine and can be seen in the picture.
Figure 26 Construction of a 900 MW steam turbine for a nuclear unit (courtesy of Eskom)
The design philosophy for fixed and moving blades is naturally different. All large turbines have
pressure compounded impulse blading or reaction blading so that there is a pressure drop
across the fixed blading. The blades must withstand this pressure drop, and if the design
provides for a large diaphragm, as is usually the case in low pressure turbines, the blades must
also support the pressure difference across the diaphragm. The blades may therefore be quite
wide (as measured in the axial direction), especially towards the outer diameter. On the other
hand, moving blades are subject to high centrifugal forces and have a slender shape decreasing
in width towards the outer diameter.
When steam expands below the saturation line, some condensation takes place, and the steam
becomes wet. A very fine mist of droplets forms and passes over the blading, and some is
deposited on the blades to create a film of water. This film is swept off the blades and en-
trained by the steam. The size of these entrained drops is governed by surface tension rather
than by condensation phenomena. They are therefore somewhat larger than the original
condensation droplets and, when travelling at high velocity, can damage any material on which
they impinge. Being larger and heavier, they are also separated more easily from the main
steam flow during changes in direction. When these larger drops are swept off the fixed blades,
their velocity tends to be lower than that of the steam, and the moving blades tend to run into
them at high speed. The impact of the drops on the blades (more accurately the blades on the
drops) causes material erosion from the blades. Minute amounts of material are removed,
leaving a rough etched like surface. If severe, this material removal can change the blade profile
and weaken the blade. After impact on the moving blades, the water film tends to be thrown
off towards the periphery by centrifugal action. Turbine blades must be designed to withstand
moisture erosion in the affected areas. This is done using inserts of hard erosion resistant
material such as tungsten or stellite or by specialized heat treatment to harden the blades in
selected areas.
6.5 Turbine Sealing Principles

Steam turbines receive steam at high pressure and discharge it at high vacuum conditions.
Provision must therefore be made to prevent steam from leaking out of the turbine where the
steam pressure is high and air from leaking into the turbine where the steam pressure is below
atmospheric. The turbine casing can be made steam tight with accurately machined flange
joints, but where the rotating shaft passes through the stationary casing, leak tight joints cannot
be made without direct contact between fixed and rotating parts. Similarly, within the turbine
itself, there is a stepwise steam pressure drop through the various stages. Each diaphragm
carrying the fixed blades must withstand a pressure differential and have seals around the hole
through which the rotating shaft passes. In a reaction turbine, where there is a pressure drop
across the moving blades as well, these must have seals around their periphery to reduce steam
leakage around the top of the blades.
Various types of glands have been devised, but large steam turbines all use the labyrinth type of
seal. The distinct advantage of this type of seal is the lack of direct mechanical contact. Under
normal circumstances, this eliminates mechanical friction and wear and heat buildup. The
objective is not to eliminate leakage, but to reduce it to manageable proportions. Labyrinth
seals are therefore designed to pass steam through the gap between the fixed and moving
parts. Abnormal shaft deflection, however, can cause contact between the shaft and seals. The
seals are therefore spring mounted so that they can be pushed back and are thin so that they
can deform or wear away quickly in the event of contact to avoid overheating and damage to
the turbine shaft.
Figure 27 Spring mounted labyrinth seal to accommodate shaft deflection
6.6 Turbine Seals

Pressure in a turbine drops from the inlet to the exhaust, with the exhaust being under subat-
mospheric conditions. Moreover, as the turbine load changes, so does the steam pressure. The
steam flow through the turbine is proportional to the inlet steam pressure. Therefore, as the
turbine is throttled and steam flow reduced, the inlet steam pressure drops and, in the limit,
will drop to condenser vacuum conditions under no flow. In any case, the entire steam space of
the turbine is connected to the condenser, and at start-up, before steam is admitted to the
turbine, this space throughout the turbine will be at the same pressure as the condenser, that
is, very nearly a complete vacuum. Under these conditions, air will leak into the turbine
through the shaft seals, creating difficulty in maintaining the proper vacuum conditions.
To overcome this problem of air in-leakage, provision is made to supply low pressure steam to
the shaft seals when they are subjected to subatmospheric internal pressures. This steam is fed
into the seals in such a way that some leaks through the labyrinth rings into the turbine and
some leaks through the opposite rings to the outside atmosphere, as shown in Figure 28. As
long as steam flows outwards, air cannot enter the turbine against this flow, and the condenser
is protected against air ingress. The steam that leaks inwards into the turbine space is simply
condensed in the condenser and rejoins the steam-water circuit. The steam that leaks out of
the turbine to atmosphere is lost unless special provision is made to capture it. With the advent
of larger turbines, provision has been made to recover this leakage steam.
Figure 28 Principle of operation of LP turbine shaft seal (courtesy of NB Power)
6.7 Steam Sealing System

A typical steam sealing system must prevent inward air leakage at all turbine shaft seals during
start-up and very low load operation and prevent outward steam leakage at the high pressure
turbine and inward air leakage at the low pressure turbine during normal operation. Therefore,
auxiliary steam from outside the turbine must be supplied during start-up, but once a certain
load has been reached, the steam leaking through the high pressure turbine seals can be
directed to the low pressure turbine seals, and the whole system becomes self-sufficient.
Figure 29 shows a simplified steam sealing system for a large turbine.
Figure 29 Turbine steam sealing system
6.8 Turbine Bearings

Large steam turbines are made up of several rotating sections, with each section coupled rigidly
with the adjacent one. Normally, each rotor is mounted on a pair of journal bearings, one at
each end, with the couplings between the bearings of adjacent rotors. The bearings are
mounted on foundations to secure the rotating elements in the proper vertical and transverse
positions. The bearings are supplied with lubricating oil, and the shaft rides on a hydrodynamic
oil film without metal to metal contact. When starting from rest with the oil film having been
squeezed out, there is metal to metal contact. To avoid rubbing and scoring of the bearings,
high pressure jacking oil is supplied to a point under the shaft to lift the shaft off the bearings
before commencing rotation.
To secure the shaft in the proper axial position, a single thrust bearing is provided, usually close
to the high pressure turbine, but always in a location that will minimize shaft movement in
other parts of the turbine due to thermal expansion. The shaft will naturally expand outwards
from the thrust bearing. The turbine casings are also fixed to the foundations in a way that
permits free thermal expansion. These fixing points must be positioned to minimize differential
movement between the casing and rotor and must maintain proper radial alignment of the
rotating parts within the fixed parts. Typical locations of fixing points for a large turbine are
shown in Figure 30.
Figure 30 Fixing points for turbine rotors and casings (courtesy of NB Power)
All turbine rotors are slightly flexible due to the elastic nature of the material from which they
are made. They therefore bend slightly under the influence of gravity. As they rotate, this bend
remains always in the direction of the force of gravity. The couplings are consequently tilted
slightly upwards. To ensure proper alignment at the couplings, the adjacent rotors must be
installed at a slight upward angle. They in turn deform slightly under gravity, and their outer
bearings must be elevated further to compensate for bending. The further a rotor is from the
centre, the greater the compensation required. The result is that the entire turbine shaft is
mounted along a catenary, as shown in Figure 31. The end points may be some 20 mm to 30
mm higher than the midpoint in a large steam turbine and generator.
Figure 31 Turbine generator shaft catenary
6.9 Bearing Lubrication

Lubricating oil is supplied to all bearings to maintain hydrodynamic lubrication and to cool the
bearings. Heat is generated in the oil by friction in the hydrodynamic oil film and by turbulence
within the oil itself. Furthermore, a substantial amount of heat is conducted from the steam in
the turbine to the bearing through the shaft.
To ensure continuity of oil supply, there are three or four oil pumps driven by different power
sources. The main oil pump is usually driven directly by the turbine through a hydraulic connec-
tion to an oil driven turbine and sometimes by an uninterruptable electrical connection from a
small separate generator on the turbine shaft. As the turbine runs up to speed, more power is
provided to run the pump, which increases its speed, and at some intermediate turbine speed,
the main oil pump output reaches the required value.
The auxiliary oil pump is electrically driven and is supplied with power from the station or unit
auxiliary electrical systems. It is used during turbine run-up and rundown when the main oil
pump does not provide adequate oil pressure. It can also be used as an emergency oil pump at
any time, should the main oil pump fail. At low turbine speeds, however, this pump itself could
fail, and therefore an emergency pump is required.
At least one electrically driven emergency pump is provided. Its power supply is usually from an
uninterruptable electrical source such as the station emergency electrical system, from a DC
battery system, or both if two emergency pumps are provided.
There are also jacking oil pumps to lift the shaft off the bearings before rolling the turbine after
a shutdown in which the turbine has been stopped. These are positive displacement pumps
with a separate pump supplying each bearing to ensure a definite flow to each bearing. Figure
32 shows a typical lubricating oil system for a large steam turbine.
Figure 32 Turbine lubricating oil system
7 Reheating and Feed Heating

7.1 Steam Separators and Reheaters
Reheating is required to ensure adequate steam quality at the turbine exhaust to minimize
moisture erosion and is done locally, as part of the steam system, in independent reheaters
using some high pressure steam. Reheating with high pressure steam limits the maximum
temperature of the reheated steam, but sufficient superheat is still present to ensure that the
low pressure turbine exhaust steam wetness is not more than about 10%. The reheater gener-
ally consists of finned tubes carrying high pressure steam, which condenses fully to water inside
the tubes while transferring its heat to the lower pressure steam flowing over the outside.
Reheating is preceded by removal of excess moisture from the steam leaving the high pressure
turbine in separators. The separators may be separate from or integral with the reheaters. If
separate, they usually operate using centrifugal principles, where the steam is given a strong
swirling motion to throw out the moisture. If integral, sudden changes in the direction of the
steam cause moisture to be trapped by chevron type plates.
The flow of reheating steam is governed by the flow of reheated steam. As the reheated steam
extracts heat from the reheating steam, the latter condenses. The greater the flow of reheated
steam and hence the rate of heat extraction, the greater is the condensation rate and hence the
required flow of heating steam to take its place. The system is therefore self-regulating pro-
vided that the heating steam is fully condensed in the tubes and that the condensate is released
in a controlled manner. It is either pumped back into the steam generator (CANDU) or cascaded
into the highest pressure feedwater heater (PWR).
7.2 Deaerator
The deaerator is part of the feedwater heating system and receives extraction steam from the
turbine. The condensate to be heated and the extraction steam are intimately mixed in the
deaerator by a system of spray nozzles and cascading trays between which the steam percolates
as shown in Figure 33. The condensate is heated to saturated conditions and the steam con-
densed in the process. Any dissolved gases in the condensate are released in this process and
removed from the deaerator by venting to the atmosphere or to the main condenser. This
ensures removal of oxygen from the system particularly during turbine start-up and minimizes
the risk of corrosion within the system. Venting to the atmosphere reduces the load on the
condenser vacuum pumps, but results in some steam loss unless provision is made to condense
it and return it to the condensate system. Venting to atmosphere is only possible if the deaera-
tor pressure is above atmospheric as it is at higher loads. In the CANDU system, auxiliary
(pegging) steam is supplied at low loads to maintain elevated temperatures in the deaerator
storage tank.
Figure 33 Deaerator and deaerator storage tank (courtesy of NB Power)

Immediately below the deaerator is the deaerator storage tank, where a large quantity of feed
water is stored at near saturation conditions. In the event of a turbine trip, the steam generator
will require an assured supply of feed water to maintain the required water inventory during
subsequent stabilizing conditions, during which residual heat must be removed. During such
conditions, the loss of extraction steam to the high pressure feedwater heaters renders them
ineffective, and water from the deaerator storage tank is pumped into the boiler or steam
generator without further heating. If deaerator pressure is maintained between 0.5 MPa and
1.0 MPa, then the corresponding temperature of this stored feed water will be between 150C
and 180C. With an adequate supply of water at this temperature in the deaerator storage
tank, damaging thermal shock to the steam generator can be avoided.
The deaerator storage tank is usually located at a high elevation between the reactor contain-
ment and the turbine hall to ensure an adequate net positive suction head at the inlet to the
feedwater pumps, thus minimizing the risk of pump cavitation.
7.3 Feedwater Heaters

The low pressure feedwater heaters receive extraction steam from the low pressure turbine for
heating the feed water. Condensate extraction pumps pump condensate from the condenser
hotwell through the low pressure heaters to the deaerator.
The high pressure feedwater heaters receive extraction steam from the high pressure turbine.
Feedwater pumps pump feed water from the deaerator storage tank through the high pressure
heaters to the steam generator.
The conventional design for a feedwater heater is that of a horizontal cylindrical shell inside
which is a bank of U-tubes connected to a divided header at one end, as shown in Figure 34.
Feed water enters one side of the header, passes through the U-tubes, and leaves from the
other side of the header. Extraction steam from the turbine enters the shell and passes over the
outside of the tubes, where it is condensed. The condensed steam collects in the bottom of the
shell and is drained away. Generally, the drains from the high pressure heaters are cascaded
through lower pressure heaters to the deaerator, and those from the low pressure heaters
likewise to the condenser.
Figure 34 Feedwater heater with integral drain cooler (courtesy of NB Power)

Like the reheaters and deaerator, the feedwater heaters are self-regulating and draw only as
much extraction steam from the turbine as they require to heat the feed water. The greater the
feedwater flow, the greater is the rate of heat absorption from the steam, and the greater is the
flow of extraction steam.
In the event of a turbine trip, the steam supply to the turbine is cut off, and pressures through-
out the turbine drop to condenser pressure. Extraction steam pressures follow suit, inducing a
reverse flow from the feedwater heaters to the turbine. The low pressure in the feedwater
heater shell initiates vigorous flashing of any condensed steam. This may cause some water to
become entrained in the reverse steam flow to the turbine. Any water entering the turbine in
this way could cause severe damage to the turbine blading. As a precaution, therefore, non-
return valves are placed in the extraction steam lines between the feedwater heaters and the
turbine. These may be motor assisted to ensure sufficiently rapid closure at the onset of
reverse flow.
8 Condenser
The purpose of the condenser is to condense the exhaust steam from the turbine so that it can
be returned to the system for reuse. In the Rankine cycle, the condenser is complementary to
the steam generator in that it condenses the steam while the steam generator evaporates the
water. Like the steam generator, it has a free water surface that interfaces with the steam, and
some form of level control is required. Steam leaving the turbine enters at the top of the
condenser and circulates around the outside of the tubes, where it is condensed by cooling
water passing through the tubes. The resulting condensate rains down to collect in a hotwell at
the bottom of the condenser. Figure 35 and Figure 36 show a typical arrangement for a large
condenser.
Figure 35 Condenser longitudinal section (courtesy of NB Power)
Figure 36 Condenser cross section (courtesy of NB Power)

Cooling water from a lake or ocean is supplied to the condenser by circulating or cooling water
pumps and passes through the condenser tubes. The volume circulated is very large and the
supply pipes quite long, and therefore damaging inertial effects must be avoided when starting
and stopping the pumps. In the event of a pump trip, vacuum breakers enable the water
column to separate at the condenser and provide an air cushion to avoid damaging reverse flow.
The condenser hotwell serves as the entry point for demineralized make-up water, which is
required to compensate for leakage losses. Level sensors indicate when there is a need for
additional water. Note that the level in the hotwell rises naturally as the system is heated and
loaded and water is displaced from other components.
An air extraction system removes air that is in the system before start-up and that may leak into
the system during normal operation. This system extracts air from the end of the steam path,
that is, at the centre of the condenser tube bundle, where it collects as steam is condensed.
Shrouded air cooling tubes in this region reduce the temperature locally to a few degrees below
saturation to increase the ratio of air to steam extracted. The air is extracted by vacuum pumps
which run continuously to maintain condenser vacuum.
In the CANDU system, water spray pipes and nozzles are located in the condenser shell outside
the tube bundle, so that during condenser steam bypass operation, cool condensate can be
sprayed into the steam entering the condenser from the bypass system. This steam is at a high
temperature and in a superheated condition and therefore must be cooled before it enters the
tube bundle.
9 Electrical Generator
9.1 Generator Rotor
The generator rotor is machined from a single long slender forging of high strength steel. Slots
are cut on two opposing sides for a two pole rotor and on four sides for a four pole rotor to take
the rotor windings. The latter are for large half speed turbine generators used in nuclear plants.
The poles between the slots produce the rotating magnetic field as the rotor turns. The copper
windings are laid in the slots in pairs so that the current flow in them produces opposite north
and south poles in a two pole rotor and alternating north and south poles in a four pole rotor.
When rotating, these magnetic poles induce an alternating current in the generator stator. The
rotor windings carry a large current at low voltage. Hence, they must be cooled and insulated
while being held securely in the slots against very high centrifugal forces.
The individual windings are separated from one another by insulating separators and from the
rotor itself by insulating slot liners. These are held in place by insulating blocks and aluminum
wedges that fit into tapered grooves at the top of the slots. The windings have spaces between
them to allow free flow of hydrogen coolant through the windings. A space is left at the bottom
of each slot to create a duct to enable coolant to flow into the rotor from the ends. The flow is
driven by centrifugal action, with the rate determined by holes in the wedges. This ensures
uniform distribution of coolant and consistent cooling of all rotor windings. Hydrogen under
moderate pressure (to increase its density) is a good coolant and very suitable for this applica-
tion.
The winding connections, which protrude at the ends of the slots, wrap around, and enter slots
further around the rotor, are held in place by end bells. These are ring shaped and are shrunk
onto the ends of the slotted part of the rotor. They cover the end windings and secure them
under the high centrifugal forces arising during operation. Centrifugal or axial flow fans are
fitted to the rotor shaft at each end to feed coolant to the ends of the rotor and to ensure
proper circulation of hydrogen through the stator. Each fan serves one half of the generator.
Rotor current is provided by an exciter connected to and driven by the generator shaft. Current
is transferred to the rotor by fixed brushes and slip rings mounted on the shaft. This current is
on the order of 5000 A and the voltage around 500 V for a 660 MW generator.
9.2 Generator Stator

The generator stator, with its laminated iron core and heavy copper windings inside a robust
fabricated steel frame, is invariably the heaviest component of the turbine generator. The stator
core is made up of laminated steel sheets having low loss magnetic properties and with a
thickness of 0.5 mm or less. The sheets are coated with an insulating varnish to prevent circu-
lating currents. Slots are arranged in the core for the stator windings, and provision is made for
circulation of hydrogen coolant.
The stator current of a typical 660 MW generator operating at 24 kV and a power factor of 0.85
is on the order of 19,000 A, or about 950 A per bar. The current generates considerable heat,
and therefore additional cooling must be provided. Water is an ideal coolant but must be pure
enough to be effectively nonconducting. It flows to and from the stator windings through
nonconducting polytetrafluorethylene (PTFE) hoses so that the windings remain fully insulated
from the coolant pumps and heat exchangers. The water pressure is less than the hydrogen
pressure to prevent water ingress to the generator in the event of coolant system leakage.
Hydrogen, driven by the fans, circulates over the outside of the windings and then through
water coolers mounted longitudinally around the periphery of the stator core or vertically at the
corners of the stator core.
9.3 Excitation System

A separate exciter coupled directly to the generator rotor shaft provides excitation current to
the rotor through brushes and slip rings. The exciter is, in effect, a small generator, which itself
requires a small excitation current. This current in turn is provided by a pilot exciter and deliv-
ered to the main exciter rotor through its own brushes and slip rings. The pilot exciter has a
permanent magnet on its shaft to provide the necessary rotating magnetic field.
The development of compact rectifiers enabled them to be incorporated into the rotating
elements of the exciter. This means that the rotor and stator windings of the main exciter can
be interchanged. The resulting brushless excitation system with a rotating armature main
exciter is less complex and requires less maintenance than the conventional system.
The further development of thyristors enabled the conventional excitation system to be re-
placed by a current control system which feeds the required excitation current directly to the
generator from an external power source. Although a feed from the generator electrical output
circuit through an excitation transformer can be used as this source, it does not have the same
intrinsic electrical security as a shaft mounted excitation system.
9.4 Generator Systems

The most important auxiliary systems associated with the electrical generator are those for
cooling, sealing, and lubricating.
Virtually all generators of the size used in nuclear power plants are cooled by hydrogen at a
pressure of 0.4 MPa to 0.5 MPa. Hydrogen under this pressure, which increases its density, has
good heat transfer properties. Compared with air under similar conditions, it has twice the heat
transfer capability and several times the thermal conductivity and specific heat. This makes it
ten times more effective at heat removal than air. The hydrogen atmosphere also prevents
degradation of insulation and other materials by oxidation processes.
Hydrogen, however, is flammable and explosive in concentrations from 4% to 76% of hydrogen
in air. Hence, it must be contained in the generator and maintained at a high purity of about
97%. Special seals are required where the rotor shaft penetrates the casing, and appropriate
provisions must be made to prevent any mixing with air during gassing or degassing of the
generator. Most seals have a very fine running clearance between the fixed and moving parts,
into which seal oil is injected. The seal oil is usually lubricating oil from the main turbine gen-
erator oil tank, but is purified and pressurized for this purpose. When gassing the generator
before operation, all air is displaced by admitting heavier carbon dioxide at the bottom of the
generator casing and allowing the air to escape at the top. Hydrogen is then admitted at the
top and, being lighter, displaces the carbon dioxide downwards. When all the carbon dioxide
has been expelled, the generator casing is sealed and more hydrogen admitted to pressurize it.
Degassing is done by following a reverse procedure, with hydrogen discharged to atmosphere
from the top of the generator while being displaced by carbon dioxide admitted at the bottom.
Like the turbine, the generator runs on bearings at each end, with the exciter having its own
bearings. These are lubricated by the main turbine oil system. The bearings, however, must be
removable because assembly of the generator is different from that of the turbine. Because the
generator stator cannot be split horizontally, the rotor must be threaded into the stator from
one end of the generator, as shown in Figure 37. Once in place, the bearings are mounted and
aligned to ensure that the rigidly bolted generator couplings maintain the required shaft cate-
nary under gravity. This means that the generator bearings must be slightly higher than those
of the turbine due to the natural bending of the turbine and generator rotors under the influ-
ence of gravity at all times, even at speed. Furthermore, the generator rotor must be able to
move axially in the bearings by about 25 mm or more to accommodate axial expansion of the
turbine rotors as their temperatures increase to those of normal operation.
Figure 37 Insertion of rotor into stator of generator (courtesy of NB Power)
10 Power Transmission
10.1 Generator Main Connections
The neutral point of the generator star connection is grounded through a current limiting device
to minimize ground fault currents in the generator, should a flashover occur in the stator wind-
ings. A convenient manner of providing a low resistance, yet current limiting ground connection
is to use a transformer where the inductance limits the current. In the event of a ground fault,
the lower voltage secondary side reflects a high current. This signals the fault to the generator
protection system, which initiates a trip of the generator circuit breaker.
10.2 Generator Circuit Breakers

Generator circuit breakers are usually provided and located along the busbars between the
generator and the tee-off to the unit transformer. They provide the necessary protection to the
generator in the event of electrical faults. If not provided in the busbars, they are located on
the high voltage side of the generator transformer in the electrical switchyard. The generator
circuit breakers must interrupt heavy currents within a fraction of a second, and therefore
provision must be made to extinguish the arc which forms as the circuit breaker contacts
separate. Usually, this is done with a jet of air in air blast switchgear. A development of this
concept is the use of sulphur hexafluoride (SF6) instead of air. It is odourless, colourless, non-
toxic, and nonflammable and has a density five times that of air, a thermal transfer coefficient
greater than that of air, and a dielectric strength comparable to oil under moderate pressure. It
is therefore ideally suited for use in circuit breakers, particularly where high voltages are used
and limited space is available.
10.3 Generator Transformers

Generator transformers are required to increase the voltage before transmission of electrical
power through the grid system. By increasing the voltage, the current is correspondingly
reduced. Because electrical losses are proportional to the square of the current, there are
significant advantages in operating a grid system at high voltages.
10.4 Transformer Connections

Three phase transformers may have star connected or delta connected circuits on the primary
and secondary sides. This leads to four possible transformer arrangements between the power
source and the load: star-delta, delta-star, star-star, and delta-delta. Practical considerations
dictate the use of certain combinations in preference to others. All high voltage three phase
circuits require a ground connection. This ensures that the potential differences between lines
and grounded structures are maintained at design values. A star connected circuit provides
such a connection at the neutral point, whereas a delta connected circuit does not. Further-
more, third harmonic currents are generated by hysteresis in the core. The magnitude of these
can be reduced by a delta connected circuit. For this reason, star-star arrangements with star
connected circuits on both sides are not common. Hence, it follows that, because most electri-
cal generators have star connected circuits, the generator transformers and unit transformers
supplied by the generator have delta connected circuits on the primary side and star connected
circuits on the secondary side to provide a new grounded neutral point for the next circuit at a
higher voltage in the case of the generator transformer and at a lower voltage in the case of the
unit transformer. In very large generating units, separate single phase transformers are pro-
vided for each phase. In such a case, a delta connected circuit may be provided on both primary
and secondary sides. The advantage of such a delta-delta arrangement is that one of the single
phase transformers can be isolated for maintenance or removed for replacement without
disrupting the continuity of the three phase system, as shown in Figure 38. The system can
therefore continue to operate, although at a reduced power level which is only about 58 per-
cent of rated power. This is not possible with a star connected secondary circuit.
Figure 38 Arrangement of three single phase generator transformers
11 Problems
1. List the reactivity control devices for a CANDU reactor and for each describe briefly its
purpose and how it operates.
2. Explain the difference between Shutdown System 1 (SDS1) and Shutdown System 2
(SDS2). Explain how each operates and under what conditions each would operate. Cla-
rify how each would be reset to the primed condition.
3. Explain how the inventory and pressure of the heat transport system are maintained.
4. Describe how heat is removed from the reactor under all operational conditions includ-
ing the shutdown state.
5. If the exit quality of the steam leaving a fuel channel is 4%, calculate the void fraction at
this point assuming no slip (vapour velocity equal to liquid velocity) in the channel. As-
sume that the pressure in the channel at the exit is 10 MPa.

6. Consider heating of the heat transport system from the cold shutdown state at 30°C to
the hot zero load condition (assumed to be 260°C) and then power raising to the full
load condition.
a. If, during heating to zero load, the heat transport system inventory is 200 Mg, de-
termine how much heavy water is displaced into the pressurizer during this op-
eration. For a pressurizer 2 m in diameter and 10 m in height, determine the rise
in pressurizer level that would occur if heavy water were not bled from the sys-
tem.
b. If, during power raising, a further 10 m3 is displaced into the pressurizer, give and
explain two different reasons for this expansion.
7. Determine the percentages of sensible heat and latent heat transferred to the secondary
circuit in the steam generator as well as the total rate of heat transfer given the follow-
ing parameters:
a. Steam flow rate 1 000 kg/s
b. Steam pressure 4 MPa
c. Feedwater temperature 175°C.
8. Suppose that the steam generator has an integral preheater and that during normal full
power operation the temperature of the feedwater entering the preheater is 170°C.
a. Calculate the difference between the temperature in the steam generator and
the feedwater inlet temperature (steam generator ΔT) during normal full power
operation if the steam generator pressure is 5 MPa.
b. Calculate the steam generator ΔT during poison prevent operation (turbine
tripped) if deaerator pressure is 360 kPa and steam generator pressure is 5 MPa.
9. Suppose that all feedwater flow to the boilers of a CANDU unit is suddenly lost when
operating at full power. Explain why this would result in an increase in heat transport
system temperature and pressure before the boiler levels would have dropped signifi-
cantly.
10. In the event of a turbine generator trip or load rejection from full load, the reactor pow-
er is decreased to a lower level to prevent xenon buildup in the reactor and the surplus
steam is dumped to the condenser.
a. Explain what limits the maximum amount of steam that can be rejected and state
the usual limit. At this limit, state the maximum load that the reactor can sus-
tain.
b. Explain why water is mixed with the steam before release into the condenser.
Give reasons why this must be done.
11. A labyrinth shaft seal is used to ensure proper sealing where the shaft passes through
the casing of a low pressure steam turbine.
a. Explain briefly the design philosophy of such a seal.
b. Sketch a cross section of the shaft and seal showing labyrinths and external pipe
connections. Show the flow of steam and air through the pipes and labyrinths in
normal operation.
c. Sketch the pressure profile across the seal showing clearly the conditions inside
and outside the turbine casing.
d. Explain what would happen to the pressure and flow in the seal if the external
steam supply to it should fail. How would this affect the operation of the tur-
bine?
12. Describe, with the aid of a sketch, the arrangement of a typical thrust bearing. Explain
the purpose of thrust bearings and state, with reasons, where they are located in a typi-
cal large multi-cylinder turbine.
13. Explain, with the aid of a sketch, how the shaft of a large multi-cylinder turbine genera-
tor is aligned. Clarify how the bearings are set and what effect would be experienced at
the couplings between the rotors if the shaft were not aligned properly.
14. Explain the general philosophy of providing lubrication to the turbine and generator
bearings. Explain what pumps are used and how they operate. Clarify how they are dri-
ven and what happens should their driving power fail.
15. Explain the purpose of a deaerator and the importance of maintaining its pressure in the
event of a turbine trip.
16. Explain the difference between feedwater heating in a deaerator (open heater) and a
feedwater heater (closed heater). Consider the temperature differences between the
heating and heated fluids and their respective pressures. On this basis, explain why
there is usually only one deaerator, but multiple feedwater heaters.
17. Describe the structure of the main condenser and explain the purpose of its key internal
components.
18. Explain why hydrogen is used for cooling in the electrical generator and what precau-
tions need to be taken during gassing and de-gassing. Clarify the operational procedure
for these processes.
Answer Guide for Chapter 8
1. See Section 2.6 Reactor Control Systems and Devices.
2. See Section 2.6.5 Mechanical rod shut-off absorbers and Section 2.6.6 Liquid poison in-
jection system.
3. See Section 3.3 Pressurizer as well as Section 3.4 Inventory Control System and Figure 15
Heat transport purification and feed system.
4. See Section 3.5 Shutdown Cooling System and Figure 15 Heat transport purification and
feed system as well as Section 3.5 Shutdown Cooling System and Figure 16 shutdown
cooling system. See also Section 4.1 Steam Generator Function.
5. Use steam tables. Void fraction α = 0.34 (using light water steam tables).
6. Use steam tables. (a) Volume displaced ΔV = 53 m3 (using light water steam tables).
Rise in level ΔH = 17 m much greater than available height so some inventory must be
bled from the system. (b) Increase in ΔT in steam generator due to increased rate of
heat transfer and hence increased average temperature of reactor coolant. Generation
of some steam in the fuel channels which has a greater volume than the water from
which it was formed.
7. Use steam tables. Sensible heat 17%. Latent heat 83%. Total heat transfer rate Ω =
2058 MJ/s.
8. Use steam tables. (a) Steam generator ΔT = 94°C. (b) Steam generator ΔT = 124°C.
9. See Figure 18 Temperature profiles in steam generator. Consider change and restoration
of average temperature difference between primary coolant and secondary steam.
10. See Section 5.2 Steam Bypass System and Section 5.3 Condenser Steam Discharge
Valves.
11. See Section 6.5 Turbine Sealing Principles and Figure 27 Spring mounted labyrinth seal to
accommodate shaft deflection as well as Section 6.6 Turbine Seals and Figure 28 Prin-
ciples of operation of LP turbine shaft seal. Consider relative pressures and note that
flow is always down the pressure gradient.
12. See Section 6.8 Turbine Bearings and Figure 30 fixing points for turbine rotors and cas-
ings.
13. See Section 6.8 Turbine Bearings and Figure 31 Turbine generator shaft catenary.
14. See Section 6.9 Bearing Lubrication and Figure 32 Turbine lubrication oil system.
15. See Section 7.2 Deaerator.
16. See Section 7.2 Deaerator and Figure 33 Deaerator and deaerator storage tank as well as
Section 7.3 Feedwater Heaters and Figure 34 Feedwater heater with integral drain coo-
ler.
17. See Section 8.1 General Arrangement and Figure 35 Condenser longitudinal section and
Figure 36 Condenser cross section.
18. See Section 9.4 Generator Systems.

Atomic Energy of Canada Limited, Canada Enters the Nuclear Age, AECL, McGill-Queen’s Univer-
sity Press, 1997.
Chaplin, R. A., “Nuclear Reactor Steam Generation”, Chapter 3.10.2.10 in Thermal Power Plants,
Encyclopedia of Life Support Systems by UNESCO, EOLSS Publishers, Oxford, UK, 2006.
Chaplin, R. A., “Steam Turbine Components and Systems”, Chapter 3.10.3.3 in Thermal Power
Plants, Encyclopedia of Life Support Systems by UNESCO, EOLSS Publishers, Oxford, UK,
2006.
Chaplin, R. A., “Steam Turbine Steam System”, Chapter 3.10.3.4 in Thermal Power Plants,
Encyclopedia of Life Support Systems by UNESCO, EOLSS Publishers, Oxford, UK, 2006.
Chaplin, R. A., “Electric Power Generation”, Chapter 3.10.3.9 in Thermal Power Plants, Encyclo-
pedia of Life Support Systems by UNESCO, EOLSS Publishers, Oxford, UK, 2006.
13 Acknowledgements
Acknowledgements are extended to Atomic Energy of Canada Limited and to NB Power for the
use of information, diagrams, and photographs to support the text of this chapter.
George Bereznai
Simon Pang
1
CHAPTER 9
Nuclear Plant Operation
Prepared by
Dr. Robin A. Chaplin
Summary:
This chapter deals with the operating concepts of a CANDU nuclear power plant. It combines
some theoretical aspects with basic operating procedures to explain how the plant operates.
Key aspects related to plant control are addressed. Space allows only the primary energy
generation, transport and conversion components to be covered, but there are many other
components whose operation is vital for the efficient and safe operation of the plant. There is
approximately an equal division of detail between the nuclear reactor, the heat transport and
steam systems, and the steam turbine.
Table of Contents
1 Power Production ................................................................................................................... 5
1.1 Power Output Regulation ............................................................................................... 5
1.2 Operational Constraints .................................................................................................. 7
2 Plant Control and Protection .................................................................................................. 7
2.1 CANDU Plant Control Systems ........................................................................................ 7
2.2 Reactor Protection Systems .......................................................................................... 10
2.3 Reactor Shutdown Systems........................................................................................... 10
2.4 Emergency Coolant Injection ........................................................................................ 11
2.5 Turbine Protection System............................................................................................ 12
3 Nuclear Reactor Operation ................................................................................................... 13
3.1 Reactivity Characteristics .............................................................................................. 13
3.2 Source Multiplication.................................................................................................... 13
3.3 Approach to Criticality .................................................................................................. 15
3.4 Approach Technique ..................................................................................................... 18
3.5 Reactor Start-up............................................................................................................ 19
3.6 Reactivity Changes ........................................................................................................ 22
3.7 Fuel Burnup................................................................................................................... 23
3.8 Refuelling Considerations ............................................................................................. 24
4 Heat Transport System.......................................................................................................... 25
4.1 Operational Considerations .......................................................................................... 25
5 Steam Generators ................................................................................................................. 26
5.1 Plant Control ................................................................................................................. 26
5.2 Steam Generator Pressure Control ............................................................................... 27
©UNENE, all rights reserved. For educational use only, no assumed liability. Nuclear Plant Operation – July 2016
5.3 Swelling and Shrinking .................................................................................................. 28

5.4 Steam Generator Level Control..................................................................................... 30
6 Steam System........................................................................................................................ 34
6.1 Characteristics and Function......................................................................................... 34
6.2 Atmospheric and Condenser Steam Discharge Valves ................................................. 35
6.3 Steam System Operation............................................................................................... 36
7 Main Condenser.................................................................................................................... 37
7.1 Thermodynamics and Heat Transfer............................................................................. 37
7.2 Turbine Limitations ....................................................................................................... 40
7.3 Environmental Limitations ............................................................................................ 41
7.4 Condenser Tube Fouling................................................................................................ 41
7.5 Cooling Water Circuit .................................................................................................... 42
7.6 Flow Transients ............................................................................................................. 43
7.7 Variation in Condenser Vacuum.................................................................................... 44
7.8 Causes of Poor Condenser Vacuum .............................................................................. 45
7.9 Condenser Vacuum Raising........................................................................................... 46
7.10 Condenser Vacuum Breaking ........................................................................................ 47
8 Steam Turbine ....................................................................................................................... 48
8.1 General Operational Considerations............................................................................. 48
8.2 Turbine Start-up ............................................................................................................ 49
8.3 Turbine Losses............................................................................................................... 51
8.4 Turbine Load Variation .................................................................................................. 51
8.5 Differential Expansion ................................................................................................... 57
8.6 Turbine Governing......................................................................................................... 58
8.7 Turbine Responses ........................................................................................................ 59
8.8 Reactor Trip ................................................................................................................... 59
8.9 Turbine Trip ................................................................................................................... 60
8.10 Steam Flow Control....................................................................................................... 61
8.11 Load Rejection............................................................................................................... 61
8.12 Turbine Overspeed Protection ...................................................................................... 63
9 Problems ............................................................................................................................... 64
List of Figures
Figure 1 Lagging and leading plant operation................................................................................. 5
Figure 2 Basic plant control system ................................................................................................ 8
Figure 3 Source multiplication by subcritical multiplication factor .............................................. 15
Figure 4 Stabilization time for different subcritical multiplication factors ................................... 16
Figure 5 Approach to critical - subcritical multiplication factor change ....................................... 17
Figure 6 Approach to critical - inverse count rate change ............................................................ 18
Figure 7 Major activities during start-up and loading .................................................................. 21
Figure 8 Effects of fuel burnup on reactivity................................................................................. 24
Figure 9 Steam generator temperature profile............................................................................. 27
Nuclear Plant Operation 3
Figure 10 Steady state swelling due to different load conditions................................................. 29

Figure 11 Transient swelling due to changing load conditions..................................................... 30
Figure 12 Constant steam generator water level.......................................................................... 31
Figure 13 Ramped steam generator water level........................................................................... 32
Figure 14 Three element steam generator level control system .................................................. 33
Figure 15 Steam generator level alarms and set points (courtesy of NB Power) ......................... 34
Figure 16 Main steam system (Point Lepreau) (courtesy of NB Power) ....................................... 36
Figure 17 Condenser temperature profile .................................................................................... 38
Figure 18 Changes in condenser temperature profile .................................................................. 39
Figure 19 Variation in condenser pressure ................................................................................... 40
Figure 20 Condenser performance curves (adapted courtesy of Eskom)..................................... 41
Figure 21 Typical condenser cooling water system....................................................................... 43
Figure 22 Simplified condenser cooling water system ................................................................. 44
Figure 23 Effect of change in condenser pressure (courtesy of NB Power).................................. 45
Figure 24 Effect of change in turbine load and condenser pressure ............................................ 46
Figure 25 Condenser cross section (courtesy of NB Power) ......................................................... 47
Figure 26 Major activities during start-up and loading ................................................................ 50
Figure 27 Steam pressure variation through the turbine (courtesy of NB Power)....................... 52
Figure 28 Steam pressure variation with load change.................................................................. 53
Figure 29 Steam pressure versus steam flow and turbine load.................................................... 53
Figure 30 Partial load turbine expansion lines (courtesy of NB Power) ....................................... 54
Figure 31 Change in whirl velocity at low loads............................................................................ 56
Figure 32 Recirculation of steam at low loads (courtesy of NB Power)........................................ 57
Figure 33 Differential axial expansion of turbine rotors and casings............................................ 58
Figure 34 Turbine generator non-sequential and sequential trips ............................................... 62
Figure 35 Turbine-generator load rejection.................................................................................. 63
Figure 36 Turbine generator overspeed trip ................................................................................. 64
Nomenclature for Equations

A area m2
cp specific heat at constant pressure kJ/kg°C
f thermal utilization factor
k neutron multiplication factor
M mass flow rate kg/s
U overall heat transfer coefficient kJ/sm2°C
S0 number of source neutrons
S∞ number of measured neutrons
p pressure MPa
p resonance escape probability
v specific volume m3/kg
V velocity m/s
Δk change in neutron multiplication factor
ΔT rise in temperature (of fluid) °C
ε fast fission factor
η reproduction factor
η efficiency
θ temperature difference (between fluids) °C
Λt thermal neutron non-leakage probability
Λf fast neutron non-leakage probability
ν neutrons emitted per neutron absorbed
ρ density kg/m3
Σf macroscopic fission cross section m-1
Σa macroscopic absorption cross section m-1
Ω rate of heat transfer kJ/s
1 Power Production
1.1 Power Output Regulation
Consider a very simple system with a nuclear reactor, steam generator, and turbine generator
supplying electrical power to an isolated electrical grid, as shown in Figure 1. The power must
be generated at the moment it is required by the consumers connected to the grid. Power
production must follow demand exactly, and any mismatch will cause the grid frequency to fall
or rise as demand increases or decreases.
Figure 1 Lagging and leading plant operation

A basic control system works as follows to maintain appropriate power output from the plant.
In the event of an increase in demand for electrical power, the mismatch will cause the grid
frequency to fall. Because the turbine generator is synchronized to the grid, its speed will drop
accordingly. This will be sensed by the turbine governor, which will open the governor valves to
admit more steam to increase the power output of the turbine generator. The additional flow
of steam to the turbine will cause a reduction in steam pressure in the steam generator. This in
turn will be sensed by the reactor regulating system, which will lower the liquid zones or with-
draw control rods from the reactor core until the increased fission rate generates sufficient
additional heat to restore the steam generator pressure. In the event of a decrease in demand,
the reverse occurs. This is known as the reactor following or turbine leading mode of operation
(or normal mode in some plants because it is a natural way of maintaining stable conditions).
Such a system, however, cannot maintain the specified frequency (60 Hz in North America)
exactly without large unstable oscillations, and therefore a certain speed droop is incorporated
into the turbine governor. This enables a progressive increase in governor valve opening (steam
flow) as the turbine speed (grid frequency) falls. A typical droop setting is 4%, which means
that, if the turbine were initially at zero load and at full speed, its speed would have to drop to
96% before the governor valve would be fully open. Such a speed is not acceptable for the
turbine due to possible blade vibration, nor to the grid due to loss in speed of connected
motors. Therefore, the governor is adjusted to bring the speed back to 100% at full load. In the
event of a turbine trip or load reduction to zero under these conditions, the reverse would
occur, and the turbine speed would rise to 104% of full speed.
In the operating mode just described, the nuclear reactor output follows the electrical grid
demand, and therefore its power level oscillates continuously. This can have certain adverse
effects, depending upon the type and design of the reactor. Excessive oscillations impose
temperature transients on the fuel, which could cause premature failures of the fuel cladding.
Large oscillations near full power could cause power limits to be exceeded, thus tripping the
reactor and losing power production as well as imposing restart transients on it and the turbine.
Due to high capital cost and low fuel cost, it is desirable to run nuclear reactors at full power
most of the time. An alternative mode of operation is therefore often used at nuclear plants, in
which reactor power output is fixed. This is known as reactor leading or turbine lagging opera-
tion. To maintain stable operation, reactor power is controlled at a given value by measuring
the neutron flux and adjusting the control rods or liquid zones accordingly. Pressure in the
steam generator must be maintained at the proper value to ensure stable conditions in the
reactor coolant circuit. This is done by opening or closing the turbine governor valves to control
the steam flow from the generators. The turbine then delivers power according to the steam
flow, and the generator sends this power into the grid system, regardless of the grid frequency.
The grid frequency must then be controlled by other turbine generators which feed into the grid
system and operate in the turbine leading mode.
By referring to Figure 1 showing the two modes of operation, it can be seen that steam genera-
tor pressure is a key controlling parameter in both modes. This highlights the importance of the
steam generator, where a balance of heat input and heat output must be maintained to main-
tain its pressure. Furthermore, the difference in temperature between the primary coolant and
the secondary working fluid determines the rate of heat transfer. Hence, the reactor coolant
temperature is determined by the saturation temperature and thus by the pressure in the
steam generator.
1.2 Operational Constraints

During operation, parts of the reactor, steam system, and turbine are subjected to high tem-
peratures. If these parts have thick walls or a substantial solid mass, they will likely suffer
thermal stress during the heating and cooling that arises during start-up and shutdown and also
during load transients. If a thick component is heated on one side, this side will tend to expand.
If constrained by the still cold base material so that it cannot expand, an internal stress will be
set up. The reverse happens during cooling. This means that large rigid components which are
subject to transient and uneven heating and cooling will suffer low cycle fatigue damage and
may ultimately fail. This effect can be minimized by slow heating and cooling to reduce tem-
perature differences in single components such as reactor pressure vessels, steam generators,
steam pipes, turbine casings, and turbine rotors. This means that all these components must be
preheated slowly before start-up and the unit loaded slowly. Similarly, load changes on the unit
should be done slowly. This imposes operating restrictions on reactors and turbines. Generally,
the larger the unit, the longer will be the time to start it up and load it. This makes large units
less flexible in operation than smaller units.
In the reactor, temperature transients on the fuel cause structural changes within the fuel and
stress on the cladding and therefore must be minimized to avoid premature fuel failures. In the
turbine, uneven heating and cooling of the rotor can cause bending, which in turn causes
excessive vibration and in the extreme case, contact with the casing. Therefore, heating and
cooling must be monitored very carefully.
A nuclear reactor is also subject to xenon transients which may inhibit operation for a period.
When the load on a reactor is suddenly reduced, xenon, a fission product that absorbs neutrons
very strongly, builds up in the fuel and may force a reactor shutdown. The xenon eventually
decays after about 40 hours, and the reactor can be restarted. During this time, there is a loss
in electrical power production, and the larger the unit, the more serious will be this loss in
revenue generating output.
2 Plant Control and Protection

2.1 CANDU Plant Control Systems
A CANDU nuclear plant can operate as a base load plant or as a load following plant. The latter
is the conventional mode because the plant responds naturally to demand changes on the grid
system, as explained above. Fossil fuel fired plants operate in this way. In nuclear plants, there
are special considerations, the main one being the high capital cost and the low fuel cost
compared with conventional plants. Nuclear plants should therefore operate continuously at
maximum load to ensure the lowest generation cost. A steady load also avoids thermal tran-
sients, prolongs the life of all high temperature components, and minimizes the possibility of
premature fuel failures. Hence, most nuclear plants operate as base load plants, and the
control system in a CANDU unit provides for this, as shown in Figure 2.
Figure 2 Basic plant control system

The plant can operate as a load following plant where the turbine responds to grid conditions
and the reactor follows the turbine to produce the required output (turbine leading and reactor
following). In the alternate mode the reactor power is fixed, and the turbine follows to produce
the equivalent output (reactor leading and turbine following). This alternate mode with reactor
power fixed enables the reactor to operate close to its power limits without risk of tripping due
to control system oscillations from the turbine or other downstream component perturbations,
thus maximizing plant output. In some plants, this alternate mode is called normal mode
because the plant normally operates this way, and load following is the alternate mode.
The amount of steam produced in the steam generator depends upon the amount of heat
received from the reactor. If the water level and pressure in the steam generator are to remain
constant, the heat inflow and heat outflow must be perfectly balanced. Either one may be held
constant (leading) and the other varied (lagging) to maintain conditions in the steam generator.
This determines the mode of operation of the plant (turbine leading and reactor lagging, or
reactor leading and turbine lagging). Steam generator pressure controls both modes of opera-
tion. In turbine leading or reactor following, the steam pressure determines the required
reactor output by signalling the reactor regulating system to increase or decrease fission heat
release. In reactor leading or turbine following, the steam pressure determines the power
output by signalling the turbine governing system to admit more or less steam to the turbine.
2.1.1 Steam generator pressure control

In the turbine leading mode, the steam generator or boiler pressure control (BPC) system
measures steam pressure and manipulates reactor power by altering the reactor regulator
setpoint. In the reactor leading mode, it measures steam pressure and manipulates steam flow
by adjusting the turbine governor speeder gear.
2.1.2 Steam generator level control

The steam generator or boiler level control (BLC) system measures water level, compares steam
and feedwater flows, and manipulates feedwater flow. In addition, reactor power or turbine
power is measured and the level setpoint ramped up or down accordingly.
2.1.3 Unit power regulator

The unit power regulator (UPR) measures the electrical output and controls this output by
manipulating steam flow, which is done by adjusting the turbine governor speeder gear. This is
only applicable in turbine leading mode.
2.1.4 Reactor regulating system

The reactor regulating system (RRS) measures reactor neutron flux and thermal power and
manipulates the liquid zone control absorbers and adjuster rod positions to control neutron
flux.
2.1.5 Heat transport system (not shown in Figure 2)

The heat transport system pressure and level (HTSP&L) control system measures the reactor
coolant pressure and pressurizer water level and manipulates the feed and bleed system as well
as the pressurizer heaters and water sprays and also the steam release if pressure rise is exces-
sive.
2.1.6 Turbine governing system

The turbine governor senses the turbine speed and adjusts the steam flow to maintain the
required speed. This action is applicable only when the turbine generator is not connected to
the grid system. Before synchronization, the speed is adjusted by the speeder gear to match the
frequency of the grid system exactly. Once synchronized, the turbine generator speed is fixed
by the grid frequency. If the speeder gear is then manipulated, it will open or close the gover-
nor valves to admit more or less steam and hence increase or decrease the turbine load. Once
the load has been set on the turbine, any change in frequency on the grid system will alter the
turbine speed, and the turbine governor will respond accordingly and change the governor
valve position to adapt to the change in grid demand. This will alter the turbine generator
output according to the speed droop setting. Further manipulation of the speeder gear by the
unit power regulator (turbine leading) or the boiler pressure control (reactor leading) will
restore the desired power output.
2.2 Reactor Protection Systems

The reactor has protection systems that prevent excessive operating conditions which could
cause a reactor trip and that trip the reactor if certain operating conditions are exceeded. A
reactor trip is to be avoided if unnecessary because it causes a loss in electric power production
and imposes thermal transients on components. Therefore, provision is made for the reactor
power to be reduced by a setback or stepback to avoid potential trip conditions. Certain
conditions, however, require an immediate reactor trip because continued operation could put
the reactor into a dangerous condition. Such conditions require assurance that the reactor will
trip, and therefore dual shutdown systems are provided, where each system is completely
independent of the other, operates on different principles, and is activated by different inputs.
2.2.1 Reactor setback

A reactor setback reduces reactor power in a ramped manner to a predetermined level. The
rate of power reduction and the end point of the power level depend upon the initiating condi-
tions. For example, a high local neutron flux or flux tilts beyond a certain level will trigger a
setback to 60% or 20% of full load, depending upon severity, to avoid overrating the fuel locally.
A turbine trip or load rejection initiates a setback to 60% of full power because the condenser
can absorb this level of heat load. A high steam generator pressure requires a setback to 10%
because it indicates an inability of the turbine and condenser to absorb the steam flow ade-
quately. Failure to meet other conditions which are essential for safe reactor operation, but
which could be rectified with the reactor on load, requires a setback to 2%.
2.2.2 Reactor stepback

A reactor stepback reduces reactor power immediately to approximately 60% of full power by
initiating a drop of the mechanical control absorbers. It is therefore similar to a reactor trip, but
does not invoke the reactor shutdown systems, although it is a backup to the shutdown sys-
tems. For example, a trip of the heat transport pumps or low steam generator water level will
trigger a stepback because these indicate a loss of heat removal capability. A high rate of
neutron power increase will also initiate a stepback because it indicates a lack of safe reactor
control.
2.3 Reactor Shutdown Systems

The reactor must be protected from the worst type of failure or combinations of failures. A loss
of coolant accident is one such scenario. In the event of loss of coolant from one or more fuel
channels, the voidage will initially cause an increase in reactivity. This must be counteracted by
immediate insertion of a greater amount of negative reactivity to ensure subcritical conditions.
There are two independent shutdown systems, SDS1 and SDS2, which are capable of doing this.
2.3.1 Shut-off rod insertion

Shut-off rods controlled by the SDS1 system drop by gravity into the reactor core to absorb
neutrons. They are made of hollow cadmium rods encased in stainless steel and are guided into
the reactor by vertical guide tubes. Cadmium is a strong neutron absorber and quickly takes the
reactor power to a very low level. Activation occurs by release of electrically energized clutches,
and the initial fall is accelerated by springs.
2.3.2 Liquid poison injection

Liquid poison controlled by the SDS2 system is injected by pressurized helium into the modera-
tor to absorb neutrons. The gadolinium nitrate solution absorbs neutrons strongly, thus effect-
ing a rapid drop in reactor power to a very low level. To ensure adequate distribution of the
liquid poison, it is injected through a pattern of horizontal tubes within the reactor core.
Activation occurs by the release of air pressure which holds the helium release valves in the
closed position.
Both systems are activated within one second of failure detection, and sufficient negative
reactivity is inserted within two seconds. The two systems are entirely independent of one
another, and their operation depends on different principles and mechanisms. Although a loss
of coolant accident is obviously very serious, there are several other scenarios which could put
the reactor at risk, some of which may be transient due to operating disturbances. Several
control parameters therefore feed into the SDS1 and SDS2 systems to shut down the reactor if
these parameters deviate too far from what is deemed to be a safe operating condition. A
dropping steam generator level, for example, will initiate first a stepback, then the SDS1 system,
and finally the SDS2 system, so that if the stepback does not arrest the loss in inventory, the first
shutdown system should, but if it fails, the second shutdown system will respond accordingly.
Quick recovery after an SDS1 trip is possible because the shut-off rod control mechanism can be
reset and the rods withdrawn without residual reactivity effects. However, this is not possible
after an SDS2 shutdown because the liquid poison will have thoroughly mixed with the modera-
tor, and it will take several hours for the moderator purification system to remove the gadolin-
ium. The SDS1 system is therefore the first to be actuated, and if this fails to arrest the fault
condition, the SDS2 system will follow. Therefore, transient conditions or minor problems which
initiate only an SDS1 trip and which can be corrected quickly can be recovered from before the
xenon transient causes the reactor to poison out. Major faults or excessive deviations from
required parameters would result in both systems tripping in quick succession.
2.4 Emergency Coolant Injection

In the event of a loss of coolant accident, an assured supply of coolant must be supplied to the
fuel channels to maintain cooling immediately after the resulting reactor shutdown. This is
effected by the emergency coolant injection system, which injects high pressure water into the
heat transport system. This water is maintained under pressure in tanks by pressurized nitro-
gen. After this has been exhausted and the heat transport system pressure has fallen to a lower
level, a pump-driven low pressure injection system takes over. Ultimately, lost coolant is recov-
ered, circulated through heat removal exchangers, and returned to the heat transport system to
stabilize the situation. Typically, the high pressure system is actuated at a heat transport system
pressure of about 5 MPa to give sufficient margin from normal operating pressure to minimize
spurious injections during normal reactor cooldown.
2.5 Turbine Protection System

Like the reactor, the turbine has protection systems to prevent damage to components or
catastrophic failure. As with the reactor, avoidance of unnecessary tripping is achieved by a
runback to reduce turbine power sufficiently to regain safe operating conditions. In this way,
speed and temperature transients on the turbine can be avoided, as well as the risk of faulty
synchronization. Serious faults, however, require an immediate turbine trip.
2.5.1 Turbine runback

A turbine runback is the reduction of turbine load at a preset rate to zero load or until the
initiating condition has been cleared. The rate is typically in the range of 1% to 10% of full load
per second. The rate may be fast or slow depending upon the condition. A latched runback is
one in which the governor speeder gear returns the turbine to the no-load speed setpoint. An
unlatched runback terminates when the initiating condition has been cleared. As an example, a
turbine trip would cause a fast latched runback to ensure that the governor speeder gear was
reset to its no-load speed condition to avoid rapid reopening of the governor valves when
clearing and resetting the trip (after rectifying the fault). A fast unlatched runback would be
initiated by high condenser pressure (poor vacuum) because the condenser was unable to
handle the exhaust steam flow adequately. This runback would terminate when the pressure
had been restored to an acceptable value, and the turbine load would be maintained at this
level.
2.5.2 Turbine trip system

In the event of a serious problem on the turbine generator, the turbine is tripped. This action
causes a fast runback to the no-load set point on the governor and releases hydraulic pressure
from the governor and stop valve actuators. This enables the valves to close rapidly and shut
off steam flow to the turbine. Signals causing a trip come from various sources such as exces-
sive rotor vibration, high bearing temperature, a generator electrical fault, and other parame-
ters which have the potential to damage the turbine.
2.5.3 Overspeed governor

A major concern with regard to the turbine generator is the risk of overspeeding. This can
happen only when the turbine is disconnected from the electrical grid, which means that critical
times are before synchronization and after a turbine generator trip or load rejection. Some
degree of overspeeding occurs with a generator trip or a load rejection when the steam valves
are still closing after electrical output has been terminated. To provide protection against
excessive overspeed, an overspeed governor is mounted on the turbine shaft. This is activated
by increasing centrifugal force on overspeed bolts which independently trigger hydraulic pres-
sure release from the steam valve actuators to shut off the steam flow.
3 Nuclear Reactor Operation

3.1 Reactivity Characteristics
To maintain the neutron multiplication factor k equal to unity in an operating reactor, small
adjustments must be made continually to one of the factors in the six-factor formula given as
follows:
k =  p  f f t, (1)
where  is the fast fission factor, p the resonance escape probability,  the reproduction factor, f
the thermal utilization factor, f the fast neutron non-leakage probability, and t the thermal
neutron non-leakage probability. The manipulated factor is usually the factor f, the thermal
utilization factor, which includes the neutron absorption in the reactor, a. This affects the
number of neutrons and hence the neutron flux and reactor power.
Control systems must be designed to maintain the desired neutron power following short term
changes initiated by natural perturbations or imposed transients. In the longer term, changes
occur due to buildup of neutron absorbers and burnup of neutron producers in the reactor core.
Because these changes are slow, they do not affect the control system as such, but the overall
reactor configuration must be adjusted to maintain the desired balanced condition for a steady
neutron chain reaction. The control system then maintains an equilibrium about this balanced
condition.
An important consideration is the extremely high velocity of neutrons (2200 m/s when thermal-
ized at 20°C). With the close spacing of fuel elements in the reactor, neutrons do not travel very
far. The overall neutron lifetime from production due to fission until absorption to produce
fission is therefore extremely short. The average neutron lifetime in a heavy water moderated
CANDU reactor is about 1 millisecond, and in a light water moderated reactor such as a PWR or
a BWR, it is much less. The complete neutron cycle from one generation to the next takes only
this amount of time. It is evident, therefore, that any small deviation from the equilibrium
situation, in which the number of neutrons in one generation is equal to that of the previous
generation, will very rapidly grow in magnitude.
If all neutrons had a lifetime of about 1 millisecond, it would be almost impossible to design a
control system that would be able to sense changes and effect control before the neutron
population grew or shrank out of the control range. Fortuitously, some neutrons are produced a
short time after the actual fission process. The kinetic behaviour of a reactor is critically de-
pendent upon the existence of these delayed neutrons because they have the effect of increas-
ing the average lifetime of the neutrons arising from fission. This increase in average lifetime to
about 1 second because fewer than 1% of the neutrons have much longer effective lifetimes
enables control systems to maintain stable operation provided that certain limits are not
exceeded.
3.2 Source Multiplication

The delayed neutrons also influence the reactor under shutdown conditions because they are
produced by the decay of certain precursors. While the reactor is shut down, this decay process
continues, and some neutrons are always present because the decay curve is asymptotic. Both
uranium-235 and uranium-238 are naturally unstable and decay mainly by α-particle emission.
Both also fission spontaneously, giving off neutrons in the process. Any fuel, even unused fuel,
therefore produces small numbers of neutrons. Furthermore, in heavy water moderated
reactors such as the CANDU, neutrons can be created by interaction of high energy γ-rays from
certain fission products with deuterium atoms. These are known as photoneutrons. If, after a
very long shutdown, these natural sources of neutrons do not produce sufficient neutrons to be
detected by the reactor instrumentation, then artificial neutron sources are inserted into the
reactor.
These source neutrons can cause fission in fissile fuel, producing more neutrons and establish-
ing a chain reaction. When the reactor is shut down, the neutron multiplication factor k is,
however, less than unity. This means that the chain reaction will decay, but in the meantime,
other source neutrons will start new chain reactions. The result is that, along with these
neutrons, there will be some fission neutrons, so that the total number of neutrons will be
greater than would have arisen from the neutron sources only. The factor by which the total
number of neutrons S∞ is greater than the number of source neutrons S0 is known as the
subcritical multiplication factor. The relationship between these and the value of k is given by
the following formula:
S∞ = S0 / (1- k) (2)
From the above equation, it can be seen that, as k approaches unity, the measured number of
neutrons S∞ becomes many times greater than the source number of neutrons S0. Further-
more, it takes longer for an equilibrium condition to be reached. A very simple numerical
example for k = 0.5 and S0 =100 is given in Figure 3 for illustrative purposes.
As long as S∞ stabilizes at a fixed value after an increase in k, the reactor remains in a subcritical
condition, with k less than unity. When k is exactly unity, however, the fraction of source
neutrons is negligible compared with the total number of neutrons, and the number of neu-
trons in each generation will be the same. The system is then said to be critical.
Figure 3 Source multiplication by subcritical multiplication factor
3.3 Approach to Criticality

When a nuclear reactor is started up from the shutdown condition, the neutron flux must be
increased by several orders of magnitude. In the shutdown condition, the neutron flux is
determined by the intensity of the neutron sources and the subcritical multiplication factor. The
value of the neutron multiplication factor k is substantially less than unity, and the reactor is
subcritical. When starting up the reactor, the value of k is adjusted to bring it to unity, making
the reactor critical and capable of sustaining a continuous fission chain reaction. In so doing,
the subcritical multiplication factor increases the neutron flux level because the difference
between k and unity, Δk, becomes smaller and smaller:
S∞ = S0 / (1 - k), (3)
S∞ = - S0 / Δk. (4)
As Δk approaches zero, the measured neutron flux S∞ theoretically goes to infinity. This is
obviously not practical because critical conditions must be established at a measurable neutron
flux level. The real situation is that Δk can only be determined and controlled to a certain
degree of accuracy. Once this limit has been reached, slight perturbations cause control system
responses that mask the actual value of Δk, and the reactor can be considered to be oscillating
very slightly above and below criticality. Under these conditions, the reactor is considered to be
critical, with the average value of k equal to unity.
Even though a nuclear reactor may be shut down, the reactivity within the reactor may change
with time due to buildup or decay of fission products. This means that one cannot simply
restart a reactor by inserting an amount of positive reactivity equal to the amount of negative
reactivity inserted to shut it down. Approach to critical is therefore a delicate manoeuvre, like
an approach to a moving and invisible target. At every step, therefore, an assessment must be
made of how far from critical the reactor actually is. Each step must also be small enough so as
not to overshoot the point of criticality.
While Δk has a large negative value, a change in reactivity will be reflected quickly on the
instruments because the fission chain reactions initiated by the source neutrons die away
rapidly since the value of k is far below unity. Following insertion of some positive reactivity,
the reactor power will rise quickly to a new equilibrium value. When, however, Δk has a small
negative value and k is very close to unity, the reactor will respond very much more slowly.
Under these circumstances, the fission chain reactions initiated by the source neutrons persist
for many generations. Following insertion of some positive reactivity, the reactor power will
rise slowly, but more significantly, and take much longer to settle at its new equilibrium value.
When the reactor becomes critical due to insertion of additional positive reactivity, the reactor
power will rise steadily and continuously without levelling off at an equilibrium value. Because
the reactor will be very slightly supercritical, the power will in fact rise exponentially. If power
versus time is plotted as shown in Figure 4, criticality will have been achieved when the meas-
ured power does not level off, but increases exponentially. Generally, such a crude mechanism
of monitoring the approach to criticality is not suitable for commercial reactors.
Figure 4 Stabilization time for different subcritical multiplication factors
If it is assumed that the measured reactor power P∞ (equivalent to S∞) doubles at each step,
then various parameters such as k and Δk can be calculated to show their changes. Figure 5
shows how k and Δk change with each doubling of power from P∞ = 0.0001 at k = 0.9.
Figure 5 Approach to critical - subcritical multiplication factor change

The neutron count rate on the instruments can be assumed to be proportional to the measured
neutron flux S∞ and the inverse count rate then determined. These changes may be plotted
graphically. The most useful plot is the inverse of count rate versus reactivity added, as shown
in Figure 6. This shows that criticality is achieved when the inverse of count rate becomes zero.
The plot is nearly linear, enabling the amount of reactivity to be added to achieve criticality to
be easily determined by extrapolation.
Once the reactor is confirmed to be critical and under control of the reactor regulating system,
a specific power level can be set, and the reactor regulating system will reduce the liquid zone
levels slightly to insert a little more reactivity and maintain a stable rise in power to the desired
level. Typically, the reactor regulating system never needs more than about 0.1 mk of reactivity
change to maintain stable conditions or to change the power level.
Figure 6 Approach to critical - inverse count rate change
3.4 Approach Technique

The actual process of approach to criticality depends upon how long the reactor has been shut
down. After a period of several weeks, the source of neutrons from delayed neutrons and
photoneutrons will have fallen to a sufficiently low level so that, even with the subcritical
multiplication factor, the reactor regulating system instruments will have drifted off scale and
will not be able to measure the neutron flux. Supplementary start-up instrumentation is
therefore used as a temporary measure until the regular instrumentation can read the neutron
flux. During power raising under these conditions, manual manipulation of reactivity and hence
power is required. This requires careful monitoring and plotting of inverse count rate versus
reactivity. Once the power is within the range of the regular instruments, they can be relied
upon to hold the power at a particular level. The time for power to stabilize becomes longer as
the reactor moves closer to criticality, and therefore using the reactor regulating system to
maintain a particular power level minimizes the risk of adding more reactivity before the power
has stabilized. If the reactor is started up after a short shutdown period, the reactor regulating
system will be able to sense the neutron flux and react accordingly.
Assuming that the regular instruments can be used for reactor control, the general procedure is
as described below. If not, the same procedure applies, but must be done manually. The
reactor regulating system can be used with the reactor in reactor leading mode of operation.
With poison removal stopped, the operator requests a doubling of power and monitors the
liquid zone control absorber levels. When these have reached about the 20% level, power
raising is stopped and the power held at this level by the reactor regulating system. The poison
removal process is then started or restarted with the RRS holding the power level. As the
reactivity tends to increase, the liquid zone levels rise to compensate, keeping reactivity and
hence power level constant. When the liquid level is high, the poison removal process is
stopped. Again the operator requests a doubling of power, and the process is repeated. Be-
cause smaller and smaller amounts of reactivity addition are required as criticality is ap-
proached, the drop in liquid level becomes smaller and smaller, and after a few iterations, the
liquid zone absorber levels can accommodate a full doubling of power. In a typical CANDU
reactor, a drop in liquid level from 65% to 20% represents a reactivity change of about 3 mk. If
the power were doubled with this change, the reactor would be about 6 mk below criticality.
Further iterations of the process can bring the reactor to within 1 mk of criticality. The final step
in the process is verification that the reactor is sufficiently close to criticality for the RRS to
maintain control of power. Power raising can then follow by the operator adjusting the power
setpoint to the required value.
The advantage of this method of approach to criticality is that the reactor does not actually go
critical and, when it is nearly critical, power is increased by adjusting power level and not by
direct manipulation of reactivity devices.
3.5 Reactor Start-up

A simplified procedure for a reactor start-up is shown in Figure 7. Besides the important proc-
ess of achieving criticality, all auxiliary systems must be available and in operation as required,
and the heat transport system must be pressurized and heated. Auxiliary systems such as the
feed and bleed system, moderator cooling system, shutdown cooling system, and emergency
core cooling system must be available as required. The emergency core cooling system is
disabled during a cold depressurized state to prevent inadvertent initiation in this plant state.
The cooling systems, however, will be in operation to remove decay heat under cold shutdown
conditions. There is naturally some overlap in these activities, as well as with those of the
steam turbine and the steam and feedwater heating systems (which are described later).
Warming up of the pressurizer and degasser can begin before the approach to criticality.
Initially, the pressurizer is isolated from the heat transport system and warmed until it reaches a
pressure of 4 Mpa while the degasser is warmed to a pressure of 1.1 MPa. At this point, ap-
proval is required to proceed with connecting the pressurizer to the heat transport system
because plant conditions will be changed during pressurization of the heat transport system.
The pressurizer and heat transport system are then pressurized separately to 6.7 MPa by the
pressurizer heaters and the feed and bleed system respectively. They are then connected
together by opening two motorized isolation valves. Once connected, the main heat transport
pumps can be started, and the heat transport system and pressurizer can be warmed together
as one system by controlling the shutdown cooling system to remove excess heat as required.
Warming of the heat transport system is done using pump heat from the main pumps because
their energy is dissipated in the system. Initially, a significant challenge exists for the operator
to maintain the heat transport system temperature below 100°C until the emergency core
cooling system has been enabled. Normal procedure is to maintain the heat transport system
temperature at less than 80°C. Regulations require that the emergency core cooling system be
enabled anytime that the heat transport system temperature rises above 100°C. With the heat
transport system temperature at 80°C, the approach to critical would typically commence, along
with further warming of the heat transport system.
During the approach to criticality, initiation of reactor power increase by removing moderator
poison and continued warming of the heat transport system proceed as parallel activities. The
concern about warming the heat transport system during approach to criticality is the effect
that it has on core reactivity when reactor power is being increased and criticality has not been
achieved. During the approach to criticality, only one reactivity addition mechanism should be
in use at any one time. The primary method is moderator purification by removing poison.
However, by adding pump heat from the heat transport system pumps, the reactivity is being
adjusted due to the temperature coefficients, including those of the fuel, at play during warm-
up. This reactivity effect must be considered regardless of whether an equilibrium core or a
fresh core is in place. As the heat transport system is warmed, the temperature coefficients add
overall positive reactivity to an equilibrium core due to the buildup of Pu-239 which fissions
more readily as temperature rises within this range, whereas in a fresh core, this reactivity
effect is negative due to the lack of Pu-239. Therefore, warm-up of the heat transport system is
stopped when the core is 30 mk subcritical, which is approximately equivalent to the presence
of 1 ppm of moderator poison (gadolinium nitrate). This concentration can be confirmed by
monitoring the online moderator conductivity, which is proportional to the concentration. A
subcritical reactivity balance is conducted at 30 mk to verify the prediction for criticality and to
ensure that all reactivity factors have been accounted for. For the remaining approach to
criticality, no other actions which adjust reactivity other than that of the operator removing
moderator poison by moderator purification are permitted. This guarantees that the core
achieves criticality as predicted. In practice, criticality is defined as the point where the reactor
regulating system has sufficient positive and negative reactive addition available by means of
the liquid zone control absorbers to maintain reactor power at its setpoint. This occurs in an
equilibrium core when the reactor regulating system can double reactor power within a 10%
change in liquid zone level.
Figure 7 Major activities during start-up and loading
3.6 Reactivity Changes

The previous section considered short term reactivity changes due to operational manoeuvring,
including start-up. This requires relatively small changes in reactivity to effect the desired
result. In fact, it is important that, except in the case of a reactor trip, such changes be small to
ensure proper control of the reactor.
Of further concern, however, are certain other effects which, during normal operation, result in
reactivity changes much greater in magnitude than are required for operational control pur-
poses. These occur over a much longer time frame and as such do not upset normal control.
They do, however, require significant reactivity adjustment to maintain criticality in the reactor.
These effects change one or more factors in the six-factor formula, given again as follows:
k =  p  f f t. (5)
The result is a drift of k from unity in a critical reactor. To restore criticality, that factor must be
returned to its original value or another adjusted to compensate for the change so as to bring k
back to unity.
The effects causing these changes fall into three main categories:
Fuel burnup
Xenon transients
Temperature effects.
The fuel effects are due to fuel depletion, buildup of transuranic nuclides, and production of
fission products. Eventually, the combination of these factors prevents criticality from being
maintained, and refuelling with new fuel is required.
Xenon-135 in particular is troublesome because its equilibrium level in the reactor is maintained
by its being destroyed by neutron absorption as fast as it is created. A reduction in reactor
power decreases the neutron flux and hence its rate of destruction. This results in a surge of
xenon production and an increase in neutron absorption. The consequent reactivity changes
may be difficult for the control system to overcome, especially after a significant power reduc-
tion.
The temperature effects are related to the neutron absorption characteristics of the fuel and
moderator. The temperature of the whole reactor increases from the cold shutdown condition
to the hot zero load condition. The temperature of the fuel then increases further as power is
raised from the zero load condition to the full load condition. It is desirable that the overall
effect on reactivity due to temperature be negative so that the natural tendency of the reactor
is to go subcritical (k less than unity) during a temperature excursion. This enhances the stabil-
ity of the reactor, giving it a degree of self-regulation and hence increased safety.
3.7 Fuel Burnup

During operation, fissile fuel such as uranium-235 burns up continuously and eventually reaches
a point where it is so depleted that it cannot sustain a nuclear chain reaction. At this time, the
fuel is replaced with new fuel and the process continued. As the uranium-235 is burned up,
fission products build up in the fuel, and some of these are neutron absorbers. On the other
hand, some uranium-238 is converted into plutonium-239 by neutron absorption and two
successive β-particle emissions:
Plutonium-239 builds up quite rapidly at first. Some of it absorbs neutrons in two stages to
become plutonium-241, another fissile fuel, but the concentration of this never reaches signifi-
cant proportions within the normal fuel lifetime. Initially, the buildup of plutonium-239 actually
causes an increase in core reactivity. This is due to its higher fission cross section as well as the
greater number of neutrons produced per fission compared with uranium-235. Ultimately,
depletion of uranium-235 and buildup of fission products in the reactor result in a decrease in
core reactivity.
The effect of fuel burnup on reactor reactivity is felt primarily through the factor η, which
combines the number of neutrons emitted per neutron absorbed ν, the macroscopic fission
cross section of the fuel, Σf fuel, and the macroscopic absorption cross section of the fuel, Σa fuel:
η = v Σf fuel / Σa fuel. (6)
As the fuel is burned up, Σf fuel decreases, and as fission products accumulate, Σa fuel increases.
The overall effect is that η decreases with time. The effects of samarium-149 and plutonium-
239 as mentioned above, together with uranium-235 depletion and fission product buildup, can
be shown graphically as in Figure 8 by plotting the value of  with respect to time and noting
that the neutron multiplication factor k varies accordingly. At the very beginning, there is a drop
in η (and hence k) due to buildup of samarium-149, a neutron absorber. Somewhat later, but
still early in the fuel cycle, the value of η (and k) increases and reaches a peak due to formation
of plutonium-239. Then the effects of uranium-235 depletion and fission product buildup
become significant, and η (and k) falls progressively.
The effects of xenon-135 are not shown on this graph because the associated degree of reactiv-
ity change is due mainly to reactor load (neutron flux) and is greater in magnitude than the
effects shown here, being as much as -30 mk under full load conditions.
Figure 8 Effects of fuel burnup on reactivity
3.8 Refuelling Considerations

Except when a reactor is loaded with fresh fuel, reactor refuelling is a continuous process, with
about a dozen channels being refuelled each week. In selecting the channels to be refuelled, a
number of factors should be considered. Channels with the highest burnup should be refuelled
first, the reactor power distribution should be kept symmetrical, and the channels with the
largest reactivity gain should be selected if the overall core reactivity is low. However, channels
close to recently refuelled channels should be avoided, as well as those with known abnormal
conditions which could force an unplanned reactor shutdown if a refuelling failure should occur.
Any channels containing defective fuel should be refuelled as soon as possible.
The preferred state during refuelling is with the reactor critical, on load, and under control of
the reactor regulating system. Control absorbers must not be in the core because this is an
abnormal and transient operating condition. There should be no reactivity changes during
refuelling, meaning that adjuster positions should not be altered nor poison addition or removal
be in progress. This ensures that reactivity changes during refuelling are readily observable.
4 Heat Transport System

4.1 Operational Considerations
When at full power, the reactor derives approximately 92% of its heat from the fission process,
7% from decay of fission products, and 1% from the heat transport system pumps. Immediately
after a reactor trip, the heat generated by fission products, even though it decreases rapidly
initially, and that generated by the pumps continues. The heat transport system must remove
heat at all times through the steam generators. This requires that a temperature difference
between the reactor coolant and the steam circuit be maintained. At full load, this temperature
difference is large, but at low load and under shutdown conditions, it must be small enough to
control the rate of heat discharge, because too high a rate would impose excessive cooling and
large thermal transients. Steam generator pressure is the main temperature control parameter.
During warm-up from the cold condition, pump heat input and residual decay heat, even after a
long cooldown, are sufficient to bring the system up to the hot standby condition. During
shutdown, with a higher decay heat load, steam is discharged from the steam generators under
controlled conditions to bring down the pressure and temperature slowly enough to avoid
excessive thermal stress on large components. This steam is usually discharged to the con-
denser through the condenser steam discharge valves. These valves open progressively, but
because the steam specific volume increases with decreasing pressure, they eventually reach
their fully open condition, and the rate of cooling decreases. Beyond this point, the shutdown
cooling system can control the cooling rate. During this cooldown period, one concern is
delayed hydride cracking of the zirconium alloy pressure tubes at an intermediate temperature
in the range of about 100°C to 200°C, and therefore the system must not be allowed to remain
within this temperature range for longer than necessary.
In the event that the condensers are unavailable, steam can be discharged to the atmosphere
through the steam reject valves, although this results in loss of treated water. If the steam
generators are unavailable or the pressure control system faulty, the heat transport system can
be cooled directly by the shutdown cooling system, which has sufficient capacity to handle the
decay heat. It can also maintain sufficient flow in the fuel channels under shutdown conditions
to remove the heat generated should the heat transport pumps not be available.
The inventory and flow of the heat transport system must naturally be maintained for effective
removal of heat from the fuel. Possible failures of various valves which admit or release fluid to
and from the system must be considered and allowance made for the system to accommodate
such failures without jeopardizing safety. One situation which initially gives a wrong indication
is a failed-open pressurizer steam vent valve. This represents a loss of inventory, but the de-
creased pressure causes boiling in the pressurizer and a consequent increase in level, which
indicates excessive inventory. All losses of inventory through valves and fittings can be accom-
modated by the normal control systems, which either compensate for the problem or reduce
reactor power.
An accidental loss of inventory or loss of coolant accident due to a system rupture requires
different actions depending upon its severity. A major rupture could result in loss of coolant
flow to one or more fuel channels. An emergency coolant injection system is provided to supply
water to the reactor under these circumstances. This system, however, operates at a lower
pressure than that of the heat transport system, and therefore rapid depressurization is neces-
sary. This implies reducing temperature and pressure rapidly by discharging steam from the
steam generator through the steam reject valves. This procedure, known as crash cooling or
crash cooldown, imposes severe thermal stress on the components, which would subsequently
require intensive inspection.
The configuration of a CANDU reactor is such that all feeder pipes from the fuel channels are
directed upwards to headers, which in turn are connected to the steam generators at an even
higher level. Hence, any buoyancy effect due to heating or vapour generation will tend to draw
coolant upwards from the reactor to the steam generators, promoting circulation by natural
convection. In the event that the heat transport pumps fail, this provision for thermosyphoning
between the reactor and steam generator is a safety feature of the reactor. The difference in
elevation between the reactor and the steam generators promotes natural circulation, and the
built-in inertia of the pumps, giving a rundown time of between 2 and 3 minutes, enables this
circulation to be established during the initial disturbing transients. This natural circulation is
much slower than the normal forced circulation provided by the heat transport system pumps
and therefore can be used only under low power conditions.
A typical scenario under which thermosyphoning would be required is in the event of electrical
power loss to the heat transport pumps. This would result in an immediate reactor trip. As
reactor power drops to about 5% of full power within about 10 seconds (as a result of decay
heat), the circulating pumps run down slowly due to built-in rotational inertia, and flow de-
creases progressively to the point where thermosyphoning takes over naturally and maintains
flow in the same direction, but at a much reduced flow rate.
5 Steam Generators
5.1 Plant Control
As is evident from the introduction which deals with power output regulation and from Figure
1, the steam generator is a key component in satisfactory plant control. In both modes of
operation, the steam generator pressure is the measured parameter, and the control system
acts accordingly.
The rate of heat transfer Ω is governed by the following equation:
Ω = U A θ, (7)
where U is the overall heat transfer coefficient, A is the surface area through which heat is
transferred, and θ is the temperature difference between the fluids on each side of the heat
exchange surface that exists between the primary side reactor coolant and the secondary side
steam-water circuit. Because U and A do not vary much, it is evident that any change in θ will
affect Ω, and vice versa. Figure 9 shows typical conditions of the reactor coolant and steam-
water mixture in the steam generator under full load.
Figure 9 Steam generator temperature profile

The steam pressure affects the saturation temperature of the secondary side steam-water
mixture, so the latter can be manipulated by adjusting the steam generator pressure. This in
turn will affect the primary side reactor coolant temperature and is an important control
parameter in reactor warm-up and cooldown procedures. Under these conditions, the heat
transfer rate Ω is very low, although it may be somewhat higher following a trip from a high load
condition due to a larger decay heat load.
When supplying power to the grid system under a set steam generator pressure, as load and
hence Ω is increased or decreased, the primary side reactor coolant temperature will rise or fall
as θ changes.
Steam generator pressure is therefore a key parameter under all operational conditions.
5.2 Steam Generator Pressure Control

Figure 2 shows the basic plant control system. It may be seen that the steam generator pressure
is controlled by the boiler pressure control (BPC) system, which senses the steam generator
pressure. In the reactor leading mode of operation, the BPC signals the turbine governing
system to adjust the governor valves (steam flow) to the turbine to maintain the set pressure.
In the turbine leading mode of operation, the BPC signals the reactor regulating system to
adjust the neutron flux (heat output) of the reactor to maintain the set pressure.
Under abnormal conditions when there is a mismatch between reactor heat production and
turbine power output, such as would occur immediately after an electrical load rejection or
turbine trip, boiler pressure is controlled by the condenser steam discharge valves (CSDVs in
Figure 16) to be described later in the text. There is also provision for steam discharge directly
to the atmosphere through the main steam safety valves and the atmospheric steam discharge
valves (MSSVs and ASDVs in Figure 16) to avoid excessive pressure in the steam generator.
These valves are a backup in the event that the CSDVs malfunction or are not available due to
condenser restrictions.
5.3 Swelling and Shrinking

It is important to maintain an adequate water level in the steam generators to ensure proper
heat transfer and heat removal capability. A further reason to maintain the water level within
certain limits is to maintain efficient separation of steam from water in the cyclonic separators.
Too high a level can cause flooding of the separation region, and too low a level can result in
surging within that region. The water level, however, is affected by the operational conditions
within the steam generator and can change markedly without any change in the total water
inventory due to the phenomena of swelling and shrinking.
When steam is generated in the steam generator, vapour bubbles are formed on the surfaces of
the tubes and rise with the circulating water. These vapour bubbles occupy more space than
the water from which they were formed, resulting in an apparent swelling of the total water
inventory in the steam generator. The effect of water displaced by the vapour bubbles is seen
as an increase in level in the system. This is commonly known as swelling, and the converse as
shrinking. There are steady state and transient effects with regard to swelling and shrinking.
5.3.1 Steady state swelling and shrinking

During steady state operation, the system is in a state of thermal equilibrium, that is, the rate of
heat removal from the system by the turbine is equal to the rate of heat input to the system by
the reactor. At zero load, no vapour bubbles are formed, and the water level is at a certain
point. As load increases, vapour bubbles are formed. These displace water and cause a rise in
level, as shown in Figure 10. For each load, there is an appropriate water level in the steam
generator corresponding to the number of vapour bubbles in the water. At full load, the water
has the highest concentration of vapour bubbles, and the water level is at its highest. This
analysis assumes that there has been no attempt to control the water level and that the feed-
water flow has matched the steam flow throughout the load change. On a reduction in load,
the water level would return to its original value. This increase and decrease in level are respec-
tively known as steady state swelling and steady state shrinking.
Up to this point in the analysis, it has been assumed that steam generator pressure remains
constant under all conditions. This is not necessarily the case. Certain load manoeuvres on the
turbine and operation of certain steam valves will cause pressure changes in the steam lines and
hence in the steam generator itself. Thus, transient conditions known as transient swelling and
transient shrinking arise.
Figure 10 Steady state swelling due to different load conditions
5.3.2 Transient swelling and shrinking

At a particular turbine load there is to a certain rate of steam generation and a certain number
of vapour bubbles in suspension in the water. In the event of an increase in power demand, the
turbine will require more steam. The increased opening of the governor valves draws additional
steam from the steam line and steam generator, causing a temporary pressure drop in the
system. The drop in pressure in the steam generator allows the vapour bubbles to expand and
increase their volume. In addition, because the new saturation temperature is below that of
the water in the steam generator, some water flashes to steam, creating even more vapour
volume. The combined effect is reflected as a swelling of the steam generator contents and a
rise in water level, as shown in Figure 11. Once the reactor responds to this increased demand
for power and thermal equilibrium has been restored, the pressure in the steam generator will
return to its set value, and the transient swelling will subside.
Figure 11 Transient swelling due to changing load conditions

In the event of a decrease in turbine power demand, the effect will be a temporary rise in steam
generator pressure. This will be followed by compression of the vapour bubbles and shrinking
of the steam generator contents until thermal equilibrium is restored.
A rapid increase in reactor power due to certain manoeuvres, such as recovering from a reactor
trip to prevent poisoning out, can also produce transient swelling. The increased power in-
creases the rate of steam generation and vapour production in the steam generator. This causes
a temporary swelling effect until the turbine load is matched and thermal equilibrium restored.
A rapid decrease in reactor power, for example following a reactor trip, will reduce vapour
production, causing a temporary shrinking effect until thermal equilibrium is restored.
Note that transient swelling and shrinking occur only during rapid load changes. During normal
load changes, only steady state effects will be evident, and these will be related to the particular
power level. Transient swelling and shrinking effects, when they occur, are superimposed on the
steady state swelling and shrinking.
5.4 Steam Generator Level Control

Whenever there is a free liquid surface in a system, the system inventory must be controlled to
ensure that the level remains within limits. In the steam generator, this is done by measuring
the level and controlling the feedwater flow rate. Any deviation of the actual level from the
desired or set level results in a correction to the flow rate.
Swelling and shrinking have an influence on level control in that, for a given water inventory in
the steam generator, the level will rise and fall as the load increases or decreases. Alternatively,
for a given water level, the inventory will go down or go up as the load increases or decreases.
Both situations can lead to difficulties in maintaining desired conditions in the steam generator,
particularly during rapid changes in load when transient swelling and shrinking effects are
superimposed on the steady state effects.
Consider first the situation in which the water level is to be kept constant, as shown in Figure
12. For fixed level control, a certain margin is required to allow for transient swelling and
shrinking because, during rapid load changes, the level control system cannot respond quickly
enough to maintain the level. At low load, swelling will occur as load increases, and at high
load, shrinking will occur as load decreases. The steam generator must therefore be large
enough to accommodate these fluctuations in volume and hence in level. This is done by
increasing the diameter of the upper portion and making it tall enough so that there is a margin
above and below the desired water level.
Figure 12 Constant steam generator water level
If, however, the water inventory is maintained at a fixed value and the set level is allowed to rise
and fall with load, then this range of level can be used to accommodate transient swelling and
shrinking partially, as shown in Figure 13. Transient swelling will occur only from a low load, and
hence the required margin must be above this level. Transient shrinking likewise will occur only
from a high load and requires a margin below the level at high load. If the set level at high load
is already above the set level at zero load, the margins will overlap, as shown in Figure 13. This
enables the vessel to be made smaller while still providing appropriate margins for transient
conditions. To achieve this, the desired or set level must be increased with load, resulting in so-
called ramped level control. This is a more natural method of control because the total inven-
tory remains nearly constant and the control system does not have to drive towards a new
inventory every time the load changes.
Figure 13 Ramped steam generator water level

Transient swelling and shrinking have influenced the design of level control systems in all types
of steam generators and boiler drums. A sudden increase in turbine load will increase steam
flow and reduce pressure in the vessel. This will result in transient swelling and a high water
level. A controller that responds to water level only will reduce feedwater flow to bring down
the water level just at a time when in fact more water is required. After the transient, the
steam generator can be left with a severe shortage in inventory. Such a single element control-
ler tends to respond the wrong way during transients. The problem may be overcome by using
a three element controller which measures steam flow and feedwater flow as well as water
level, as shown in Figure 14. This figure is an expansion of the boiler level control (BLC) system
shown as part of the overall plant control system in Figure 14, which shows just the inputs and
outputs of the boiler level control system. For the case given above, it would sense high steam
flow, compare this with feedwater flow, and send a flow error signal to indicate a required
increase in feedwater flow. At the same time, the transient swelling would send a level error
signal to indicate that a decrease in feedwater flow was required. Initially, the two signals
would cancel one another until the transient swelling had subsided, and then the feedwater
regulating valve would open to balance the steam and feedwater flows to restore the water
level to the setpoint. The level setpoint would in the meantime have been raised by the level
control program to correspond to the new turbine load. The use of a three element level
control system in conjunction with a ramped level setpoint results in relatively smooth control
of water level during all types of transients.
Figure 14 Three element steam generator level control system

(courtesy of NB Power)
Figure 15 shows how the level setpoint is ramped up with load. The initial steep slope roughly
matches the degree of steady state swelling. This is desirable because the low ratio of through-
put to inventory makes the control system take longer to adjust the inventory to correct level
errors. Inventory adjustment is faster at higher throughputs, and therefore the level ramp slope
can be less steep, as shown in Figure 15. This lesser slope indicates that in fact the steam
generator inventory is decreased as load increases in the higher load range. This is desirable
with regard to the efficiency of the cyclonic steam separators, which do not operate at optimum
efficiency at very high or very low water levels.
Figure 15 also shows how the low level alarm and reactor stepback as well as the SDS1 and
SDS2 trips are also ramped with reactor power. The latter slopes are steep to match the degree
of swelling or shrinking closely to ensure adequate inventory after a reactor trip from any power
level.
Figure 15 Steam generator level alarms and set points (courtesy of NB Power)
6 Steam System
6.1 Characteristics and Function
The main purpose of the steam system is to convey steam from the steam generators to the
steam turbine. In a nuclear plant, the steam generators also serve as a heat sink for the reactor
under all operating conditions, and because steam is produced, the steam system becomes the
major part of the reactor heat rejection system. Provision must therefore be made to handle
steam in large quantities even when not required by the steam turbine.
Steam pressure control in the steam generators is important with regard to their function as a
heat sink. As steam pressure varies, so does the prevailing saturation temperature. Because
the rate of heat transfer depends upon the difference in temperature between the primary
coolant and secondary steam systems, a variation in steam temperature will affect the overall
capacity for heat removal. Assuming that the primary coolant pressure is fixed and controlled
by the pressurizer, thus limiting the maximum primary coolant temperature, an excessive steam
pressure will limit the heat rejection capabilities of the steam generator. However, excessive
steam pressures, which would likely arise from insufficient steam discharge, would be pre-
vented by steam discharge through the main steam safety valves. Conversely, an excessively
low steam pressure and corresponding temperature will produce excessive heat flow across the
steam generator and pull down the temperature of the reactor coolant. This is done in a
controlled manner during a reactor cooldown.
For proper control of the reactor state, the steam system must therefore be able to handle the
amount of steam produced under all operating conditions and to control and maintain steam
generator pressure within prescribed limits. The steam system must be designed to handle
decay heat generation following a reactor shutdown, and it is also desirable that it be able to
maintain a high heat discharge rate to prevent xenon poisoning during a short-duration un-
planned turbine shutdown.
In the turbine steam cycle, approximately one third of reactor heat is converted into electricity,
with the remaining two thirds rejected to the environment through the turbine condenser. The
rejection path is a convenient way of disposing of heat when the turbine is unavailable. Thus, if
a turbine trip forces a short term shutdown of the turbine generator system, the reactor can be
maintained at approximately two thirds power, with the excess steam production dumped to
the condenser. Two thirds of full power on the reactor is sufficient to maintain the neutron flux
at a high enough level to prevent poisoning out of the reactor and thus permit a restart at any
time.
In the event that the turbine condenser is unavailable, an alternative steam discharge path is
required. This is directly to the atmosphere, but naturally results in an enormous loss of fluid
inventory from the system. For this reason, it is an undesirable mode of operation, but an
essential safety feature.
Figure 16 shows a typical steam system with various control, isolation, and relief valves.
6.2 Atmospheric and Condenser Steam Discharge Valves

The atmospheric steam discharge valves (ASDVs) and condenser steam discharge valves (CSDVs)
are used to control steam generator pressure rather than to protect against overpressure. For
this reason, under certain conditions, the setpoint at which they operate may be lowered to
avoid an excessive pressure transient when their operation is known to be inevitable. This
lowering of the pressure setpoint occurs upon a turbine trip or load rejection and possible
subsequent poison prevention operation. As stated earlier, the atmospheric steam discharge
valves and condenser steam discharge valves divert steam from the turbine and respectively
discharge excess steam to the atmosphere or dump it to the condenser to balance the heat flow
from the steam generators against the heat flow into the steam generators.
Operation of the atmospheric steam discharge valves and condenser steam discharge valves
may occur under operational conditions such as a controlled cooldown of the heat transport
system, a turbine trip or load rejection followed by poison prevent operation, during start-up
conditions when more steam is produced than is required by the turbine, or during any opera-
tion when an excess of steam is produced.
Figure 16 Main steam system (Point Lepreau) (courtesy of NB Power)

Discharge of steam to the atmosphere is of course wasteful because the loss must be made up
with appropriately treated water. In any case, the total quantity of steam discharged is limited
by the available make-up water capacity. If the condenser is unavailable, this becomes the
major path for heat rejection on the steam side. Discharge of steam in this way should be
limited to emergency or unusual circumstances.
6.3 Steam System Operation

One important aspect of the steam pipework is valve operation. In the event of a fault on the
turbine, generator, or electrical system, the steam supply must be isolated very quickly, usually
in a fraction of a second, and therefore the stop valves must operate quickly and reliably.
Another aspect of steam system operation is the possibility of pipe vibration due to sudden
changes in steam flow or flow disturbances due to bends and fittings. Steam pipes must be
suspended on flexible supports to allow for small changes in length due to thermal expansion.
These would permit vibration in the absence of proper restraints and dampers, and therefore
their design is an important aspect of steam pipework. Thermal effects are not confined to
thermal expansion. To withstand high steam pressures, pipe walls must be thick enough to
ensure a safe working stress. When a pipe is heated from a cold condition with hot steam, the
inside wall temperature rises faster than the outside wall temperature. This causes the inside of
the pipe to try to expand while restrained by the outside. The inside is therefore subject to a
compressive stress and the outside to a tensile stress (in the absence of stress due to internal
pressure). If subjected to internal pressure at the same time, parts of the pipe may be sub-
jected to stresses greater than their permissible working stress. The result may be deformation
of or damage to the pipe. This can be avoided by limiting the heating rate of the pipe, and all
systems therefore have provision for slow heating by admitting a small quantity of steam before
subjecting the pipe to full steam temperature and pressure.
During steam line heating, considerable quantities of steam condense on the inside surfaces of
the pipes and must be removed through drains at the low points of the system. The drains are
open during pipe warming to ensure that collected water is properly drained away and are then
closed during normal operation. If not properly drained, water could be entrained with the
flowing steam and carried into the turbine, where it could cause severe thermal shock or impact
damage to the turbine blading.
7 Main Condenser
7.1 Thermodynamics and Heat Transfer
From a thermodynamic point of view, the steam-water circuit receives heat at a high tempera-
ture in the boiler and rejects heat at a low temperature in the condenser. The greater the
difference in temperature, the greater will be the drop in enthalpy and the more work produced
per kilogram of steam. If the initial steam conditions remain unchanged, it is evident that a
reduction in the exhaust temperature will increase the work done in the turbine. This will
ultimately improve overall station efficiency. Exhaust conditions are therefore important and
have a direct impact on plant performance.
Exhaust steam temperature is determined largely by cooling water temperature. The heat flow
per unit time Ω from the exhaust steam to the cooling water is governed by the following
equation:
Ω = U A θ, (8)
where U is the overall heat transfer coefficient, A the surface area through which the heat is
transferred, and θ the difference in temperature between the water and the steam. Once the
condenser has been designed, A is fixed, and, for given flow rates on both steam and water
sides, the value of U is unchanged if the condenser tubes remain clean. The amount of heat
transferred is therefore proportional to θ, the difference in temperature between the steam and
the water.
The cooling water, in passing through the condenser tubes, picks up heat Ω at a certain rate and
increases in temperature ΔT according to the following formula:
Ω = M cp ΔT, (9)
where M is the mass flow rate and cp the specific heat of water. The outlet temperature of the
cooling water is generally some 10C higher than the inlet temperature for the design mass flow
rate. Note that Ω in this formula is the same as Ω in the previous formula because all the heat
transferred across the tubes is absorbed by the cooling water. The result is a temperature
gradient in the water along the tubes from one end of the condenser to the other.
On the steam side, there is no temperature gradient because the saturation temperature and
pressure are interdependent and the pressure everywhere in the condenser is practically the
same. A temperature profile through the condenser illustrating the meaning of θ and ΔT may
therefore be drawn, as shown in Figure 17. Note that θ is the average difference in temperature
between the steam and the water. More correctly, it should be the log mean temperature
difference, but using the average simplifies the analysis.
Figure 17 Condenser temperature profile
Figure 18 helps to visualize the effect of various changes. One effect is a change in inlet cooling
water temperature. This has a direct and proportional effect on steam temperature. Another
effect is a change in turbine load. This in turn changes the rate of heat transferred, Ω, and
hence both θ and ΔT. Some simple numerical examples make it possible to verify these effects.
Generally, the variation in steam temperature is proportional to load.
Figure 18 Changes in condenser temperature profile

The steam pressure in the condenser is governed by the temperature of condensation. The
steam pressure, however, is not proportional to load because the relationship between satura-
tion temperature and pressure is not linear, as indicated in Figure 19. By combining this varia-
tion in steam properties with the temperature variation with load, a condenser performance
curve may be obtained, as shown in Figure 20. Using this figure, the pressure in the condenser
can be predicted for various cooling water conditions. Superimposed on this figure are the
turbine operating limits for a 900 MW nuclear unit.
Figure 19 Variation in condenser pressure
7.2 Turbine Limitations

In most power plants, the condenser pressure is governed by the prevailing ambient tempera-
tures. These pressures range from about 0.004 MPa to about 0.005 MPa, corresponding to a
range of exhaust steam temperatures from 29C to 33C respectively. Steam turbines are
designed for this range of backpressures and operate well within this range. If, however, opera-
tion is extended outside this range, operating problems may arise. These are primarily due to
the change in the specific volume of steam as the pressure deviates from the design value. At
high backpressures (high temperatures), steam specific volume decreases, so that the steam
passes through the last stage blading at lower velocities. Conversely, at low backpressures (low
temperatures), steam specific volume increases, so that the steam passes through the last stage
blading at higher velocities. This may be confirmed by reference to the mass flow rate equa-
tion:
M = ρ V A = V A / v, (10)
where the density ρ is the inverse of the specific volume v. For a given mass flow rate M passing
through a fixed area A, the velocity V increases as the specific volume v increases, and vice-
versa.
Figure 20 Condenser performance curves (adapted courtesy of Eskom)

The turbine blades are designed only for a certain velocity range. Low velocities (due to high
backpressure) result in incorrect blade entry angles and buffeting of the steam on the blades.
This causes vibration and high stress (due to increased density) on the turbine blades and
shortens turbine life. High velocities (due to low backpressure) result in larger kinetic energy
losses in the steam leaving the turbine. This ultimately limits the advantage gained in reducing
the backpressure because the increased leaving losses eventually become greater than the
increased work done in the blades. High velocities (due to low backpressure) may also result in
choking of steam flow due to sonic conditions in the turbine blades. In any case, no additional
work is obtained from the turbine once the flow has choked. Figure 20 shows the upper back-
pressure limits for a typical large turbine.
7.3 Environmental Limitations

Cooling water leaving the plant carries with it some 2/3 of the heat produced by a nuclear
reactor or fossil fired boiler. Most of this is rejected through the condenser and is reflected as a
rise in the temperature of the condenser cooling water. The discharged warm cooling water
subsequently mixes with that in the environment and may affect aquatic life. For this reason,
limits may be imposed on the cooling water temperature rise ΔT and on the effluent tempera-
ture TE.
7.4 Condenser Tube Fouling

Fouling of condenser tubes by material deposition, biological growth, or physical blockage will
reduce condenser performance by reducing either the overall heat transfer coefficient U or the
effective surface area A if some tubes are completely blocked. The following equation illus-
trates these phenomena:
Ω = U A θ. (11)
If either U or A is reduced, then the temperature difference θ between the steam and the
cooling water will have to increase to maintain the same total heat transfer rate Ω. The net
result is an increase in steam temperature and exhaust pressure and a consequent drop in the
work done by the turbine and a loss in plant efficiency.
To avoid fouling, provision is made in most plants to filter the incoming water through screens
and in some plants to dose the water with a biological growth inhibitor. Some fouling is inevita-
ble, and condenser tubes must be cleaned periodically either online or offline.
7.5 Cooling Water Circuit

A typical condenser cooling water system consists of an intake structure with suitable rotating
or travelling screens, two or three pumps in parallel, ducts to and from the condenser, and an
outfall to discharge the water far enough from the intake to prevent external recirculation.
Figure 21 shows diagrammatically a system with three pumps and screens, control valves to
allow any pair of pumps to be in operation, and isolating gates at the intake and outfall to allow
drainage of the system for maintenance. This system supplies three pairs of condensers which
are associated with a turbine having three low pressure cylinders. Each pump can normally
supply 50% of the design water flow. If only a small derating of the unit is necessary when
operating on one pump, then two pumps will suffice, but if a large derating is required, a third
pump is usually installed. The former case applies to plants where cooling water temperatures
during the year are particularly low, whereas the latter case usually applies to plants using
warmer water. The third pump also enables the plant to operate at full power when one pump
or screen is out of commission for maintenance.
Generally, some control of condenser conditions is possible by operating one or two condenser
cooling water pumps. A change in flow will affect exhaust steam temperature and condenser
pressure, as explained in the previous section. At plants operating in areas with particularly
cold cooling water, for example, if condenser pressure becomes too low and choking occurs in
the turbine blades, one pump can be shut down to raise condenser pressure. This in fact is
desirable because it saves auxiliary power and improves overall plant efficiency. At plants
operating in areas with high cooling water temperatures, operating limitations are usually at the
other end of the scale, and the maximum flow of cooling water is required to avoid excessive
condenser pressures.
Figure 21 Typical condenser cooling water system
7.6 Flow Transients

The condenser cooling water system has very long and large pipes carrying water at high
velocities. Such a large mass of moving water has an enormous amount of kinetic energy.
Precautions must therefore be taken when starting and stopping this flow of water. It is impor-
tant that sufficient time be allowed for the water to accelerate or decelerate before initiating
another operation.
The condenser is usually the highest point in the system and operates like a syphon, as shown in
Figure 22. In the event of a pump trip, the water upstream of the condenser is retarded by
gravity, whereas that downstream of the condenser tends to continue flowing under the influ-
ence of gravity. This causes separation of the water column at the condenser due to elevation
and pressure drop in the condenser tubes. The resulting vacuum in this part of the system will
decelerate the flow downstream of the condenser and eventually suck it back towards the
condenser. This reverse flow, on striking the upstream water column near the condenser, will
cause severe shock and damage to the condenser or even destroy its tubes and tube plates.
Such disastrous consequences can be prevented by incorporating vacuum breakers at the
condenser outlet water boxes. These open under the vacuum conditions caused by water
column separation and allow air to enter the system and provide an air cushion to prevent or
minimize reverse flow of the downstream water column.
Figure 22 Simplified condenser cooling water system

During a complete system trip, priming may be lost due to admittance of air into sub-
atmospheric regions. To restore priming, a vacuum priming system is connected to the con-
denser water boxes. This draws air out of the system before start-up and ensures that all
condenser tubes are filled with water and that there are no air pockets in the water boxes which
could persist during normal operation.
7.7 Variation in Condenser Vacuum

An increase in condenser pressure has obvious effects. The enthalpy drop across the turbine is
reduced, and the steam flow through the last stage blades departs from the optimum, thus
reducing their efficiency. Both these phenomena cause a drop in turbine output and a loss in
plant efficiency.
A decrease in condenser pressure has opposite effects on turbine output. A lower pressure
causes an increase in the enthalpy drop across the turbine, thus increasing the turbine output.
It also causes the steam flow pattern to depart from design conditions, thus decreasing the
work done by the turbine. This latter effect is smaller, so that overall there is an increase in
turbine output. Other effects, however, become significant at very low pressures. Choking may
occur in the last stage turbine blades, thus limiting the increase in turbine output. Even if this
does not occur, the increased steam specific volume creates high steam velocities, which
increase the kinetic energy loss in the exhaust steam and increase the friction loss in the steam
passing through the blades. Both losses increase with the square of the velocity and rapidly
overtake the gain due to increased enthalpy drop across the turbine. Beyond a certain point,
therefore, the turbine suffers a net loss in output.
Increased steam moisture content at very low condenser pressures can damage the turbine,
and certain limits may be imposed on operation under such conditions to avoid excessive blade
erosion. Moisture in the steam also affects turbine output because the impact of the moisture
drops retards the blades and slightly reduces turbine output.
The net result of these effects is shown in Figure 23. With rising pressure, turbine output
decreases almost linearly, but with falling pressure, the increase in turbine output is less, and a
maximum is reached. Beyond the maximum, the turbine suffers a decrease in output with
falling pressure.
Figure 23 Effect of change in condenser pressure (courtesy of NB Power)

The effects of increased or decreased condenser pressure on the turbine have been discussed.
Reference to a Mollier diagram will clarify these effects and enable them to be visualized in
conjunction with load changes on the turbine. A change in condenser pressure moves the
exhaust point along the turbine expansion line, whereas a change in turbine load moves the
exhaust point along a constant pressure line, as shown in Figure 24. Such changes affect the
temperature, density, wetness, and enthalpy of the exhaust steam and hence the turbine
output.
7.8 Causes of Poor Condenser Vacuum

Cooling water temperature and turbine load variations will drive the condenser pressure one
way or the other. Such changes are normal, but may cause the condenser pressure to deviate
beyond normal limits. Other causes of rising pressure usually require operator intervention to
restore the condition of the system. Condenser tube fouling, for example, will ultimately
require that the tubes be cleaned to restore conditions. Tube flooding on the steam side, tube
draining on the water side due to air accumulation at high points in the cooling water system,
and accumulation of incondensable gases blanketing the tubes in the steam space are further
causes of poor vacuum which require an analysis of conditions to determine the reason.
Figure 24 Effect of change in turbine load and condenser pressure
7.9 Condenser Vacuum Raising

On start-up, the entire steam space of the condenser as well as those of the turbine and the
separator-reheaters must be evacuated. Initially, the extracted air is of near-atmospheric
density, but eventually it is of exceedingly low density as high vacuum conditions are ap-
proached. As vacuum raising progresses, the mass flow rate drops as the density decreases.
During vacuum raising, therefore, a high flow, low vacuum system is desirable, whereas under
normal operating conditions, a low flow, high vacuum system is required. Note that some air
leakage into the system always occurs during operation.
Air entering the condenser is carried by the steam until the steam condenses on the tubes. The
air cannot flow back against other steam coming in and remains near the tubes at the end of
the steam flow path. Obviously, air must be removed from these parts of the condenser. An air
removal system is therefore required to pick up air from the appropriate places in the con-
denser tube bundle. Air extraction efficiency depends upon the partial pressures of air, pair, and
steam, psteam, in the condenser. The total pressure, ptotal, is the sum of these partial pressures:
ptotal = psteam + pair. (12)
The total pressure is the prevailing pressure in the condenser as dictated by the saturation
conditions. Under these conditions, the partial pressure of air is negligible, and an air extraction
system would extract mainly steam, which is not very effective. If, however, the partial pressure
of air is increased such that the ratio pair/psteam is increased from near zero to a reasonable
value, air removal effectiveness is much increased. This can be done by local cooling within the
condenser. Local cooling to a lower temperature will reduce psteam to a pressure corresponding
to the saturation pressure at the lower temperature. Because ptotal is dictated by the overall
conditions in the condenser, the value of pair must rise to make up the difference. Thus, the
ratio of pair/psteam is increased, and the ratio of air to steam by mass in that area is also in-
creased. The greater the mass of air in the steam, the greater will be the efficiency of air
extraction.
Local cooling in the condenser is effected by shielding sections of tubes in the tube bundle from
the incoming steam flow. The air extraction points are located in these zones where the partial
pressure of air has been increased, as can be seen in Figure 25.
Figure 25 Condenser cross section (courtesy of NB Power)
7.10 Condenser Vacuum Breaking

Condenser vacuum can be relieved on shutdown by condenser vacuum breaking valves which
allow atmospheric air to enter the condenser shell and the steam turbine. Vacuum may be
broken during or after turbine rundown depending upon operating recommendations and
circumstances. When the vacuum breakers are opened, condenser pressure increases rapidly.
If the turbine is still rotating, the denser atmosphere around the rotor resists the turbine blade
motion, causing the rotor to decelerate quickly. These retarding forces are very high and may
stress the blades severely. On the other hand, the rundown time is shortened dramatically,
which may be desirable in the event of a turbine fault such as a bearing failure.
8 Steam Turbine
8.1 General Operational Considerations
Steam turbines generally operate very reliably and for lengthy periods between shutdowns for
major maintenance. During these periods, however, slow performance degradation may occur
due to wear, erosion, or fouling of critical components. This has a slight detrimental effect on
overall plant efficiency, but this translates into a loss in electrical power output and loss in
revenue, which is significant in the long term, especially in a nuclear plant which normally
operates at full load. To optimize plant efficiency, these performance losses must be catego-
rized and monitored with a view to correcting deviations where possible.
Moisture in the steam in the turbine has a detrimental effect, but it tends to decrease as load is
decreased. At very low loads, the exhaust steam may actually become superheated, which is
even worse. Therefore, slightly wet exhaust steam provides some operating flexibility without
changes in exhaust steam temperature.
Turbine back-pressure as determined by condenser conditions is also critical to turbine opera-
tion. Deviations in back-pressure affect steam flow through the last stage turbine blades,
resulting in various undesirable effects. The turbine must therefore operate within prescribed
back-pressure limits, and the condenser must be capable of maintaining these limits.
Steam turbines are also subject to thermal transients, which result in thermal stress and differ-
ential expansion. These require monitoring to ensure that they do not exceed prescribed limits
and endanger turbine integrity.
The heavy turbine and generator rotors run at high speed, and therefore any slight mass imbal-
ance or shaft bending will cause serious and potentially damaging vibration. When a turbine
has been shut down and stationary for a long period, the shafts begin to sag. This sagging is in
addition to the natural deflection under gravity and is the beginning of creep, but is reversible
by slowly rotating the shaft and reversing the sag. This takes time, and, until reversed, the sag
bows the shaft, creating an imbalance. When a hot turbine is shut down after a period of
operation and comes to rest, the natural cooling process results in a temperature difference
between the top and bottom of the casing and rotor. The cooler bottom part of the rotor
shrinks slightly relative to the top, bowing the shaft upwards. This hogging also creates an
imbalance should the turbine be restarted and may in an extreme case jam the rotor against the
casing. This can be avoided by putting the turbine onto turning gear and turning the rotor until
uniformly cool.
Another critical aspect of turbine operation is governing to maintain speed and load. The
turbine is the link between the energy input as steam and the energy output as electricity. Any
mismatch between these will cause the turbine to accelerate or decelerate. Uncontrolled
acceleration is very dangerous, and precautions must be taken to ensure rapid closure of the
steam valves in the event of disconnection from the electrical grid system.
8.2 Turbine Start-up

A simplified turbine start-up procedure is shown in Figure 26. Before turbine runup, several
lengthy processes are required, but these can be done concurrently to some degree. The first
key operation is to start the lubrication system and to put the turbine onto turning gear to roll
out any sag and to ensure uniform heating once steam is admitted. Before any heat can be
received, the condensate system and feedwater system must be put into service, followed by
the condenser cooling water system to remove reject heat. When steam is available, warming
of the steam lines can commence, and the turbine gland sealing system can be put into service.
With sealing steam in place, vacuum raising can commence. When sufficient vacuum is avail-
able, steam can be admitted to the turbine to commence the runup. The turbine is normally
held at a low speed while shaft eccentricity, bearing vibration, etc. are checked and then accel-
erated to pass quickly through the critical rotor speeds so as not to excite excessive rotor
vibration. Full speed must, however, not be reached until full vacuum is available to avoid
putting excessive stress on the long low pressure turbine blades due to windage. When at full
speed, connection to the grid system can be requested. For satisfactory synchronization, the
generator voltage must be the same as that of the grid, the generator frequency, determined by
the turbine speed, must be equal to the grid frequency, and the generator phase angle must
match the grid phase angle. If synchronization is done manually, it is better to err on the side of
the turbine being slightly fast so that it picks up some load immediately as the turbine speed
drops to match that of the grid. This avoids an unnecessary reverse torque on the turbine
generator shafts, and a small load should immediately be applied to prevent fluctuating torque.
Once synchronized, the turbine can be loaded at a rate dependent upon the cylinder tempera-
tures. The initial rate is usually slow to allow metal temperatures to rise slowly to minimize
thermal stress due to steep temperature gradients in the cylinders and rotors. The reheaters
are valved in above about 25% load, and the loading rate can usually be increased after that,
provided that various checks verify that key parameters such as differential expansions are
within limits.
Figure 26 Major activities during start-up and loading
8.3 Turbine Losses

Fluid friction is the most significant of all turbine losses. High velocity steam encounters surface
friction and turbulence in passing through the turbine blading. Fluid friction losses amount to
about 10% of the total energy input to the turbine. This is a major loss factor which is ulti-
mately reflected in turbine internal efficiency.
Moisture loss occurs when condensed moisture passes through the turbine blades and the
moisture drops impinge upon the moving blades. This retards the moving blades, reducing
turbine output and efficiency. There is a loss in stage efficiency of about 1% for each 1% mois-
ture content of steam at that point in the machine.
The residual kinetic energy in the steam leaving each stage is usually recovered in the following
stage. At the last stage, however, the specific volume of the steam is very large, and to obtain
manageable flow areas at the turbine exit, high velocities must be tolerated. This results in a
high residual kinetic energy and a significant leaving loss in the order of 2% to 3% of total
energy input.
Heat losses occur directly from the turbine. This results in a small overall loss due to convection
and radiation to the surrounding atmosphere.
Bearing losses arise due to oil friction in the bearings. Some auxiliaries such as the main oil
pump are usually directly driven from the turbine shaft and add to these losses.
Windage loss occurs in the generator where fans mounted directly on the generator rotor
circulate the hydrogen coolant within the generator housing.
Some electrical losses occur within the generator windings, but are quite small.
8.4 Turbine Load Variation

As the load on a turbine varies, the pressure profile throughout the turbine changes, as shown
in Figure 27. At any one point along the steam flow path, the pressure varies according to
steam flow in a linear manner, as shown in Figure 28. This has been demonstrated in practice
on a number of machines and is an accepted result. At different points in the turbine, a similar
relationship is found. At zero load, however, some steam flow is required to drive the turbine
against bearing friction and generator windage. Turbine load (generator output) is therefore
not quite proportional to steam flow. The overall result can be summarized as in Figure 29,
which shows steam pressure, steam flow, and turbine load or output all on the same plot. Note
that in many problems, the zero load steam flow is not known. In such cases, this zero load
steam flow may be assumed to be small and the overall steam flow taken to be proportional to
turbine load. Note that the value for zero load steam flow is much exaggerated in the diagram.
Figure 27 Steam pressure variation through the turbine (courtesy of NB Power)
Figure 28 Steam pressure variation with load change

Figure 29 Steam pressure versus steam flow and turbine load
The Mollier diagram is an ideal way to represent varying load conditions. Figure 30 shows the
changes in the turbine expansion line for different loads on a large nuclear unit. Some interest-
ing characteristics are apparent in this figure. Because reheating is effected by live steam, the
temperature of the reheated steam remains essentially the same with decreasing load. In
practice, the reheat temperature actually rises with decreasing flow because the temperature
difference for heat transfer becomes less with reduced heat flow. At the low pressure turbine
exhaust, the moisture becomes less with reducing load, and eventually the exhaust steam
actually becomes superheated. To avoid this situation, the reheating steam flow is throttled at
very low loads to reduce steam temperatures in the low pressure turbine. Even with no reheat,
at very low loads, the decrease in internal efficiency due to incorrect steam flow in the blades
causes the exhaust temperature to rise to excessive values. This is aggravated by possible
recirculation of exhaust steam within the turbine blading.
Figure 30 Partial load turbine expansion lines (courtesy of NB Power)

During partial load operation, the incoming steam is throttled to reduce its pressure. As a
consequence, the specific volume of the steam increases. Except in the high pressure areas of
the Mollier diagram, the constant temperature lines in the superheated regions are nearly
horizontal. This means that the throttling process, which is at constant enthalpy, is also almost
at constant temperature. Therefore, the expansion index n in pvn equals a constant is approxi-
mately unity. The change in specific steam volume therefore almost exactly offsets the change
in pressure. Flow velocities in the superheated region of the turbine are therefore practically
constant over a wide range of loads. This means that the velocity triangles are largely unaf-
fected and that the steam continues to pass through the blades at the correct angle. This is
fortuitous and promotes high efficiency. In contrast, hydro turbines operating with incom-
pressible water suffer marked losses in efficiency at low loads due to off-design flow directions.
As the steam pressure becomes less in the low pressure parts of the turbine, this no longer
applies because the turbine exhausts against a positive (even though very low) absolute pres-
sure. The absolute pressures around the last stages are therefore not quite proportional to
steam flow, and the pressure ratio across the last stages (especially the last) drops. This results
in reduced steam velocity entering the moving blades.
In large modern turbines, the changes in the velocity diagram at the last stage can be quite
significant, as illustrated in Figure 31. Here, the velocity diagram has been redrawn to show the
whirl velocity VW. As long as the whirl velocity is positive, that is, Y is on the right hand side, the
steam does work on the blade. At very low loads, the steam incoming velocity VS1 is reduced,
and at the blade tip, the whirl velocity becomes negative, that is, Y is on the left hand side,
indicating that the blade now does work on the steam. Observation of the blade profile at the
tip indicates that, once VS1 becomes substantially less than VB, the blades will act as a fan and
will drive the steam onwards. This fanning action at low steam flows occurs towards the tips of
the blades, that is, around the outer periphery of the turbine rotor. When the fanning action
exceeds the normal steam flow, a recirculation in the last stage will be set up, with steam
returning through the blades near their roots where the impulse type profile simply plows
through the steam at low loads. Such a recirculation is shown in Figure 32. Once set up, this
recirculation results in excessive steam heating by friction and turbulence. If no corrective
action is taken, steam temperatures can become so high as to cause thermal damage to the
turbine.
Two types of corrective action are possible, assuming that the turbine is obliged to continue
operation at this abnormally low load. One is simply to displace the recirculating hot steam
with a small flow of “cooling” steam through the turbine. The other is to spray cooling water
into the recirculating steam at the turbine exhaust so as to reduce its degree of superheat
before it circulates back into the turbine blades. Large nuclear units may have to operate at
very low loads for extended periods during operating manoeuvres on the reactor. These
turbines are usually fitted with exhaust hood sprays or a cooling steam system. Furthermore, a
nuclear reactor such as a CANDU may trip inadvertently or be shut down for a quick repair. In
this case, it would have to be brought back on load within about 40 minutes to avoid suffering
long term (40 hour) unavailability due to buildup of neutron absorbing xenon in the reactor core
while off load. During such events, it is desirable to maintain the turbine at full speed and ready
to be loaded up once reactor power has been restored. This may be accomplished by keeping
the generator connected to the grid system and allowing it to drive the turbine. Such operation
is known as motoring. When motoring, recirculation of steam in the turbine blades is most
severe, and cooling steam and water sprays are required to maintain the temperature of the
low pressure turbine blades within limits.
Figure 31 Change in whirl velocity at low loads
Figure 32 Recirculation of steam at low loads (courtesy of NB Power)

During turbine operation, it is important to minimize thermal transients and to avoid damaging
combinations of thermal stress and mechanical stress. For this reason, turbines are warmed
thoroughly before being run up to full speed, and, during operation, major load changes are
made progressively and slowly so that the turbine casing and rotor can adapt to the changes in
internal steam temperatures. Small load changes within certain limits can naturally be made
quickly to enable the turbine to follow system load requirements or to control frequency. The
turbine may, of course, suffer a trip and immediate initiation of the automatic shutdown se-
quence. This does not necessarily impose a severe thermal transient because steam flow
through the turbine is stopped and it remains hot for an extended period. A restart during this
period must, however, be accompanied by rapid loading so as to minimize the period of admis-
sion of cooler throttled steam to the turbine casings under low load conditions.
All start-ups, shutdowns, and major load changes subject the turbine to some thermal stress.
Such operations reduce the ultimate life of various turbine components and, depending upon
their severity, may be considered as equivalent to a certain number of normal steady state
running hours on the turbine.
8.5 Differential Expansion

When turbine casings and rotors are heated, they expand generally in all directions. During this
expansion, the centre lines of both must remain coincident while the casings and rotors expand
in an axial direction, as shown in Figure 33.
The rotor, having less mass than the casing and being surrounded by steam, tends to expand
more rapidly than the casing, leading to so-called differential expansion between them. This is
apparent mainly in the axial direction and blading, and seals must be designed to accommodate
this relative axial movement as the temperature changes. Furthermore, each cylinder casing is
separately supported on its foundations, but the rotors of all the cylinders are linked together to
form a continuous shaft. There is one thrust bearing to locate the shaft in an axial direction,
usually near the high pressure cylinder. From this point, the expansion of each rotor contributes
to the total expansion, so that there is an additional differential expansion which is cumulative
and greatest at the cylinder furthest from the thrust bearings. Such differential expansion must
be monitored during warm-up and loading to ensure that there is no risk of contact between
the fixed and moving parts.
Figure 33 Differential axial expansion of turbine rotors and casings

(courtesy of NB Power)
8.6 Turbine Governing

The turbine governor valves control the flow of steam to the turbine and, as such, are part of
the main steam system. As the governor valves are opened from the closed position during
start-up, the energy input to the turbine increases. The turbine then accelerates until the
energy losses due to windage and friction balance the energy input. Further admission of
steam increases the speed of the turbine. Turbine speed is therefore dependent upon governor
valve position, and the governor acts to control turbine speed. In this way, the full speed zero
load condition can be maintained.
Once the turbine generator has been synchronized to the electrical grid system, the turbine
speed is locked to the grid frequency. Additional opening of the governor valves increases the
energy input to the turbine, but the speed can no longer increase. At a constant speed, the
windage and friction losses remain constant, and therefore the excess energy appears as
generator output. Further admission of steam increases the load on the generator. Turbine
load is therefore dependent upon governor valve position, and the governor acts to control
turbine generator load.
The main characteristic of the governor is the speed droop. This indicates the amount that the
speed will change as the load is varied. A speed droop of 5% indicates that, if the speed is 100%
at zero load, it will be 95% at full load. Such a change would not normally be permitted, and the
speed would be progressively reset as the load increased. The speed droop, however, deter-
mines how turbines operating in parallel will maintain stability and share load. With turbines
operating in parallel, the speeds must all be the same and match the system frequency. Then,
for a given change in frequency, a turbine with a smaller speed droop will change its load by a
greater amount.
8.7 Turbine Responses

In the event of various transients and plant upsets, the turbine will respond in different ways to
minimize risk to the plant and to facilitate return to normal operating conditions. The following
plant upsets are considered:
Reactor trip
Turbine trip
Electrical load rejection.
For each of these, the main steam system valves must operate in an appropriate way to protect
the turbine.
Once the generator has been disconnected from the grid system, any mismatch between power
input to the turbine and electrical output from the generator will result in a change in speed. In
a modern steam turbine, the mechanical inertia of the rotors is relatively small compared with
its rated power output. This means that it can accelerate very quickly under an excessive
mismatch between power input and power output. Overspeeding is potentially the most
dangerous condition that can occur on a turbine because it can have the most disastrous
consequences. Past experience has shown that a turbine will literally explode under very high
overspeed conditions. The overspeed accident of the 600 MW Unit 4 at Duvha Power Station in
South Africa in February 2011 is a stark reminder of this.
8.8 Reactor Trip

In the event of a reactor trip in a nuclear unit, the mismatch in power input to and power
output from the steam generators will result in a rapid drop in steam generator pressure and
fast cooling of the primary heat transport system. This fast cooling will cause shrinking and
reduce pressure in the secondary steam system.
To avoid these adverse consequences, turbine steam flow must be quickly reduced by closing
the governor valves. Ideally, steam flow to the turbine should be reduced at the same rate as
the reduction in steam generator steam production. The drop in steam generator output
slightly lags the drop in reactor fission power because of the heat stored in the primary heat
transport fluid and because of decay heat generated in the nuclear reactor.
To ensure a proper response, the boiler pressure control system ensures that the governor valve
will be closed progressively and power input from the steam generators to the turbine reduced
to zero while excess steam is diverted to the condenser through the condenser steam discharge
valves. When the turbine power falls below that required to overcome windage and friction,
the generator will draw the necessary power from the grid system. Under these conditions, the
turbine generator is driven at synchronous speed by the electrical grid and is said to be in
motoring mode. A quick recovery from motoring mode is possible so that, once the reactor trip
has been cleared, the unit can be brought up to full power quickly to minimize the effect of the
xenon transient. For this reason, other valves in the system, such as the intercept valves, may
be automatically positioned in readiness for a turbine restart. The advantage of the motoring
mode is to avoid disconnecting the turbine from the grid system, thereby eliminating partial
turbine rundown and runup as well as resynchronization, all of which require careful manoeu-
vring.
Once tripped, however, a reactor requires checking to verify the cause of tripping and careful
restart procedures, all of which take time. Therefore, motoring may not be a practical option,
and the turbine may need to be tripped automatically in response to a reactor trip.
8.9 Turbine Trip

The purpose of a turbine trip is to protect the turbine in the event of adverse operating circum-
stances or component malfunction or failure. A turbine trip should be initiated whenever
continued operation poses a risk of damage to the turbine.
When a turbine trip occurs, power input and power output must be stopped. On the generator
side, power output is stopped virtually instantaneously as the generator circuit breaker opens.
On the turbine side, however, the steam valves take a finite time to close. During this brief
period, some steam passes through the valves, carrying with it substantial energy. This steam,
along with the steam already trapped within the turbine, expands and produces a transient
power input to the turbine. The mismatch between power input and power output results in a
temporary rise in speed until the effects of windage and friction cause the speed to fall.
This is known as a non-sequential trip. The temporary speed rise can be avoided if the trip on
the generator side is slightly delayed to balance the power output against the power input. The
excess energy in the steam entering the turbine is then passed into the grid system during the
tripping process. Once this energy flow ceases, the generator circuit breakers are opened. The
turbine then coasts down from its normal running speed under the influence of windage and
friction. This is known as a sequential trip.
8.9.1 Non-sequential trip

As indicated above, in a non-sequential turbine trip, the generator circuit breakers open at the
same time as turbine steam valve closure is initiated. The valves take a finite time (perhaps half
a second) to close, and the incoming steam along with the trapped steam in the turbine accel-
erates the turbine into an overspeed condition (about 10% above normal speed), as shown in
Figure 34. Note that the horizontal time scale is not linear. The speed rise is in seconds, and the
speed drop is in minutes.
A non-sequential trip must be initiated whenever there is an electrical fault on the generator or
generator transformer because the heavy currents arising from faults can cause severe damage
very quickly.
8.9.2 Sequential trip

In a sequential trip, the generator circuit breakers remain closed while the turbine steam valves
are closing, so that the surplus power from the steam can be passed into the electrical grid. As
soon as the electrical power flow from the generator falls to zero or reverses, the generator
circuit breakers open. In this way, overspeeding of the turbine is prevented, as shown in Figure
34. This is inherently safer for the turbine generator than a non-sequential trip.
A sequential trip is initiated whenever there is a mechanical fault on the turbine or associated
equipment. A sequential trip actually promotes a slightly faster shutdown in an emergency
because the overspeed transient is avoided. Moreover, if the mechanical fault is due to bearing
problems or excessive vibration, it is highly desirable not to subject the turbine to overspeed
conditions.
8.10 Steam Flow Control

Mention has been made of the steam trapped inside the turbine after the steam inlet valves
have closed. This steam, on expansion to vacuum conditions in the condenser, will continue to
accelerate the turbine. Modern steam turbines contain a large volume of steam in their casings
and associated pipework and components such as moisture separator reheaters and feedwater
heaters. To minimize the effects of this trapped steam, certain other valves such as the inter-
cept valves and the extraction steam line check valves also close to restrict steam flow into the
turbine.
8.11 Load Rejection

A load rejection is, in effect, an electrical trip. In the event of certain external faults on the
electrical grid system or transmission line leaving the plant, the generator transformer circuit
breaker will open, leaving the unit isolated from the system, but still supplying its own unit
service load. Under these circumstances, the unit suffers an instantaneous rejection of virtually
its entire load, but it must be kept running and ready for resynchronization with the grid system
once the fault has been cleared. A load rejection is therefore not a mechanical trip, and the
steam flow must not be cut off permanently. The turbine does, however, suffer the same type
of transient as during a non-sequential trip.
Figure 34 Turbine generator non-sequential and sequential trips

In a load rejection, the power output from the generator is drastically reduced, and closure of
the steam inlet valves must be initiated immediately. Even so, the mismatch between power
input to the turbine as steam and power output from the generator will accelerate the turbine
into an overspeed condition (about 8% above normal speed). This speed transient is not quite
as severe as in a non-sequential turbine trip because some surplus power is diverted into the
unit services, which account for between 6% and 7% of the full generator power output.
Naturally, this speed transient affects the electrical frequency of the supply to the unit services,
and the motors of the auxiliaries suffer the same speed transient. Once the speed transient has
passed and the turbine speed returned to normal, steam must once again be admitted to the
turbine to provide enough power to drive the auxiliaries and to maintain the normal speed
condition. The governed speed when supplying only the station load is, however, somewhat
above the normal running speed due to the speed droop setting on the governor. The differ-
ence between the two speeds is known as the permanent speed rise. Returning the speed to
normal requires adjustment to the governor. The speed transient arising from a load rejection is
illustrated in Figure 35. This assumes no adjustment to the governor settings, but modern
control systems initiate a turbine runback to bring the speed back to approximately normal
running speed. As before, the speed rise is in seconds and the speed drop in minutes.
Figure 35 Turbine-generator load rejection
8.12 Turbine Overspeed Protection

All turbines are fitted with direct acting overspeed protection. This usually takes the form of
mechanical spring loaded eccentric rings or bolts mounted on the turbine rotor. At an over-
speed of about 110% or slightly more of normal speed, they operate and trip the hydraulic
circuit of the steam inlet valves. This action is entirely independent of the governing system or
other tripping circuits.
In the event of a load rejection and failure of the governing valves to close, the turbine speed
would continue to rise. At an overspeed in excess of 110% of normal speed, the overspeed trip
would be activated to close the emergency stop valves, and also the governing valves if their
failure to close was a governing system failure.
An important aspect regarding the above scenario is that, while a load rejection is initiated at
100% of normal speed, the overspeed trip is activated at about 110% of normal speed. In the
event of a governing system failure, the turbine therefore accelerates under full steam flow
from 100% to 110% of normal speed before the steam inlet valves begin to close. They take a
finite time to close, and after that, the trapped steam expands. During this period, the turbine
continues to accelerate into an even higher overspeed condition, as shown in Figure 36. This
can be as high as 118% to 120% of normal speed depending upon the turbine design. The
turbine overspeed protection does not therefore limit the turbine speed to the settings se-
lected. It is the last line of defence and should be treated accordingly. Any overspeed in excess
of the overspeed trip setting should be regarded as dangerous because turbine rotors are
overspeeded to only about 125% of normal speed at the manufacturers’ facility to confirm their
integrity.
The transient overspeed arising from a non-sequential turbine trip or a load rejection should
always be below the overspeed trip settings to demonstrate proper operation of the governing
system and to avoid a turbine trip following a load rejection.
Figure 36 Turbine generator overspeed trip
9 Problems
1. List the key controlling units of a CANDU nuclear plant and explain their function and
how they operate. Clarify the difference between normal and alternate modes of opera-
tion and explain which controlling units are affected and how they are affected by a
change from one mode to the other.
2. Explain how the reactor is protected from abnormal operating conditions and potential
accident conditions arising from process excursions or equipment failures.
3. Explain how the turbine is protected from abnormal operating conditions and from po-
tential accident conditions arising from control malfunctions or equipment failures.
4. Explain the theoretical aspects related to the approach to criticality in a reactor, noting
in particular the relationship between stabilization time and reactivity added. Explain
also how the point of criticality can be predicted and how one knows when criticality has
been achieved.
5. Describe the approach to criticality technique used in a CANDU reactor and clarify why
this step-wise method has an inherent safety feature.
6. Explain how reactor reactivity changes during operation, noting which key fission prod-
ucts affect reactivity over the life of the fuel. Illustrate these changes in a diagram and
show on the diagram at which point refuelling is necessary. Assume that the reactor ini-
tially has new fuel only.
7. Assuming that a hypothetical steam generator operates under the following conditions:
Steam temperature 260C (saturated steam)
Feedwater temperature 260C (saturated water)
Steam flow 500 kg/s
Primary coolant pressure 10 MPa
Primary inlet temperature 290C
Primary outlet temperature 270C.
a. Calculate the primary coolant mass flow rate.
Assuming that the primary coolant mass flow rate is constant and that the heat transfer
coefficients remain constant, determine the following:
b. Primary system temperatures when the steam flow rate is increased to 1000
kg/s, assuming that the steam pressure is kept constant.
c. Secondary steam pressure when the steam flow rate is increased to 1000 kg/s if
a temperature of 300C in the primary system is not to be exceeded.
For both (b) and (c) above, sketch the temperature variation along the length of the tube
bundle and show terminal temperatures. Assume a linear rise in temperature of the
primary coolant.
8. Illustrate the two modes of operation of a typical CANDU unit as follows:
a. Sketch a typical pressure control system for a steam generator when operating in
the reactor leading mode.
b. Sketch a typical pressure control system for a steam generator when operating in
the turbine leading mode.
Each sketch should show the nuclear reactor, steam generator, and turbine generator as
well as the primary sensors and control elements. For each, describe how the system
would respond to a transient load condition (increase or decrease in load).
9. Explain the terms “swelling” and “shrinking” when applied to a boiler or steam genera-
tor. Clarify the difference between steady state and transient conditions and explain
what causes each to occur. As clarification, review the sequence of events following a
load change on the system. Explain briefly what effect this has on a single element level
control system.
10. If the pressure in the steam generator drops from 4.0 MPa to 3.9 MPa, determine what
percentage by mass of the water in the steam generator will immediately be converted
to steam. If this steam remains in suspension in the water in the form of bubbles, by
what percentage will the water swell (transient swelling)?
11. In large CANDU plants, the steam generator level set point is usually programmed so
that the water level is lower at low power levels than at high power levels. Explain why
the water level set point is different for different power conditions. Clarify how the wa-
ter inventory in the steam generator changes as the level set point is ramped up and
down.
12. Show in a sketch how steam generator level varies with load in a typical CANDU steam
generator. Show also in a separate sketch how steam generator inventory would be ex-
pected to vary with load. Discuss the relative merits of maintaining constant level or
constant inventory in the steam generator and clarify to what extent these are achieved
with the current arrangement.
13. Consider a typical three element controller for a steam generator.

a. Sketch the control system used to control the water level. Show what measure-
ments are taken and how the signals are processed by the control system. The
diagram must be sufficiently detailed and fully labelled so as to be self-
explanatory.
b. Explain why a three element controller is used instead of a single element (level
only) controller. Support the explanation with an example demonstrating when
and why a single element controller may give an incorrect response.
14. In the event of a turbine-generator trip or load rejection from full load, the reactor
power is decreased and maintained at a certain level to prevent excessive buildup of xe-
non in the reactor and surplus steam is dumped to the condenser.
a. Explain what limits the maximum amount of steam that can be rejected and state
the usual limit.
b. Explain why water is mixed with the steam before release into the condenser. Es-
timate the condition of the steam before mixing with water and hence determine
how much water must be added per kilogram of steam.
15. Consider a condenser operating under the following conditions:
Cooling water inlet temperature 10C
Cooling water outlet temperature 20C
Steam inlet (saturation) temperature 30C.
Sketch the anticipated temperature profiles for both cooling water and steam across the
condenser (that is, along the tubes) for each of the following conditions:
a. Given or normal conditions.
b. Cooling water inlet temperature increased to 16C.
c. Cooling water flow reduced to 1/2 its original value.
d. Turbine load reduced to 1/2 its original value.
Assume a linear change in cooling water temperature along the tubes.
16. Consider a condenser which operates at 0.005 MPa and condenses steam with a mois-
ture content of 10% at a rate of 600 kg/s. Cooling water enters the condenser at 13°C
and leaves at 23°C. Determine the following:
a. Temperature profile (sketch) across the condenser (along the tubes from inlet to
outlet).
b. Average temperature difference between steam and cooling water.
Having established the basic condenser parameters given above, determine what
changes in these parameters would occur and what their magnitude would be for the
following new conditions:
c. Cooling water inlet temperature increase to 18°C.
d. Cooling water flow rate decrease to 50%.
e. Fouling of condenser tubes which decreases overall heat transfer coefficient by
25%.
f. Reduction in steam flow to condenser to 400 kg/s.
For each case, sketch the new temperature profile across the condenser, assuming a lin-
ear change in cooling water temperature along the tubes and average temperature dif-
ferences.
17. The condenser of a large steam turbine receives exhaust steam at varying flows depend-
ing upon the prevailing turbine output. On axes of condenser pressure versus condenser
steam flow, sketch the following:
a. Variation in condenser pressure with steam flow into the condenser (due to
variation in turbine output).
b. Variation in condenser pressure with steam flow into the condenser at a cooling
water temperature higher than that in (a) above.
Explain why the condenser performance curve has the shape shown and why the varia-
tion shown in (b) is different from that shown in (a).
18. Explain how the flow of steam through the turbine is affected by low load operation.
Clarify what adverse conditions arise in the last stage blading under low load conditions
and how this affects turbine performance. Explain how these adverse effects can be
minimized.
19. Describe motoring and explain under what conditions the turbine would be subject to
motoring. Give the advantages of motoring and note the precautions that must be
taken when motoring.
20. Consider the start-up phase of turbine operation.
a. Explain in detail the cause of sagging and clarify under what circumstances sag-
ging of the turbine shaft can occur.
b. Explain in detail the cause of hogging and clarify under what circumstances hog-
ging of the turbine shaft can occur.
c. Explain why such states are not desirable and the adverse consequences arising
from such deformation.
d. Explain how such states can be prevented from occurring or alleviated if such
conditions exist.
21. Explain axial differential expansion in a large steam turbine. Explain how this can be
minimized by proper location of the thrust bearing in the case of a large nuclear unit.
Support the explanation with a suitable sketch showing the direction of expansion of ro-
tor and casing. Show the fixing points of the casing and the location of the thrust bear-
ing. Clarify under what conditions excessive axial differential expansion can occur. De-
scribe the adverse consequences caused by excessive axial differential expansion.
22. Explain the operation of the following and clarify under what circumstances each would
be used:
a. Regular governor when controlling speed.
b. Regular governor when controlling load.
c. Emergency overspeed governor.
23. Describe the sequence of events and the turbine response during the following tripping
modes. Clarify under which circumstances each would occur and sketch the speed re-
sponse of the turbine in each case.
a. Sequential turbine trip.
b. Non-sequential turbine trip.
c. Turbine load rejection.
Show on each sketch the period during which the governor valve closes and the period
of steam expansion in the turbine.
24. Explain how a turbine is protected against excessive overspeed. Describe a potential se-
quence of events with associated physical phenomena that could lead to the total de-
struction of the turbine.
Answer Guide for Chapter 9
1. See Section 2.1 CANDU Plant Control Systems and Figure 2 Basic plant control system.
2. See Section 2.2 Reactor Protection Systems, Section 2.3 Reactor Shutdown Systems and
Section 2.4 Emergency Coolant Injection.
3. See Section 2.5 Turbine Protection System.
4. See Section 3.3 Approach to Criticality as well as Figure 4 Stabilization time for different
subcritical multiplication factors and Figure 6 Approach to critical – inverse count rate
change.
5. See Section 3.4 Approach Technique and Section 3.5 Reactor Start-up.
6. See Section 3.7 Fuel Burnup and Figure 8 Effects of fuel burnup on reactivity.
7. See Section 5.1 Plant Control and Figure 9 Steam generator temperature profile.
Use Steam Tables. (a) Flow rate M = 8008 kg/s assuming light water. (b) Inlet tempera-
ture t = 290°C. Outlet temperature t = 310°C. (c) Steam Temperature t = 250°C. Steam
pressure p = 3.97 MPa.
8. See Section 5.2 Steam Generator Pressure Control and Figure 2 Basic plant control sys-
tem.
9. See Section 5.3 Swelling and Shrinking as well as Figure 10 Steady state swelling due to
different load conditions and Figure 11 Transient swelling due to changing load condi-
tions.
10. Use Steam Tables. Amount of water converted to steam = 0.26%. Degree of swelling =
10%.
11. See Section 5.4 Steam Generator Level Control as well as Figure 12 Constant steam gen-
erator water level and Figure 13 Ramped steam generator water level.
12. See Section 5.4 Steam Generator Level Control and Figure 15 Steam generator level
alarms and set points.
13. See Section 5.4 Steam Generator Level Control and Figure 14 Three element steam gen-
erator level control system.
14. See Section 6.1 Characteristics and Function and Section 6.2 Atmospheric and Con-
denser Steam Discharge Valves. Use Steam Tables. Assume a steam generator pressure
of 5 MPa and a condenser pressure of 0.005 MPa. Water required = 0.10 kg water per kg
steam.
15. See Section 7.1 Thermodynamics and Heat Transfer and Figure 18 Changes in condenser
temperature profile. (a) Steam temperature = 30°C. (b) Steam temperature = 36°C. (a)
Steam temperature = 35°C. (a) Steam temperature = 20°C.
16. See Section 7.1 Thermodynamics and Heat Transfer and Section 7.4 Condenser Tube
Fouling as well as Figure 18 Changes in condenser temperature profile. (a) Steam tem-
perature = 33°C. (b) Average temperature difference θ = 15°C. (c) Steam temperature =
38°C. (d) Steam temperature = 38°C. (e) Steam temperature = 38°C. (f) Steam tempera-
ture = 26°C.
17. See Section 7.1 Thermodynamics and Heat Transfer as well as Figure 19 Variation in con-
denser pressure and Figure 20 Condenser performance curves.
18. See Section 8.4 Turbine Load Variation as well as Figure 31 Change in whirl velocity at
low loads and Figure 32 Recirculation of steam at low loads.
19. See Section 8.4 Turbine Load Variation.
20. See Section 8.1 General Operational Considerations.
21. See Section 8.5 Differential Expansion and Figure 33 Differential axial expansion of tur-
bine rotors and casings.
22. See Section 8.6 Turbine Governing.
23. See Section 8.9 Turbiine Trip.
24. See Section 8.12 and Figure 36 Turbine generator overspeed trip.

Bereznai, G., Nuclear Power Plant Systems and Operation, Revision 4, UOIT, Canada, 2005.
British Energy International, “Modern Power Station Practice”, Volume C, Turbines, Generators,
and Auxiliary Plant, Third Edition, Pergamon Press, UK, 1991.
Chaplin, R. A., “Fundamentals of Nuclear Energy”, Chapter 3.6.1 in Nuclear Energy and Reactors,
Encyclopedia of Life Support Systems by UNESCO, EOLSS Publishers, Oxford, UK, 2010. [This
chapter gives a brief outline of basic nuclear physics and how a reactor operates].
Chaplin, R. A., “Nuclear Reactor Kinetics”, Chapter 3.6.1.5 in Nuclear Energy and Reactors,
chapter gives the theory related to reactor kinetics and gives simple numerical examples of the
effect of delayed neutrons, subcritical multiplication factor and approach to criticality].
Chaplin, R. A., “Reactivity Changes”, Chapter 3.6.1.6 in Nuclear Energy and Reactors, Encyclope-
dia of Life Support Systems by UNESCO, EOLSS Publishers, Oxford, UK, 2007. [This chapter gives
explanations of the effect of fission products such as xenon and samarium as well as the effect
of temperature on the reactivity of the reactor].
Chaplin, R. A., “Nuclear Reactor Steam Generation”, Chapter 3.10.2.10 in Thermal Power Plants,
chapter gives explanations of the dynamics of steam generators and describes level and pres-
sure control and consequences of inadequate control].
Chaplin, R. A., “Steam Turbine Steam System”, Chapter 3.10.3.4 in Thermal Power Plants,
chapter describes the steam system and explains the operation of the condenser].
Chaplin, R. A., “Steam Turbine Operational Aspects”, Chapter 3.10.3.5 in Thermal Power Plants,
chapter describes the thermodynamic phenomena occurring in the turbine which affect its
operation as well as the external influences such as those due to the condenser. It also includes
thermal effects and governing].
11 Acknowledgements
Acknowledgments are extended to NB Power for the use of information, diagrams and photo-
graphs to support the text of this chapter.
George Bereznai
Simon Pang
1
CHAPTER 10
Instrumentation and Control
prepared by
Dr. G. Alan Hepburn
Independent contractor (AECL Retired)
Summary:
This chapter describes the role of instrumentation and control (I&C) in nuclear power plants,
using the CANDU 6 design as an example. It is not a text on the general design of instrumenta-
tion and control algorithms, a subject which is well covered by many textbooks on the subject.
Rather, it describes the architectural design of these systems in nuclear power plants, where the
requirements for both safety and production reliability are quite demanding. The manner in
which the instrumentation and control components of the various major subsystems co-operate
to achieve control of the overall nuclear plant is described. The sensors and actuators which are
unique to the nuclear application are also described, and some of the challenges facing design-
ers of a future new build CANDU I&C system are indicated.
Table of Contents
1 Introduction ............................................................................................................................ 3
1.1 Overview ......................................................................................................................... 4
2 Nuclear Safety and Production Requirements for I&C Systems ............................................. 5
2.1 Requirements for the Special Safety Systems................................................................. 8
2.2 Safety Requirements for the Process Systems................................................................ 8
2.3 Production Requirements for the Process Systems ........................................................ 9
3 Overall I&C Architecture ....................................................................................................... 10
3.1 Architectural Design of the Special Safety Systems I&C............................................... 10
3.2 Architectural Design of the Process Systems I&C ......................................................... 12
3.3 Features of the CANDU DCC Design to Enhance Production........................................ 12
3.4 Safety-Related Functions of the Process Systems......................................................... 12
4 Overall Plant Control Functionality ....................................................................................... 14
4.1 RRS control logic............................................................................................................ 15
4.2 BPC Control Logic .......................................................................................................... 18
4.3 UPR Control Logic.......................................................................................................... 18
5 Special Safety Systems Functionality .................................................................................... 18
5.1 Shutdown Systems 1 and 2 ........................................................................................... 18
5.2 Emergency Core Cooling System................................................................................... 19
5.3 Containment System..................................................................................................... 19
6 I&C Systems Layout and Equipment ..................................................................................... 20
6.1 Sensors .......................................................................................................................... 23
6.2 Actuators....................................................................................................................... 25
6.2.1 Instrument Air System .................................................................................................. 27
6.3 Actuators for the Special Safety Systems...................................................................... 28
6.4 Control Logic Technology in the Process I&C Systems.................................................. 29
©UNENE, all rights reserved. For educational use only, no assumed liability. Instrumentation and Control – September 2014
6.5 Logic Technology in the Special Safety Systems............................................................ 32

7 Regional Overpower (ROP) Trip Logic ................................................................................... 33
8 Design Verification ................................................................................................................ 35
9 Bringing the CANDU I&C Design into the 21st Century ......................................................... 36
9.1 Technological Obsolescence ......................................................................................... 37
11 Problems ............................................................................................................................... 39
12 References............................................................................................................................. 42
List of Figures
Figure 1 CANDU overall plant control........................................................................................... 15
Figure 2 Arrangement of the 14 control zones............................................................................. 15
Figure 3 Absorber-rod drive switching.......................................................................................... 17
Figure 4 Adjuster-rod drive switching........................................................................................... 17
Figure 5 Rod drive speed .............................................................................................................. 17
Figure 6 Layout of a CANDU 6 control room................................................................................. 21
Figure 7 Pictorial view, CANDU 6 control room ............................................................................ 22
Figure 8 SIR detector geometry. ................................................................................................... 24
Figure 9 Liquid zone control compartment (simplified). .............................................................. 26
Figure 10 Reactor building instrument air system (simplified)..................................................... 27
Figure 11 Two-thirds voting logic for SDS1 trip ............................................................................ 29
Figure 12 CANDU 6 DCC architecture ........................................................................................... 31
Figure 13 Effect of compensation on Pt-clad detector response.................................................. 35
Instrumentation and Control 3
1 Introduction
This chapter describes the role of instrumentation and control (I&C) in nuclear power plants,
using the CANDU 6 design as an example. The scope of I&C includes:
 Implementation of control strategies for those control functions which are auto-
mated,
 Presentation of information to the operator and receipt of operator inputs for those
functions which are under operator control,
 Initiation of reactor shutdown, emergency coolant injection, and containment isola-
tion in the event of failure of the above control functions, and
 Data acquisition.
The automated functions include:
 Automatic control of the reactor, balance of plant, and auxiliary systems;
 Activation of the special safety systems;
 On-power refuelling (CANDU reactors);
 Human/machine interface.
Virtually all the systems in a nuclear power plant contain an instrumentation and control
element.
Although some of the material presented is common to other reactor types, the design details
described pertain to the CANDU reactor. The CANDU 6 design as implemented at sites in
Canada has been chosen as the reference because it is the most widely deployed CANDU design
world-wide. While the implementation of I&C in other plants is broadly similar, the differences
between the CANDU 6 and other CANDU stations are significant. These differences will be
noted in some cases where they are of particular interest to the understanding of I&C in gen-
eral.
At the time when the CANDU 6 and many of the world’s light water reactors were designed, the
process I&C subsystems in nuclear plants were implemented using a combination of individual
analog control loops and discrete Boolean logic using relay technology. The design of the analog
loops is based on classical linear frequency-domain control theory as described in any textbook
on the subject.
As a consequence of the large core of the CANDU reactor and of its many fuel channels, reactor
inlet and outlet headers, and other components, the CANDU design is very extensively instru-
mented. As CANDU reactors began to exceed about 250 MWe, the size of the reactor core and
the on-power refuelling combined to make manual supervision of the spatial distribution of
power in the core more and more unwieldy. The result was a strong motivation to introduce
computer control of the reactor and of key process control loops. The resulting control algo-
rithms are quite complex, with many related inputs and outputs. Implementing such a system
using conventional equipment would have been impractical. Therefore, a central dual-
redundant digital control computer (DCC) system, in which the control logic is defined by
software, was introduced in the CANDU design very early on relative to the I&C industry as a
whole, to say nothing of the nuclear power industry. All CANDU reactors, starting with Douglas
Point, have used computers in their process I&C systems. At least from Pickering on, the use of
this technology was a matter of design necessity.
This chapter is intended to provide an introduction to the role of I&C in nuclear power plant
applications and is consequently oriented towards issues that are unique to the nuclear context.
It is not a text on the basics of instrumentation and control, nor is detailed knowledge of this
field necessary to understand the material presented. The details of the control strategies used
in each plant system are not covered. The only detailed discussions of these strategies pre-
sented are of the subsystems involved in reactor and overall plant control and of the regional
overpower trip logic.
For those interested in studying the individual systems in more detail, there is a plethora of
material available on the CANTEACH Web site (https://canteach.candu.org). In reviewing this
material, the author noted that it contains many references to design features that are unique
to individual sites as if they were part of the generic CANDU design. For those who need to
know the details of a specific station, the most reliable source is the system design manuals
pertaining to that station.
1.1 Overview
The reliability and availability requirements for nuclear power plants tend to differentiate them
from many other applications. The implications of these requirements for the major I&C
subsystems are discussed from the point of view of both nuclear safety and plant production.
References to current standards and regulatory documentation are provided.
The overall architecture of I&C systems is described, using the reference CANDU 6 design as an
example. There are close ties here with the architecture of the electrical power systems de-
scribed in Chapter 11.
The implementation technologies used in nuclear I&C are also described. As an aid to under-
standing the design decisions evinced by the design, a brief discussion of the technologies
available to the designers in the early 1970s is provided.
To give the reader an insight into the more detailed I&C logic functions typical of the CANDU
design, the operation of the main I&C systems involved in overall control of the plant (reactor
regulating system, boiler pressure control, and unit power regulation) is described in Section 4.
The design of the I&C subsystems of the special safety systems is described in Section 5. The
tools and techniques used to verify the design are discussed in Section 8.
I&C technology has arguably experienced the most dramatic change of any of the technologies
used in power plant design in the years since most of the world’s nuclear fleet was constructed.
However, at least in the case of CANDU, although a number of I&C subsystems have been
replaced by more modern equipment at various stations over the years, the lack of a truly new
build design means that the existing I&C design is now very dated. If and when the next new
build is undertaken, the designers will face an extremely challenging task in bringing the design
concepts and implementation technologies forward into the 21st century.
Several new-build designs have been initiated in the interim, notably for the CANDU 3, CANDU
9, and the Advanced CANDU reactor, but none of these has been carried out to the point where
detailed design documentation was produced and implementation planning was documented,
let alone where a reactor had been constructed, commissioned, and licensed. Furthermore, I&C
technology moves forward perhaps one generation every ten years, so new designs have a
rather limited shelf life.
Given the small number of new builds elsewhere in the world, many plants share this tendency
to have outdated I&C technology. A discussion of some of the issues facing the designers of an
upcoming new build (as opposed to a new replicate design) is included in Section 9, although, to
be sure, any such discussion will be overtaken rather rapidly by technological developments.

 The role of I&C systems in the safety and process systems of the CANDU design,
 The high-level safety and production requirements applicable to these systems,
 The role of system architecture in achieving safety and production reliability,
 The technologies used to implement these systems and the constraints present at
the time of their design that influenced the design choices made,
 The unique features of the CANDU I&C design and why they were adopted, and
 Some of the major challenges facing the designers of future CANDU I&C systems.
2 Nuclear Safety and Production Requirements for I&C Systems

A frequent reaction of the general population to the idea of nuclear power is, “What happens if
it goes out of control”? They immediately have visions of mushroom clouds and grainy black-
and-white movies from the Second World War. In fact, a nuclear explosion of this sort is simply
not possible with a power reactor. To create a nuclear explosion requires a small mass of highly
enriched uranium to be held very firmly together using conventional explosives for the short
time it takes for the chain reaction to go massively supercritical and the consequent release of
energy to occur.
In a power reactor, maintaining the configuration of the fissile material is also key to maintain-
ing the chain reaction, but the fuel is not nearly as highly enriched (or in the case of CANDU, not
enriched at all), and the fissile material must be maintained at a predetermined separation of
the order of a few tens of centimetres, with some moderating material, such as graphite, water,
or heavy water in the intervening space. If these conditions are not met, the chain reaction will
not be sustained.
In any nuclear reactor at constant power, the rate of production of neutrons by fission is exactly
matched by the rate of re-absorption in new fissions and by various losses. The neutron multi-
plication factor k = 1. If k >1 for some time, power will increase rather quickly to the point that
so much heat is produced that the fuel melts and the core disassembles, thereby destroying the
geometry required for criticality. But there will not be a nuclear explosion. There may be a
steam explosion, as happened at Chernobyl, but that, while undeniably violent, is not in the
same league as a nuclear explosion.
Although a nuclear explosion is not possible, in the absence of effective process control and
other defense mechanisms, the reactor would be destroyed, and radioactive material could be
released to the environment, causing an ecological disaster like that which occurred at Cherno-
byl. Because there is far more fissile material in a power reactor than in a weapon (tonnes vs.
kg), the result would be widespread contamination, which is not an acceptable outcome.
For these reasons, a control system is needed to bring the chain reaction to a useful power level
and then to hold it there. For thermodynamic efficiency, the hotter the energy source (fuel
elements), the better. However, if dryout is ever allowed to occur, cooling deteriorates quite
rapidly, resulting in fuel melting. Therefore, power has to be raised to some point below dryout,
then held close to constant by maintaining k = 1. This manoeuvering and control is achieved by
adding or removing neutron-absorbing control elements: poison, rods, and in the case of
CANDU, light water. Insertion of each reactivity-control mechanism results in a reduction in the
neutron multiplication factor. The absorption value of these devices is expressed in milli-k (mk).
Insertion of one mk of negative reactivity decreases the neutron multiplication factor by 0.001.
In a light water reactor, fresh fuel requires maximum negative reactivity (rod insertion). As fuel
burns up, rods must be gradually withdrawn to sustain the chain reaction. Power distribution
within the core is highly predictable because the rods are withdrawn in a predetermined pat-
tern. In CANDU, with continuous refuelling, fuel reactivity can be maintained indefinitely, but
local peaks will occur when fresh fuel is inserted. This, together with the larger size of the core,
which in the absence of control action can lead to local flux oscillations due to Xe135 (a neutron
absorber with a fairly short half-life which is produced as a result of the fission process), means
that, in a CANDU reactor, flux has to be controlled both as an average value, for the entire core,
and spatially, to avoid unwanted flux tilts.
Heat must also be removed from the fuel as it is produced. This is achieved by circulating a
heat-transport fluid (e.g., light or heavy water) over the fuel. The energy removed is transferred
to the secondary coolant circuit, where it produces steam to drive the turbine. Failure to
remove heat can result in damage to the fuel or to heat-transport system components, so
control of the heat-transport system is also critical.
The safety role of the process control systems is to keep the various reactor systems operating
within predetermined safe limits. Given the potential consequences of process-system failure,
however, the defenses against such failure must be extremely robust—much more robust than
could credibly be achieved by the process systems themselves. Several additional layers of
defense are used in any modern power reactor. The probability that each layer will not be
available to accomplish its function is expressed as some unavailability figure, e.g., 10-3 years
per year. In the case of a defense mechanism, this is equivalent to saying that the mechanism
will operate as expected 999 times out of every 1000 challenges.
In any modern reactor design, the additional levels in this defense-in-depth approach are:
 The shutdown systems,
 The emergency core cooling system (ECCS), and
 The containment system.
In CANDU plants, these systems are referred to as the “special safety systems”, while in PWRs,
they are referred to as “reactor protection systems” and “engineered safety features”. In the
CANDU context, anything else is a “process system”. During plant operation, the special safety
systems are poised ready to perform their function when called upon to do so, while the proc-
ess systems are generally in continuous operation. However, some process-system functions are
also normally dormant, but take action to preclude the need for special safety system interven-
tion.
The role of the shutdown system is to stop the chain reaction very rapidly (generally in less than
two seconds) if there is an indication that the process parameters are going outside acceptable
limits. Shutdown systems achieve their function by rapidly inserting large amounts of negative
reactivity. This is typically done in all reactors by inserting neutron-absorbing rods into the core
using a highly reliable power source—the force of gravity (or pressurized gas accumulators, in
the case of boiling water reactors), augmented initially by springs in the case of CANDU.
The CANDU core has a slightly positive void coefficient. This means that, as power increases
through the point that the primary coolant starts to boil, the reactivity would also increase in
the absence of any control-system action. In other words, a positive feedback situation exists.
On initial consideration, this does not appear to be a desirable characteristic, because an event
that results in a power increase is not self-limiting. The Canadian industry’s response to this has
been to provide a second separate, diverse shutdown system (SDS2). In all designs from Bruce
A on, both shutdown systems are equally capable. The second system injects gadolinium nitrate
poison into the moderator, using compressed helium gas as the motivating force.
Every effort is made to ensure that the two shutdown systems are independent (different
reactivity mechanisms, different design teams, complete electrical and spatial separation of
instrumentation and mechanisms, different I&C devices wherever possible and practical), so
that the design unavailabilities of 10-3 years per year for each system can be multiplied to
achieve an overall unavailability of the shutdown function of 10-6 years per year.
Proponents of light water reactors tend to make much of the CANDU’s positive void coefficient,
but the comparison with respect to inherent power dynamics is not all one-sided. Clearly, in
addition to the direction that power tends to move, the rate at which it moves is also important.
Although a milli-k may not sound like much, the average lifetime of a neutron is less than one
millisecond. Considered simplistically, this would mean that an excess reactivity of only 1 mk
would result in an approximate tripling of neutron power within one second. Fortunately, this
simplistic approach is not applicable in a power reactor core. Due to the effects of a relatively
small number of delayed neutrons (about 6%), the increase in power in one second would, for a
CANDU, be around 1%. In a light water reactor, it would be around 10%. This is one key differ-
ence between the CANDU core and a typical LWR. For a more complete explanation, see
[Rouben2002] and Chapter 5, Section 6. Moreover, once power reduction starts, the strong
negative power coefficient of a light water reactor tends to resist the desired power reduction,
prolonging heat generation in the fuel. In a light water reactor, the shutdown system alone
does not insert enough negative reactivity to shut down the reactor. These reactors also rely on
an inherent feature of the physics of the light water core—the Doppler resonance phenome-
non—to help terminate the reaction [Rouben2008].
Stopping the chain reaction is necessary, but not sufficient to mitigate the effects of a process-
system failure. The decay heat, which initially amounts to about 6% of the pre-shutdown value,
still has to be removed. In a fossil-fuel plant, cutting off the source of new energy (the fuel feed)
instantly removes the heat source. In a power reactor, even after the chain reaction has been
terminated, decay heat could cause the fuel to melt, with the resulting undesirable conse-
quences, for many days after the reactor has been shut down, unless continued cooling of the
fuel is provided. Usually, the heat-transport system is available to do this, but to guard against
any failures in this system, a backup cooling system is required—the emergency core cooling
system, whose role is to provide an alternate path for removal of decay heat from the fuel if
there is an indication that the process systems responsible for doing this have failed.
The last line of defense is the containment system, whose role is to provide an envelope around
the parts of the plant that contain fission products so that this material will not be released to
the environment. The containment system also condenses any steam released into this enve-
lope, thus limiting any upward pressure excursion following such a release.
2.1 Requirements for the Special Safety Systems

Oversight of the nuclear industry in Canada is the responsibility of the Canadian Nuclear Safety
Commission (CNSC). The forerunner of the CNSC was the Atomic Energy Control Board (AECB).
One of the CNSC’s responsibilities is to develop the regulations governing the design and
operation of nuclear power plants. The regulatory requirements for nuclear power plant safety
in Canada are contained in the AECB Regulatory Documents listed in the references.
The reliability requirements for the special safety systems, which have a dominating influence
on both the architecture and detailed design of these systems, are:
 SDS1 and 2: All Canadian power reactors are required to have two independent
shutdown systems. As stated earlier, this requirement came about as a result of a
design solution to an inherent characteristic of the CANDU core. This is now en-
shrined in a regulatory requirement, AECB regulatory document R-10 [AECB1977],
which requires two shutdown systems “unless otherwise approved by the Board”.
The overall unavailability for each of these systems is required to be less than
10-3 years per year.
 ECCS and containment: The relevant AECB regulatory documents [AECB1991b,
1991c] require that each of these systems have an unavailability of less than 10-3
years per year.
These numerical requirements are derived from the need to achieve what is deemed to be an
acceptably low release rate of radioactivity to the public following a number of postulated
accidents, or design basis events (DBEs). In other words, these are not just arbitrary numbers.
2.2 Safety Requirements for the Process Systems

The process I&C systems also play a role in plant safety. They have to keep the process systems
operating within their designed operating envelope. If a convincing case could be made that the
process systems would have an extremely high probability of success in doing this, there would
be no need for the special safety systems. However, the reality is that, unless the design of the
reactor and its associated process systems is inherently fail-safe, independent special safety
systems are the only credible way to achieve an acceptable level of safety.
The unavailability targets for the special safety systems are predicated on their not being
challenged very frequently by failure of the process systems. For this reason, the process
systems do have some safety requirements, depending on the safety-related role of the process
system in question. The system singled out for special mention in the case of CANDU reactors is
the reactor regulating system, which has its own CSA Standard [CSA2011a]. The requirement
for this system, per 4.3.1.1 of that Standard, is that:
“The design target for failure of reactor power control shall be established by probabilistic
safety analysis methods.
Notes:
In CANDU nuclear power plants, the design target frequency for loss of regulation is his-
torically less than 1 in 100 years.”
A “loss of regulation” (LOR) is defined to be “a failure resulting in an unplanned increase in

bulk reactor power”.
Unlike the AECB/CNSC regulatory documents, this CSA Standard has recently been updated to
reflect the possibility of non-CANDU designs being licensed in Canada. That “1 in 100 years”
number used to be a requirement. Now, whatever design number is used for failures to control
reactor power has to be reflected in the overall plant safety analysis.
Because any “unplanned increase in reactor power” should be terminated by the shutdown
system, one can conclude that a loss of regulation has by definition occurred if either of the
shutdown systems trips, unless it can be shown that the shutdown system itself activated
inadvertently. In many cases, it is possible to analyze the event in retrospect and to demon-
strate that the shutdown-system intervention was unnecessary, but this analysis still has to be
carried out to confirm that the process I&C systems design was not at fault.
In earlier versions of the referenced CSA Standard, which were in force at the time the current
Canadian reactors were designed, there was a requirement that “each reactor unit shall be
designed and operated such that the combined frequency of all serious process failures does
not exceed 1 in 3 years”. This number applied to the entirety of all process systems. In an ideal
world, the I&C component of each process system would be allocated a portion of that once in
three-year budget, and a reliability analysis would be performed as part of the design to dem-
onstrate that this requirement was met.
Because an LOR is by definition attributable to RRS, it is apparent that the Standard has already
taken care of the allocation of the RRS portion of the budget—namely 1 in 100 years. This
includes the entire RRS, from sensors to reactivity control mechanisms. Breakdown of this 1 in
100-year number to each RRS subsystem, including the I&C subsystem, is quite correctly left to
RRS designers.
Systems design in the 1970s and ‘80s was not pursued in as rigorous a manner as is now consid-
ered to be best practice, and the derivation of the safety-related and other reliability require-
ments for the various process I&C subsystems is typically not fully documented in the design
record of CANDU plants. Until recently, design of process systems in the Canadian nuclear
industry did not generally follow the detailed process of requirements analysis and allocation
which is standard practice in the aerospace and software engineering industries, for example,
although these techniques are used in the key areas of plant safety analysis and in the develop-
ment of safety-critical software.
There are many qualitative requirements included in N290.4, but the reliability requirement
quoted above is the only performance parameter stated.
2.3 Production Requirements for the Process Systems

Clearly, for production reasons, quite apart from safety implications, it is desirable that the
process systems be sufficiently reliable that they do not result in frequent power reductions or
shutdowns. It should be mentioned at this point that the physics of the CANDU reactor are such
that, if the reactor is shut down from full power, the buildup of Xe135 in the fuel will result in the
accumulation of so much negative reactivity that, if the reactor is not restarted within about 30
minutes, it will not be possible to start it up again for about 40 hours. This is known as a “poi-
son-out” and is discussed in more detail in Chapters 5 and 13 (Section 5.3.1). Because many of
the control systems must be available if the plant is to run for more than a few seconds, it can
be seen that anything longer than a brief outage of one of these key systems will have a severe
economic impact. Poisoning out is not an issue for the PWR/BWR.
Production availability is not, of course, a regulatory requirement, and therefore there are no
regulatory documents which pertain to this key attribute. It will be seen in subsequent sections
that the control systems are designed to fail safe, which will result in a shutdown of the reactor
without the need for special safety-system action. It will also be seen that much of the key
process control logic is implemented in a pair of dual-redundant digital control computers
(DCCs). Although there is no specific requirement for availability of the process systems in
general, these systems were designed to be highly reliable and tolerant of individual component
failures. The DCCs were designed not to fail in a manner that leads to a poison-out more than
once in three years. Note that, while this reads somewhat like the “serious process failure”
requirement described in the previous section, the two requirements are quite distinct because,
due to the fail-safe design of the process I&C systems, most process I&C outages will not result
in a serious process failure as defined in Section 2.2.
3 Overall I&C Architecture

3.1 Architectural Design of the Special Safety Systems I&C
The architectural design of the special safety systems reflects the requirements for low unavail-
ability discussed in Section 2.1. It was stated earlier that the CANDU design includes two
independent shutdown systems. Independence is assured by geographical separation of the
systems, including their I&C components, and by equipment and design diversity. For example,
nucleonic sensors and actuators for SDS1 penetrate the core vertically from above the reactor,
on the reactivity mechanisms deck. Nucleonic sensors for SDS2 are located at the side of the
reactor, where the poison injection takes place. The actuation technologies are completely
diverse, with mechanical neutron-absorbing rods used on SDS1 and injection of a neutron-
absorbing solution into the moderator for SDS2.
To maximize physical separation between systems which must be physically independent, all
safety-related systems in the CANDU plant are divided into two groups. Group 1 systems
include most process systems and SDS1. Most of the other special safety systems, including
SDS2, are allocated to Group 2. Within containment, because both groups must, after all,
interface with a single reactor, it is possible to separate systems only by adopting approaches
such as the horizontal/vertical separation described in the previous paragraph, but outside
containment, the two groups are assigned completely separate locations. SDS1 I&C equipment,
for example, is located in the control equipment room, adjacent to the process systems I&C
equipment, but SDS2 I&C equipment is located in the secondary control area, on the other side
of the reactor building. Penetrations through the reactor building wall for the two groups are
separated by 90°. For a more complete discussion of grouping, see Chapter 13, Section 5.2.8.
This physical separation guards against relatively localized common-mode events such as
mechanical destruction, fires, and flooding which might otherwise disable equipment in both
groups simultaneously.
Not all common-mode events can be addressed by geographical separation and technological
diversity. The sensors and actuators must be able to perform their functions during the acci-
dent conditions against which they are required to provide protection, which in many cases will
impact both systems simultaneously (e.g., high radiation fields, temperature, and humidity).
The equipment must therefore be qualified to function in the anticipated post-accident envi-
ronment. Equipment must also be qualified to survive and function during more widespread
common-mode events such as electromagnetic disturbances and earthquakes, against which
only limited physical protection is possible.
The low unavailability of each special safety system is achieved in part by redundancy of the
equipment within each system. To ensure independence of the instrumentation, it is channel-
ized. This involves physical separation of the instrumentation, cable routes, logic equipment,
and actuators, and the electrical supplies that power them. Typically, the special systems
instrumentation and actuation logic is divided into three or more separate channels. The
separation of the electrical supplies closely reflects the separation within the I&C systems, as
described in Chapter 11.
Channelization minimizes the probability that many classes of events will disable more than one
channel at a time. Because safety-system action requires two of the three channels to call for
activation of the safety function, the system will continue to perform its design function even if
the third channel has failed in an unsafe manner. The design of the individual channel logic is
such that unsafe failure of even a single channel is unlikely.
Channelization also enables the logic in each channel to be tested periodically. A channel under
test is placed in a state where it votes for safe action. In calculating the system unavailability,
each component failure is assumed to be random and is detected by a test carried out at a
specific test interval. On average, then, failures may go undetected for one-half this test inter-
val. Therefore,
unavailability = failure frequency x test interval / 2.
It can be seen that the test frequency of functions that are normally poised but inactive is a key
input to the unavailability calculation. Unless such functions are tested periodically, any system
which depends on them must be assumed to be unavailable.
The detailed requirements for channelization in CANDU plants are documented in a series of
safety design guides which are not in the public domain.
In PWR plants, a single shutdown system is normally used, with four channels of sensor and
actuation logic. A reactor trip will result if any two of the four channels call for a trip. For
CANDU plants, two three-channel shutdown systems are used. A reactor trip will result if any
two of the three channels in a given shutdown system call for a trip. Either shutdown system
will trip the reactor if the logic for any trip parameter calls for a trip in at least two channels.
This is known as “general coincidence” logic. It is also possible to design a system in which a trip
will occur only if the logic for the same parameter in at least two channels calls for a trip. This
design is referred to as “local coincidence logic” and is used in some other CANDU plants. It can
be seen that, although local coincidence logic is less prone to spurious trips, it requires more
inter-channel communication, which compromises channel separation to some extent and adds
complexity, particularly when the logic is hard-wired. In practice, the spurious trip rate has
been found to be sufficiently small that the additional complexity of local coincidence logic is
not warranted.
3.2 Architectural Design of the Process Systems I&C

The redundancy approach used in the special safety systems to minimize their unavailability is
also used in the process systems to enhance the availability of key control functions, thus
minimizing losses of production. Key control functions typically use triplicated sensors, with
voting logic to reject sensors which have failed. Control logic is duplicated or triplicated, and
much of the process equipment itself (pumps, valves, etc.) is either triplicated, using three 50%-
capacity units, or duplicated using two 100%-capacity units. However, the separation require-
ments applicable to special safety systems do not apply to the process systems because the
latter are not required to operate following an accident.
The architecture of the power supplies for the process I&C reflects that of the I&C equipment.
3.3 Features of the CANDU DCC Design to Enhance Production

As stated in Section 2.3, the economic penalties of even a brief process I&C outage in a heavy
water reactor may well extend far beyond the duration of the equipment outage itself, and
therefore, for both safety and production reasons, triplicated or duplicated equipment is exten-
sively used.
The DCCs are a prime example of equipment duplication. The DCCs normally act in a mas-
ter/standby configuration, with only one DCC in control of a specific function at any given time.
The individual control logic subsystems (or “control programs”) running on the DCCs perform
their own internal self-checks. Normally, the programs running in the master DCC will be in
control of the plant. If a control program in the master DCC determines that it does not have
sufficiently reliable input signals to enable it to compute its outputs, that program will stop
running, causing control of that specific function to transfer to the standby DCC. Plant control
can continue with certain restrictions, with some functions being controlled from the master
DCC and some having switched over to the standby machine. This feature is also used to reduce
the risk of plant outages when a change is made to the control logic. The new logic can be run
in the master DCC only, with the old logic available to take over in the standby DCC at the
operator’s discretion if problems are encountered.
3.4 Safety-Related Functions of the Process Systems

The frequency of loss-of-regulation events can be reduced by incorporating logic into the
process systems design to detect potential loss of control situations and to take action to reduce
reactor power independently of the shutdown systems. In the CANDU design, this is accom-
plished by fail-safe design of the process I&C logic and by two layers of safety-related logic—the
setback and stepback functions—implemented in the control computers. “Failing safe” in this
context means failing the output devices in the direction of shutting down the reactor and
positioning the process systems for post-shutdown conditions.
3.4.1 Setback function

The setback logic continuously monitors a number of process parameters. Typically, these are:
 High local neutron flux;
 Spatial control outside the normal range of operation;
 Low de-aerator level;
 High steam generator pressure; and

 Upsets in moderator temperature or pressure.
If there is an indication that these parameters are straying outside acceptable limits, a reactor
power setback is initiated by gradually ramping back the power set-point generated by the set-
point logic (see Section 4.1). In many cases, reducing reactor power will return the process
system to its normal operating condition. Once a setback is terminated, operator intervention is
required before any further automatic adjustment of reactor power set-point is possible.
3.4.2 Stepback function

Some abnormalities in the plant control systems require more rapid action than is available
using a power set-point manoeuvre. The following conditions are monitored:
 Reactor trip;
 Turbine trip;
 Loss of line (grid connection);
 Heat-transport pump trip;
 High heat-transport pump pressure;
 High flux power or high flux rate; and
 Low steam generator level.
If any of these conditions is found to be present, an immediate power reduction, or “stepback”
in reactor power, is initiated by releasing the clutches that hold the mechanical control absorb-
ers out of the core. The stepback is normally permitted to continue to zero power, but in the
cases of turbine trip or loss of line, it is arrested at an intermediate power level by catching the
rods in mid-fall. This partial power reduction permits continued reactor operation at a power
level that prevents the growth of xenon in the fuel from shutting down the reactor.
Stepback is the exception to the transfer-of-control approach described in Section 3.3. Both
DCCs must initiate a stepback before one will take place. This reduces the risk that the stepback
function will act spuriously. If a given DCC is shut down or the stepback function is not running
on it, then the remaining machine can initiate a stepback on its own.
Although the stepback initiation logic resides in the DCC, it is independent of the RRS logic and
therefore can mitigate some high flux power or flux rate events caused by failure of RRS itself.
The stepback function reduces the frequency of demands on the two safety shutdown systems,
SDS1 and SDS2.
3.4.3 DCC self-check function

Each DCC incorporates self-checking logic to confirm the availability of certain key functions.
For minor losses of capability, such as loss of individual instrument measurements, the system
will continue to function normally. However, if a major loss of capability is detected, the DCC in
question will be shut down by allowing the computer’s watchdog timer to time out. This results
in all its outputs being de-energized, which, if the DCC in question is the master, transfers
control to the standby machine.
An external watchdog timer, which must be reset every few seconds, is incorporated into each
DCC. If this fails to happen, as would be the case, for example, if the DCC software got stuck in a
loop, the watchdog timer would time out, and the DCC’s outputs would again be de-energized.
The sense of the actuation logic for all systems controlled by the DCC is chosen such that de-
energization of the DCC outputs will tend to move the affected process system in a safe direc-
tion. If both DCCs fail, for example, the mechanical control absorbers will drop into the core,
and the light water zones will fill, resulting in a reactor shutdown.
Although the DCC itself is not seismically qualified, this watchdog timer function is. Hence, it is
sometimes referred to as a “seismically qualified stepback”. This ensures that, the DCCs will be
reliably disconnected from the process equipment they control should they fail as a result of a
seismic event.
4 Overall Plant Control Functionality

The CANDU approach to overall plant control will now be described as an example of the overall
plant control required for a nuclear power plant. Note that CANDU reactors are relatively more
manoeuvrable than LWRs because they do not have to respect the rather slow manoeuvring
rates imposed by a very thick-walled pressure vessel. The manoeuvrability of LWRs also tends
to be restricted at the start and end of their fuel cycle.
Although most of the 200-plus systems in the CANDU plant contain an I&C element, three main
I&C subsystems are involved in the control of the nuclear steam supply: the reactor regulating
system (RRS), boiler pressure control (BPC), and the unit power regulator (UPR). Due to the
complex nature of the control logic used, all three of these functions are implemented primarily
in the DCCs. The term “boiler”, by the way, was changed at some point to “steam generator”,
but the name of the DCC program has stuck.
The plant can be operated in two modes, as specified by the operator, the choice being repre-
sented by the position of the A/N (alternate/normal) switch shown in Figure 1:
1. In Normal mode, the operator specifies a plant electrical output in megawatts. In this
mode, the UPR logic adjusts the governor valves supplying steam to the turbine to main-
tain the specified electrical output. The BPC function monitors boiler pressure, and any
error will result in changing the requested reactor power set-point which is sent from
BPC to RRS.
2. In Alternate mode, the operator specifies the reactor power output, in percentage of full
power, and a manoeuvring rate. The alternate mode set-point is then moved to meet
this request at the specified rate. RRS sets the reactor power, which determines the en-
ergy delivered by the primary heat-transport system to the boilers. BPC then manipu-
lates the turbine governor valves to maintain boiler pressure at a fixed set-point. BPC
also has control of the condenser and atmospheric steam discharge valves, which can be
opened to receive steam if the turbine load is lost, thus isolating the remainder of the
system from abrupt changes in generator load and permitting continued reactor opera-
tion without the turbine as a load, thus avoiding a reactor shutdown due to buildup of
Xe135.
Figure 1 CANDU overall plant control
4.1 RRS control logic

RRS control logic in CANDU accomplishes the following:
1. Because the CANDU core is relatively large and subject to spatial instability due to
the dynamic action of Xe135, it has been found necessary to divide the reactor into 14
zones for control purposes (see Figure 2, which shows the arbitrarily labelled “A” and
“C” ends of the reactor). There are seven zones at each end of the reactor core.
Figure 2 Arrangement of the 14 control zones

The power measurement and calibration logic determines the current reactor power
in each of these zones, based on the readings provided by 28 platinum-clad Inconel®
flux detectors (two in each zone) that penetrate the reactor vertically from the reac-
tivity mechanisms deck. The flux power is estimated for each of these 14 zones.
Because the platinum-clad Inconel detectors (see Section 6.1.1) do not provide an
absolute value for neutron power, the zone power estimates have to be calibrated
against thermal power. This is determined from measurements of temperature rise
between the reactor inlet and outlet headers at low power, and from secondary-side
measurements of steam flow, feed-water flow, and feedwater temperature at higher
power (> 70%), because boiling in the channels renders the primary-side measure-
ments unreliable at these power levels.
2. The appropriate power set-point is selected based on plant mode (Normal/Alternate,
see Section 4). This set-point will be overridden if a setback is required (see Section
3.4). In this event, the reactor control mode is automatically switched to alternate
mode, and the power set-point is ramped down to a predetermined endpoint. If the
operator presses the HOLD POWER button on the RRS panel, automatic adjustment
of the power set-point will also be suspended.
3. The demand power logic compares the overall reactor power measured in item (1)
above with the required set-point and generates a power error signal Ep, which be-
comes the basis for the control of the reactivity mechanisms described next.
Note that internally RRS works with logarithmic power, expressed in decades. This is
appropriate because the reactor is a multiplicative device and the reactivity-control
mechanisms control the rate of neutron multiplication. The reactivity mechanisms
are driven based on Ep, which is also a logarithmic variable, measured in decades.
However, in many explanations, and indeed in the displays seen by the operator, the
Ep axis is labelled as “Power Error %” (see, for example, Figure 3). What is implied is
that the value is a percentage of current power, not of full power.
4. During normal steady-state operation, the reactivity in each of the 14 zones is ma-
nipulated by adjusting the level of light water in each of 14 cylindrical compartments,
one located in each zone (see Figure 2). In a heavy water reactor, light water is a
neutron absorber. Each of the 14 compartments has a fixed outflow and a valve
which controls the inflow of light water. These valves are manipulated in response to
a combination of the overall power error, Ep, and any deviation of zone flux from the
average over all 14 zones. The total worth of all 14 light water zones is about 7 mk.
The logic to control overall power is run twice per second. Tilt control is updated
every two seconds.
In a zone where the overall error and the spatial components of the error signal sum
to zero, inflow will equal outflow, and the light water level will remain constant at
some intermediate level. The level of light water in each zone does not contribute to
valve opening. However, as the level approaches either completely full or com-
pletely empty, the controlled variable changes from zone power to light water level
so that the zones never completely flood or drain. These light water zone controllers
are described in Section 6.2.
5. If more negative reactivity is required than the liquid zones are able to provide, as
indicated by the zones becoming close to full, then this is accomplished by the me-
chanical control absorber logic, which drives banks of neutron-absorbing rods into
the core. These absorber rods are normally located outside the core above the reac-
tor. There are four mechanical control absorbers, with a total worth of 11 mk. The
switching logic for them is shown in Figure 3.
Figure 3 Absorber-rod drive switching

6. Similarly, when there is a lack of positive reactivity, the adjuster control logic can
withdraw banks of adjuster rods from the core (see Figure 4). These rods normally
reside in the core. There are 21 adjuster rods arranged in seven banks, with a total
worth of 15 mk. All adjusters in a given bank are driven simultaneously, but only one
bank is driven at a time.
Figure 4 Adjuster-rod drive switching

The speed at which adjuster and absorber rods are driven is determined by a vari-
able-frequency power supply and is a function of power error, as shown in Figure 5.
Figure 5 Rod drive speed

Once drive of either adjuster or absorber rods is started, it proceeds to completion to
avoid large top-bottom flux tilts.
The absorber rods also incorporate clutches, similar to those used on the SDS1 shut-
down rods. When a stepback is required, these clutches are disengaged, allowing
the absorbers to fall into the core under gravity. In this way, a much faster power re-
duction is achieved than if the rods were driven electrically.
In the long term, if the light water zones tend to move significantly away from the
50%-full point, the operator can adjust the available reactivity by adding poison to or
removing poison from the moderator.
4.2 BPC Control Logic

During operation at significant power levels, the goal of BPC is to maintain a constant, prede-
termined boiler pressure. The BPC program also incorporates logic to handle the warming up
and cooling down of the plant, but this will not be discussed here in the interests of brevity.
In Normal mode, BPC will pass a power set-point to RRS to maintain constant boiler pressure
despite possible variations in the steam taken by the turbine. Hence, the reactor is said to
follow the turbine in this mode. The set-point includes proportional and integral terms based
on pressure error.
If boiler pressure becomes excessive, BPC can manipulate a combination of condenser steam
discharge valves (CSDVs) and atmospheric steam discharge valves (ASDVs) to control boiler
pressure. The CSDVs cause steam to bypass the turbine and go straight to the condenser. The
ASDVs discharge steam directly to the atmosphere.
In Alternate mode, BPC also controls the steam fed to the turbine by manipulating the turbine
governor. In this mode, the reactor power set-point is specified by the operator, and the result-
ing thermal power generates steam in the boilers. The objective of BPC is to maintain constant
boiler pressure. Any deviation in boiler pressure will result in the governor valves being ma-
nipulated to maintain boiler pressure, and hence the reactor is said to lead the turbine in this
mode. The algorithm in this case uses proportional and derivative terms based on pressure
error, plus a feed-forward term based on the rate of change of reactor power.
In addition to manipulating the governor, BPC drives the turbine load limiter, which reads the
current load set-point and drives the load limiter to a point 100 MW above this value.
4.3 UPR Control Logic

When the plant is operating in Normal mode, UPR manoeuvres the turbine load towards the
target load specified by the operator. It maintains a variable LR (load reference), which is
ramped towards the target load at a rate which is also chosen by the operator. The actual
turbine load is compared to this load reference, and the turbine speeder and load limiter are
adjusted accordingly.
In Alternate mode, UPR has no control function. It fulfils a monitoring role only.
5 Special Safety Systems Functionality

The four CANDU special safety systems will now be described.
5.1 Shutdown Systems 1 and 2

Safety analysis determines a safe operating envelope for the plant. The values of a number of
plant parameters are then monitored in each shutdown-system channel to confirm that this
envelope has not been exceeded. If any parameter exceeds a predefined trip set-point, then the
channel votes for a reactor trip. To guard against modelling errors, the CANDU design uses two
diverse parameters in each shutdown system where practical to protect against each postulated
initiating event. The case of impracticality has been invoked in the regional overpower protec-
tion (ROP) logic, which provides the primary defence against a slow loss-of-regulation accident.
The ROP sensors use the same technology in both SDS1 and SDS2, and there is no backup
nucleonic parameter for a slow loss of regulation, although the heat-transport high-pressure
parameter does provide some protection in this case. The log-rate trip provides some diversity
for fast LORs.
Most of the parameters that trip the shutdown systems are sensed using conventional process
instrumentation, which provides fairly accurate indications of process values. However, in the
regional overpower detection function, the platinum-clad Inconel sensors provide a much less
direct indication of the physical parameter of interest: the onset of dryout in each bundle in the
core. The logic used in the ROP trip will be described in Section 7, after the limitations of the
sensors used have been introduced in Section 6.1.1.
5.2 Emergency Core Cooling System

The purpose of the Emergency Core Cooling System (ECCS) is to provide an alternate means of
cooling the fuel when a loss-of-coolant accident (LOCA) has occurred. For a discussion of LOCAs
of various degrees of severity, see Chapter 13, Section 5.4.
Emergency cooling occurs in two phases:
1. The ECCS monitors heat-transport system pressure and initiates high-pressure coolant
injection whenever this pressure drops below the set-point. Valves are opened, which
results in light water being forced into the reactor inlet and outlet headers, using pres-
surized helium as the motivating force. This is followed by a medium-pressure coolant
injection phase in which light water from the dousing tank becomes the source of cool-
ant.
2. When the supply of light water is exhausted, low-pressure recovery is initiated. This in-
volves collecting spilled liquid from the reactor-building sump and recirculating it into
the core through the reactor inlet and outlet headers. Both medium pressure and low
pressure coolant injection are effected by pumps powered by Class III electrical power.
The sensors and logic used by the ECCS are conventional, but must of course be qualified to
function in a LOCA environment.
5.3 Containment System

For a full description of the containment system, see Chapter 13, Section 5.5. The containment
itself is the envelope designed to contain any release of radioactive material. In normal opera-
tion, it is maintained at sub-atmospheric pressure by vacuum pumps which exhaust to the
atmosphere through a filtration system. Following a significant release within the reactor
building, which is detected by high pressure, radiation, or both, the containment must be
isolated. Any ensuing rise in reactor-building pressure is limited to facilitate the task of keeping
leakage within allowed limits. On detection of a release, the containment is isolated by means
of dual valves or dampers incorporated in every line or duct that penetrates the containment
envelope. The dampers isolate the reactor building ventilation ducts and are pneumatically
operated. Examples are the ventilation ducts, the spent fuel port, and the feed-water and
steam lines.
The pressure rise is limited by the dousing system. In the upper area of the building, a dousing
tank contains water which is released as a spray to condense steam resulting from a heat-
transport system or steam-line break. Dousing is initiated by opening a combination of electri-
cally activated and pneumatically actuated valves located beneath the dousing tank. Series
valves are used to minimize the probability of inadvertent dosing actuation. Diverse actuation
sources enhance the probability that dousing will occur when required. (This applies to single-
unit CANDU stations. Multi-unit stations use a vacuum building common to all four units. In
the event of a LOCA, the affected containment is connected to this vacuum building, and the
steam emanating from the LOCA site is doused there).
A cooling system limits the temperature rise within containment, thus maintaining the integrity
of the building in the long term. The cooling system I&C is conventional. A hydrogen ignition
system ignites any free hydrogen gas before concentrations can become high enough to be
hazardous.
6 I&C Systems Layout and Equipment

In everything that has been said up to this point, the reader may have gained the impression
that operation of the nuclear unit is entirely automatic. In the short term (up to about 30
minutes), this is indeed the intent of the design. Experience has shown, however, that there is a
role in plant operation for both automated and human control. Humans are not particularly
good at performing repetitive, routine tasks without error, and of course most of the control
loops involved are simply too fast and have too many interactions with other loops for a human
operator to be involved in controlling the loop in real time. However, the human operator has
proved to have useful capabilities when things depart from the routine, where perhaps the state
of the plant is not completely clear, and when available information is either incomplete or
conflicting. He can then assume an overview role and direct the progression of events towards
a stable conclusion. To do this, he has to be involved in the operation of the plant in a supervi-
sory role and be presented with available information in a useful way.
The control room is the centralized point for reactor operation. A geographically separate
secondary control area is provided to shut down the plant and to provide monitoring of those
parameters which confirm safe shutdown following the event, should the main control room
become uninhabitable. The Group 2 I&C equipment is located in this area.
The chief operator’s station in the CANDU control room is the operator’s desk, shown in Figure
6. A separate shift supervisor’s office is also provided, but he is not immediately involved in unit
operation. The operator’s desk is equipped with video display units (VDUs) driven by the DCCs.
The operator can select from a large number of displays which provide information on the
status of various systems.
Figure 6 Layout of a CANDU 6 control room

Behind the operators’ desk are a number of panels designed for stand-up operation. These
panels are organized by system:
1. Containment
2. SDS #2
3. Emergency core cooling
4. SDS #1
5. Moderator
6. Reactor regulation
7. Primary heat-transport system
8. Steam generators
9. Turbine
10. Electrical panels
11. Fuel handling.
Figure 7 shows a view of the CANDU 6 control room, looking from behind the operator’s desk.
This figure shows a somewhat more modern version of the control room than that depicted in
Figure 6. In the centre of the group of stand-up panels are two panels containing large video
displays used to display the overall status of the plant. Above these displays are two large
monitors which present annunciation information.
Figure 7 Pictorial view, CANDU 6 control room

At the top of each function panel is a matrix of alarm windows, some of which are driven by the
DCCs and some of which are hard-wired to so-called “contact alarm units” in the conventional
control logic. These are used to provide system-specific indications of alarm conditions and
include essential alarms which will be needed should the control computers fail. The stand-up
panels themselves contain a mixture of conventional operating controls (switches, lights,
meters, and dedicated PID controllers) and VDUs and keyboards interfaced to the DCCs.
Virtually all CANDU stations have been retrofitted with safety-system monitoring computers to
give the operator warning of diminishing margin to trip. These computers monitor buffered
signals from the shutdown systems and have a dedicated VDU on the operator’s desk. This
enables the operator to take steps to avert a reactor trip. For example, the introduction of new
fuel can result in an unanticipated reduction in ROP trip margin during refuelling operations.
The trip-monitoring computer alerts the operator, who can then manually reduce the RRS set-
point if the margin becomes uncomfortably small.
The CANDU 6 control room uses a “dark panel” approach. When no lights are visible, things are
normal.
Behind the control panels is the control equipment room, housing the physical I&C for Group 1
I&C equipment (see Figure 6). Separate equipment rooms are provided for the DCCs.
The detailed control logic for each system is subject to change in the early design stages of a
new nuclear plant, and therefore a means has been devised to enable much of the required
equipment to be installed before these details are available. Much of the logic equipment, as
well as the sensors and actuators, can be installed and wired to the control distribution frame
(CDF) before details of the interconnecting logic are known. The interconnection logic is imple-
mented later by wiring discrete connecting wires at the CDF. The CDF is, in effect, a giant
junction box about two metres high and several tens of metres long and is a major feature in
the CANDU control-equipment room (see Figure 6).
Sensors are typically located close to the parameters being measured. This results in a large
number of sensors being located within containment, although the electronics may still be
separated from the sensor elements to facilitate maintenance. The electrical signals are taken
through the containment wall by penetrations which are designed to preclude leaks from within
containment along the cables.
Actuators are likewise close to the process equipment being controlled. They may use electrical
power directly or use electrical/pneumatic converters in the case of loads driven by instrument
air. Most of the electrical loads are driven by switchgear located in the turbine building.
6.1 Sensors
For the most part, the I&C systems use sensors that are used in other process control applica-
tions. However, because the subject of this text is a nuclear reactor, a description of the sensors
used for measuring neutron flux is presented below, with particular emphasis on the in-core flux
detectors, which are unique to the CANDU reactor. To ensure the necessary independence, the
sensors used by each special safety system are dedicated to the system in question and are
separate from the sensors, wiring, and associated equipment used by the process systems.
6.1.1 Neutron flux sensors

When starting up and operating a nuclear reactor, it is necessary to measure neutron power
over a very wide range, of the order of 10 decades. During start-up, special neutron counters
are used to monitor neutron flux as poison is removed from the moderator using ion exchange.
After a prolonged shutdown, special fission chambers are also used temporarily. At approxi-
mately 10-7 x full power, ion chambers located outside the reactor core come on-scale. In the
CANDU reactor, control is transferred to the DCCs at this point, and the neutron counters are
removed. The ion-chamber readings provide a logarithmic reading of neutron power, which RRS
uses for control up to around one decade below full power.
In an ion chamber, two concentric boron-coated cylinders capture thermal neutrons. The
resulting alpha emissions cause ionization of hydrogen gas in the space between the cylinders.
A high voltage is applied between the cylinders, which captures electrons caused by this ioniza-
tion. The resulting current (~100 μA at full power) is amplified and provides signals proportional
to bulk linear, logarithmic, and log-rate power which are usable over a wide range (10-7 to 1.5 x
full power).
In the last decade, between 10% and 100% power, the thermal power in the fuel itself is the
variable of interest. A prompt measurement of spatial power is required because any control or
shutdown action must take place before local fuel damage can occur. Unlike most process
variables, there is no single instrument that provides even a close approximation to the power
in each channel of the CANDU core. Yet it is crucial to be able to determine this power because
the thermal power in each channel determines the point at which dryout will start to occur.
Thermal power is closely related to neutron flux, and therefore determining the latter provides
a good estimate of the former.
Neutron flux within the core varies continually, even when the reactor is at constant power, due
to the rather slow oscillations (in the order of hours) in the distribution of Xe135 throughout the
core and to local flux peaks resulting from online refuelling. Although it is not possible to
control flux down to the location of a few fuel bundles, it is feasible to control flux tilts between
the 14 reactor zones. To do this, the flux in each of these zones must be measured locally.
In-core flux detectors are used in the CANDU 6 reactor to measure local flux power for both
control and shutdown systems. They resemble coaxial cables with an Inconel sheath and a
central core made of either platinum-clad Inconel or vanadium in the sensitive section, as
shown in Figure 8.
Figure 8 SIR detector geometry.

Vanadium detectors respond primarily to neutrons, but are too slow for either control or
shutdown purposes (time constant τ = 335 seconds).
Platinum-clad Inconel detectors provide a signal which is less than fully prompt and varies with
age. They respond to both neutrons and γ-rays (proportion approximately 50:50). Because
under steady-state conditions γ-flux is proportional to neutron flux, the detector current may be
regarded as being proportional to neutron flux. However, although the response of the detector
itself to both neutrons and γ-rays is essentially prompt, 30% of the γ-rays which arise indirectly
from β-active fission products are delayed, and therefore the overall detector response to a step
change in neutron flux is basically about 84% prompt, with the remainder being delayed.
All detectors provide currents of a few microamps and are connected to amplifiers which
convert this current to a proportional voltage before passing the signal to the DCCs or to shut-
down-system logic. In the case of the SDS detectors, the amplifiers incorporate variable gains,
which are adjusted daily to reflect the SDS ROP set-points, and electronic compensation to
enable the output signal to compensate for the less than fully prompt detector response (see
Section 7).
On Canadian CANDU 6 reactors, all flux detectors are now of the SIR type (straight individually
replaceable), incorporating a helium atmosphere. The SIR design facilitates maintenance, and
the helium atmosphere has been found to improve reliability.
Neither vanadium- nor platinum-clad detectors provide an absolute response to flux. Their
response changes gradually with age and flux exposure in ways that are not yet totally predict-
able. However, in the short term (e.g., days), their response is repeatable, and therefore if they
are calibrated against thermal power, they can provide values which are useful for estimating
the thermal power generated in the fuel. This lack of an absolute power measurement contrib-
utes significantly to the complexity of both the ROP protection and power control algorithms. A
detailed explanation of this is beyond the scope of this text.
The flux detectors used and their locations are as follows:
 102 vanadium detectors are inserted vertically from the reactivity mechanism deck.
These are used by the flux mapping program in the DCCs. The vanadium detectors
are also the basis for the setback on high local neutron flux (see Section 3.4.1)
 28 platinum-clad Inconel detectors (two for each of the 14 zones) are inserted verti-
cally from the reactivity mechanism deck. In each zone pair, one is in channel A and
one in channel C. These are used by the RRS program, the flux mapping program,
and the stepback program in the DCCs.
 34 platinum-clad Inconel detectors are inserted vertically from the reactivity mecha-
nism deck. These are used by SDS1. The detectors are divided between channels D,
E, and F.
 24 platinum-clad Inconel detectors are inserted horizontally and are used by SDS2.
The detectors are divided between channels G, H, and J. Future builds are expected
to require the same number of detectors as for SDS1.
6.1.2 Other sensors

Most process analog signals are transmitted using 4-20 mA current loops (to help minimize the
influence of electrical noise) from the point of measurement to the central control area. Tem-
peratures are measured using resistance temperature detectors (RTDs) and to a limited extent
thermocouples.
Given the large ground currents to be expected in generating stations, particular care is needed
in sensor wiring to ensure that electrical noise and offset voltages do not corrupt these analog
signals.
Contact sensing in the CANDU design uses 48VDC logic, although 24VDC is more common in
modern process control equipment. This relatively high voltage has been found to minimize
problems with degraded field contacts.
Although some I&C systems have been replaced at various sites with more modern digital
equipment, there are still very few, if any, instances in CANDU plants of sensors which use digital
communication right from the point of measurement. Because the individual stations make
their own design decisions about equipment replacement, they would have to be consulted for
accurate information in this regard.
6.2 Actuators
Actuators in CANDU are primarily electrically driven by on/off 48VDC outputs from the control
system logic. Small loads may be driven directly, while heavy loads are driven by motor control
centres which receive their inputs as 48VDC signals. Some larger valves are pneumatically
actuated by electrical-to-air converters, and therefore in addition to the electrical supply
system, there is an instrument air system with odd and even channels. Because this system is
not described elsewhere, an overview of it is included in Section 6.2.1. Air-driven valves can
typically move much more rapidly than electrically driven ones. For example, on CANDU, large
air-driven valves are used to isolate containment. Air-driven valves typically use local air accu-
mulators to provide a local power source following loss of the main air supply. This power can
be used to drive air-driven actuators to a predetermined safe position following such an event.
There are a number of proportional actuators, such as the boiler feed-water control valves and
the valves which control the flow of light water to the 14 zone controllers.
The CANDU design predated devices that are commanded directly by digital signals sent over a
communications network, and therefore such devices will be found only in areas where the
original equipment has been replaced.
Although most actuators used in the CANDU design are conventional, the actuators which
control spatial power are unique. The reactor is divided for spatial control purposes into 14
zones (see Section 4.1). Each zone contains a cylindrical compartment with a variable amount
of light water (see Figure 9). The level of light water in each compartment is controlled by
varying the flow of water into it, using a valve whose position is determined by a 4-20 mA
controller (normally an output from the DCC). The light water outflow is essentially constant.
The level of water in each compartment is measured by bubbling helium in at the bottom of the
compartment and sensing the difference in pressure between the incoming gas and the pres-
sure in the helium balance header connected to the tops of all 14 compartments (see Figure 7).
This system has the advantage of requiring no active components inside the reactor core. It
works well as long as no compartment either floods or drains. However, flooding or draining
does occur occasionally, and an alternative design which does not have this disadvantage might
be worth investigating.
Figure 9 Liquid zone control compartment (simplified).

The absorber rods used for reactor control are very similar to the shut-off rods used on SDS1
(see Section 6.3), except that they are not spring-assisted and use hydraulic damping to limit
their rate of movement, allowing them to be caught in mid-fall should a stepback to other than
zero power be required,

The adjuster rods have no rapid movement capability. The winches that control their position
are driven by electric motors connected to a variable-frequency supply.
6.2.1 Instrument Air System

Pneumatically operated valves are very common in the process control industry in general.
These valves include on/off and continuous control valves and are typically commanded by an
electrical signal. Compressed air for these valves is supplied by the Instrument Air System, a
service system which, in a manner somewhat analogous to the electrical system described in
Chapter 11, distributes compressed air to the numerous end-user devices throughout the plant.
Like the electrical system, the instrument air system provides channelized sources and distribu-
tion paths to ensure that the appropriate degree of redundancy and separation is available. The
compressed air provides a powerful force multiplier, and operation of the end devices also tends
to be relatively fast. The requirements for both instrument air and electrical power systems are
covered in a common document [CSA2011b].
This description will concentrate on the reactor-building instrument air system because it is of
greatest interest in terms of nuclear safety and reliability. Within the reactor building, there are
about 130 on/off valves and around 70 continuous-action control valves powered by instrument
air. Prominent among the former are the valves which initiate liquid poison injection for shut-
down system #2, the dousing valves, the containment isolation dampers, and the moderator
temperature-control valves, all of which play significant roles in ensuring plant safety. The
reactor-building instrument air system is seismically qualified.
Refer to the greatly simplified schematic in Figure 10.
Figure 10 Reactor building instrument air system (simplified).

During normal plant operation, the reactor-building instrument air system receives its supply of
compressed air from the same bank of compressors that provide instrument air for the remain-
der of the plant. These are located in the turbine hall and run off Class 3 electrical power, so
that air supply is subject to interruption for a five-minute period in the event of a loss of Class 4
power. The air passes into the reactor building through a single line equipped with two isolation
valves in series, which are operable from the control room and provide containment isolation,
typically following a LOCA. A number of the pneumatic actuators inside the reactor building are
required to operate in the immediate post-LOCA period and must be able to operate during the
five-minute interruption caused by loss of Class 4 power. Therefore, a set of three air-storage
tanks is provided inside the reactor building to provide the necessary compressed-air reservoir.
Note that many of the loads are on/off and have to execute only a single movement in this time
frame. The single supply line branches within the reactor building to feed these tanks, each
feed being equipped with its own dual isolation valves and a check valve to preclude loss of the
tanks’ contents in the event of a loss of supply air. Therefore, the three tanks provide a short-
term source of triplicated, uninterruptable instrument air within the reactor building. The air,
once it has been used, is vented to the reactor-building atmosphere and will eventually be
exhausted through the building ventilation system.
Following a LOCA, containment ventilation will be isolated, and continued use of the normal
instrument-air source described above would tend to pressurize the reactor building gradually.
To address this concern and permit continued long-term use of pneumatic actuators in con-
tainment following a LOCA, a separate post-LOCA instrument-air source is provided on recent
CANDU 6 plants. This post-LOCA instrument-air (PLIA) system consists of a compressor located
outside the reactor building and two redundant dryers located inside the building, together with
a number of isolating valves. This equipment is not seismically qualified. The PLIA system feeds
the same reactor-building instrument-air tanks described above through appropriate isolating
valves. The compressor draws its inlet air supply from within the reactor building, thus avoiding
the tendency to pressurize the reactor building. The PLIA compressor will be brought on-line
before the valves isolating the supply of normal instrument air to the reactor building are
closed. The air handled by the PLIA system is potentially contaminated, and therefore the
system components outside containment are located in isolated rooms. Water removed from
the flow, which collects in a tank downstream of the compressor, is returned to the reactor
building active drainage through an isolated line.
6.3 Actuators for the Special Safety Systems

In the case of SDS #1, 28 cadmium shutoff rods are used, which drop into the reactor core under
gravity, assisted initially by springs. They are held out of the core by electromagnetic clutches
which are energized by two-out-of-three logic driven by all three channels, as shown in Figure
11. When any two of the three channels vote for a reactor shutdown, the clutches will be de-
energized, and a reactor shutdown will take place.
To SOR Clutches
Figure 11 Two-thirds voting logic for SDS1 trip
In the case of SDS2, the contents of six tanks containing gadolinium nitrate/D2O solution are
injected into the moderator using pressurized helium as the motive source. The pneumatically
operated valves which connect this pressurized gas to the tanks containing the poison are
arranged in a two-out-of-three configuration which is logically similar to that of the electrical
contacts shown in Figure 11. Again, a trip in any two of the three SDS2 channels will open the
corresponding helium supply valves and trip the reactor.
6.4 Control Logic Technology in the Process I&C Systems

The CANDU design predated the widespread use of digital technology in I&C by perhaps a
decade. Nevertheless, the design relies very heavily on digital control technology, which will be
described below.
Although the digital control computers (DCCs) are arguably the heart of the CANDU I&C system,
they are augmented by a mass of conventional control equipment. For binary control, 48VDC
telephone relays were used in the early designs because these were readily available and highly
reliable. By the time the detailed design of the CANDU 6 was undertaken, the telephone
industry had moved on to solid-state technology, and the telephone relays were replaced by
relays designed specifically for the process control industry. By the time Darlington was built,
programmable logic controllers (PLCs) were becoming available, and that station used a custom-
built PLC to implement much of the Boolean logic.
Much of the closed-loop analog control of the CANDU subsystems is accomplished by general-
purpose analog loop controllers for which the parameters are set up as required for the individ-
ual application when the plant is commissioned. Because these applications are fairly conven-
tional, they will not be described further here.
6.4.1 Digital I&C technologies in CANDU I&C process systems

The computers used to control the CANDU plant do not use off-the-shelf products designed
specifically for digital process control, as is common practice today. At the time of the first
CANDU designs, in the mid-1960s to mid-1970s, the number of successful digital control appli-
cations world-wide could be counted in no more than double digits, and the available technol-
ogy was a general-purpose minicomputer with custom-designed process input/output hard-
ware. “Mini” was a relative term. A minicomputer was the size of a small refrigerator and with
the associated input/output equipment could well expand to be 10 metres long. These com-
puters were not inexpensive, nor were they noted for their reliability. One did not procure
computers in large numbers and deploy them where they were required. Rather, a pair of
centrally located computers configured as master and standby was used, and the necessary
sensors and actuators were connected to them using discrete pairs of wires (as opposed to
communications networks). Networking of computers was unknown. Once the decision was
made to use computers, one looked at what other functions, in addition to those that had to be
there, could be implemented using this expensive but very flexible resource.
The digital computers on the CANDU 6 implement the following functions:
 Closed-loop control for:
o the reactor,
o steam generator level
o steam generator pressure,
o turbine loading,
o heat-transport pressure and inventory, and
o moderator temperature;
 Control of the fuelling machines, which enable on-power refuelling of the reactor;
 Run-up of the turbine;
 Presentation to the operator of annunciation information for all process systems in
the plant;
 Data logging for all process systems in the plant.
A detailed description of the logic implemented in the DCCs is beyond the scope of this text, but
as an example, see the description of Overall Plant Control given in Section 4. The architecture
of the dual-DCC system is shown in Figure 12.
Figure 12 CANDU 6 DCC architecture

All the software was written in assembly language. Although this ensured that the software was
as compact and as fast as possible, the result was a program that relied totally on the program-
mer for its integrity and which was not directly reviewable by other than software cognoscenti.
6.4.1.1 Annunciation and status display

In addition to the inputs needed by the software which controls the plant, several hundred
analog and contact inputs are connected to the DCCs to enable alarm information to be pre-
sented to the operator and to provide him with status information on the various systems, using
the VDU controllers and printers shown in Figure 12.
Alarms are generated within the control logic when abnormal conditions are detected. These
conditions are detected either by hardware logic in the control equipment or by tests built into
the DCC control software. The alarms are presented to the operator using a combination of
illuminated “windows” above the control panels and DCC-driven VDUs in the central area of the
stand-up panel. In addition to the alarms generated by the control functions resident in the
DCCs, the computers monitor thousands of contact and analog signals, enabling their alarms to
be displayed on the VDUs and recorded on a printed log in chronological sequence.
The CANDU alarm annunciation system works well for routine operation and for minor plant
disturbances. One can imagine, however, that during major plant disturbances (e.g., loss of
Class IV power), the operator is faced with a sudden flood of alarm messages, flashing windows,
and horn tones. When Class IV power is lost, many subsystems will no longer be available and
will initiate an alarm in consequence, but all that the operator really needs to know is the major
event: loss of Class IV power. To mitigate this situation, the CANDU alarm-annunciation scheme
uses some very basic prioritization logic. Each alarm is designated as “major” or “minor”. When
certain predefined plant upset conditions are in progress, minor alarms are temporarily sup-
pressed on the VDUs. Although this does help, there is no doubt that a more intelligent alarm
presentation system would be useful. Such schemes have been contemplated in the past and
will probably be incorporated in any new CANDU build, but progress has been slow because the
conditions under which it is appropriate to suppress any given alarm typically require very
careful examination of the operation of the systems involved, meaning that the construction of
the database behind such a system would be very labour-intensive. This would appear to be an
ideal application for some kind of expert system. However, there is great reluctance to imple-
ment any system whose response to a given plant upset is not completely deterministic. The
approach to such intelligent annunciation systems has always been very conservative. One can
imagine the repercussions if it were determined in some post-event investigation that suppres-
sion of a particular alarm had resulted in incorrect operator action.
6.4.1.2 Fuelling-machine control

The on-line refuelling function is unique to the CANDU design. On the CANDU 6 stations,
fuelling-machine control is implemented in the DCCs. On many of the Ontario 4 unit stations,
this function has separate, dedicated computer controls provided by the fuelling-machine
vendor, but again, minicomputer technology is used. Unlike the closed-loop process-system
control discussed earlier, control of the fuelling machines is an example of sequential control.
The computer carries out a step in a sequence and then checks that the expected feedback
indications have been returned from the plant before moving on to the next step. Each step
consists of a single predefined action and a set of post-action checks. A job is made up of a
series of steps. Jobs may be executed automatically from start to finish or may progress one
step at a time under operator supervision. In any event, if the expected feedbacks do not
materialize, execution stops, and the operator is alerted.
The sequential actions required to complete each fuelling activity (i.e., a “job”) are defined in an
interpretive language developed specifically for the CANDU fuelling application. Although not
as readable as a modern sequential control language, this language does provide the reader
with some level of isolation from the underlying assembly-language logic. The operator can
allow the job to proceed entirely automatically or can elect to require his permission between
each step and the next.
Because fuelling is not required on a continuous basis, the fuelling-machine function is only
implemented in a single DCC.
6.4.2 Historical perspective

As an indication of how the technology has changed in 40 years, the CANDU 6 DCCs had 64K
bytes of memory and a one-megabyte disc drive. The central processor cycle time was around
one microsecond. To maximize performance, the machines were programmed in assembly
language. Every byte of software was custom developed. A previously developed operating
system was not used, for example. The result is a system which, although very flexible, requires
outdated skills to maintain and is also particularly prone to errors in software maintenance.
However, all attempts to migrate to more modern equipment have been rejected because of
the financial consequences of any delays or errors in implementation of a replacement system.
Although aging DCC equipment has been replaced, changes have been made at the individual
hardware module level, with custom replacements being engineered to match the fit, form, and
function of the old equipment, and particularly to have minimal impact on the existing soft-
ware. The DCC equipment does not support the standard techniques which have evolved over
the years for the development of highly reliable computer systems. Standards such as
[CANDU1999] would be applied to the systems and software engineering of any new-build
design. The replacement of the DCC platform is undoubtedly one of the greatest challenges
facing the designers of a new-build CANDU.
6.5 Logic Technology in the Special Safety Systems

On all CANDUs before the CANDU 6, the algorithms for making the trip decision were imple-
mented entirely in hardware (primarily by analog comparators). This tends to mean that the
trip set-points for each parameter are constant values because implementation of set-points
which are themselves functions of another plant variable becomes rather complex.
6.5.1 Digital I&C technologies in CANDU 6 shutdown systems

With the need to have two diverse trip parameters for each postulated accident, it became
difficult to identify suitable parameters with fixed set-points without imposing a large penalty
on reactor power or incurring an increased frequency of unnecessary reactor trips. The decision
was made to use set-points for some process trip parameters that were functions of other plant
variables, notably reactor power. Means of generating these set-points using analog electronic
hardware were examined, but ultimately the decision was made to implement the trip logic for
process trips (i.e., trips not based on reactor power measurement) using digital computers.
They were not, however, called computers, but were referred to as “programmable digital
comparators” (PDCs).
6.5.2 Digital I&C technologies in Darlington shutdown systems

Given the success of the PDCs on the CANDU 6, the decision was made on Darlington to go
further. Not only would both nucleonic and process trips be implemented in computers, but
separate computers would be added to support automated testing of the shutdown systems.
Darlington would be the first reactor in the world to feature completely digital shutdown-system
logic. In recognition of this, the term “PDC” was dropped. At Darlington, the equipment was
called “trip computers”. It was recognized that additional steps would have to be taken to
ensure high quality in the software engineering which implemented the trip logic. What was
not recognized was just how onerous this process would become.
In the ten years that elapsed between CANDU 6 and Darlington, the software industry had
matured immensely, and computers were being used to an increasing degree in applications
involving public safety. The experience had not been entirely positive. Lives had been lost as a
result of software errors. In fact, one of the landmark case studies cited in software engineering
texts involved a radiation-therapy machine engineered by AECL which resulted in several deaths
before the software design errors were identified and corrected. Interested readers can Google
the “Therac-25”.
When the design of the Darlington shutdown-system software was undertaken, it was recog-
nized that there was a problem, but the appropriate systems engineering standards had not yet
been written that would establish acceptable ways to address the problem. The Darlington
shutdown systems became a stepping stone in the development of these standards, but the
program encountered serious delays and cost overruns. A product of this effort was a series of
standards written by AECL and Ontario Hydro governing the engineering of software for safety-
related applications at various levels of criticality [CANDU1999].
The international safety-related systems industry has since introduced a standard which is
applicable not only to software development, but also to systems engineering of safety-related
systems using digital technology in general [IEC2001a]. A derivative of this standard applicable
specifically to the nuclear industry also exists [IEC2001b], but it has yet to be broadly adopted
by the Canadian nuclear industry.
Despite the early experience at Darlington, the end result is a system that is well regarded by
operating and regulatory staff.
7 Regional Overpower (ROP) Trip Logic

Description of the ROP logic has been deferred to this point to enable the characteristics of the
sensors involved to be discussed before describing how these are accommodated in the design.
The following discussion is applicable to both SDS1 and SDS2.
The CANDU 6 ROP systems currently have 34 and 24 flux detectors in shutdown systems 1 and 2
respectively. Hence, there are only 8 to 12 detectors per channel, yet each channel is required
to protect against fuel dryout in any one of the 380 channels in the core. As discussed in
Section 6.1.1, the detector signals represent only point values in the distribution of flux
throughout the core. Moreover, the outputs decline with age and are not completely prompt.
The trip comparator associated with each ROP detector is set to a value determined offline by a
computer code “ROVER” to ensure that at least two detectors in each SDS will call for a trip
before dryout occurs at any location in the core. The computer code considers a large number
of normal and abnormal reactor configurations, but does not account for non-equilibrium
fuelling. Typically, with new pressure tubes, this comparator set-point will be around 122%,
although it declines as the pressure tubes age. The code is re-run and the comparators are
adjusted infrequently (e.g., once every three years) to account for changes that affect thermal
hydraulics, such as pressure-tube creep. The comparator setting includes the desired margin to
trip.
The RFSP (Reactor Fuelling Simulation Program) code is run frequently (e.g., once every three
days). It synthesizes a flux map based on the reactor’s power and fuelling history, on reactivity
mechanism positions, and on the 102 vanadium detector readings. The power in each channel j
in the channel-power peaking factor region (all but the outer extremity of the core), CPj, is then
computed and divided by the reference power for that channel, CPref j, to yield the normalized
channel power:
Pnorm j = CPj/CPref j.
The highest Pnorm j becomes the channel-power peaking factor (CPPF, typically around 1.08).
Therefore, the CPPF represents the worst-case power peaking factor due to refuelling ripple
over all channels. The detector amplifier gains are then adjusted so that they read (current
reactor power x CPPF). It can be seen that this approach is quite conservative because it
assumes that all detectors will be affected equally by the channel at greatest risk. This purpose
of this calibration is to compensate for changes in position of the reactivity mechanisms and for
local flux changes due to fuelling. It also incidentally compensates for decreasing detector
sensitivity with age. Typically, one or two detectors per channel will be found to be out of limits
in any particular adjustment cycle.
Each detector amplifier also includes two filters which compensate for the non-prompt re-
sponse of the detector itself. These are not normally adjusted. The compensated signals passed
to the comparators represent a fairly close approximation to the thermal power in the fuel,
which is what causes dryout, and hence is the signal on which the reactor needs to be tripped.
The effect of this compensation is shown in Figure 13.
The set-points for both SDS1 and SDS2 are based on the same criteria, including trip margins,
and therefore it is a matter of chance which system will trip first. Indeed, in the case of a slow
LOR, it is likely that SDS2 will trip first because it has fewer detectors and hence its set-points
must be somewhat more conservative. In the past, there was an incentive to arrange things so
that SDS1 would trip first because it could be re-poised much more quickly and potentially avoid
a reactor poison-out. Nowadays, the analysis associated with restarting the reactor after any
trip is sufficiently time-consuming that a poison-out is inevitable in either case.
Figure 13 Effect of compensation on Pt-clad detector response

Clearly, there are many uncertainties involved in determining the trip set-points, both in the
comparator settings determined by ROVER and in the calibration of the detectors to account for
CPPF. Typical uncertainties are:
 simulation uncertainties
 detector uncertainties
 uncertainties specific to a given SDS channel
 common uncertainties
 CCP uncertainties
 refuelling uncertainties
 uncertainties due to the effects of moderator poison.
These uncertainties are combined statistically and incorporated into the comparator settings,
the intent being to ensure that at least one detector in each SDS channel will trip in a given
shutdown system before dryout occurs, 98% of the time,. As noted in Section 7, two channels
must trip before a reactor trip occurs. The CANDU 6 and Ontario reactors use different methods
of justifying compliance with this 98% criterion, but achieve similar results.
Uncertainty in any trip set-point results in its being assigned a conservative value. In the case of
ROP set-points, this results directly in a loss of plant capacity, and therefore any improvement in
understanding the uncertainties is likely to result in higher trip set-points, which will have an
immediate economic payoff. Even a 1% increase, for example, results in an increased power
capacity of 60,000 MWh per year for a 600-MW station. Depending on the going rate for
power, that is worth several millions of dollars every year.
8 Design Verification
All CANDU systems were subject during the design process to a design review carried out by
their designers’ peers. This review was based on the detailed design documentation produced
for the customer. With each succeeding design generation, the documentation became more
comprehensive. However, the process-systems review was never based on the detailed com-
prehensive list of system requirements which is typical of best-practice systems design in a
modern engineering context.
In the case of the special safety systems, the situation is better because it was necessary to
demonstrate detailed compliance with the safety requirements.
For closed-loop control systems, the dynamic performance of the control algorithms is of major
interest because any instability can require close operator attention at best and opera-
tor/automatic action up to and including shutting down the plant in the worst case. The design
of the CANDU controls was carried out before computers that could simulate plant dynamics in
real time were available. The performance of the proposed algorithms was checked using plant
models and models of the control algorithms run on a mainframe computer in batch mode. For
readers who never knew the days when telephones had wires, a mainframe computer occupied
a large room buried in the bowels of the design office, consumed many kilowatts of power, was
attended by a staff of dozens of technicians, and executed jobs fed to it in the form of decks
consisting of thousands of punched cards. You submitted your job and waited a day or two for
the results to be returned. These took the form of endless columns of numbers printed by a line
printer on reams of fanfold paper.
The plant models used for such verification were not themselves subject to any formal verifica-
tion.
In the case of the special safety systems, dynamic performance was of critical interest because
any unexpected delays in initiating a trip could lead to fuel damage. In this case, the codes used
for verification, such as Cathena, have been subjected to formal validation.
Over the life of the CANDU 6 plants, all sites have added operator training simulators which
closely mimic the performance of the subject plant in real time. Although these models have
not been formally validated to serve as engineering simulations (i.e., as accurate representa-
tions of the process dynamics), it has become common practice to verify any significant control
system changes on the operator training simulator before implementation. This not only
provides some assurance that the changes to the control algorithms will work, but checks any
associated changes to the human/machine interface. The training simulator has been particu-
larly effective in enhancing the level of confidence in changes to the DCC software because the
simulators typically emulate the DCC processor, which means that the new DCC software can be
loaded and run on the simulator. It should be noted, however, that due to possible lack of
fidelity in the plant models used in training simulators, successful operation of the changes on
the simulator does not offer a guarantee that the change will be trouble-free on the plant itself.
In the past, no new-build CANDU plant has had a training simulator installed when it was
originally commissioned. In today’s regulatory environment, availability of an operator training
simulator before commissioning has become a requirement. This will lead to a situation where
it becomes very important to be able to adapt the operator training simulator rapidly to reflect
changes made to the plant design before and during commissioning.
9 Bringing the CANDU I&C Design into the 21st Century

The most advanced CANDU design to be built up to this time is Darlington, which went into
service in 1990. Since that time, the instrumentation and control industry has seen a revolu-
tion. Key to this change has been the enormous progress in digital electronics and communica-
tions. Distributed control system technology is now predominant in I&C system implementa-
tion, and standard computer workstations bring almost unlimited computer power to the
information management functions supporting plant operation.
During the intervening years, at least three new CANDU designs have been embarked upon, yet
none has been developed to the point of implementation. These were the CANDU 3, CANDU 9,
and the Advanced CANDU Reactor (ACR). All these featured I&C technology which was current
at the time. One might think that, given the advances in computing power, the design of the
control systems for these plants would present few problems. Yet this was far from the case.
In a modern distributed control system (DCS), use of which is standard practice in the I&C
industry today, the physical inputs are typically digitized close to the field instrumentation. The
digital data are then carried over a communications network to one of a number of processing
units in which the control algorithms are executed. The control commands are then networked
to output stations located near the controlled actuators. The algorithms in the input, process-
ing, and output modules typically run asynchronously, and inter-processor communication is
also typically asynchronous. Therefore, the time delays from input to output are no longer
deterministic, as is the case for a DCC. If the resulting uncertainties are to be negligible, every-
thing has to run many times faster in a modern system if the same dynamic performance is to
be guaranteed in the worst case.
Modern DCS systems have the advantage of being programmed in application-specific high-level
languages, frequently using graphical representations of the logic, which make the design much
more susceptible to review by outsiders. However, many vendors’ products use proprietary
languages which are not readily portable to a different vendor’s product, should it become
necessary to replace the control system at some stage in the life of the plant. An international
standard does exist for programmable controller languages [IEC1993], but adoption of this
standard has not yet been widespread.
Periodically, the prospect of using some of the more exotic modern I&C technologies such as
expert systems, optimal control, neural networks, and fuzzy logic controllers is raised, typically
by academics. Many of these techniques have been around since the 1960s, after all. The
problem with many of these approaches is that they are typically non-deterministic and there-
fore difficult to qualify. The regulators quite rightly frown on any design that cannot be guaran-
teed to produce the exact same response every time it is presented with the same sequence of
input conditions.
The old adage, “If it ain’t broke, don’t fix it”, is very good advice. Any change to the existing
design must show that it can provide enhanced power output, greater plant availability, and/or
lower operating costs if it is to be seriously considered. In this chapter, the author has at-
tempted to point out some features of the existing design that may benefit from re-design in
any future CANDU new build.
9.1 Technological Obsolescence

Although Darlington is the last clean-sheet CANDU design developed, there have been several
new builds of the CANDU 6, notably in Wolsong, Qinshan, and Cernavoda. Moreover, a number
of existing plants have faced the need to replace control equipment as its reliability declined
due to age and lack of spare parts.
Perhaps the highest-profile equipment items involved which were not susceptible to plug-in
replacement were the shutdown-system computers for the new-build CANDU 6 plants and the
DCCs.
The PDCs were replaced with more up-to-date distributed control system hardware. However,
due to the relatively small amount of logic involved and the fact that the PDCs are not geo-
graphically distributed, the new design was able to avoid reliance on digital communications.
Architecturally, then, the new PDCs were little different from the systems they replaced. The
functions of the safety-critical software running on the existing equipment were re-engineered
for the new platforms. Although this re-engineering was by no means a trivial task, the exis-
tence of well-documented requirements and accepted procedures [CANDU1999] for carrying
out this work made it fairly routine.
The replacement approach for the DCC hardware took a very different approach. The obvious
approach to DCC replacement is to use a modern DCS, as discussed in the previous section.
However, DCC replacement has the potential to lead to a very protracted and costly plant
outage if anything goes wrong during the replacement or if the software implementation proves
to have defects once the plant is brought back online. To minimize this risk, the fundamental
requirement of the plant owners has been to replicate the existing control logic in precise detail.
The existing DCCs contain a large body of purpose-designed software, ranging from the operat-
ing system to dozens of applications whose details are specific to each CANDU site. The most
reliable definition of the functionality of this software is the program listing itself, which is
written in a machine-specific assembly language.
As stated in the introduction, the design and implementation of the analog control loops
implemented in the DCCs are based on classical frequency-domain techniques. At a detail level,
the control logic consists of an interconnected set of gain elements, integrators, and differenti-
ators. However, because these elements are ultimately implemented in assembly language and
are specified in the form of difference equations, the basic structure of the control logic is far
from obvious to anybody reviewing the detailed design. Translating the existing control logic
into the context of a modern DCS platform with a high degree of confidence that the platform
change will not introduce latent errors would be a formidable task. Not only is there a risk of
errors in the interpretation of the software functionality, but the different architectures of the
two platforms make emulation of real-time performance very difficult to guarantee.
Therefore, if a DCS were used to replace these DCCs, the DCC logic would have to be reverse-
engineered and a new DCS-based design developed and re-validated from the ground up. No
plant owner or CANDU 6 replication project manager has accepted this risk. The alternative was
to custom-build replacement computer equipment which would be fit-, form-, and function-
compatible with the existing hardware and which would host the existing software with rela-
tively minor changes.
As it happened, the CANDU industry was not the only one to be confronted with this problem,
and replacement computer hardware based on more modern components had been developed
for both the Pickering A and CANDU 6 DCCs. These “clone” computers served as a starting point
for the refurbishment of these units, a process which is still ongoing.
All this means that all existing CANDUs will run with aging I&C technology for the next couple of
decades and that the first CANDU new build, should there be one, will be the proving ground for
the design, licensing, and implementation of what is now standard practice for the I&C industry,
assuming that the “clone” DCC is not retained.
The design life of a CANDU plant is around 40 years. With refurbishment, this could well extend
to 70 years or more. Each generation of I&C technology becomes obsolete in about 15 years,
particularly if one is speaking about digital equipment. The manufacturers of the equipment
typically will not guarantee to support their product beyond about 20 years, though if one
happens to pick a particularly successful product line, this date might be extended by another
10 years. Therefore, any nuclear plant will have to face replacement of its digital I&C equipment
at least twice in the lifetime of the plant. In the case of the original CANDU DCCs, the original
equipment was specifically developed in Canada by a custom systems integration supplier who
owned much of the design data and was able to purchase the rest from the original equipment
manufacturers. Manufacturers of off-the-shelf I&C products are not interested in this kind of re-
engineering business. However, by far the most economic and reliable solution in the future
will be to base the I&C on off-the-shelf equipment. Design for replacement has not so far been
a requirement on next-generation reactors. This is a challenge that should be accepted. The
existing refurbished CANDU plants will probably have to face this issue with their DCCs at least
once more during their extended lifetimes because it is hard to imagine the present vendor still
providing support in 2040. In future replacements, it is unlikely that development of a replace-
ment platform which can run the existing software will be practical, and therefore the issue of
rewriting the software when the equipment is replaced will have to be faced. If a hardware
platform which supports the industry-standard languages [IEC1993]is used, this offers some
hope that the existing software will be portable. However, the industry in general has shown
little inclination to adopt these languages, meaning that there is no guarantee that this ap-
proach will yield the desired advantages.

By this point, the reader will be aware that the architectural design of the control systems is
heavily dependent on the safety requirements that these systems implement, and on the
necessity to avoid unnecessary interruptions of the plant’s raison d’être – the production of
electrical power. Chapter 13 provides an excellent description of the safety requirements, and
should be read in conjunction with the material presented herein.
Chapter 5 provides a discussion of the reactor physics, including the control actuators which the
RRS and shutdown systems use to control reactor flux.
I&C systems are only as good as the motive power sources that drive them—in the case of
CANDU, primarily Class I and II electrical power and instrument air. The design of the electrical
power distribution system is presented in Chapter 11. A brief description of the instrument air
system is included in Section 6.2.1 of the present chapter.
11 Problems
1. Starting with an equilibrium reaction, assume that a control rod is moved out of core by
a fixed distance. What form will the curve of reactor power vs. time assume?
2. Given the approximate numbers quoted in Section 2 and assuming that the light-water
zone controllers are initially half full, in the absence of any other control or shutdown ac-
tion, what value would the power in a CANDU reactor reach in one second if all the light-
water zone controllers suddenly drained? Assume that initial power = 100% and that the
total reactivity worth of the zone controllers = 7 mk.
3. From the point of view of reliability, why is it important in a reactor with two shutdown
systems that the two systems be independent?
4. In the CANDU design, what is the probability of occurrence of an uncontrolled increase

in reactor power which is not stopped by at least one of the shutdown systems?
5. What is the cost, in lost power production, of an outage leading to a 40-hour poison-out
of a CANDU reactor? Assume that the electrical output = 700 MW and that the utility is
paid 3.5 cents/ kW hour.
6. The two-out-of-four architecture used on many PWRs has one major advantage over
two-out-of-three logic. What is it?
7. Some of the factors that can cause a shutdown system to fail to perform its design func-
tion are:
1. Earthquakes
2. Flooding
3. Localized physical damage to the system
4. Electromagnetic effects
5. Design/analysis shortcomings.
Steps which are taken to minimize the impact of these factors on the system as a whole
include:
a) Multiple equipment groups (wide physical separation)
b) Channelization (physical and electrical separation)
c) Equipment/technology diversity
d) Equipment qualification
e) Multiple activation parameters
f) Fail-safe design (i.e., the safe state is the de-energized state).
Prepare a table assigning one or more of these lettered steps to the numbered factors
above.
8. Prepare a succinct statement which captures the key differences between a setback and
a stepback in the CANDU I&C design.
9. Self-check is not listed as one of the stepback parameters. So why is a self-check some-
times referred to as a "seismically qualified stepback"?
10. If it is assumed that every dual DCC failure leads to a poison-out, that DCC failure is re-
quired to lead to a poison-out not more than once in three years, and that, on average, it
takes 20 minutes to get a DCC back in operation following a failure, what is the design
target for mean time between failures of a single DCC?
11. Name a factor that could cause an MTBF calculation based on hardware failure rates
alone to be very optimistic.
12. The logic to avoid flooding of the light water zones precludes filling them beyond about
90%. What action will RRS take if the zones fill, but a positive power error persists?
13. What sensors drive the logic providing primary protection against slow LORs in a
CANDU? What is the rationale for not providing diverse logic to protect against this
event?
14. What is a key difference between the ECCS and containment systems with respect to
continued availability of electrical power?
15. What would be some key requirements for the secondary control area?
16. It has been suggested that the control system be made capable of initiating a power set-
back on low margin to trip. What possible objection could there be to such a proposal?
17. What are the potential effects of a failure of the air-conditioning system which services
the room housing the DCCs?
18. Why might the CDF be less relevant in the context of modern I&C technology?
19. List some pros and cons of digitizing instrument readings inside containment.
20. When using thermal measurements to calibrate flux-detector readings, what processing
will have to be applied to the flux-detector readings before they are compared to ther-
mal power measurements?
21. Why are ion chambers not used for control in the upper decade (10% to 100%) of reac-
tor power?
22. Why do you think digital sensors have seen limited application in retrofit projects on ex-
isting CANDU power plants?
23. What would be the most demanding requirement facing a potential replacement for the
level measurement in the light-water zone compartments? Suggest some possible tech-
nologies which might be used.
24. What long-term factors might require the operator to adjust bulk reactivity using mod-
erator poison?
25. What characteristics of the CANDU design drove the choice to use digital computer con-
trol before it was a widely accepted technology?
26. What are some negative consequences of having the control logic defined in software
written in assembly language?
27. What would be a major reason that an annunciation scheme based on expert system
technology might be difficult to license?
28. Give an example of a non-proprietary language which would be appropriate to the defi-
nition of sequential control logic.
29. List some advantages to making shutdown-system testing an automatic function.
30. List two reasons why software is often considered to be particularly prone to design er-
rors.
31. Which ROP adjustment accounts for the changing shape of flux within the reactor core?
32. Why will the ROP trip comparator set-points need to be adjusted downwards as the
pressure tubes age?
33. Give an example of a non-proprietary language which would be appropriate to the defi-
nition of (a) analog and (b) Boolean control logic.
34. In what way does the architecture of a distributed control system, as opposed to that of
a centralized computer system, affect the execution timing of the control logic?
12 References
[Rouben2002] Rouben, B. Introduction to Reactor Physics, Atomic Energy of Canada Ltd., Sep-
tember 2002.
[Rouben2008] Rouben, B., Reactivity Coefficients, McMaster University course EP 4D03/6D03,
September 2008.
[AECB1977] AECB. Regulatory Document R-10: The Use of Two Shutdown Systems in Reactors,
AECB, January 11, 1977.
[AECB1991a] AECB. Regulatory Document R-8, Requirements for Shutdown Systems for CANDU
Nuclear Power Plants, AECB, February 21, 1991.
[AECB1991b] AECB. Regulatory Document R-7, Requirements for Containment Systems for
CANDU Nuclear Power Plants, AECB, February 21, 1991.
[AECB1991c] AECB. Regulatory Document R-9, Requirements for Emergency Core Cooling
Systems for CANDU Nuclear Power Plants, AECB, February 21, 1991.
[CSA2011a] CSA. N290.4-11: Requirements for the Reactor Regulating Systems of Nuclear Power
Plants, Canadian Standards Association, October 2011.
[CSA2011b] CSA N290.5-06: Requirements for Electrical Power and Instrument Air Systems of
CANDU Nuclear Power Plants.
[CANDU1999] CANDU. Standard for Software Engineering of Safety Critical Software, CE-
10010STD, Revision 2, CANDU Computer Systems Engineering Centre of Excellence, De-
cember 1999.
[IEC2001a] IEC. IEC 61508, Functional Safety of Electrical/Electronic/Programmable Electronic
Safety-Related Systems, International Electrotechnical Commission, June 2001.
[IEC2001b] IEC. IEC 61513, Nuclear Power Plants – Instrumentation and Control for Systems
Important to Safety – General Requirements for Systems, International Electrotechnical
Commission, March 2001.
[IEC1993] IEC. IEC 61131-3, Programmable Controllers, Part 3: Programming Languages, Inter-
national Electrotechnical Commission, March 1993.
13 Acknowledgements
Werner Fieguth
Dave Fournier
Norm Ichiyen
Alek Josefowicz
1
CHAPTER 11
Electrical Systems
Prepared by
Dr. Jin Jiang1
Department of Electrical & Computer Engineering
Western University
Summary
This chapter covers grid requirements, station power systems, and major electrical components
in CANDU nuclear power plants (NPP). Grid requirements at an NPP location are discussed in
terms of reliability and availability of off-site power, the need for a secure electricity supply for
the electrical generation process, and the role of electricity in ensuring the safety of CANDU
nuclear power plants. The chapter also describes the operating principles of the major pieces of
electrical equipment found in a CANDU plant.
The chapter is divided into four parts. In the first part, general and nuclear safety-based
principles and practices for the design of electrical systems in CANDU 6 plants are listed.
In the second part, the main electrical connection to the power grid is explained. The concepts
of switchyard, protection schemes, grid connection, and synchronization are also addressed
from a CANDU NPP point of view. The chapter considers situations involving electric power
production during normal operation, as well as power consumption for maintaining plant safety
during shutdowns. The relationship between internal station power, generated power, and grid
power is clarified in light of reactor safety.
The third part discusses the internal plant electrical system. The section offers a detailed
classification of power sources by their reliability levels and explains the interrelationships
among them. The section also provides a justification for the classification of these power
sources and introduces the concepts of DC power sources, standby power supplies, and
emergency power systems.
The final section briefly introduces the major electrical systems and devices in a CANDU plant,
including the generator, transformers, voltage/current transducers, and circuit breakers. The
section first explains the operating principles of these systems and devices and then provides
their specific ratings and designs in a CANDU plant.
To facilitate learning, a list of exercises has been compiled at the end of the chapter. The reader
1 with contributions of Sections 1.1, 1.2, 3.3, 3.7, 3.8, and 3.9 from Mr. Alek Josefowicz, P.Eng., CANDU Energy
(Retired)
©UNENE, all rights reserved. For educational use only, no assumed liability. Electrical Systems – August 2015
should attempt to answer these questions to gain further understanding of the materials
presented. Additional information on electrical systems in nuclear power plants can also be
obtained through the list of key references provided at the end of the chapter.
It is important to note that electrical systems may vary slightly in different CANDU plants. For
example, some diagrams may show elements of shared systems, the CANDU 6, as a single unit
design where the design principles exclude sharing except for the switchyard. The main goal of
this chapter is to provide a basic knowledge of electrical systems in a CANDU plant, rather than
to examine details of a specific plant.
Learning outcomes
 The goal of this chapter is to provide students with a clear understanding of the importance of
the availability of electrical power for maintaining the safety of a nuclear power plant under
conditions different from the normal mode of operation, but which are, however, within the
conditions evaluated in the safety analysis report.
 Students should be able to explain why grid power is as important to the safety of an NPP as
the power output from the NPP is to the grid.
 Students should be able to identify any deficiency in the reliability of the power grid at the
power station location.
 Students should be able to read the station power distribution diagram by identifying
different classes of power sources, i.e., Class I through Class IV. They should also be able to
match the names of the safety-related systems with the corresponding power classes.
 Students should be able to describe the relationships among the different classes of power
sources.
 Students should be able to explain the functionalities of both standby generators and the
emergency power system.
 Students should be able to list the major systems involved in power generation and
transmission.
 Students should be able to explain the principles of energy conversion from mechanical
energy to electrical energy through synchronous generators.
 Students should be able to describe the functionality and working principles of the excitation
and cooling systems of the synchronous generators.
 Students should be able to describe the working principles of transformers and

voltage/current transducers, as well as to identify where in the plant they are used.
Electrical Systems 3
 Students should be able to identify and describe the different types of circuit breakers and
disconnect switches.
 Finally, students should be able to explain how the generated electricity is delivered to
millions of customers.
Table of Contents
1 Introduction ............................................................................................................................ 6
1.1 General............................................................................................................................ 6
1.2 Nuclear Safety-Based Design Principles and Practices for a CANDU EDS ...................... 6
2 Electrical Power Grids and their Connection with an NPP ..................................................... 7
2.1 A Holistic View of Electrical Systems between an NPP Station and the Grid................. 7
2.2 Unique Grid Power Requirements for NPP Safety.......................................................... 8
2.3 Switchyard between the Grid and a CANDU NPP Station .............................................. 9
2.4 Summary ......................................................................................................................... 9
3 Electrical Systems Internal to a CANDU Plant....................................................................... 10
3.1 Sources of Electrical Power for CANDU NPP Station Use............................................. 10
3.2 Class Definition of Power Sources ................................................................................ 11
3.3 Channelization .............................................................................................................. 13
3.4 Electrical Power Sources under Different Classes ........................................................ 14
3.4.1 Class I......................................................................................................................... 14
3.4.2 Class II........................................................................................................................ 15
3.4.3 Class III ...................................................................................................................... 16
3.4.4 Class IV ...................................................................................................................... 16
3.5 Load Transfer among Different Buses .......................................................................... 17
3.6 Standby Generators (SGs)............................................................................................. 19
3.7 Emergency Power Systems (EPS).................................................................................. 20
3.8 Grounding and Lightning Protection ............................................................................ 20
3.9 Control of Electrical Loads ............................................................................................ 21
3.9.1 Loads powered from switchgear .............................................................................. 21
3.9.2 Loads powered from the MCC .................................................................................. 21
3.9.3 Class IV and Class III loads......................................................................................... 22
3.9.4 Class II and Class I loads ............................................................................................ 22
3.9.5 EPS loads ................................................................................................................... 22
3.10 Summary ....................................................................................................................... 22
4 Main Electrical Components in a CANDU Plant.................................................................... 23
4.1 Generators .................................................................................................................... 23
4.1.1 Basic principle ........................................................................................................... 23
4.1.2 Generators in a CANDU plant ................................................................................... 26
4.1.3 Excitation system ...................................................................................................... 27
4.1.4 Excitation transformer in a CANDU plant ................................................................. 28
4.1.5 Cooling and protection systems ............................................................................... 29
4.2 Transformers................................................................................................................. 30
4.2.1 Basic principles.......................................................................................................... 30
4.2.2 Major transformers in a CANDU plant...................................................................... 32

4.3 Voltage and Current Transducers ................................................................................. 33
4.3.1 Principles................................................................................................................... 33
4.3.2 Voltage and current transducers in a CANDU plant ................................................. 35
4.4 Switches, Circuit Breakers, and Disconnect Switches................................................... 35
4.4.1 Concepts and operating principles ........................................................................... 35
4.4.2 Switchgear in a CANDU plant.................................................................................... 37
4.5 Summary ....................................................................................................................... 39
5 Summary and Relationship to other Chapters ..................................................................... 39
6 Exercise problems ................................................................................................................. 40
7 References ............................................................................................................................ 41
8 Further Reading .................................................................................................................... 42
List of Figures
Fig. 1 - Relationships between the station power and the power grid [1]..................................... 7
Fig. 2 - Voltage levels within a CANDU plant. ............................................................................... 11
Fig. 3 - Allowable interruption time vs. capacity of the different power classes. ........................ 11
Fig. 4 - Interconnections of different classes of power supplies. ................................................. 12
Fig. 5 - Dual-bus configuration for power distribution systems. .................................................. 13
Fig. 6 - Load distribution on the Class IV bus in a CANDU plant................................................... 18
Fig. 7 - Turbine and generator set. ............................................................................................... 24
Fig. 8 - Illustrative diagram of a synchronous generator.............................................................. 24
Fig. 9 - A two-pole (one pole pair) synchronous generator.......................................................... 24
Fig. 10 – Three-phase synchronous generator. ............................................................................ 25
Fig. 11 - Conceptual diagram of a static thyristor-based excitation system. ............................... 28
Fig. 12 - Basic operating principle of a transformer. .................................................................... 30
Fig- 13 - External appearance of a typical power transformer..................................................... 32
Fig. 14 - Principles of (a) voltage transducers; and (b) current transducers. ............................... 34
Fig. 15 - An over-current protection circuit breaker..................................................................... 36
Fig. 16 - An arc-extinguishing circuit breaker. .............................................................................. 39
List of Tables
Table 1 - Classification of power sources...................................................................................... 12
Table 2 – Channelization............................................................................................................... 14
Table 3 - Equipment supported by Class I power supplies........................................................... 15
Table 4 - Equipment supported by Class II power supplies.......................................................... 15
Table 5 - Equipment supported by Class III power supplies......................................................... 16
Table 6 - Equipment supported by Class IV power supplies......................................................... 17
Table 7 - Characteristics of the main generator in a CANDU plant. ............................................. 26
Table 8 - Characteristics of the emergency power system generator. ........................................ 27
Table 9 - Characteristics of the generator in the Class III power system. .................................... 27
Table 10 - Characteristics of an excitation transformer. .............................................................. 29

Table 11 - Ratings of a main output transformer (MOT).............................................................. 32
Table 12 - Ratings of a station service transformer (SST)............................................................. 32
Table 13 - Ratings of a unit service transformer (UST)................................................................. 33
Table 14 - Characteristics of SF6 circuit breakers at the generator output. ................................ 37
Table 15 - Characteristics of vacuum circuit breakers on an 11.6 kV bus. ................................... 38
Table 16 - Characteristics of vacuum circuit breakers on a 6.3 kV bus. ....................................... 38
1 Introduction
1.1 General
Even though the sole objective of a nuclear power plant is to generate electricity, it takes
electricity to run the entire plant. The electrical power system in a nuclear power plant is the
subject of this chapter. The electrical systems are designed not only for normal plant
operation, but also for conditions other than normal operation, so that plant safety can be
maintained by ensuring continuity of electrical power supplies regardless of transient
disturbances or faults during operation and post-shutdown. The power for an NPP comes from
diverse and reliable power sources that are physically and electrically isolated, so that any
single failure will affect only one source of supply and will not propagate to alternative sources.
Even after the reactor has been shut down, a significant amount of heat is still being produced
by the decay of fission products (decay heat). The amount of decay heat is sufficient to cause
fuel damage if not removed effectively. Therefore, systems must be designed and installed in
the plant to remove decay heat from the core, even in a plant shutdown condition and in the
absence of off-site power sources.
The electrical power distribution system (EDS) is a complete load group distribution system
with two independent off-site power sources, the main turbine generator, and on-site standby
power sources (standby and EPS diesel generators and, in some cases, a station blackout
generator).
1.2 Nuclear Safety-Based Design Principles and Practices for a CANDU EDS
 The EDS needs to be designed in accordance with its safety functional requirements as
defined in the safety analysis, including independent and diverse provisions aligned with
independent safety functions and including provision to supply electrical power to secure
plant safety during both normal operation and accident conditions without losing all on-
site power.
 The divisions of the power supply systems should be physically and electrically separated
from each other, thus ensuring independence among the divisions as much as possible.
 The Group 1 and Group 2 power supply systems should be physically and electrically
separated from each other as much as practically possible.
 The design of the EDS and associated support systems, including I&C, HVAC, and
cooling systems, should follow the classification, independence, redundancy, and
diversity requirements placed on Structures, Systems, and Components (SSCs).
 The EDS should be designed for a wide range of electrical transients which can be
assumed to occur during plant operation and for the assumed environmental conditions.
 The EDS should be designed for or protected from effects of both internal and external
hazards, such as short circuits or loss of the power grid.
The EDS uses commercially available conventional hardware with provisions as dictated by the
need for seismic qualification (SQ), qualification for operation in a harsh environment (EQ), and
radiation hardening. Electrical containment penetrations (ECP) form part of containment.
2 Electrical Power Grids and their Connection with an NPP

2.1 A Holistic View of Electrical Systems between an NPP Station and the Grid
Even though each specific plant may have its own unique characteristics, a typical set of
electrical connections between a CANDU station and the power grid is illustrated in Fig. 1.
Fig. 1 - Relationships between the station power and the power grid [1].
A station can have either a single reactor unit or multiple reactors. During normal operation,
the generated power is fed to the power grid through main output transformers (MOTs). In
addition, a portion of the generated power is also fed back to the units to support electricity
production through unit service transformers (USTs). Furthermore, it is good practice to cross-
link multiple units at the switchyard to increase self-reliance within the station, particularly in
situations where one shutdown unit may need to draw power from other units within the
station to remove decay heat from the reactor, to maintain essential operating services, or to
re-start the reactor as long as it has not been poisoned out by xenon.
When the power from the station units is no longer sufficient or available to meet internal
demand, the station can draw additional power from the grid through station service
transformers (SSTs). This is also the case during a start-up process.
It is assumed that the power grid is stable and that there are other power sources connected to
the grid, which are available when needed to provide power to the nuclear station site itself.
The power flow on the grid can be effectively controlled through grid interconnection and
management systems. The NPP may contribute to voltage and power control in the grid.
However, most existing CANDU power plants operate in a constant-power output mode to
support the base load supplied by the grid.
2.2 Unique Grid Power Requirements for NPP Safety
The main objective of a nuclear power plant (NPP) is to produce electricity to support industrial,
commercial, and residential loads. Electricity is therefore the final product for most NPPs.
However, it is important to realize that about 8% of the electricity produced by the plant is
consumed internally to support power production. This is true for most power plants, such as
coal or gas, although their internal consumption may be significantly lower (<4%). NPPs,
however, have unique requirements for electric power availability. It is particularly important
to have a secure electrical supply when an NPP is in a shutdown state and is not producing any
electricity of its own. Even when the fission process in a nuclear reactor stops, a significant
amount of decay heat continues to be generated from the fission products. The amount of
heat is typically so large that continued cooling is absolutely necessary to protect the fuel
sheath from melting. Pumping cooling fluid through the core removes this excess heat, but
requires an external electrical power source. Hence, the availability of electrical power (from
other units or from the grid) is crucial for the safety of CANDU and other nuclear power plants
both under normal operations and in a shutdown state. This includes situations where
thermosiphoning is used. Electrical power is required in this case to maintain water in the
steam generators, although pumping of primary coolant is not needed.
This unique requirement for electricity requires consideration of different scenarios at the
design stage of NPP electrical systems. In CANDU power plant design, the NPP site must be
chosen so that the power grid at the site has multiple feeders from different and independent
(often geographically separate) sources, as shown in Fig. 1. This requirement ensures that off-
site electrical power sources are available to the station for removing decay heat when the
reactor is shut down and is no longer producing electrical power of its own. In addition to
Canadian standards [1,2] for a CANDU NPP, the International Atomic Energy Agency has also
issued guidelines for selecting suitable sites for other types of NPPs based on the reliability and
availability of off-site power [3,4], as has also The Institute of Electrical and Electronics
Engineers (IEEE) [5]. As explained in Chapter 13, the availability of off-site electrical power will
affect NPP safety analysis.
As a part of the site selection process, the reliability of the grid must be assessed when some of
its generating capabilities are assumed to be no longer available. This is often referred to as the
(N-1) problem [6], where N is the number of available units. A desirable site for an NPP is one
where power delivery to the NPP site is still guaranteed when only (N-1) or (N-2) suppliers are
available.
The main cause of the 2011 disaster at Japan’s Fukushima Daiichi nuclear power plant was a
lack of off-site power due to the earthquake and inadequate protection of on-site standby
power systems against a tsunami. All the plant’s on-site diesel generators operated until they
were damaged by the water brought in by the earthquake-induced tsunami. Hence, the leading
cause of the disaster was the lack of power after the successful shutdown and an initial period
of reactor cooling.
Together with other facilities in a CANDU plant, the electrical systems must also meet the
seismic design requirements and qualification processes as outlined in [7] in Canada.
2.3 Switchyard between the Grid and a CANDU NPP Station
Note that even though a single line is used to show the flow of power in Fig. 1, all lines carry
three-phase power (except DC power lines). All transformers, circuit breakers, and
transmission lines in an AC power grid are three-phase devices. When delivering the generated
power to the grid, the station power must be synchronized with the grid, including the phase
sequence, voltage levels, and AC power frequency. Voltage and current transducers are used
for monitoring and control, and several high-voltage, high-current circuit breakers are placed
between the MOT and the grid connection points.
The switchyard contains numerous control and protection devices to ensure that any faults on
the grid side will not induce major disturbances to the station, and vice versa. There are also
various interlocks to prevent the incorrect operation of power devices, as well as lightning
arresters, grounding protection systems, and switchyard control systems.
2.4 Summary
Even though the main function of an NPP is to produce electricity to supply power to the grid,
unlike other types of thermal power generation systems, an NPP requires an external power
source with on-site backups to remove decay heat from the reactor when the plant is in
shutdown mode and is not producing its own electricity. Therefore, significant design
considerations have been formulated for the electrical systems within a nuclear power plant.
Furthermore, the availability of off-site power also plays a crucial role in nuclear power plant
safety and is one of the most important considerations in the site selection process when
constructing a new NPP.
3 Electrical Systems Internal to a CANDU Plant

3.1 Sources of Electrical Power for CANDU NPP Station Use
Almost all systems within an NPP rely on electrical power to operate. A “defence-in-depth”
strategy for electrical power supplies is to rely on diverse, multiple, and independent sources.
These sources for a CANDU unit are: (1) power generated from the unit itself; (2) power
generated from other units within the same station; (3) off-site power obtained from the grid;
(4) the emergency power supply; (5) the standby power supply; and (6) batteries. The power
sources in a CANDU NPP consist of both AC (alternating current) and DC (direct current) power.
“Defence-in-depth” as applicable to the electrical systems can be stated as follows:
 1st line – normal operation (grid + main generator)

 2nd line – mitigation (standby generators + batteries)
 3rd line – station blackout (batteries + designated alternative source(s))
 4th line – severe accident management (additional, diverse, alternative sources).
These sources are arranged in such a way that they supply power to station systems during
normal operation, as well as during emergency conditions to maintain NPP safety. The
equipment in the station is also graded according to its importance to safety. In an event that
electrical generation is lost, limited alternative power sources will be used first and foremost to
keep the essential safety-related systems operating.
A CANDU plant contains several buses at different voltage levels. The selected voltage levels
might be different in different plants to meet certain country-specific requirements. One
example is shown in Fig. 2, where the output voltage level of the generator is at 22 kV, and the
voltages at the unit service transformer (UST) and the station service transformer (SST) are at
11.6 kV and 4.16 kV as secondary voltages. However, in other designs, these voltages could be
13.8 kV and 4.16 kV. Also shown in Fig. 2 are two connections to off-site power at the NPP site,
one at 500 kV and the other at 220 kV.
Fig. 2 - Voltage levels within a CANDU plant.
Several other low-voltage buses exist throughout the plant and will be discussed further in the
next section.
3.2 Class Definition of Power Sources
Electrical power sources in a CANDU plant can be divided into four levels, according to the
allowed duration of voltage interruption that can be tolerated by the loads they supply. Class I
power supplies loads that cannot be interrupted. Loads on Class II power can tolerate ~4
millisecond interruptions. Loads on Class III can withstand power interruptions of up to 5
minutes, whereas Class IV loads can tolerate loss of power indefinitely. The most critical and
safety-related control and protection systems are powered from Class I and II sources.
Different classes of power supplies provide power to different systems, depending on the
amount of power the systems require and their relative importance to safety.
Typically, the cost per kW will decrease as the power supplies move from Class I through to
Class IV. The power capacities also increase from Class I to Class IV. The allowable interruption
times and capacities of the various power sources are summarized in Fig. 3.
Fig. 3 - Allowable interruption time vs. capacity of the different power classes.
To determine which class of power should be used to supply a specific system, the safety
functionalities of the system must be examined, as well as the economic impact if that supply
were unavailable. General criteria for matching the class of power supply to the load that it
supports are summarized in Table 1. They are expressed in terms of the longest power
interruptions that will not affect the safety of either the NPP or its personnel.
Table 1 - Classification of power sources

Class of power System load characteristics
Class I Power can never be interrupted under postulated conditions
Class II Power can be interrupted up to 4 milliseconds
Class III Power can be interrupted up to 5 minutes
Class IV Power can be interrupted indefinitely
Different stations may have slight variations in electrical power system configurations. An
illustrative diagram showing interconnections in the electrical power system for the different
classes of power sources in a CANDU station is presented in Fig. 4.
Fig. 4 - Interconnections of different classes of power supplies.
To increase reliability further, Class II, III, and IV power are distributed through two separate
power divisions. If a failure occurs on one division, the equipment connected to the other bus
will still be available. In CANDU plants, these two divisions are typically denoted as “Bus A” and
“Bus B” or as “Odd Bus” and “Even Bus”. During design, loads are distributed evenly between
these two divisions.
An example of such a split-bus connection is shown in Fig. 5. A symbol with two circles and an
arch over them represents a circuit breaker. Circuit breakers are used to connect or disconnect
the systems (denoted as loads) and to protect them whenever a fault occurs. The connection
between the Odd and Even buses on the diagram represents two circuit breakers, one on each
bus. To accomplish the connection, both breakers must be manually commanded to close.
Fig. 5 - Dual-bus configuration for power distribution systems.
3.3 Channelization
Important functions use three instrument channels to provide immunity against single instrument
faults. A control channel consists of interconnected hardware and software components that
process one of the duplicated or triplicated signals associated with a single parameter. A control
channel may include sensors, data acquisition, signal conditioning, data transmission, bypasses,
and logic circuits. This defines a subset of instrumentation that can be unambiguously tested or
analyzed from end to end. For safety and high-reliability applications, I&C system design uses
three instrumentation channels with a two-out-of-three voting strategy (i.e., two of the three
channels must be outside the acceptable limits to trip or actuate the system).
To perform on-line tests in such a design, the operator will place the tested channel in a trip state,
resulting in the actuation logic performing a one-out-of-two test on the remaining channels.
Process and safety systems channels are assigned as shown in Table 2.
Table 2 – Channelization.
System(s) Safety Odd (A) Third (B) Even (C)
Group Associated Associated Associated
Channels Channels Channels
RRS and Process 1* A B C
SDS1 1 D E F
ECC (NSQ) 1 K L M
SDS2 2 G H J
ECC (SQ portion) 2 KK LL MM
Containment 2 N P Q
 The channel association also applies to separation of power supplies and cabling. During
normal operation, channels A, B, and C of the UPS supply all their associated channels.
 Group 1 is primarily for power production, and Group 2 is only for safety systems.
Physical separation is required between the two groups.
 Group 2 systems can also be powered from the EPS. Functional and physical separation
is maintained even though in this situation, only one EPS generator supplies one bus from
which the three channelized power sources are derived.
 ‘1*’ denotes non-safety, however, it is associated with Group 1, and
 NSQ means “not seismically qualified”, and SQ means “seismically qualified”.
3.4 Electrical Power Sources under Different Classes

3.4.1 Class I
Class I power is used to supply loads that cannot be interrupted. It is a DC power source with
three independent distribution channels, each backed with battery banks to provide
uninterrupted power to critical loads. To maintain adequate charge on the batteries, each bus
in Class I is connected to power rectifiers, which convert AC power from Class III power sources
to DC to charge the batteries, as shown in Fig. 4. During normal operation, power from the
rectifiers is used to support the load on this bus while charging the batteries at the same time.
Hence, the batteries always remain fully charged when power is available. DC/AC inverters are
also used to convert DC power from Class I to Class II. In the event of a loss of Class III power,
batteries provide a seamless transfer to support the loads without any interruption. Note that
the batteries are capable of supplying the load on the DC buses for only about 60 minutes,
depending on the particular plant design. This is a very critical time window because all Class I
and II power would be lost if Class III power could not be restored within the interval provided
by the batteries.
The loads supported by the Class I power source are very sensitive and are critical to NPP safety
and operation. A partial list of system equipment powered from Class I is provided in Table 3.
Table 3 - Equipment supported by Class I power supplies.
Class II inverters
DC seal oil pumps for generator
DC lube oil pump for turbine generator bearings
Turbine trip circuits
Turbine turning gear
DC stator cooling pumps
Control and protection systems for station electrical distribution systems
Logic, control, command circuits, and operator interfaces for process and
safety systems (48 VDC)
The capacity of the Class I power source is based on the connected load. CANDU plants use
several different voltage levels for this DC power supply, including 48V, 220V/250V, and 400V,
all to meet the needs of the NPP’s various systems. Note that loss of Class I power is one of the
conditions that trigger the shutdown systems.
To prevent service interruption caused by a “single line-to-ground” fault, the 48V DC and 250V
DC systems are ungrounded. Ground fault detectors, which produce an alarm whenever a
ground fault occurs, are provided for each bus.
3.4.2 Class II
Class II power sources are critical to reactor operation. If Class II power is lost, the reactor will
be shut down immediately. Under normal operation, Class II power is obtained from Class I
sources through power inverters to convert DC power to AC power, as can be seen in Fig. 4. If
for any reason the inverters cannot supply a given bus, the Class III power source will be used to
support Class II power distribution.
The Class II power source supports those devices and systems that can tolerate power
interruptions on the order of milliseconds. Some typical systems supported by Class II power
source are listed in Table 4.
Table 4 - Equipment supported by Class II power supplies.
Digital control computers
Reactor regulation instrumentation
Electrically operated process valves (600 V power distribution)
Auxiliary oil pumps on the turbine and generator (600 V power
distribution)
Emergency lighting (600 V power distribution)
Three independent channels of single-phase inverters ensure complete supply independence to

the triplicated instrumentation and I&C. Class II power sources are relatively low-capacity, have
two voltage levels: 120V and 600V, and are available only in AC form.
3.4.3 Class III
Class III power supports large process loads that are unsuitable for Class II power supplies. They
are used mainly to maintain fuel cooling when the reactor is in a shutdown state and Class IV
power is unavailable. It is important to note that the duration of the loss of Class III power
consists of only the time required to start up a standby generator and re-load the Class III
power system, which is normally about five minutes.
Class III power is taken from Class IV power. In the event of total loss of auxiliary power from
off-site sources, the auxiliary power required for safe shutdown will be supplied from physically
and electrically independent diesel generators located on-site. Each power source (the feeds
from Class IV and the diesel generators) is physically and electrically independent up to the
point of connection to the Class III buses. This improves the reliability of Class III power, making
it available even in the presence of partial loss of Class IV power sources.
If the Class IV power source for a unit fails completely, it is still possible to obtain Class IV power
from other units in a multiple-unit station. Once the standby generators are started, they will
provide power to systems supplied by the Class III power source, ensuring that these critical
systems remain functional.
Some typical systems supported by Class III power sources are listed in Table 5.
Table 5 - Equipment supported by Class III power supplies.

Auxiliary boiler feed pumps
Auxiliary condensation extraction pumps
Shutdown system cooling pumps
Turbine turning gear
Heat transport feed pumps
Moderator circulating pumps
Class I power rectifiers
Fire water pumps
Emergency core coolant injection pumps
Instrument air compressors
End shield cooling pumps
Service water pumps
The voltage level of Class III power is 4.16 kV, and its capacity can range from 6 to 8 MWe.
3.4.4 Class IV
Of the four classes of power sources in a NPP, Class IV supplies loads that can tolerate infinite
interruption. This power can come from two sources. During normal operation, Class IV power
is obtained from the main generator through the unit service transformer (UST). Using power
produced internally by the plant’s own generator minimizes the potential impact of
disturbances from the grid. Class IV power can also be obtained from the grid through the
station service transformer (SST) when the UST becomes unavailable.
It is important to mention that even though Class IV power supplies the entire station during
operation, it is not actually required for safe reactor shutdown, although the unit will be shut
down immediately upon experiencing the loss of its Class IV power source.
The loads normally supplied by Class IV power are systems which can tolerate long-term power
outages without affecting the safety of equipment, personnel, or the public. These loads are
not essential to satisfy fuel cooling requirements following a reactor or turbine trip, but are
essential for operation of heat sinks above the shutdown level of reactor power. Some typical
systems supported by Class IV power sources are listed in Table 6.
Table 6 - Equipment supported by Class IV power supplies.
Main boiler feed pumps

Main heat transport circulating pumps
Condenser cooling water pumps
Generator excitation
Heating and ventilation equipment
Normal lighting systems
As can be seen from Table 6, many important systems in a CANDU plant are supplied by Class IV
power sources, and the loss of Class IV power is considered to be a major incident. The
preferred voltage levels for Class IV systems are 13.8 kV, 4.16 kV, and 600 V.
3.5 Load Transfer among Different Buses
As shown in Tables 3 to 6, NPP system loads are distributed among the various buses based on
their size and importance to system safety. Although a detailed examination of each load is
beyond the scope of this chapter, Fig. 6 provides an illustrative load diagram for the Class IV
and Class III power buses.
Fig. 6 - Load distribution on the Class IV bus in a CANDU plant.
Under certain circumstances, it is desirable to shift loads from one source to another. There
are three modes of load transfer:
 parallel transfer
 fast transfer, and
 slow transfer.
These specific transfer schemes are used at the upper voltage level of Class IV to prevent
reactor trip and maintain generation production.
The parallel transfer mode consists of two steps: (1) parallel the new power source to the
existing one, and (2) remove the existing one to complete the transfer. A fast transfer switches
the load quickly (within two power cycles) so that little interruption is observed. The slow
transfer operates after the voltage has decayed to approximately 40% to limit the maximum
voltage that could be applied to a connected load upon re-energization and can be used only if
the supply transformers can tolerate the inrush currents and if the voltage drop does not
prevent loads from being re-accelerated to nominal speeds.
Class IV transfers are manually initiated for normal transfers after start-up or before shutdown
and are automatically initiated for reactor trips, turbine-generator trips, or loss of the
transmission system. These transfers are accomplished by operating the incoming circuit
breakers on the primary Class IV distribution buses to transfer the sources between the unit
service transformer and the system service transformer.
Automatic transfer systems are also incorporated into Class II. They monitor the operation of the
power inverters and under certain conditions, transfer Class II distribution buses to alternative
supplies directly from Class III. These transfers operate within each channel or division of Class
II.
There are no transfers in the Class I system because each channel’s batteries are charged through
two 100%-capacity rectifier-chargers which share the load.
There are no transfers in Class III or in the EPS systems, although it is possible to connect the
Odd and Even main distribution buses manually when, following a loss of Class IV power, only
one standby generator in the system is operating.
Manual source selection is provided for Class I and II power conversion and distribution to
address the condition when, after a loss of Class IV power, only one standby generator is
available to power one Class III division.
3.6 Standby Generators (SGs)
To maintain power to safety, safety support, and heat-sink systems following loss of Class IV
power sources, CANDU stations contain additional on-site power sources. One type is known
as standby generators. These generators are not required to be seismically qualified.
This power source is based on two or more generators driven by diesel engines or combustion
turbines (in the case of Ontario Power Generation). As shown in Figs. 4 and 6, a generator
supplies Class III AC power to each Odd and Even bus at a 6.3 kV level. These generators are
supplied with enough fuel to keep the diesel engines running continuously for an extended
period of time (up to one week depending on a continued supply of fuel). Standby generator
systems have their own compressed air and DC power sources for start-up and will start
automatically upon loss of Class IV sources to maintain power to safety and safety support
systems. The SGs could form a seismically qualified distribution system, but the design has
evolved to create a separate seismically qualified distribution system. The seismically qualified
systems are connected to Class III because that is their preferred source of power and are
isolated from Class III only when the seismically qualified power sources can provide the
required power. The standby generators will also start whenever a loss-of-coolant accident
(LOCA) signal is issued, but will not connect to the buses until a loss of Class IV power occurs.
Standby generators should be up and running within 30 seconds after receiving a LOCA signal,
picking up all designated loads within a further three minutes. One standby generator has
sufficient capacity to supply the required loads.
Because of the critical roles played by standby generators, regular maintenance is critically
important. This typically consists of starting each diesel generator periodically from the local
control panel, paralleling it with the respective division of the Class IV supply, and letting the
generators run for a specified minimum period of time.
3.7 Emergency Power Systems (EPS)
The second set of alternative power sources in CANDU plants is known as emergency power
systems. Unlike standby generator systems, these power sources must be seismically qualified
[7], and they function completely independently of other power sources. Similarly to standby
generators, the emergency power systems start automatically upon the loss of Class IV power
and will also start on a LOCA signal. Under such circumstances, back-up generators provide
power to the NPP’s critical systems to enable reactor shutdown, monitoring, and decay heat
removal. It is expected that the system should be up and running with its intended loads within
three minutes.
The following background is relevant to design decisions affecting the EPS:
 Based on plant licensing conditions, a loss-of-coolant accident (LOCA) is a random event

because the heat transport system is fully seismically qualified and a seismic event,
another random event, is not postulated to occur in the first 24 hours after a LOCA. With
24 hours of operation of emergency core cooling (ECC) and other required safety support
systems, a 20–30 minute break can be tolerated in ECC operation. This time is sufficient
for the operators to transition from the main control room (MCR) to the SCA and to
restart the ECC and the associated systems.
 A total loss of Class IV power coincidental with a subsequent loss of Class III power,
both random events (except at Fukushima where Class III was incapacitated by the
tsunami, which was induced by the earthquake, but this is a different set of design
conditions), but without a LOCA, is a condition in which residual heat is removed from
the reactor by means of steam generators and water from the dousing tank.
Depressurization of the heat transport system is a precondition for this mode of heat
removal. Valves for implementing depressurization and maintaining the required
monitoring are powered from a UPS or by compressed air for some valves. There is
sufficient time for the operators to initiate the EPS to supplement the dousing tank
reserve with an emergency water supply (EWS).
3.8 Grounding and Lightning Protection
The grounding system is required to prevent physical injuries and equipment damage in case of
a fault and to minimize electromagnetic effects from ground fault currents as well as to prevent
interference and to protect equipment from lightning strikes.
Lightning protection is required so that equipment related to the safety of the nuclear power
plant continues to operate and important monitoring devices continue to function when
lightning hits facilities or power lines.
3.9 Control of Electrical Loads
Generally, in a typical CANDU power generating station, the electrical loads are remotely
controlled using the control logic (relay logic) and interposing circuits, both powered from 48 V
DC Class I. The output from the control logic is hard-wired to the switchgear and motor control
centre (MCC) control circuits or to the terminals of a solenoid valve when the valve is controlled
directly.
Major loads have their mode of operation (ON, AUTO, or STANDBY) selected by the operator
from the main control room (MCR) or the secondary control room (SCA). In the AUTO mode,
the load will augment the already running load(s) when the process demand exceeds the
capacity of the running load(s). In the STANDBY mode, the load will replace the normally
running load when the latter fails to operate.
3.9.1 Loads powered from switchgear
Power to the various loads is switched ON and OFF by an individual circuit breaker at the
selected voltage level. The circuit breaker protective relays may be mounted within the
breaker cell, and the relays interposing between the breaker control circuit and the load’s
control logic are located in separate cubicles or cells, called the relay and terminal (R&T)
section, adjacent to each group of circuit breakers.
A typical switchgear control circuit operates from the 250 V DC power source provided by the
Class I batteries. The circuit is used to:
 Provide power to the operation of stored-energy devices which operate on the close and
trip mechanisms of the circuit breaker.
 Close and trip the circuit breaker in response to commands from the:
o Unit operator;
o Process control system;
o Power circuit protective relays.
Operation of the close and trip circuits requires momentary signals. The circuit breaker
controls require manual local reset following a trip due to the operation of power circuit
protective relays.
3.9.2 Loads powered from the MCC
Power to these loads will be switched ON and OFF by contactors in individual combination
starters. The relays interposing between the contactor control circuit and the load’s control
logic are located in the relay and terminal (R&T) section adjacent to each group of combination
starters.
The circuit breaker in the combination starter is manually operated and, except for
maintenance, remains in the closed position. A typical MCC control circuit operates from the
120 V AC power source provided by the starter’s step-down transformer. The circuit is used to
energize and de-energize the contactor in response to commands from the:
 Unit operator;
 Process control system;
 Circuit breaker protection and overload relays.
To remain energized, the contactor requires a signal to be maintained. The circuit breaker
requires a manual local reset following a trip due to the operation of power circuit protective
functions built into the breaker.
3.9.3 Class IV and Class III loads
Typical types of interfacing circuits are:
 MCC and switchgear:

o Off/On;
o Off/Auto/On;
o Off/Standby/On;
o Off/Standby/Auto/On.
 MCC only:
o Motorized valve with non-auto control;
o Motorized valve with auto control.
 Other:
o Solenoid valve with non-auto control;
o Solenoid valve with auto control.
3.9.4 Class II and Class I loads
Loads energized from the Class II and Class I (UPS) MCCs or panels perform either special
safety-related or personnel/equipment protection-related functions. The control modes are
therefore limited to OFF/ON or OFF/AUTO/ON and, in the case of motorized valves, to
OPEN/CLOSE or OPEN/AUTO/CLOSE and operate in the same way as the Class III and Class IV
loads with the same type of controls.
3.9.5 EPS loads
Loads energized from the EPS are controlled in the same way as when they were energized
from Class III, II, and I or will be limited to manual OFF/ON controls.
3.10 Summary
The safety and operating reliability of a CANDU NPP depend heavily on availability of electrical
power to ensure proper operation of its various systems. The electrical power system inside
the plant is divided into four classes: Class I, II, III, and IV. Energy for Class I is stored in batteries
and can be obtained from the rectified power of Class III sources. Class II power is obtained
from Class I through DC/AC inverters or directly from Class III. Standby generators provide
alternative power to Class III and EPS systems. Normally, the plant obtains power from its own
unit through a UST. It is also possible and permitted to operate the plant with Class IV power
supplied through an SST. When a unit stops producing electrical power, power is drawn from
neighbouring units through switchyard connections. This may require manual re-configuration
(depending on the event) to supply the shutdown unit(s) from the running unit(s) to remove
decay heat. When these power sources are not available, grid power can be used to power
Class IV through the SST. Once Class IV power is lost, the reactor must be shut down
immediately, and heat sink systems are powered from Class III standby generators or the EPS.
In addition, CANDU stations are also equipped with two sets of long-term on-site power
supplies, at least one of which is seismically qualified, which are driven by diesel engines.
Within the same class, the Class IV, Class III, and some Class II loads in the plant are distributed
on multiple and separate buses depending on the number of loads, their power requirements,
and the plant’s Odd/Even bus philosophy. Class I and Class II power to I&C circuits is supplied
through three channelized distribution systems from channelized and independent energy
storage and conversion systems. Because Class IV buses are capable of receiving power from
either of two supplies and because automatic transfer of supplies is provided on sensing loss of
power, reliability of power is ensured, and plant operating safety is increased.
4 Main Electrical Components in a CANDU Plant

4.1 Generators
4.1.1 Basic principle
Electricity output from a CANDU nuclear power plant is generated by a synchronous generator.
The generator shaft is directly coupled to that of the steam turbine. The function of the
generator is to convert mechanical energy from the turbine to electrical energy to supply
electrical loads. A simple illustrative diagram is shown in Fig. 7.
Fig. 7 - Turbine and generator set.
The principle of a generator is based on Faraday’s law of electromagnetic induction. The main
parts of a generator are a stationary iron core and winding, known as the stator, and a rotating
iron core and winding, known as the rotor. When the rotor winding is energized through the
field excitation circuit, as the turbine rotates the rotor, a rotating magnetic field is created. The
excitation current is supplied to the rotor winding through slip rings. The rotating magnetic flux
induces a potential in the stator winding. An illustrative diagram is shown in Fig. 8.
Fig. 8 - Illustrative diagram of a synchronous generator.
Due to the relative positions of the magnetic flux and the stator winding, as the rotor turns, the
induced voltage will take on a sinusoidal form. The frequency of the generated voltage will be
directly related to the rotational speed. For the two-pole (N-S) machine shown in Fig. 9, one
full revolution will produce one full cycle of a sinusoidal wave. If the number of pole pairs on
the rotor is increased, a full revolution of the shaft will produce multiple cycles at the electrical
output. In other words, it is possible to reduce the rotational speed of the turbine, but still to
generate the desired frequency in the electrical output, by increasing the number of pole pairs.
Fig. 9 - A two-pole (one pole pair) synchronous generator.
The relationship among the speed of rotation (݊ rpm), the output frequency (݂ Hz), and the
number of pole pairs (‫ )݌‬can be stated as follows:
60݂
݊= .
‫݌‬
The word synchronous means that the magnetic field rotates in synchronism with the rotor.
When the stator windings are placed 120° apart as shown in Fig. 10, a three-phase voltage can
be generated.
Fig. 10 – Three-phase synchronous generator.
When a load is connected to the output of the stator winding, the generator will transfer the
power to the load.
Assume that the currents from each phase can be represented as:
݅஺ = ‫ܫ‬ெ sin ߱‫ݐ‬
݅஻ = ‫ܫ‬ெ sin(߱‫ݐ‬− 120°)
݅஼ = ‫ܫ‬ெ sin(߱‫ݐ‬− 240°).
The active power output delivered to the load at each phase can be calculated as:
ܲ = ݅ெ ‫ݒ‬ெ cos ߠ௠ MW
The reactive power is
ܳ = ݅ெ ‫ݒ‬ெ sin ߠ௠ MVar
where iM and vM are the phase current and phase voltage.
The angle ߠ௠ is the phase difference between the voltage and the current at the generator
output. Hence, the total real and reactive power output from all three phases can be expressed
as:
ܲ௧௢௧௔௟ = 3ܲ = 3݅ெ ‫ݒ‬ெ cos ߠ௠ MW

ܳ௧௢௧௔௟ = 3ܳ = 3݅ெ ‫ݒ‬ெ ‫ߠ݊݅ݏ‬௠ MVar
The rated power is

ܲ௥௔௧௘ௗ = 3݅ெ ‫ݒ‬ெ MW
The power factor (pf) is

‫ߠݏ݋ܿ = ݂݌‬௠
Typically, the power factor is maintained between 0.8 and 0.9. The frequency of the generated
power is controlled by a governor on the turbine, and the generator output voltage is
controlled by the field excitation through an automatic voltage regulator.
4.1.2 Generators in a CANDU plant
There are several generators in a CANDU plant: (1) the main generator; (2) the standby
generators; and (3) the generators in the emergency power system.
The main generator converts the mechanical power from the turbine to electric power that is
delivered to the grid to supply power to customers. Typical specifications of a main generator
are listed in Table 7.
Table 7 - Characteristics of the main generator in a CANDU plant.

Capacity 817 MVA
Rated output power 728 MW
Output terminal voltage 22 kV
Power factor 0.9 lagging
Frequency 60 Hz
Number of phases 3
Number of poles 4
Rated speed 1800 rpm
Excitation system Static thyristor exciter
Stator cooling Water
Rotor cooling Hydrogen
Typical specifications of the standby generators and the emergency power generators are listed
in Tables 8 and 9 respectively.
Table 8 - Characteristics of the emergency power system generator.

Rated output power 1.6 MW
Rated current 183 A
Output terminal voltage 4.16 kV
Frequency 60 Hz
Number of phases 3
Number of poles 6
Rated speed 1200 rpm
Table 9 - Characteristics of the generator in the Class III power system.

Rated output power 8.2 MW
Output terminal voltage 4.16 kV
Frequency 60 Hz
Number of phases 3
Number of poles 12
Rated speed 600 rpm
4.1.3 Excitation system
To create a magnetic field inside a synchronous generator, separate windings and an electrical
power source must be used. This part of the generation system is known as the excitation
system and is shown in Fig. 12. The excitation system is essentially a controllable DC source. By
adjusting the excitation system output voltage, the output voltage level of the generator can be
controlled, and hence the reactive power output. Because the excitation current must be
delivered to the windings on the rotor, slip rings are used.
Once the generator is running, power for the excitation system can be obtained from the
excitation transformer, which is energized from the Class IV distribution system. The power
source can be either the SST or the UST. However, AC power from the generator must be
converted to DC before it can be delivered to the rotor windings. In the past, a DC generator
coupled to the synchronous generator shaft was used to produce DC power for the excitation
system. Nowadays, this conversion is accomplished by a thyristor-based rectifier. Unlike a DC
generator, this conversion process has no moving parts; hence, it is often referred to as a static
thyristor-based excitation system.
During normal operation, the excitation system is often used to (1) control the output voltage
level of the generator, and (2) adjust the reactive power output of the generator. A feedback
control system, known as an excitation control system, is also used to ensure that adequate
excitation voltage is applied to maintain the desired generator output voltage level and the
reactive power output. These functionalities are essential to improve the reliability of the
generator system.
In the event of an emergency, the excitation system can also be used to provide additional
means to improve system stability. For example, when a fault has occurred on the transmission
system, the output voltage of the generator can decrease unexpectedly. The excitation system
can be used to slow down this voltage collapse, thus improving system stability. If a short
circuit in the generator or at the generator output terminal is detected, the excitation system
can cut its power immediately to drive the generator output voltage to zero, preventing further
damage to the generator.
In a CANDU plant, an excitation transformer is used to step down the generator output voltage
from 22 kV to 850 V before sending it to the thyristor-based rectifier. However, different plants
may have different output voltage levels. An illustrative diagram of an excitation system is
shown in Fig. 11.
Fig. 11 - Conceptual diagram of a static thyristor-based excitation system.
To start the generator, a separate excitation system must be used. The details will be omitted
here. Once the generator starts to operate, a portion of the generated power is used to
provide the excitation for its magnetic field. The excitation power is obtained by converting a
portion of the 22 kV generator output to 850V AC voltage. This voltage is further regulated
through an automatic voltage regulator (AVR) and subsequently sent to a thyristor-based static
rectifier to convert the AC voltage to DC voltage before sending it to the rotor through the slip
ring. The excitation system for the SGs and EPGs is different in that they must start when no
additional sources of AC power are present.
4.1.4 Excitation transformer in a CANDU plant
The excitation transformer is a three-phase transformer. Delta (Δ) connections are used on
both primary and secondary sides. Specifications of one such transformer are listed in Table 10.
Table 10 - Characteristics of an excitation transformer.

Rated capacity 6,300 kVA
Voltage ratio 22,000 V/850 V
Rated currents 165 A/4,279A
Loss at no load 10,858 W
Loss at full load 46,678 W
Cooling method Air-cooled
4.1.5 Cooling and protection systems
As electric current passes through the generator windings, heat is produced in both the rotor
and the stator. To maintain a safe operating temperature, adequate cooling must be provided.
For generators in CANDU plants, water cooling is used for the stator winding, whereas
hydrogen is used to cool the rotor winding and the iron cores of both the rotor and the stator.
The main part of the water cooling system consists of two centrifugal pumps and two heat
exchangers. The pumps maintain a steady flow of cooling water through the stator windings.
The water pressure is controlled by means of a pressure control valve which keeps the loop
pressure around 150–200 kPa. The water temperature is adjusted by means of a proportional
valve which mixes hot water from the outlet with water cooled by the heat exchanger. The
objective is to ensure that the temperature of the water coming out of the stator is around
46°C. The two pumps, rated at 75 kW, are powered by the Class IV electrical system. If cooling
water is lost, the generator will shut down immediately. Demineralized water must be used,
and dissolved oxygen must be controlled.
The reasons for using hydrogen as a coolant for the rotor and generator are its relatively high
thermal conductivity, low density and viscosity. The former property allows effective cooling,
and the latter property reduces the windage losses associated with the generator rotor
rotation. The main parts of the cooling system are the hydrogen supply unit, the hydrogen
cooling heat exchanger units, and the hydrogen leakage detection system. Another critical part
is the generator oil seals, which prevent hydrogen from escaping and causing a fire or
explosion.
Hydrogen stored at high purity (98%) is injected into the generator air gap between the stator
and the rotor. The pressure is controlled through a pressure regulator set at 414 kPa. The
humidity of the hydrogen is controlled by a gas dryer heater. Four heat exchanger units are
located at the four corners of the generator to maintain the outlet temperature of the
hydrogen at 40°C. To prevent leakage, the generator is tightly sealed. Due to the flammable
nature of hydrogen, care must be taken to avoid any chance of friction-induced sparks during
filling and emptying of hydrogen. Systems removing or adding hydrogen must be grounded
(possibly even the individuals using them) to eliminate the potential for sparks due to build-up
of static electricity or from energized equipment. Several hydrogen leakage detectors are
installed in the vicinity of the generator.
4.2 Transformers
4.2.1 Basic principles
The main function of a transformer is to convert AC electric energy from one voltage level to
another while minimizing the losses in the transformation process. A typical transformer has
two independent windings. One is referred to as the primary winding, and the other as the
secondary winding. These windings are coupled through a magnetic circuit in the iron core of
the transformer. Ferromagnetic materials are used to construct the core to confine the
magnetic flux inside. An illustrative diagram of a transformer is shown in Fig. 12. It is
interesting to point out that, between the primary and the secondary, there is no direct
electrical connection.
Fig. 12 - Basic operating principle of a transformer.
The operating principle of a transformer can be described as follows: the current in the primary
winding creates an alternating magnetic flux, ϕ, inside the core. The strength of this flux is
proportional to the current, Ip, as well as to the number of turns in the primary winding, Np. On
the secondary side, based on Faraday’s law of induction, a potential, Vs, will be induced in the
secondary winding. The level of this induced potential is proportional to the strength of the
magnetic flux, ϕ, which is a function of the current, Ip, as well as of the number of turns in the
secondary side, Ns. Therefore, if Ns is larger than Np, the voltage at the secondary will be higher
than that at the primary; such a transformer is often referred to as a step-up transformer. A
transformer with the winding turned the other way around is known as a step-down
transformer.
Most electrical power systems are three-phase systems. The power generated from a three-
phase synchronous generator must be connected to three-phase transmission lines through a
three-phase transformer. In fact, a three-phase transformer will have three primary windings
and three secondary windings. A three-phase transformer is formed by proper connection of
these windings on both the primary and secondary sides. For simplicity, only single-phase
transformers are described in this chapter.
If the circuit in the secondary side is closed (through a load directly, or through transmission
lines), a path will be formed for the current, Is, to flow through. Assume that all the flux serves
to couple the primary and secondary windings; therefore, the flux, ϕ, will be equal on both
sides:
ܰ௣ ‫ܫ‬௣ = ܰ௦‫ܫ‬௦.
Furthermore,
ܸ௣ ‫ܫ‬௣ = ܸ௦‫ܫ‬௦.
Therefore, it becomes clear that:
ܸ௣ ܰ௣
= ,
ܸ௦ ܰ௦
or
ܰ௦
ܸ௦ = ܸ.
ܰ௣ ௣
The ratio (Ns/Np) is known as the turn ratio. When the turn ratio is greater than unity, the
voltage level on the secondary side will be higher than that on the primary side, and vice versa.
Because a transformer is a passive device, the current is inversely related to the turn ratio; the
current decreases as the turn ratio increases and increases as the current ratio decreases.
The product of the current and the voltage permitted to be applied to the transformer is known
as the transformer rating. The rating relates directly to the conductor size, core, and heat
dissipation capability.
Like any other electrical apparatus, a practical transformer will be less than 100% efficient.
Several sources contribute to these losses. The first is the ohmic losses in both primary and
secondary windings due to the resistance of the coils. These are also called copper losses. The
second loss occurs as a result of hysteresis and eddy currents in the core. This type of loss is
normally independent of the currents in the transformer and is commonly referred to as iron
loss. These losses normally take the form of dissipated heat. In practice, the heat must be
evacuated through cooling systems. Transformer windings are often submerged in mineral oil
to carry away the heat to be dissipated at the fins on the transformer covers. To accelerate the
heat dissipation rate further, forced air, forced oil, or water circulation can be used to increase
heat transfer effectiveness. However, these added power devices will consume additional
energy.
Even though most transformers work under principles similar to those described above, their
appearance can vary greatly. A typical transformer found in a nuclear power plant is illustrated
in Fig. 13. The high-power terminals are located at the top of the transformer, where three
isolated connections can be seen. Electric fans are used to create forced air circulation to
increase the heat dissipation rate.
Fig- 13 - External appearance of a typical power transformer.
4.2.2 Major transformers in a CANDU plant
In a CANDU plant, there are many transformers serving different purposes. However, three
main transformers deserve special attention:
 Main output transformer (MOT)

 Station service transformers (SSTs)
 Unit service transformers (USTs).
Their functionalities have been explained in Section 2, and their specifications are given in
Tables 11 through 13.
Table 11 - Ratings of a main output transformer (MOT).
Rating 3 × 277 MVA
Primary-side voltage 22 kV
Secondary-side voltage 500 kV
Temperature (oil) 45°C
Temperature (winding) 65°C
Cooling method Forced oil and forced air
Table 12 - Ratings of a station service transformer (SST).

Rating 60 MVA (natural cooling)
80 MVA (forced air cooling)
Secondary-side voltage 11.6 kV
Operating temperature 50°C–75°C

range
Heat transfer medium Oil
Cooling method Natural or forced air
Table 13 - Ratings of a unit service transformer (UST).

Rating 60 MVA (natural cooling)
80 MVA (forced air cooling)
Secondary-side voltage 11.6 kV
Operating temperature 50°C–75°C
range
Cooling method Natural or forced air
4.3 Voltage and Current Transducers

4.3.1 Principles
High-voltage, high-current electrical parameters (in the kA and kV range) cannot be directly
used for control purposes. To use these parameters in control and monitoring circuits, they
must be transformed to a range suitable for these applications, generally in the ampere to milli-
ampere and volt to milli-volt range. There are two groups of electrical quantities in a CANDU
plant. The second, lower-value group is suitable for monitoring, control, and electrical
protection purposes, such as input to a meter displaying the generator power output in the
main control room, or input to a data acquisition system which captures the in-rush current of a
circulation pump. High-voltage, high-current quantities cannot be directly connected to low-
power devices without some type of conversion apparatus. To measure high voltages and large
currents effectively, their electrical parameters must be converted to voltage and current
ranges which are safe for use by measurement devices and human operators without the need
for special protective equipment.
The devices that produce the corresponding low-level signals, which are proportional in value
to the original high-power quantities, are known as transducers. Because voltages and currents
are the two most important electrical quantities in an electrical distribution system, this section
will focus mainly on voltage and current transducers. Only AC voltage and current transducers
will be discussed because most of the high-voltage, high-current quantities are of this form. An
importance difference between power transformers and voltage and current transducers is the
requirements for accuracy and linearity. These requirements are much more stringent in the
latter case.
A voltage transducer is essentially a transformer with a sufficiently small turn ratio, which
converts a high-voltage signal to a low-voltage one. The high-voltage signal is connected to the
primary side, and the low-voltage signal is generated on the secondary side. As discussed in
Section 4.2.1, transformers have the unique ability to isolate the high-voltage primary side from
the secondary side electrically. The low voltage carries the same amount of information as the
high voltage, but at a lower electrical potential, making it safer for maintenance personnel and
for equipment designed to operate at lower voltage levels.
An illustrative diagram of a single-phase voltage transducer is shown in Fig. 14(a). When a high
voltage, ܸଵ, is applied, the transducer will produce a corresponding low voltage, ܸଶ. The voltage
ratio is determined by the turn ratio of the primary and secondary windings, i.e.,
ே
ܸଶ = ே మ ܸଵ,
భ
ே
where (ே మ) is less than unity and represents the voltage reduction factor.
భ
The principle of a single-phase current transducer is shown in Fig. 14(b). The relationship
between the current on the primary side and that on the secondary side can be expressed as
follows:
ܰଵ
‫ܫ‬ଶ = ‫ܫ‬,
ܰଶ ଵ
ே
where (ே భ) determines the current reduction factor.
మ
(a)
(b)
Fig. 14 - Principles of (a) voltage transducers; and (b) current transducers.
4.3.2 Voltage and current transducers in a CANDU plant
There are many voltage and current transducers throughout the plant that provide information
on voltage and current levels in real time for control and monitoring purposes. Four sets of
current transducers are located at the generator output, each with a rating of 1,700A / 5A.
There are also two voltage transducers at the generator output, both having a reduction ratio
of 22kV/100V. Similar devices are also used for electrical protection of major transformers,
such as excitation transformers. A capacitor voltage transformer is used for high-voltage
measurement at the grid connection point to provide information necessary for plant
operation, as well as for protective relaying.
4.4 Switches, Circuit Breakers, and Disconnect Switches

4.4.1 Concepts and operating principles
There are many electrical switches in a CANDU plant. The most common are those used to turn
certain pieces of equipment such as lights, pumps, or instruments on and off. A switch that is
turned to the on position (closed) allows electricity to pass through, whereas turning it off
(opening it) breaks the electrical circuit and stops the flow of electrons. In a low-power
environment, the switches are not much different from those in everyday use.
As voltage and current levels increase, the construction and operation of these switches
becomes more complex. High-voltage or high-current switches are often known as circuit
breakers, disconnect switches, and contactors. As the names imply, one of the important
functions of such devices is to conduct or break the current flow in a circuit. There are two
main scenarios which call for such actions: (1) to execute a control command, such as to start or
stop a load, and (2) to cut off the current flow in an abnormal operating condition such as a
short-circuit fault. A circuit breaker must be able to carry and to interrupt current as needed.
An illustrative diagram of an over-current protection circuit breaker is shown in Fig. 15. The
breaker is connected in series with the circuit. If there is no manual tripping signal, or if the
current is within the operating limit, the breaker remains closed. If either a manual tripping
signal is issued (by pushing a button) or if the measured current exceeds the threshold, the
tripping coil will generate a trip signal to open the breaker, thereby interrupting the current
flow.
Fig. 15 - An over-current protection circuit breaker.
A significantly high voltage can be induced between the two contacts of the circuit breaker
when it interrupts the current flow. As the contacts separate, the resistance between them
increases rapidly, producing hot spots between the contacts. The high voltage between the
contacts can also form a very strong electrical field. As the particles between the contacts
become ionized, electric arcing occurs, which will prolong the time taken for the current to
reach zero. To minimize the impact of short circuits and reduce wear and tear on the breaker
contacts, the arc must be extinguished quickly. The breaker and bus bars must be designed to
withstand the mechanical forces resulting from short-circuit currents. Depending on the
duration of the short circuits, the amount of mechanical bracing may need to be increased.
Depending on the method of arc extinction, circuit breakers can be classified as:
 Air-break circuit breakers

 Air-blast circuit breakers
 Vacuum circuit breakers
 SF6 circuit breakers.
An air-break circuit breaker relies on the high-resistance interruption principle by rapidly

lengthening the arc through an arc runner. The arc resistance is increased to such an extent
that the arc can no longer be sustained. Such types of breakers are mainly used in low- and
medium-voltage circuits.
In an air-blast circuit breaker, high-pressure air is blasted into the arc, blowing away the ionized
gas between the contacts to extinguish the arc. The voltage and current that can be
interrupted by an air-blast circuit breaker are normally higher than in an air-break circuit
breaker. As its name implies, the contacts in a vacuum circuit breaker operate in a vacuum
interrupter chamber. The arc is generated by ionization of the contact material, whereas in an
air breaker, it is generated by the arc material as well as air ionization. Hence, in a vacuum, the
arc is immediately extinguished once the voltage can no longer sustain the plasma created at
the contact. The cost of the vacuum circuit breaker is relatively high, and they are often used in
circuits at less than 38 kV.
SF6 circuit breakers are the most common for high-voltage and high-current circuits. SF6 is
short for sulphur hexafluoride, a gas which satisfies the requirements of an ideal arc-
interrupting medium. SF6 gas has high dielectric strength and is colourless, odourless, and non-
toxic, with high thermal conductivity. It is also highly stable and non-flammable and does not
cause corrosion when in contact with the metallic parts of a circuit breaker. SF6 circuit
breakers can be found in circuits with voltages ranging from 3 kV up to 1000 kV.
Circuit breakers should not be confused with disconnect switches. Disconnect switches do not
have any arc-extinction capability and therefore cannot be used to interrupt a current flow.
Such switches are instead used to provide another layer of protection for repair or maintenance
crews, enabling them to isolate the section of a circuit being serviced. Disconnect switches can
be operated either manually or automatically (in the case of motorized switches, such as a
starter).
4.4.2 Switchgear in a CANDU plant
There are many types of circuit breakers in a CANDU plant that provide control and electrical
protection functions. Circuit breakers are installed in the plant to facilitate operation and
electrical protection of transformers or electrical distribution bus bars. An illustrative drawing
of an arc-extinguishing breaker is shown in Fig. 16. SF6 circuit breakers are used for high-
voltage switchyard circuits, whereas vacuum breakers can be used at medium voltage levels
(11.6 kV/6.3 kV). These breakers can be operated manually, automatically, or by remote
control. The major types of high-voltage circuit breakers in a CANDU plant are described in
Tables 14 to 16.
Table 14 - Characteristics of SF6 circuit breakers at the generator output.

Voltage level 24 kV
Rated voltage 30 kV
Rated current 24 kA
Rated capacity 1,000 MVA
Breaker closing time Less than 42 ms
Breaker opening time Less than 42 ms
Arc extinction time Less than 60 ms
Table 15 - Characteristics of vacuum circuit breakers on an 11.6 kV bus.

Voltage level 11.6 kV
Rated voltage 15 kV
Rated current 2,000–3,000 A
Rated capacity 1,000 MVA
Breaker closing time 45–60 ms
Breaker opening time 30–45 ms
Arc extinction time 5–17 ms
Table 16 - Characteristics of vacuum circuit breakers on a 6.3 kV bus.

Voltage level 6.3 kV
Rated voltage 8.25 kV
Rated current 1,200 A
Rated capacity 500 MVA
Breaker closing time 45–60 ms
Breaker opening time 30–45 ms
Arc extinction time 5–17 ms
Disconnect switches are often found in series with circuit breakers to provide additional
protection for workers. A disconnect switch is used in series with the circuit breaker at the
generator output. The rated voltage of this switch is 24 kV, the current is 24 kA, and the
maximum short-circuit current is 160 kA. Note that disconnect switches are not used to
interrupt current in circuits. They are used for isolation purposes (worker safety and load
isolation), as well as to reconfigure a network. They are installed to provide additional safety
measures for maintenance crews working on the power line. There are also fuses installed in
many electric systems throughout the plant to isolate short circuits or unforeseen faults.
Fig. 16 - An arc-extinguishing circuit breaker.
4.5 Summary
A generator is an energy conversion device that converts mechanical energy from the turbine
to electrical energy to supply the load. Together with the generator, there are several other
auxiliary electrical systems in an NPP, such as circuit breakers, transformers, and
voltage/current transducers. In this section, the general principles of these systems have been
first explained, followed by information specific to CANDU NPPs. After completing this section,
the reader should have a good understanding of how such systems operate, including
knowledge of CANDU-specific applications.
5 Summary and Relationship to other Chapters
Even though the sole objective of a CANDU NPP is to produce electricity, this chapter is
relatively independent of the other chapters in this book. To achieve a better understanding of
the functionality of the different classes of power sources with respect to safety, the student
should read Chapter 13 on Reactor Systems first to learn about the different safety systems and
safety functions in a CANDU NPP.
6 Exercise problems
1. State the reasons why an NPP is different from a fossil-fuel power plant in terms of its
station power requirements.
2. From a power grid point of view, what criteria are used to select a suitable site for
construction of a new NPP?
3. Why is off-site power so important to the safety of an NPP?
4. In your own words, explain the meaning of “odd/even power supplies”.
5. List and define four classes of power used in a CANDU NPP.
6. State the original energy sources of the different classes of power sources.
7. State the sequences under loss of Class III power.
8. State the sequences under loss of Class IV power.
9. Why are batteries used in Class I power supplies?
10. Which class of power is used to charge the batteries in Class I power sources?
11. Explain the sources of power for each power class under both normal and emergency
conditions.
12. Explain the role of grid power during the start-up and shutdown of a CANDU reactor.
13. What role do standby generators play in an NPP, and which class of power supply do
they support?
14. What are the main differences between standby generators and an emergency power
system?
15. Explain the different power sources that a CANDU plant has and the reasons for this.
16. List the main elements inside a generator.
17. Explain the principle of a synchronous generator used in a CANDU NPP.
18. What is the relationship between the number of generator pole pairs, the rotational speed,
and the frequency of the voltage at the generator output?
19. Why is excitation important for synchronous generators?
20. How is excitation implemented in a CANDU NPP?
21. What is the function of the excitation transformer?
22. How is a generator cooled in a CANDU plant?
23. Under what conditions can the generator be connected to the power grid?
24. How are current and voltage measured in a CANDU plant?
25. What is the main function of a transformer?
26. Identify step-up and step-down transformers used in a CANDU plant.
27. Why are the coils in large-capability transformers often submerged in mineral oil?
28. What is the main function of the station transformer?
29. Why are protection devices important in a switchyard?
30. Explain the working principle of SF6 breakers.
31. What is the difference between a circuit breaker and a disconnect switch?
32. What is the function of a switchyard?
7 References
[1] Canadian Standards Association (CSA), Requirements for Electrical Power and Instrument Air
Systems of CANDU Nuclear Power Plants (N290.5-06), 2011.
[2] CNSC Regulatory Document 2.5 Physical Design, Design of Electrical Power Systems for
Nuclear Power Plants and Small Reactor Facilities (Draft stage).
[3] IAEA, Interaction of Grid Characteristics with Design and Performance of Nuclear Power
Plants - A Guidebook, 1983.
[4] IAEA, Electric Grid Reliability and Interface with Nuclear Power Plants, No. NG-T-3.8, 2012.
[5] IEEE, Standard for Preferred Power Supply (PPS) for Nuclear Power Generating Stations
(NPGS) Standard 765-2012, 2013.
[6] H. Ren, I. Dobson, B.A. Carreras, “Long-Term Effect of the n-1 Criterion on Cascading Line
Outages in an Evolving Power Transmission Grid”, IEEE Trans. on Power Systems, Vol. 23, pp.
1217–1225, 2008.
[7] CSA, General Requirements for Seismic Design and Qualification of CANDU Nuclear Power
Plants, N289.1-08, 2008.
8 Further Reading
L.L. Grigsby, Power Engineering Handbook, CRC Press, IEEE Press, 2000 (ISBN-10: 0849385784).
IAEA, Design of Emergency Power Systems for Nuclear Power Plants, NS-G-1.8, 2004 (ISBN 92–
0–103504–7).
P. Kiameh, Electrical Equipment Handbook: Troubleshooting and Maintenance, McGraw-Hill,

2003 (ISBN-10: 0071396039).
P. Kundur, Power System Stability and Control, McGraw-Hill Professional, 1994 (ISBN 0-07-
035958-X).
J.D. McDonald, Electric Power Substation Engineering (2nd ed.), CRC Press, 2007 (ISBN-10:
0849373832).
9 Acknowledgements
The following reviewers are gratefully acknowledged for their hard work and excellent
comments during the development of this Chapter. Their feedback has much improved it. Of
course the responsibility for any errors or omissions lies entirely with the author.
Pel Castaldo
Alan Hepburn
Alek Josefowicz
1
CHAPTER 12
Radiation Protection and Environmental Safety
Prepared by
Dr. Edward Waller, PhD, PEng, CAIH, CHP
University of Ontario Institute of Technology
Faculty of Energy Systems and Nuclear Science
Summary:
Crucial to operation of a CANDU nuclear plant is protection of workers, the public, and the
environment. This chapter discusses:
1. Basic fundamentals of radiation physics as they pertain to radiation interactions that
have the potential to cause biological harm in living systems.
2. Concepts of regulatory guidance which governs the “as low as reasonably achievable”
radiation dose paradigm.
3. Radiation detection and monitoring techniques used in CANDU plant environs.
4. External and internal radiation hazards, including discussions on shielding and personal
protective equipment.
5. Radiation management plans, worker dose monitoring, and control and waste man-
agement issues.
6. Radiation releases to the environment, derived release limits, and environmental protec-
tion.
Table of Contents
1 Introduction ............................................................................................................................ 6
1.1 Overview ......................................................................................................................... 6
2 Biological Effects of Radiation ................................................................................................ 7
2.1 Basic Radiation Interactions with Tissue ........................................................................ 7
2.2 Biological Radiosensitivity............................................................................................. 11
2.3 Biological Effects of Exposure ....................................................................................... 15
2.4 Radiation Risk Models................................................................................................... 20
2.5 Effects on Pregnancy..................................................................................................... 26
2.6 Summary ....................................................................................................................... 28
3 Dosimetry, Dose Limitations, and Guidance......................................................................... 29
3.1 Dose Definitions............................................................................................................ 30
3.2 Background Radiation Exposure ................................................................................... 37
3.3 International Guidance ................................................................................................. 39
3.4 Canadian Guidance ....................................................................................................... 45
3.5 Summary ....................................................................................................................... 46
4 Radiation Instrumentation.................................................................................................... 47
4.1 Basics of Detection........................................................................................................ 47
©UNENE, all rights reserved. For educational use only, no assumed liability. Radiation Protection and Envir. Safety – September 2014
4.2 Gas-Based Detectors..................................................................................................... 60

4.3 Scintillation-Based Detectors........................................................................................ 64
4.4 Semiconductor-Based Detectors .................................................................................. 66
4.5 Portable Instrumentation ............................................................................................. 67
4.6 Specialized Detectors.................................................................................................... 69
4.7 Dosimetry Techniques .................................................................................................. 72
4.8 Summary ....................................................................................................................... 77
5 External Radiation Hazards and Shielding ............................................................................ 77
5.1 External Sources and Dosimetry................................................................................... 78
5.2 External Hazards ........................................................................................................... 90
5.3 External Protection (Radiation Shielding)..................................................................... 93
5.4 Summary ..................................................................................................................... 103
6 Internal Radiation Hazards ................................................................................................. 103
6.1 Internal Pathways and Dosimetry............................................................................... 103
6.2 Internal Hazards .......................................................................................................... 111
6.3 Internal Protection...................................................................................................... 113
6.4 Summary ..................................................................................................................... 116
7 Radiation Safety Management ........................................................................................... 117
7.1 Dose Records .............................................................................................................. 117
7.2 Radiation Work Planning ............................................................................................ 118
7.3 Personal Protective Equipment .................................................................................. 120
7.4 Contamination Control................................................................................................ 120
7.5 Emergency Planning.................................................................................................... 123
7.6 Summary ..................................................................................................................... 123
8 Radiation and the Environment.......................................................................................... 124
8.1 Radiological Environmental Monitoring Program (REMP) ......................................... 124
8.2 CANDU Releases ......................................................................................................... 133
8.3 Protection of the Environment ................................................................................... 134
8.4 Summary ..................................................................................................................... 136
9 Summary of Relationship to Other Chapters...................................................................... 137
10 Problems ............................................................................................................................. 137
11 Further Reading .................................................................................................................. 147
12 References .......................................................................................................................... 148
13 Acknowledgements............................................................................................................. 153
Radiation Protection and Environmental Safety 3
List of Figures
Figure 1 Three pillars of health physics .......................................................................................... 6
Figure 2 DNA double helix: target for radiation damage ............................................................... 8
Figure 3 Process chain for radiation interaction with tissue .......................................................... 9
Figure 4 Direct and indirect radiation interaction mechanisms with DNA .................................. 10
Figure 5 Chromosome................................................................................................................... 11
Figure 6 The cell cycle ................................................................................................................... 12
Figure 7 Cell mitosis ...................................................................................................................... 12
Figure 8 DNA single-strand break ................................................................................................. 13
Figure 9 DNA double-strand break ............................................................................................... 13
Figure 10 Radiation damage to cells............................................................................................. 14
Figure 11 DNA damage endpoints ................................................................................................ 15
Figure 12 Differential blood count after severe radiation insult (adapted from [Gusev2001] and
[Cember2009]) ...................................................................................................................... 18
Figure 13 Dose-response curve showing two common models................................................... 23
Figure 14 Generalized dose-response curves............................................................................... 24
Figure 15 Alternative dose-response curves ................................................................................ 25
Figure 16 Radiation protection regulation pyramid ..................................................................... 29
Figure 17 Relationship between dose and endpoint.................................................................... 35
Figure 18 Origin of average yearly radiation exposure (adapted from [NCRP2009]) .................. 39
Figure 19 Chronology of radiation-protection guidance (adapted from [Inkret1995]) ............... 40
Figure 20 Paradigm shift from deterministic to stochastic effect protection .............................. 41
Figure 21 Progression of radiation protection principles (adapted from [Clement2009]) .......... 42
Figure 22 Relationship of international bodies for radiation protection regulations .................. 42
Figure 23 Ionization process ......................................................................................................... 47
Figure 24 Ionization in a gas-filled tube........................................................................................ 48
Figure 25 Excitation process ......................................................................................................... 48
Figure 26 Excitation in a scintillator-type detector ...................................................................... 49
Figure 27 Geometric efficiency examples (far planar, near planar, and well detectors)............. 52
Figure 28 Example energy calibration graph ................................................................................ 53
Figure 29 Example efficiency calibration graph............................................................................ 54
Figure 30 Simple particle-counting system................................................................................... 58
Figure 31 Simple spectroscopy process........................................................................................ 59
Figure 32 Operating regions of gas-filled counters ...................................................................... 61
Figure 33 GM counter sample plateau ......................................................................................... 63
Figure 34 Commercial HPGe system (top) and detail of lead castle (bottom)............................. 67
Figure 35 Portable contamination meter with pancake probe .................................................... 68
Figure 36 Neutron Bonner sphere ................................................................................................ 69
Figure 37 Portal monitor............................................................................................................... 70
Figure 38 Commercial portal monitor .......................................................................................... 70
Figure 39 Commercial whole-body counter ................................................................................. 71
Figure 40 TLD glow curve from irradiated LiF chip ....................................................................... 73
Figure 41 Commercial TLD reader (top) and tray of LiF chips (bottom)....................................... 74
Figure 42 LSC detection process ................................................................................................... 75
Figure 43 LSC electronics .............................................................................................................. 76
Figure 44 Commercial LSC counter (top) and detail of trays with vials (bottom) ........................ 77
Figure 45 Beta dose delivery geometries ..................................................................................... 78

Figure 46 Beta surface contamination.......................................................................................... 79
Figure 47 Neutral-particle radiation dose-delivery geometries ................................................... 83
Figure 48 Linear interaction coefficients for dry air ..................................................................... 84
Figure 49 Line-source geometry ................................................................................................... 85
Figure 50 Planar source geometry................................................................................................ 86
Figure 51 Volume-source geometry ............................................................................................. 86
Figure 52 Potential external exposure sources during reactor operation ................................... 92
Figure 53 Potential external exposure sources after shutdown (> 10 min) ................................. 92
Figure 54 External hazards............................................................................................................ 94
Figure 55 Inverse-square law........................................................................................................ 94
Figure 56 Narrow-beam or thin-shield attenuation ..................................................................... 96
Figure 57 Broad-beam or thick-shield attenuation ...................................................................... 96
Figure 58 Photon interaction cross sections for lead ................................................................... 98
Figure 59 Photon interaction coefficients .................................................................................... 99
Figure 60 Response functions for neutrons................................................................................ 100
Figure 61 Internal contamination pathways............................................................................... 104
Figure 62 Basic mechanisms for internal dose ........................................................................... 105
Figure 63 Intake to excretion pathways ..................................................................................... 106
Figure 64 Absorbed fraction in mass .......................................................................................... 109
Figure 65 Worker in PPE ............................................................................................................. 115
Figure 66 Relationship between background, action level, detection limit, and regulatory limit
............................................................................................................................................. 121
Figure 67 General environmental transfer model...................................................................... 131
Figure 68 Common radiological assessment approach for human and non-human biota ........ 135
Figure 69 Derived consideration levels....................................................................................... 135
List of Tables
Table 1 Radiation-induced DNA damage mechanisms................................................................. 10
Table 2 Deterministic effects [NCRP2001].................................................................................... 16
Table 3 Stochastic risk coefficients (adapted from ICRP (2007)).................................................. 19
Table 4 Relative human-organ vulnerabilities to ionizing radiation............................................. 19
Table 5 Association of cancers with radiation exposure .............................................................. 21
Table 6 Epidemiological categories for radiation risk studies ...................................................... 22
Table 7 Radiation weighting factors [ICRP2007] .......................................................................... 33
Table 8 Tissue weighting factors [ICRP1977,1990,2007].............................................................. 34
Table 9 Dose units......................................................................................................................... 34
Table 10 Sample inhalation dose coefficients [ORNL2013].......................................................... 36
Table 11 Primordial and cosmogenic radioisotopes..................................................................... 38
Table 12 ICRP103 dose limits for planned exposure scenarios [ICRP2007] ................................. 44
Table 13 Standard thickness of various beta attenuators (from [Cember2009]) ........................ 81
Table 14 Important neutron reactions ......................................................................................... 88
Table 15 Removal cross sections for neutrons ........................................................................... 102
Table 16 Parameters affecting internal dose.............................................................................. 104
Table 17 Effective half-life .......................................................................................................... 108
Table 18 Air-purifying cartridge designations............................................................................. 116
Table 19 Example of samples collected by Canadian utilities [Cole1997].................................. 128
Table 20 Some key fission and activation radionuclides expected from a reactor release ....... 129
Table 21 Order-of-magnitude CANDU radionuclide emissions.................................................. 133
Table 22 Reference animals and plants (RAP) ............................................................................ 136
1 Introduction
In Chapter 3, the fundamentals of nuclear physics, radioactive decay processes, and radiation
interactions with matter were discussed. This chapter expands upon the concepts in Chapter 3
and explores the fundamentals of radiation protection and environmental safety, which are of
vital importance to the safe operation of CANDU stations. In a holistic sense, the overarching
field that deals with radiation protection and environmental safety is the area of health physics.
Health physics is the branch of science that deals with protection of workers, the public, and
the environment from potential detrimental effects from exposure to ionizing and non-ionizing
radiation. The three primary pillars of health physics in the context of CANDU operations are
presented in Figure 1. Radiation shielding is presented as a pillar because it is an important
aspect of radiation protection that is often a field in and of itself.
Health Physics
Environmental Safety
Radiation Shielding
Radiation Protection
Figure 1 Three pillars of health physics

CANDU stations use experts in health physics, radiation protection, shielding, and environ-
mental safety to ensure that reactor operations comply with Canadian regulations in a safe and
controlled manner.
1.1 Overview
Experts in health physics generally require formal training in radiation protection, environ-
mental radioactivity, and shielding design. In this chapter, the concepts of biological effects,
radiation protection, health physics, shielding, and environmental safety are discussed with an
emphasis on CANDU nuclear plant operations. This chapter represents the principal concepts
related to the three pillars (Figure 1) which provide the well-rounded understanding of health
physics that is required for a career in CANDU nuclear operations. Specific concepts discussed
are: (1) basic fundamentals of radiation biology and the basics for understanding biological
harm in living systems; (2) concepts of regulatory guidance which govern the “as low as rea-
sonably achievable” radiation dose paradigm; (3) radiation detection and monitoring tech-
niques that may be used in CANDU plant environs; (4) external and internal radiation hazards,
including discussions on shielding and personal protective equipment; (5) radiation manage-
ment plans, worker dose monitoring, and control and waste management issues; and (6)
radiation releases to the environment, derived release limits (DRL), and personnel and envi-
ronmental protection.

 The biological basis for health impact from radiation exposure
 National and international regulatory guidance principles involving radiation exposure
 Overview of radiation-protection instrumentation used in CANDU plants
 The equations governing external exposure from point and extended sources
 Concepts of radiation shielding of external sources
 The equations governing internal exposure to radiation
 Concepts of radiation protection and management programs for CANDU plants
 Environmental radioactivity from CANDU plants
 Protection of the environment.
2 Biological Effects of Radiation

The principles of radiation biology are germane to background radiation-safety guidance for
workers and the public with respect to CANDU operations. Personnel operating in and around
CANDU plants, fuel-processing facilities, mining activities, and used nuclear fuel storage must
have a fundamental understanding of how radiation may interact with the human body and
how protective practices can be used to reduce any potential risk from exposure. In short,
radiation biology is the backbone of all radiation protection and health physics.
Before discussing the mechanisms of how radiation affects living organisms, which form the
basis of radiation protection, the basics of radiation interactions with matter must be under-
stood. These concepts were discussed in Chapter 3 (Nuclear Processes and Neutron Physics).
This section focuses on the fundamentals of how different types of radiation deposit energy in
tissue through interactions and the principal target which drives radiation-protection regula-
tion: deoxyribonucleic acid, or DNA. More information on the historical aspects of radiation
biology may be found in [Preston2005].
2.1 Basic Radiation Interactions with Tissue

As discussed in Chapter 3, radiation interacts with materials in a variety of ways. Radiation can,
for example, scatter, be absorbed, or create secondary particles as it passes through matter.
The exact way in which radiation interacts with matter is dictated by the type of radiation and
the nature of the matter being traversed. The endpoint(s) of the interaction(s) are dictated by
the material being traversed. For example, neutrons impinging upon polymers can have the net
effect of creating displaced atoms in the lattice (displacement damage), which can result in
embrittlement of the material. The same neutrons impinging upon human tissue can cause
changes to the DNA molecules, inducing chromosome aberrations which can lead to cancer. It
is worth-while to note here, however, that radiation can also be used to strengthen polymers
and to treat cancer, so that the mode of application of radiation to material becomes important
in determining the system endpoint. The fundamentals of radiation interactions with tissue
begin with the concept of energy deposition.
Energy that is deposited in biological material can lead to two principal effects on an individual
atom: excitation or ionization. Excitation raises an orbital-shell electron to a higher energy state
without ejecting it from the atom; radiation can be emitted during this process. Ionization, on
the other hand, is the ejection of an electron from an orbital shell of an atom; other kinds of
radiation can also be emitted during this process. It is generally the ionization process that
causes damage to DNA through direct or indirect action.
If DNA is the primary critical target for radiation-protection applications, then consideration
must be given to DNA damage from direct and indirect action of the radiation to which tissue is
exposed. DNA is a macromolecule which is the main constituent of chromosomes and the
material that contains and communicates genetic information about all life. The DNA molecule
is formed as a double helix and is coded with nucleotide base pairs along a twisted backbone of
alternating sugar and phosphate groups, as depicted in Figure 2.
Backbone
Base pairs (coding)
Double helix
Figure 2 DNA double helix: target for radiation damage

Direct damage to DNA can occur when ionizing radiation interacts directly with the DNA mole-
cule. For example, direct energy deposition on a backbone strand can have the effect of break-
ing the strand at that location. Direct damage is more probable for high linear energy transfer
(LET) radiation, such as -particles and neutrons. Indirect damage is a more complex process
involving radiation interactions with water molecules. On average, the human body is com-
posed of approximately 60% water [Cember2009], which is contained mostly within the body’s
cells. Therefore, most direct interactions of all types of radiation on the human body are with
water.
The process of indirect DNA damage proceeds as follows: radiation ionizes a water molecule, as
described by Eq. (1):
H 2O  H 2O   e  . (1)
The products of water dissociation, H2O+ and e-, interact further with water molecules. The
electron interaction with water yields a hydrogen free radical, H  , and a hydroxide ion, OH-, as
shown in Eq. (2):
H 2O  e   H 2O   H   OH  . (2)
The positive ion undergoes immediate dissociation as described in Eq. (3):
H 2O   H   OH  , (3)
which yields a positive hydrogen ion, H+, and a free hydroxyl radical, OH  , the neutral form of
the hydroxide ion, OH-. The free hydrogen radicals tend to form gaseous hydrogen (H2) which is
relatively harmless in the human body. The hydroxyl radical, however, is extremely reactive and
can damage DNA. These hydroxyl radicals are, however, short-lived and therefore need to be
created in the vicinity of target DNA to do damage. An alternate damage mechanism is for the
short-lived hydroxyl radicals to combine to form hydrogen peroxide as indicated in Eq. (4):
OH   OH   H 2O2 . (4)
Hydrogen peroxide is a strong oxidizing agent which is fairly stable in the human body and can
therefore migrate to distances far from the creation point and damage DNA.
A number of events on varying time scales must occur before deleterious effects from exposure
are observed. The complex chain of events from radiation exposure to a biological endpoint is
summarized in Figure 3.
Chemical
Free changes
Incident Fast radical Biological
radiation electron from bond effects
formation
breakage
Figure 3 Process chain for radiation interaction with tissue
The time scales extend from femtoseconds for initial ionization, to tens of nanoseconds related
to ion-radical lifetimes, to tens of microseconds for free-radical lifetimes, and finally to years
post-exposure for onset of cancer.
The dominant modes of interactions for ionizing radiation that are important to CANDU reac-
tors are contrasted in Table 1. The direct and indirect interaction mechanisms of the creation of
charged particles by neutral particles are depicted in Figure 4. Dose assessment will be consid-
ered in Section 5 for external irradiation and Section 6 for internal irradiation.
Table 1 Radiation-induced DNA damage mechanisms
Tissue inter-
Radiation Dominant damage mechanism
action
Production of fast-moving electrons creating free radicals
X- and γ rays
Indirect which induce strand breaks
ionization Production of recoil protons, alpha particles, and heavier
Neutrons
atoms, creating free radicals which induce strand breaks
Beta particles
(electrons) Direct ioniza- Direct action with DNA, inducing strand breaks. Directly
Alpha particles tion ionizing particles can also generate free radicals.
(helium nuclei)
Indirect Action
gamma
H ray
free radical
OH· O
H
fast
Direct Action neutron
2 nm
4 nm
Figure 4 Direct and indirect radiation interaction mechanisms with DNA
2.2 Biological Radiosensitivity

The effect of radiation on living systems varies with the magnitude of exposure, duration of
exposure, and region of the body exposed. Responses are characterized as deterministic
(threshold) and stochastic (mutagenic, carcinogenic, or teratogenic). The law describing cellular
radiosensitivity is attributed to Bergonie and Tribondeau [Bergonie1906] and states that the
radiosensitivity of a cell is directly proportional to its reproductive activity and inversely propor-
tional to its degree of differentiation. Cells most active in reproduction and cells not fully
mature will be most harmed by radiation exposure. Conversely, the more mature and specia-
lized the cell is, the less sensitive it is to radiation. One of the main reasons for the cause-effect
relationship between radiation exposure and rapidly dividing cells is that the effect is essentially
seen more rapidly when cells divide more rapidly.
2.2.1 Cell cycle

Cells are made up of various components. The cell membrane is the first line of defence for the
cell and regulates what enters or exits the cell. The endoplasmic reticulum provides a transport
mechanism from the cell membrane to the cell nucleus. Ribosomes in the cell are producers of
proteins, and Golgi apparati store, package, help transmit, and customize proteins. Lysomes
break down unwanted material in the cells, and mitochondria generate energy for the cell. For
radiation-exposure risk, the most important component is the nucleus, specifically the chromo-
somes, which are depicted in Figure 5. Chromosomes are made up of DNA segments that code
for proteins. Humans have 23 pairs of chromosomes. The DNA, an extremely small molecule in
the chromosomes, carries the genetic code for all proteins. There are approximately 3x109 base
pairs, but only ~5% are active genes (the other 95% are the “glue” that holds the active genes
together). Radiation can do harm when it hits an active gene.
Chromatid w/DNA
Centromere
(intersection of two chroma-
tids)
Figure 5 Chromosome
The cell cycle is depicted in Figure 6. The cell progresses from mitosis (cell division) to gap 1
(inactivity) for each divided cell which goes on to its own cycle, then to DNA synthesis, and then
to gap 2 (inactivity) before mitosis.
The cell has maximum radiation sensitivity during the mitotic phase of the cell cycle. The cell
cycle involves a number of stages before replication, as depicted in Figure 7.
G1
(1st gap:
G2 relative
(2nd gap: inactivity in
relative M (mitosis) cell)
inactivity in
cell)
S G1
(DNA (1st gap:
Synthesis relative
inactivity in
phase)
cell)
Cells that cease division

Figure 6 The cell cycle
Interphase Pre-mitotic Metaphase
replication
Figure 7 Cell mitosis
2.2.2 DNA damage

As previously stated, the primary target for cell killing is DNA. The primary lesion is the DNA
strand break, and in the case of permanent damage, the DNA double-strand break must be
considered. A DNA single-strand break is depicted in Figure 8. Radiation or free radicals interact
with one of the sugar/phosphate backbone strands. This type of damage is of little risk because
the DNA will either repair itself or, at worst, die upon replication.
Figure 8 DNA single-strand break

A DNA double-strand break, on the other hand, is depicted in Figure 9. This type of damage can
result either from a single, highly ionizing particle interacting with the DNA (such as an alpha
particle) or from two events occurring so close in time that they occur, for all intents and
purposes, at the same time. In this case, it is also possible for the DNA to repair itself if the
strands reattach, but also for the DNA to misrepair itself, either with its own twisted DNA or
with another DNA fragment. In this case, genetic coding can be produced that does not die
upon replication, yet is not native to normal cellular function. This endpoint can be a precursor
to cancer. This type of damage can be observed by examining chromosomes. Radiation damage
to DNA can produce aberrations such as rings (a chromosome wrapped upon itself), fragments
(small bits of “orphan” chromosomes), or dicentrics (two broken chromosomes which form
together to make up a larger macromolecule).
Single event OR Double event
Misrepair
Figure 9 DNA double-strand break

A complete depiction of how radiation damages cells can be seen in Figure 10. During the
interphase, radiation-induced strand breaks are generated. This leads to chromosome misrepair
and abnormally repaired chromosomes during DNA replication, yielding dicentrics, fragments,
(and rings). Note that there is a background level of approximately one dicentric per 1000
lymphocytes in human blood, and therefore excess dicentrics are indicative of radiation dam-
age. During metaphase, the cells divide and can undergo either apoptosis or mutation.
Interphase Mis-repair DNA replication Metaphase
g dicentric
g Death or
mutation
fragments
Figure 10 Radiation damage to cells

The possible endpoints of damaged DNA are graphically depicted in Figure 11. Upon DNA
damage from ionizing radiation exposure, the DNA may (1) repair itself without coding error
(common with single-strand breaks), (2) perform cell “suicide” (apoptosis), or (3) mutate into a
viable, yet undesirable, DNA molecule that may have oncogenic (tumour-producing) properties
upon cell division.
Various factors can impact the extent of radiation damage at the cellular level, including:
 Dose – generally speaking, as the dose increases, the surviving fraction of cells decreases. In
other words, increasing the dose increases the damage.
 Dose rate – as the dose rate increases, the surviving fraction decreases.
 Radiation quality and relative biological effectiveness – for low linear energy transfer (LET)
radiation, it requires more than one track to cause a dicentric; for high-LET radiation, one
track can cause a dicentric. As a general rule, high-LET radiation is more damaging to cells
than low-LET radiation.
 Oxygenation – aerated cells are more radiosensitive. Oxygen reacts with free radicals to
produce peroxide, which is highly toxic.
 Radioprotectants – if present, radioprotectors such as aminothiols can scavenge free
radicals before DNA damage occurs.
The biological effects of exposure are explored in Section 2.3.
[1] error free repair
[2] cell death/apoptosis
normal damaged
DNA DNA
[3] error prone repair
radiation
Figure 11 DNA damage endpoints
2.3 Biological Effects of Exposure

In the previous section, factors affecting the extent of radiation damage to cells were explored.
A variety of factors that influence the biological effects of exposure must be taken into account,
such as:
 Total absorbed energy (dose)
 Rate of dose delivery
o Acute (seconds-minutes)
o Chronic (days-years)
o Type of radiation
 Source of exposure
o External
o Internal
 Age at exposure
 Time since exposure
 Area of exposure
o Localized (cells; organs)
o Extremities (hands; feet)
o Whole body
o Superficial (skin only)
o Deep tissue.
As presented above, exposure can be categorized in many ways. Two very important concepts
related to risk are the rate of dose delivery for exposure and the endpoint.
The rate of dose delivery for exposure is conveniently divided into two principal categories: (a)
acute and (b) chronic. Acute exposure is normally considered as a single, large, and short-term
whole-body dose, and the effects are observed in a short time frame post-exposure. The effects
are categorized by four sequential stages: (i) Initial (prodromal) which lasts approximately 48
hours, (ii) latent, which lasts 48 hours to 3 weeks, (iii) manifest illness, which lasts 6–8 weeks,
and (iv) recovery, which can last weeks to months (if death does not occur). Chronic exposure,
on the other hand, is typically defined as a lower, protracted dose over a long period of time.
The two principal categories describing the endpoint effects of radiation exposure in humans
are (a) deterministic effects and (b) stochastic effects. Dose definitions (Gy, Sv, etc.) are pro-
vided in Section 3.1.
2.3.1 Deterministic effects

Deterministic effects are those that manifest at some threshold dose, with increasing severity
with increasing dose. Some estimated deterministic effects as a function of absorbed dose from
X-ray, gamma-ray, and electron exposure of target organs is provided in Table 2 [NCRP2001].
Table 2 Deterministic effects [NCRP2001]
Absorbed
Effect Target organ
dose (Gy)
Sterility (temporary) Testes 0.15
Nausea Whole body 0.35
RBC depression Bone marrow 0.5
Skin reddening (reversible) Skin 2
Sterility (permanent) Ovaries 2.5–6
Testes 3.5
Vomiting Whole body 3
Hair loss (temporary) Skin 3–5
Erythema Skin 5–6
Death (LD50/30)† Whole body 4
†LD50/30 is the Lethal Dose where 50% of an exposed population dies within 30 days.
LD50/60 is a similar definition at 60 days post-exposure.
There are three general categories of acute radiation sickness, as presented below.
1. Hematopoietic
 Dose range 3–8 Gy
 Radiation damages precursors to red and white blood cells and platelets
 Prodromal phase may occur immediately
 Symptoms include septicemia
 Mixed survival
 Examples include Chernobyl personnel (203 exhibited symptoms; 13 died)
2. Gastrointestinal
 Dose > 10 Gy
 Symptoms include abdominal pain/fever, diarrhea, and dehydration
 Death in 3–10 days (no record of human survivors above 10 Gy)
 Examples include Chernobyl firefighters
3. Cerebrovascular
 Dose > 100 Gy
 Death in minutes to hours
 Examples include criticality accidents.
Radiation exposure to high doses is well known to induce changes in blood-count levels. There
are numerous components of blood and a variety of ways to subdivide blood characteristics.
The four factors that are often examined when high radiation exposure is expected are lympho-
cytes, neutrophils, thrombocytes, and hemoglobin, described below:
 Lymphocytes are white blood cells that govern the body’s immune response (directly
fighting disease and infection)
 Neutrophils are a subset of white blood cells that fight infection
 Thrombocytes (platelets) are active in blood clotting
 Hemoglobin is the component of red blood cells that carries oxygen.
An example of a hypothetical complete blood count with differential (CBC w/ diff) after expo-
sure to a large dose (insult > 2 Gy whole body) of radiation is provided in Figure 12. It may be
seen that immediately after exposure, there is a sharp decrease in lymphocytes, followed by a
spike in neutrophils. The neutrophil spike is the body’s response to what it believes may be an
infection. Three distinct stages of illness are evident through the exposure: (1) latent period
from exposure to approximately 20 days, (2) manifest illness, where the exposed person is
extremely sick, from day 20 to 35, followed by (3) recovery, assuming the exposed person has
survived the blood counts dropping to almost zero. Note that Figure 12 represents a very high
whole-body exposure and that the actual blood counts and time frames can vary greatly from
person to person. Note also that models have been developed to determine the severity of
radiation exposure retrospectively using sequential lymphocyte counts post-exposure
[Goans1997].
24 20 400
Lymphocytes and
350
Thrombocytes x 1000
20
Neutrophils x 1000
Hemoglobin (grams) 10 300
16
250
12 200
150
8
100
4
lymphocytes 50
Insult
0 0
-10 0 10 20 30 40 50 60
Days
Figure 12 Differential blood count after severe radiation insult (adapted from [Gusev2001]
and [Cember2009])
The universal principles of radiation protection dictate that deterministic effects must be
avoided in occupational exposure. However, this is not the case with medical exposure because
a deterministic-level dose may be required for the medical procedure. A typical example is
radiation therapy for treatment of cancer, where deterministic-level doses are routinely applied
to treat the cancer.
2.3.2 Stochastic effects

Stochastic effects are those effects which are assumed not to have a threshold dose and for
which the severity of the effect is independent of absorbed dose. Stochastic effects may mani-
fest in delayed or late somatic effects. The usual endpoint stochastic effect considered for
radiation exposure is cancer induction (for example, leukemia, bone cancer, or lung cancer),
which has been inferred at absorbed-dose levels on the order of hundreds of mGy. In addition,
genetic and hereditary effects have been observed using animal studies, although these effects
have never been observed in human populations exposed to ionizing radiation. Stochastic
effects can occur from both acute and chronic exposures.
The often-quoted value for fatal cancer risk as a function of whole-body radiation exposure is
5% per Sv (dose definitions are found in Section 3.1). In other words, if a population of 100,000
were exposed to a 1-Sv whole-body dose, then 5000 excess fatal cancers would be expected
(note that there is approximately a 40% baseline fatal cancer incidence across the population).
A tabulation of the current ICRP [ICRP2007] guidance on stochastic risk coefficients is provided
in Table 3. ICRP recommends continued use of the 5% per Sv value.
Table 3 Stochastic risk coefficients (adapted from ICRP (2007)

(2007)))
Exposed population Cancer (%/Sv) Heritable (%/Sv) Total (%/Sv)
All 5.5 0.2 5.7
Adult only 4.1 0.1 4.2
A study performed by [Brenner2003

Brenner2003] indicated that, for doses within a range of 50–100
50 mSv
(protracted exposure) and 10––50 50 mSv (acute exposure), there was direct epidemiological
evidence from human populations to indicate that exposure to ionizing radiation increased the
risk of certain cancers. The primary argument to support linear non
non-threshold
threshold response at low
dose is nott from scientific observation, but from mechanistic arguments and assumptions.
Based upon the scientifically observed evidence, there is no conclusive proof of any deleterious
endpoint effect below 10 mSv and only weak evidence of an effect below about 50 mSv. In any
case, the risk factors below approximately 100 mSv are very low for all exposure scenarios.
Note that effects from radiation exposure cannot be definitively attributed to an individual.
That is to say, a deterministic effect that may be observ
observed
ed in an individual at one dose level
may not be observed in another individual at the same dose level. This occurs because indi-
viduals have different sensitivities to ionizing radiation exposure. Likewise, cancer risk from
radiation exposure cannot be ext extrapolated
rapolated to an individual for a given exposure. Radiosensitiv-
Radiosensiti
ity is organ-specific because not all organs have the same vulnerability to radiation. Table 4
depicts the relative sensitivity variation among some critical organs.
Table 4 Relative human
human-organ
organ vulnerabilities to ionizing radiation
Although cellular and animalal studies provide indicators of risk, the primary source of data for
radiation-exposure
exposure risk modeling is epidemiological studies [UNSCEAR2008].
[UNSCEAR2008] The primary
source of data used for risk models is the Japanese atomic-bomb bomb survivors. Other exposed
groups thatt have contributed data to risk modeling include radiotherapy cancer patients
(cervical, endometrial, childhood, breast, Hodgkin
Hodgkin’s lymphoma, etc.), ), radiotherapy patients
with non-malignant
malignant conditions (spondylitis, thymus, tonsils, ringworm, etc
etc.),
), diagnostic
diagnost radiol-
ogy patients (tuberculosis fluoroscopy, pelvimetry, scoliosis, etc
etc.), workers with occupational
©UNENE, all rights reserved.. For educational use only, no assumed liability. Radiation Protection and Envir. Safety – September 2014
exposure (radium dial painters, miners, radiologists, nuclear workers, etc.), and people experi-
encing environmental exposure (nuclear weapons fallout, Chernobyl, Techa River, etc.).
Radiation is an incredibly weak carcinogen, and as therefore estimation of risk from exposure at
low doses requires the use of risk models, which are described in the following section.
2.4 Radiation Risk Models

In the simplest terms, risk may be defined by Eq. (5):
Risk=  Probability of occurrence  ×  Consequence  (5)
Therefore, it is possible to have a low-probability event with high consequence with a low risk
factor, or a high-probability event with low consequence with a low risk factor. However, if
neither probability of occurrence nor consequence is non-trivial, then it is possible to have a
high risk factor. In terms of radiation risk, the BEIR Committee [BEIRVII2006] defines risk as:
“the chance of injury, loss, or detriment; a measure of the deleterious effects that may
be expected as a result of action or inaction”.
To understand risk from radiation exposure, a fundamental understanding of what the risk
detriment parameter is (for example, cancer) and the background incidence of this detriment is
required. Incidence of effect, or endpoint, can be estimated using in vitro or in vivo studies of
humans, animals, or plants (of organisms, components of organisms, or both, down to the DNA
level). Risk to populations from exposure can be estimated using the principles of epidemiology,
which is the study of the causal factors of the frequency of disease in humans.
Risk from radiation exposure is typically quantified as some probability of endpoint effect
(incidence) per unit dose exposure. The relative risk (RR) is a common measure of risk and is
defined by Eq. (6):
Incidence in exposed population
RR = . (6)
Incidence in unexposed population
An important metric for risk estimate is based on the increased likelihood of an endpoint
(cancer) as a result of exposure to a carcinogen (radiation) and is termed the excess relative risk
(ERR), defined by Eq. (7):
Incidence in exposed population
ERR = RR - 1 = - 1. (7)
Incidence in unexposed population
The excess absolute risk (EAR), sometimes called the attributable risk, is a measure of the
discrepancy in incidence rates between exposed and non-exposed populations and is given by
Eq. (8):
EAR = Incidence in exposed population  Incidence in unexposed population . (8)
For example, if the cancer incidence in an exposed population is 0.5 Gy-1 and in an unexposed
population is 0.4, the RR is 1.25 Gy-1, and the ERR is 0.25 Gy-1. The EAR requires that the inci-
dence in the unexposed population be resolved to a percentage by multiplying by the exposure
(Gy). For a 100-mGy exposure, RR = 0.125, ERR = 0.025, and EAR is negative. When RR >1, then
exposure increases the risk, and if RR <1, then exposure does not increase risk. In this hypothet-
ical case, the exposure does not increase cancer risk.
It is often desirable to estimate whether a given exposure is the cause of an observed endpoint
(for example, cancer). The metric used for this is the probability of causation (PC), which is
given by Eq. (9):
ERR
PC= . (9)
1+ERR
In the example above, the PC is calculated as 0.024.
Understanding the risks of radiation exposure means understanding the cause-effect relation-
ships between endpoints and exposure. Some cancers are more strongly linked to radiation
exposure than others. Table 5 presents a variety of cancers in the following categories: (i)
strong association with radiation exposure, (ii) moderate association with radiation exposure,
although strongly influenced by other risk factors, and (ii) little or no association with radiation
exposure.
Table 5 Association of cancers with radiation exposure
Strong Moderate Weak to none
Leukemia (excluding Childhood leukemia Bone

chronic lymphocytic)
Thyroid Bladder Brain
Breast Esophageal
Colorectal Kidney
Lung Multiple myeloma
Stomach Non-Hodgkin’s lym-

phoma
Ovarian
Stochastic risk, such as the incidence of cancer, can be plotted as a function of dose to generate
a dose-response curve that represents the incidence of an endpoint as a function of increasing
dose. There are numerous sources of data for generating risk curves; some major cohorts
where data have been obtained are presented in Table 6.
Table 6 Epidemiological categories for radiation risk studies
Population Category Comments
Japanese atomic bomb survivors Numerous endpoints studied

Events:
Environmental exposures
Chernobyl NPP4 accident
Techa River cohort
Fukushima disaster
Nuclear-weapons fallout
Natural background radiation
Cohorts:
Occupational exposures
Radium dial painters
Miners
Radiologists
Technologists
Nuclear-energy workers
Veterans
Endpoint cancers:
Radiotherapy cancer treatments
Cervical
Endometrial
Childhood
Breast
Hodgkin’s lymphoma
Conditions treated:
Radiation therapies (non-cancer)
Spondylitis
Mastitis
Thymus
Infertility
Tonsils
Menstrual disorders
Otitis Media
Ulcers
Ringworm
Hemangioma
Conditions diagnosed:
Diagnostic radiology
Tuberculosis
Scoliosis
Pelvimetry
All other diagnostic procedures
Radionuclide internalizations Isotopes of: Th, I, P, Ra, U, Pu
Generally speaking, much of our knowledge about dose response has been obtained from
relatively high dose / high dose rate exposure, whereas most non-medical exposure is low dose
/ low dose rate exposure. Figure 13 depicts a hypothetical, yet realistic, dose response curve
(incidence versus dose, where incidence could be relative risk of cancer, for example). The solid
circles represent points for which there are statistically significant data with reasonable confi-
dence levels (for example, data from Japanese atomic bomb survivors). The open circles
represent non-statistically significant data with low confidence. “High dose” would generally
refer to data obtained above a 1-Gy whole-body dose; “low dose” would generally refer to data
below approximately 0.1 Gy. Some risk estimates (for example, leukemia) are generated using a
linear quadratic curve [BEIRVII2006] through the data and through the point (0,0), which
represents “zero risk at zero dose”. Most risk estimates, however, are generated using a linear
curve through the data and through point (0,0). This is termed the linear no-threshold (LNT)
hypothesis of radiation risk and is very contentious in the radiation-protection community
[Bond1996] [Scott2008] [Siegel2012].
Incidence
Adequate fit to
data possible
in this region
(0,0)
Low Dose High Dose
Dose
Figure 13 Dose-response curve showing two common models
The primary reason why the linear no-threshold hypothesis of radiation risk is contentious is
that, as discussed in Section 2.3, there is no conclusive evidence below approximately 100 mGy
to demonstrate deleterious effects of exposure to ionizing radiation. That being said, regulators
have adopted the linear no-threshold hypothesis for radiation risk because it is considered
conservative (that is, to overestimate the risk from ionizing-radiation exposure).
The various models for dose response [Cember2009] can be represented by a generalized
expression given by Eq. (10):
  D D 

f  D    0  1 D   2 D 2 e 
2
1 2
, (10)
where the dose-response function is given by f(D), D is the dose, 0 is the natural incidence of
the effect (for example, spontaneous cancer incidence), 1 and 2 are linear and quadratic
fitting parameters, and β1 and β2 are parameters used in representation of cell killing or mortal-
ity only at higher doses (that is, the killing function dominates at high dose).
In terms of relative risk (RR), Eq. (10) becomes Eq. (11):
  D D 
 0   1 D   2 D 2  e
2
f D
1 2
RR   . (11)
0 0
A graphical representation of the various functions for Eq. (10) is provided in Figure 14. The
values for the fitting parameters have been selected arbitrarily, and a background incidence (of
a given effect) was arbitrarily selected as 10-3. At low dose, the linear, linear quadratic, and full
expressions are basically identical. At high dose, the quadratic function matches the linear
quadratic function, and clearly cell killing/mortality attenuates the linear quadratic curve.
10
1
Incidence
linear
0.1
quadratic
linear-quadratic
w/ killing
0.01
0.001
0.001 0.01 0.1 1 10
Dose
Figure 14 Generalized dose-response curves

Alternative theories of dose response include models that incorporate threshold responses
(that is, risk only increases after a threshold dose) and hormetic response (see Figure 15).
Hormesis may be defined as “a process in which exposure to a low dose of a given insult that is
damaging at higher doses induces an adaptive beneficial effect on the cell or organism” [Matt-
son2008]. There has been an abundance of literature published in support of both threshold
and hormetic responses (see, for example, the journal Dose-Response [DoseResponse2013]).
The relative risk-response curve may be adjusted to compensate for low-dose hormetic effects
using Eq. (12):
  D D 
 
2
 1 D   2 D 2 e 1 2
RR  1  B  D  * PROFAC  ,
0
(12)
0
where PROFAC is the protection factor, which is the population average probability of cancer
prevention given activated natural protection and B(D) is the benefit function, which is the
probability of activated natural protection [Scott2012], which increases at low doses far more
than f(D). At low doses, f(D)/0 is essentially unity, and at low doses, the relative risk may be
approximated by Eq. (13):
RR  1  B  D  * PROFAC  . (13)
Incidence
(0,0)
hormesis High Dose
Dose
Figure 15 Alternative dose-response curves
2.4.1 Radiation risk controversy

Almost immediately after the discovery of X-rays by Wilhelm Roentgen in late 1895, the emana-
tions were being put to use for therapeutic and diagnostic medical applications. The following
year, reports of skin damage due to X-ray exposure were published. Early radiation-protection
guidance focused on eliminating deterministic effects such as skin erythema. By the beginning
of the 20th century, evidence was appearing that radiation might be responsible for tumours in
radiologists, and supporting evidence of cancer induction due to radiation was obtained by
studying pitchblende miners and radium dial painters. No formal standards for radiation
protection existed until 1913, and at that point, they were established at thresholds for limiting
skin erythema and subsequently expanded for internalized radionuclides in terms of maximum
permissible body burden. Our current system of understanding of effects of ionizing radiation
exposure has been developed over the past 100 years by (a) performing cellular-level studies,
(b) performing animal studies, and (c) performing epidemiological studies on exposed popula-
tions [BEIRVII2006].
Scientific debate on low-level radiation effects in the radiation-protection community involves
the model used to determine risk from exposure. There is no unanimous agreement on the
model of stochastic risk as low doses. There is scientific support for the linear-non-threshold
model, threshold dose models, and also hermetic models. Although our current system of
radiation-protection guidance is based upon the linear non-threshold model of dose-response,
the concept of threshold dose is apparent as early as the writings of Paracelsus (1493-1541), in
his famous statement “What is there that is not poison? All things are poison and nothing is
without poison. Solely the dose determines that a thing is not a poison” [Deichmann1986]. In
abbreviated terms, Paracelsus’ statement becomes “The dose makes the poison”. This suggests
that, upon exposure to an anthropogenic chemical, there is a threshold dose above which
something becomes poisonous to the body. The concept of threshold dose, established in the
16th century, was abandoned for stochastic-effect radiation-protection purposes in the 20th
century. As early as the 1955 Recommendations of the International Commission on Radiologi-
cal Protection, statements were promulgated that “no radiation level higher than natural
background can be regarded as absolutely ‘safe’ ” [Clarke2005], despite there being no conclu-
sive proof that low-dose radiation causes any harm, stochastic or non-stochastic. It is this belief,
manifest in the so-called linear non-threshold (LNT) hypothesis of radiation risk described
above, that has divided public opinion and worker understanding of the effects of ionizing
radiation. Various misapplications and misinterpretations of radiation risk using the linear non-
threshold hypothesis [Bond1996] [Siegel2012] have made general understanding of the risk of
exposure to low-dose radiation confusing and controversial.
2.5 Effects on Pregnancy

In much of the literature, the terms zygote, embryo, fetus, and conceptus are used. Accepted
medical definitions are provided as follows [Mosby1990]:
Zygote [Gk zygon+sporos seed]: the developing ovum from the time it is fertilized until it is
implanted in the uterus.
Embryo [Gk en in, bryein to grow]: 1. Any organism in the earliest stages of development; 2. in
humans, the stage of prenatal development between the time of implantation of the fertilized
ovum about 2 weeks after conception until the end of the seventh or eighth week. The period is
characterized by rapid growth, differentiation of the major organ systems, and development of
the main external features.
Fetus [L, fruitful]: the unborn offspring of any viviparous animal after it has attained the par-
ticular form of the species, more specifically, the human being in utero after the embryonic
period and the beginning of the development of the structural features, usually from the eighth
week after fertilization until birth.
Conceptus [L concipere to take over]: the product of conception; the fertilized ovum and its
enclosing membranes at all stages of intrauterine development, from the time of implantation
to birth.
2.5.1 Teratogenesis
The conceptus is a blanket definition which encompasses the zygote, embryo, and fetus stages.
Teratogenic insult may be defined as any exposure that is capable of harming the conceptus
(for example, inducing birth defects). Radiation is extremely harmful to the conceptus, and to
children in general, for two simple reasons: (1) a majority of the cells are in active growth
stages in young organisms, and as was demonstrated by the law of Bergonie and Tribondeau,
rapidly dividing cells are more radiosensitive; and (2) exposure in the young allows for more
time for observable effects to manifest before death.
The human conceptus has an approximate nine-month gestation period, with different radia-
tion sensitivities over this time frame. During the first two weeks of gestation (the zygote
stage), the egg is fertilized and implanted in the uterine wall. During this phase, the cells divide
at a rapid rate and are highly sensitive to radiation. During this stage, cell killing is the most
likely outcome from large doses of radiation. During the following approximately three to seven
weeks of gestation (the embryonic stage), differentiated organ development begins, and the
embryo is susceptible to cell killing or congenital malformations at moderate to high doses of
radiation exposure. From approximately week 8 until birth (the fetal stage), there is rapid
growth which makes the fetus susceptible to birth anomalies such as mental retardation. After
approximately 17 weeks gestation, all the brain cells have formed and do not further divide,
and therefore the most sensitive period for teratogenic effects such as mental retardation is
believed to be between 8 and 17 weeks gestation for moderate- to high-dose radiation expo-
sure.
The manifestation of stochastic effects (such as cancer-cell induction) is more probable for in-
utero exposure. This is recognized by the International Commission on Radiological Protection
by suggesting that the risk is a few times that of the population as a whole. Using the value of
5.5%/Sv from Table 3, the cancer-risk coefficient for in-utero exposure is estimated at approxi-
mately 17%/Sv. Note also that most regulatory agencies have adopted a “balance of preg-
nancy” dose limit on workers, which is designed to limit in-utero exposure. Operationally, when
a worker declares a pregnancy to her employer, the dose to the conceptus must be limited to a
pre-determined value (in the 2007 ICRP recommendations [ICRP2007], this is 1 mSv).
2.5.2 Perception of risk

Of interest in pregnancy and radiation exposure is the perception of risk by the mother. There is
very little scientific evidence that points to increased risk of harm to the conceptus at low
doses. To attempt to answer the question, “How do pregnant women perceive the risk of low-
level radiation exposure?” [Bentur1991] examined the perception of teratogenic risk for expo-
sure to ionizing radiation during pregnancy, specifically considering women who were sched-
uled to receive diagnostic radiation exams during pregnancy. The control group for the study
was women exposed to non-teratogenic drugs and chemicals during their pregnancy. In this
study, [Bentur1991] claimed that, “The two probable factors in creating this misperception of
risk are pregnancy-induced anxiety and misinformation. Anxiety is enhanced by the known
effects of nuclear disasters and due to continuous flow of information in the medical literature
about the long-term effects of therapeutic irradiation on fertility and carcinogenicity”. In the
study, women were counselled about the low risk factors for diagnostic radiology procedures.
Before counselling, the pregnant women exposed to radiation assigned themselves a higher risk
than the non-exposed population, even though their doses from diagnostic procedures would
be extremely low. After counselling, the perception of teratogenic risk did not differ between
the groups. This result is very important because it suggests that education and consultation
with pregnant women who may be exposed to ionizing radiation can reduce stress and anxiety
from the exposure and possibly reduce the risk of the mother selecting therapeutic abortion
due to unfounded perceptions of the exposure.
2.6 Summary
From a toxicological standpoint, more is known about the effects of radiation on living organ-
isms than any other insult, including chemicals. As far as carcinogenic toxins go, ionizing radia-
tion is one of the weakest carcinogens on the planet. Effects have been investigated using
studies of DNA/molecules/cells, studies on animals, and human epidemiological studies. The
time frames for observable effects extend from subseconds (physical) to seconds (chemical) to
many years (biological). The endpoints for effects may be (i) radiation enters the body, but
misses important targets (a highly likely “endpoint”); (ii) radiation does not cause any damage
to the targets; (iii) radiation damages the target, but the target repairs itself; (iv) damaged cells
may die; and (v) damage cells may change (mutate). High doses of radiation may cause prompt
deterministic effects, such as skin burns. Both high and moderate doses of radiation may
increase the risk of stochastic effects such as cancer. Low-dose radiation exposure had not been
observed to cause deleterious effects in the populations studied, although the risk of exposure
has been extrapolated from high-dose observable effects to the low-dose regime using the
linear no-threshold (LNT) dose-response model.
What is known about ionizing radiation exposure can be summarized as:
 Radiation is a very weak carcinogen
 Probability of cancer is a function of dose (increased risk with increased dose; severity of
cancer is not a function of dose)
 There is weak-to-no evidence of cancer effects at doses below approximately 100 mSv.
What is NOT known about ionizing radiation exposure:
 Whether there are any negative effects of exposure below approximately 100 mSv
 Whether there are any beneficial effects of exposure below approximately 100 mSv
 Whether there are any effects at low dose other than cancer and leukemia
 Whether there are any inherited effects at any exposure.
Some practical considerations for understanding ionizing-radiation exposure:
 “Normal” cancer incidence is very high (30%–40% over the population)
o Cause-effect relationships cannot be proven on an individual basis; relative risk
can only be established for large exposed groups
 Stochastic effects have a long latent period
o Leukemia: 2–7 years post-exposure
o Other cancers: 10–50 years post-exposure
 Exposure must occur to tissue for radiation-induced cancer to form
o Cancer will not initiate in an organ of the body unless that organ has received a
dose
 Certain cancers are not associated with low to moderate doses of ionizing radiation
o Hodgkin’s disease
o Non-Hodgkin’s disease
o Chronic lymphocytic leukemia
o Cutaneous malignant melanoma
o Uterine cancer
o Prostate cancer
 Risk from radiation exposure to child and conceptus are enhanced
o Rapidly dividing cells are radiosensitive
o Longer lifespan post-exposure enables latent cancers to manifest.
The risks of ionizing radiation exposure have been quantified using a variety of dose-response
models. A commonly used model is the linear no-threshold (LNT) model, which is generally
considered to be conservative and as such is used extensively by regulators. There are numer-
ous other models, including hormesis models, which are actively debated in the scientific
literature and about which scientists have shown evidence for and against. Dose-protection
guidance, the LNT approximation, and how the “as low as reasonably achievable, social and
economic factors taken into account” (ALARA) mantra is used are discussed in Section 3.
3 Dosimetry, Dose Limitations, and Guidance

The underlying principle of any process that uses ionizing radiation is that the benefit from
potential exposure outweighs the potential risk from exposure. Radiation protection is guided
by regulations that eliminate deterministic effects and limit stochastic effects from exposure.
The relationship between scientific discovery, international synthesis of data, and adoption of
national regulations is illustrated in Figure 16.
National
Regulations
International
Recommendations
Scientific Knowledge
Figure 16 Radiation protection regulation pyramid

Scientific knowledge and discovery is the broad base from which we derive fundamental
radiation-protection quantities. This knowledge is disseminated through national reports,
industrial reports, peer-review publications, conferences, and many other scientific venues.
International bodies periodically synthesize the available data from the scientific literature and
produce recommendations on various aspects of radiation protection. International recom-
mendations form the basis of national standards, but are not themselves regulations. National
regulations are adopted by government bodies country by country, and although these bodies
tend to follow international recommendations, they are not obligated to do so. In Canada,
there are various governmental agencies involved in establishing radiation-protection guide-
lines, with different areas of responsibility depending upon the source of radiation. For exam-
ple:
 Nuclear reactor operations – Canadian Nuclear Safety Commission (CNSC)
 Radioactive sources and waste – Canadian Nuclear Safety Commission (CNSC)
 Uranium mining and milling operations – Canadian Nuclear Safety Commission (CNSC)
 High-energy particle accelerators – Canadian Nuclear Safety Commission (CNSC)
 X-ray regulations – each province regulates independently
o X-ray use – Health Canada (HC) guidelines
 Naturally Occurring Radioactive Material (NORM) - each province regulates independently
o NORM mitigation – Health Canada (HC) guidelines.
Other agencies that have input into radiation protection regulations include Environment
Canada (EC) and Transport Canada (TC). A number of national agencies in Canada are involved
in safety and security of radioactive material, including (i) all the above-named agencies, (ii)
Public Safety Canada, (iii) Canadian Security Intelligence Service (CSIS), (iv) Department of
National Defence, and (v) Canada Border Services Agency (CBSA).
This section discusses fundamental definitions related to ionizing-radiation dose, background
radiation, and national/international guidance regarding dose and dose limitations. Background
information on the historical development of the radiation protection-system can be found in
[Inkret1995] [Lindell1996] [Jones2005] [Clarke2005] [Walker2000].
3.1 Dose Definitions

The concepts of absorbed dose, equivalent dose (radiation weighting factors), and effective
dose (tissue weighting factors) are discussed here in relation to risk concepts. In addition, radon
guidance will be introduced.
3.1.1 Exposure
Before discussing dose, it is worthwhile to discuss, in a historical sense, exposure. In the early
days of X-rays, a common dosimeter was a piece of dental film attached to a paper clip. The
daily allowable exposure was an exposure that was just enough so that some “fogging” of the
film could be observed on processing. This was known as a “paper-clip unit” of radiation expo-
sure and amounted to an early dosimeter. For larger doses such as might be used in therapeutic
medicine, a “skin erythema unit” was used, which was an exposure that would just cause visible
reddening of the skin. Neither unit was what could be considered biologically meaningful,
although they did provide a measure of protection.
Exposure was defined for X- (and gamma) radiation in terms of air ionization, and the original
unit of air ionization established in 1928 was the Roentgen (R). The current definition is given
as:
1 R  2.58 104 C kg 1 of air .
Absorbed dose was subsequently defined as the energy absorbed per unit mass from any kind
of ionizing radiation in any target. As such, the absorbed dose is a physical quantity. To calcu-
late the absorbed dose in air from a given exposure, the ionization potential of air (which is 33.7
eV/ion pair or 33.7 J/C) can be used. Using this value, the absorbed dose in air from an expo-
sure of 1 Roentgen can be calculated as:
 C  J J
1 R   2.58 104  33.7   8.7 10 3
 kg   C kg .
A similar calculation of the soft-tissue ionization potential yields 9.5x10-3 J kg-1.

Because energy deposition per unit mass was deemed to be close to actual biological damage, a
new unit was developed, called the Rad (standing for Radiation Absorbed Dose). The Rad was
defined as an energy deposition of 100 erg per gram of material, where 107 erg = 1 Joule. The
Roentgen could therefore be defined as 87 erg/g in air (or 95 erg/g in soft tissue), and the Rad
was defined as 100 erg/g. For regulatory purposes, it was therefore assumed that if one could
measure the ionization of air in Roentgen (R), this would be approximately equivalent to the
absorbed dose in tissue (Radtissue). To convert exposure to absorbed dose for any medium, Eq.
(14) can be used:
m
87  m
D  Rad     E ( Roentgen) , (14)
100  a
a
where
D = absorbed dose in Rad
E = exposure in Roentgen
μm = energy-absorption coefficient for tissue (or any medium)
μa = energy-absorption coefficient for air
ρm = tissue density (or density of any medium)
ρa = air density.
Although the historical unit of Rad is still used in some instances, SI units are more common
internationally and are promulgated through all major international guidance bodies such as
IAEA, ICRP, and UNSCEAR.
3.1.2 Absorbed, equivalent, and effective dose

Three fundamental quantities are required when considering radiation protection:
 Absorbed dose (D) is energy absorbed per unit mass. 1 gray (Gy) = 1 joule per kilogram
(J/kg)
 Equivalent dose (HT, units of Sievert, Sv) takes into account the relative biological effec-
tiveness of different radiation types
 Effective dose (E, units of Sievert, Sv) takes into account the potential for detrimental
effects to the various organs and tissues.
It is important to realize that absorbed dose is a physical quantity, whereas equivalent and
effective dose are derived quantities used for radiological-protection purposes. In addition,
these derived quantities use reference “individuals” and assigned factors, roughly approximate
risk on a population basis, and are not useful for accurate estimation of risk on an individual
basis.
The absorbed dose can be expressed as in Eq. (15):
ET
DT ,r  , (15)
mT
where
DT,r is the absorbed dose to target T from radiation r (J/kg, or Gy)
ET is the energy deposited in target T (Joule)
mT is the mass of target T (kg).
The equivalent dose is calculated by summing over all different radiations interacting in the
target material using Eq. (16):
H T   wr DT ,r , (16)
where
HT is the equivalent dose to target T (Sv)
wr is the radiation weighting factor for the radiation type.
Radiation weighting factors are related to the relative biological effectiveness (RBE) of different
kinds of radiation, which is an indication of the relative amount of radiation damage done to
tissue for a given absorbed dose. The recommended radiation weighting factors from ICRP103
are provided in Table 7.
Table 7 Radiation weighting factors [ICRP2007]
Radiation Radiation weighting factor, wr

Photons (X- and gamma) 1
Electrons (incl. beta) 1
Protons 2
Alpha particles and fission fragments 20
  ln E 
2
Neutrons (continuous function of energy) 2.5  18.2e 6

E  1 MeV
  ln(2 E ) 
2
N.B.: in ICRP60 [ICRP1990], the centre expression 5.0  17.0e 6

1 MeV  E  50 MeV
represented all wr)   ln(0.04 E ) 
2
2.5  3.25e 6
E  50 MeV
The effective dose, E, is calculated using the equivalent dose and tissue weighting factors using
Eq. (17):
E   wT HT , (17)
where wT is the tissue weighting factor for target T.

Tissue weighting factors have been established as indicators of the relative sensitivities of
different tissues to radiation exposure. The ICRP103 tissue weighting factors, along with ICRP60
and ICRP26, are given in Table 8. Blank entries mean that the organ was not uniquely identified
in the guide and was included in “remainder”. The sum of tissue weighting factors over all
tissues is unity.
Table 8 Tissue weighting factors [ICRP1977,1990,2007]
Tissue weighting factor, wT

Organ
ICRP26 (1977) ICRP60 (1990) ICRP103 (2007)
Gonads 0.25 0.20 0.08
Red Bone Marrow 0.12 0.12 0.12
Lung 0.12 0.12 0.12
Breast 0.15 0.05 0.12
Thyroid 0.03 0.05 0.04
Bone surface 0.03 0.01 0.01
Remainder 0.30 0.05 0.12
Colon 0.12 0.12
Stomach 0.12 0.12
Bladder 0.05 0.04
Liver 0.05 0.04
Esophagus 0.05 0.04
Skin 0.01 0.01
Salivary glands 0.01
Brain 0.01
Sum 1.00 1.00 1.00
Unfortunately, ICRP adopted the same base unit (Sv) for both equivalent dose and effective
dose, which has the potential to lead to confusion in dose representation.
The equivalencies between classical and SI dose units (and activity, which was discussed in
Chapter 3) are provided in Table 9. Dose rate units are the same base units as a function of time
(for example, Gy/h).
Table 9 Dose units
Unit Classical unit SI unit Unit conversion
Activity Curie (Ci) Becquerel (Bq) 1 Ci = 3.7x1010 Bq
Absorbed dose Radiation absorbed dose (Rad) Gray (Gy) 100 Rad = 1 Gy
Equivalent dose Roentgen equivalent man (Rem) Sievert (Sv) 100 Rem = 1 Sv
Effective dose Roentgen equivalent man (Rem) Sievert (Sv) 100 Rem = 1 Sv
In Canada, it is not uncommon to find references to classical units used as operational quanti-
quant
ties in nuclear power plants. However, SI units are the approved units of radiation dosage
dos
according to the CNSC regulations.
As was discussed in Section 2.3.2 and demonstrated in Table 3,, after exposure to low-dose-rate
low
radiation, the ICRP [ICRP2007] suggests detriment
detriment-adjusted
adjusted nominal risk coefficients for cancer
and heritable effects in terms of risk per Sv (% per Sv).
The relationship between the various dose quant
quantities
ities and endpoint estimates
estimate is depicted in
Figure 17. The relative uncertainty in physical and derived quantities is shown by increasing
arrow thicknesses,, with the uncertainty compounding with each step to calculate the endpoint
risk. This demonstrates that, although absorbed
absorbed-dose
dose estimates may be made with relatively
good confidence, the ability to estimate population risk from low dose / dose rate exposure is
highly uncertain.
Figure 17 Relationship between dose and endpoint
3.1.3 Committed dose

The committed effective dose (CED) is the effective dose that an individual will eventually
receive after having taken a radionuc
radionuclide
lide into the body (inhalation, ingestion, injection,
injection or
skin/wound absorption). Committed usually means to 50 years, although 70 years is used when
a childhood intake has occurred.
3.1.4 Collective dose

The collective dose is the sum of individual doses over a population, with units of person-Sv.
person If
the collective dose is divided by the number of persons in the exposed population, then an
average individual dose can be estimated. Although this quantity has been used in the litera-
liter
ture, ICRP 103 does not recommen
recommend the use of collective dose.
3.1.5 Derived quantities

Derived quantities are values or results that are made up, or derived, from other values. Often,
derived quantities are made up of other fundamental quantities. Four primary derived quanti-
ties are discussed in this section: (i) dose-conversion factors (DCF), (ii) annual limits on intake
(ALI), (iii) derived air concentration (DAC), and (iv) radon progeny (WL; WLM).
3.1.5.1 Dose-conversion factors (DCF)

Dose-conversion factors (DCF) are tabulated derived values that enable conversion from one
quantity to another. For example, dose-conversion factors are tabulated for inhalation and
ingestion for workers in ICRP68 [ICRP1994] and the public in ICRP72 [ICRP1996a]. A dose-
conversion factor for an internalized radionuclide typically has units of Sv/Bq committed
effective dose (CED), and therefore if the intake (in Bq) is known, the estimated CED in Sv can
be immediately calculated. In the case of internalized radionuclides, values are available for
inhalation and ingestion, for differing particle sizes, and for the different chemical forms of
many radionuclides. Examples of tabulated dose conversion factors for 14C are shown in Table
10.
Table 10 Sample inhalation dose coefficients [ORNL2013]
Inhalation dose coefficients (Sv / Bq) from ICRP 68
Chemical form= CO2 CO CH4 Vapour
Effective (ICRP 60) 6.5E-12 8E-13 2.9E-12 5.8E-10
These conversion factors are derived using compartmental biokinetic models for internal
dosimetry. Other dose-conversion factors can include external DCFs such as those found in
ICRP74 [ICRP1996b] and discussed in Section 5.
3.1.5.2 Annual limit on intake (ALI)

The annual limit on intake (ALI) is the activity (Bq) of a radionuclide which, if taken internally,
would result in a dose equal to the annual limit. As such, the ALI is normally defined as the dose
limit divided by the effective dose per unit intake. Both equivalent and effective doses are
considered; the more restrictive limit is used. Note that simply multiplying the number of ALIs
taken in by 20 mSv to calculate committed effective dose may not be a correct estimate be-
cause some ALIs are based on effective dose and some on equivalent dose. The ALI is calculated
using Eq. (18):
Dlimit  Sv  0.02 Sv
ALI  Bq   
 Sv  Sv . (18)
 wT HT  Bq  DCF
Bq
 
3.1.5.3 Derived air concentration (DAC)
The derived air concentration (DAC) is defined as the air concentration (Bq/m3) that would lead
to the inhalation of 1 ALI by a reference person over 2000 hours (50 weeks x 40 h/week)
assuming that the reference person inhales 1.2 m3/h (light work). The DAC is calculated using
Eq. (19):
 Bq  ALI  Bq  ALI  Bq 
DAC  3   
 m  2000 hr 1.2 m
3
2400 m3 . (19)
hr
DACs enable rapid estimation of committed effective dose rate because the DAC represents 1
ALI (in the case of Canadian regulations, 20 mSv) in 2000 h, or stated in another manner, 1 DAC
= 10 μSv/h. The number of DAC-h corresponds to the CED.
3.1.5.4 Radon-derived quantities (WL; WLM)

Radon gas (222Rn) is considered harmful primarily because it has numerous short-lived daugh-
ters such as polonium, bismuth, and lead which are solids and readily adsorbed on particles of
fine dust than can be inhaled. When air containing radon is breathed in, it can be readily
expelled without inducing any damage whatsoever. However, the radon daughters, both from
gas decay and dust, can be trapped in lung tissue, where they can induce local damage to lung
tissue through alpha-particle interactions. The combined alpha energies of the daughters per
unit volume of air are the most important parameters to be considered when estimating
damage to lung tissue. There is a special unit used to define this quantity, called the working
level (WL).
The working level (WL) is defined as any combination of the short-lived decay products of radon
(218Po, 214Pb, 214Bi, and 214Po) in one dm3 (1 litre) that will result in the ultimate emission of
1.3x105 MeV of alpha energy (or 2.08x10-5 J/m3). An atmosphere containing 3.6 Bq/dm3 (~100
pCi/L) of radon in equilibrium with its daughters is 1 WL. Note that the short-lived daughters
are not necessarily in equilibrium with the radon parent, and for this reason, conversion of
radon concentration to daughter concentrations uses a factor of about 0.7 for uranium mines
(0.4 for homes). For example, finding 3.6 Bq/dm3 in a mine is approximately equivalent to 0.7
WL. It is common practice to use an equilibrium factor of 0.5, and therefore a WL would be
equal to 100 pCi/L (3.7 Bq/L).
A working-level month (WLM) is defined as the exposure to the equivalent radon daughter
concentration of 1 WL for a time period of 170 hours. For regulatory purposes, an abundance of
data was obtained from uranium miners, which showed that miners exposed to 100 or more
WLM had an increased risk of developing lung cancer. A regulatory limit of 4 WLM per year was
therefore set for uranium workers. This was set to correspond to the fact that a 25-year work
duration at a level of 4 WLM would lead to a cumulative exposure of 100 WLM. Dose-
conversion factors for radon are presented in ICRP50 [ICRP1987]. A worker dose-conversion
factor of 5 mSv/WLM is often used for radiation-protection purposes.
3.2 Background Radiation Exposure

The Earth is radioactive and has been since its formation. The Earth contains radioisotopes that
originated from planetary formation (primordial radioisotopes) and also from extra-terrestrial
origins (cosmogenic radioisotopes produced continuously in the upper atmosphere through
cosmic-ray interactions which deposit on Earth through atmospheric dispersion, deposition,
and precipitation). Some important primordial and cosmogenic radioisotopes are listed in Table
11.
Table 11 Primordial and cosmogenic radioisotopes
Primordial Cosmogenic
40 142 174 10 22
K Ce Hf Be Na
50 144 180 26 35
V Nd Ta Al S
87 147 187 36 7
Rb Sm Re Cl Be
113 148 190 14 33
Cd Sm Pt C P
115 152 204 32 32
In Gd Pb Si P
123 156 238, 235 39 28
Te Dy U Ar Mg
138 176 232 3 24
La Lu Th H Na
In addition to the two natural source categories, radioisotopes are introduced into the envi-
ronment from anthropogenic sources such as nuclear medicine, certain industrial processes,
routine nuclear power operations, historical nuclear weapons testing, and nuclear reactor
accidents. Theoretically, all radioisotopes on the nuclide chart can be introduced into the
environment through anthropogenic processes. In practice, only a few anthropogenic radioiso-
topes are persistent in the environment, for example, 137Cs. All the above-listed sources contri-
bute to the ubiquitous background that constantly surrounds us [Waller2013].
The most recent analysis of U.S. population dose due to natural and anthropogenic radionuc-
lides was conducted by the NCRP and published in Report No. 160 [NCRP2009]. The report
indicated that roughly half the yearly exposure to a member of the U.S. population is from
natural background radiation, primarily radon (originating from uranium in the ground) and
thoron (originating from thorium in the ground) gases, whereas the other half is from anthro-
pogenic (human-made) sources, of which the majority is due to medical procedures (see Figure
18). In Canada, it can be expected that the ratios will be similar to the U.S. values.
Note that in the U.S. report [NCRP2009], the average estimated yearly dose corresponding to
Figure 18 is approximately 6.2 mSv. The data indicated a large increase in average yearly dose,
which was exclusively due to medical procedures. Before the release of NCRP Report No. 160,
the average yearly dose was estimated to be approximately 3.6 mSv from all sources
[NCRP1987]. Therefore, the dose increase of 3.6 mSv to 6.2 mSv (almost 75% greater) due to
medical procedures is significant and worthy of increased surveillance.
Because Canada is a country in which advanced medical procedures such as computer tomo-
graphy (CT) are widely used, it is likely that our yearly average dose is proportionally similar to
that shown in Figure 18.
Terrestrial
3%
Internal
5%
Space
Computed 5%
Tomography
24%
Nuclear Medicine
12% Radon & Thoron
37%
Interventional
Fluoroscopy
7% Conventional
Radiography Consumer Industrial
Occupational
5% 2% < 0.1%
<0.1%
Figure 18 Origin of average yearly radiation exposure (adapted from [NCRP2009])

However, since the use of radiation for medical procedures is more strictly controlled in Cana-
da, it is also likely that the average yearly dose is proportionally smaller that U.S. values, with a
correspondingly higher percentage assigned to radon and thoron compared to, for example, CT.
A study done by [Grasty2004] for natural background radiation in Canada indicated that the
average annual radon and thoron component of the dose is 926 μSv, compared to the
[NCRP2009] U.S. estimate of 2294 μSv. In fact, taking into account all sources of natural (not
anthropogenic) radiation, it was found [Grasty2004] that the average yearly effective dose from
natural sources was 1769 μSv, which is significantly smaller than the 2422 μSv worldwide
average reported by [UNSCEAR2000]. Note, however, that the natural background can vary
greatly from location to location in Canada (and, of course, world-wide).
3.3 International Guidance

The history of radiation protection effectively begins in 1895 with the discovery of X-rays by
Wilhelm Roentgen (for which he received the Nobel Prize in physics in 1901). Radiation-induced
dermatitis was first observed by Emil Grubbé in the U.S. and Henry Drury in the U.K. in 1896. By
the end of 1896, within a year of the discovery of X-rays, the first radiation-protection advice
was proposed by Wolfram Fuchs in the United States and can be paraphrased as:
 Coat skin with VaselineTM and leave extra on the most exposed area
 Do not stand within 12 inches of the X-ray tube
 Make exposure as short as possible.
As crude as the above recommendations are, they embody the underlying protective mantra of
radiation shielding that we still use today: shielding, distance, and time. As a result, in 1902, a
first crude limit corresponding to approximately 100 mGy per day (or 30000 mGy per year) was
proposed, corresponding to the lowest radiation exposure that could be detected by fogging on
a photographic plate [Inkret1995]. However, despite an acknowledgment of potential detri-
mental effects, between the time of discovery of X-rays and the mid-1920s, numerous injuries
due to radiation exposure were reported (especially among early radiologists). Following
increasing radiation-safety concerns by radiologists and the public, the first International
Congress of Radiology was held in 1925 in London, England. Until that point, quantification of
dose was not consistent and therefore was considered a priority. As a result, the International
Commission on Radiation Units and Measurements (ICRU) was formed. In 1928, the second
Congress was held in Stockholm, Sweden, which led to the formation of the International X-Ray
and Radium Protection Committee, which was the predecessor of the International Commission
on Radiological Protection (ICRP). Throughout the years, the trend in radiation protection has
involved a lowering of the acceptable radiation dose for workers, and by the 1960s, of limita-
tions on acceptable dose for members of the public. The trend in dose standards is presented in
Figure 19.
100,000.0
30 Sv
Occupational
10,000.0 Public
Dose equivalent (mSv)
1,000.0 700 mSv

300 mSv
150 mSv
100.0
50 mSv
20 mSv
10.0 5 mSv
1 mSv
1.0
0.1
1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020
Figure 19 Chronology of radiation-protection guidance (adapted from [Inkret1995])

The change in philosophy for radiation-protection standards has shifted since 1896 as more
knowledge about the detrimental effects became known and as more opportunities for popula-
tion studies of effects (epidemiological studies) became available. A prime example is the
paradigm shift from protecting only for deterministic effects (Figure 20, left) to acknowledge-
ment, quantification, and protection standards to minimize risk from stochastic effects (Figure
20, right).
Deterministic Stochastic
Increasing Probability 
Increasing Severity 
Linear no-threshold
assumption
Threshold
Increasing Dose  Increasing Dose 
Figure 20 Paradigm shift from deterministic to stochastic effect protection

The paradigm shift in basic radiation-protection philosophy lead to numerous fundamental
changes in the way radiation protection was being administered on an international scale.
Although the original uses of ionizing radiation were predominantly in medicine, increased non-
medical uses became more common throughout the years and were reflected in adaptations of
international guidance. Practical guidance shifted to quantified dose limitation and optimiza-
tion. Although it had always been assumed that if humans were adequately protected from the
deleterious effects of ionizing radiation exposure, then the environment was also protected, a
change in philosophy moved towards the need for demonstration of adequate environmental
protection. Although the concept of the greatest good for the greatest number dominated
radiation-protection practice for many years, current philosophy also considers adequate
individual protection as well as that of the group. The historical progression of radiation-
protection principles is depicted in Figure 21.
Occupational All occupational All exposure,

exposure in medicine exposure occupational and public
Assumed protection Protection of man

Protection of man
of environment and environment
Deterministic effects Recognition of Deterministic and

only stochastic effects stochastic effects
Optimization and
Practical guidance Dose limitation
dose limitation
Focus on protection Protection of group

Avoid all harm
of group and individual
Figure 21 Progression of radiation protection principles (adapted from [Clement2009])
A number of international organizations are pivotal in setting radiological protection standards
and providing guidance. The essential organizations that provide input to national regulatory
standards are depicted in Figure 22. These organizations are discussed in the following sections.
Regulator
IAEA
Practical Guidance
ICRP
Philosophy & Policy
UNSCEAR
Basic Radiation Effects Science
Figure 22 Relationship of international bodies for radiation protection regulations
3.3.1 UNSCEAR
UNSCEAR is the United Nations Scientific Committee on the Effects of Atomic Radiation, which
was established by the United Nations in 1955 to assess and report levels and effects of expo-
sure to ionizing radiation and provides the scientific basis for evaluating radiation risk and for
establishing protective measures. UNSCEAR provides the scientific guidance that is consulted
when developing ICRP recommendations. UNSCEAR publications can be found online at:
http://www.unscear.org/unscear/en/publications.html .
3.3.2 ICRP
ICRP is the International Commission on Radiological Protection, established in 1928 to advance
for the public benefit the science of radiological protection, in particular by providing recom-
mendations and guidance on all aspects of protection against ionizing radiation. The main
objective of the Commission’s recommendations is to provide an appropriate standard of
protection for humans without unduly limiting the beneficial practices giving rise to radiation
exposure. The ICRP produces recommendations on radiological protection that are adopted
worldwide based on science and value judgements. The ICRP guidance documents form the
basis for IAEA regulatory recommendations. ICRP publications can be found online at:
http://www.icrp.org/publications.asp.
3.3.3 ICRP 103

The current recommendations of the ICRP [ICRP2007] form the basis for the radiation-
protection system. The primary aim of the recommendations is to provide an appropriate level
of protection for people and the environment without unduly limiting the desirable human
activities that may be associated with radiation exposure. The three primary goals of radiation
protection are:
 Justification – more benefit than detriment from exposure
 Optimization – exposure has been optimized to be as low as reasonably achievable
 Dose limitation – limit the dose (except for medical exposures).
The three goals form the core of the radiation-protection system and apply to three different
situations:
 Existing situations – natural events, past practices, past events; the situation already exists
when a decision on control is required, including natural background and residues from past
practices.
 Planned situations – practices, medical exposures; the situation involves planned operation
of sources, including decommissioning and disposal of radioactive waste.
 Emergency situations – preparedness, response; the situation is unexpected and requires
urgent action to mitigate consequences.
In ICRP103, three exposure types are also considered:
 Occupational exposure - all exposure incurred by workers in the course of their work, with
the exception of (i) excluded exposures and exposures from exempt activities involving rad-
iation or exempt sources, (ii) medical exposure, and (iii) normal local natural background
radiation.
 Public exposure - exposure incurred by members of the public from radiation sources,
excluding any occupational or medical exposure and the normal local natural background
radiation.
 Medical exposure - exposure incurred by patients as part of their own medical or dental
diagnosis or treatment; by persons, other than those occupationally exposed, knowingly,
while voluntarily helping in the support and comfort of patients; and by volunteers in a pro-
gramme of biomedical research involving their exposure.
Dose limits corresponding to planned public and occupational exposure are provided in Table
12. Note that the numerical values are essentially the same as reported in ICRP60 [ICRP1990],
from which the current CNSC regulations are derived. Note that the dose-limit guidance refers
to exposures above background radiation (as discussed in Section 3.2) and as such may be
considered as an additional dose over background. It should also be noted that, operationally,
action levels are often used that are a fraction of the dose limit. Exceeding an action level
would initiate an investigation into the reason for the dose and may initiate a review of proce-
dures to make doses as low as reasonably achievable, social and economic factors being taken
into account (ALARA). The ALARA principle is also found in CNSC document G-129, Rev1
[CNSC2004].
Table 12 ICRP103 dose limits for planned exposure scenarios [ICRP2007]
Occupational Public
20 mSv/a averaged
Effective Dose (Whole Body) over 5 years (50 mSv/a 1 mSv/a
max)
Equivalent Dose (Lens of the Eye) 150 mSv/a 15 mSv/a
Equivalent Dose (Skin) 500 mSv/a 50 mSv/a
Equivalent Dose (Hands and Feet) 500 mSv/a n/a
ICRP103 also considers the target for exposure in defining a “representative person”. A repre-
sentative person is an individual receiving a dose that is representative of the more highly
exposed individuals in the population. This term is the equivalent of, and replaces the notion of,
the “average member of the critical group” described in prior ICRP documents (for example,
[ICRP1990]).
Finally, ICRP103 considers explicitly the need for protection of the environment by delineating
the requirement for scientific evidence to demonstrate that protection is adequately afforded
and the need for improved protection as required. Further guidance is provided by ICRP108
[ICRP2008].
3.3.4 IAEA
The IAEA is the International Atomic Energy Agency, established within the United Nations
framework in 1956 as the world’s “Atoms for Peace” organization to promote safe, secure, and
peaceful nuclear technologies. The Agency works with its member states and multiple partners
world-wide to promote safe, secure, and peaceful nuclear technologies. Three main areas of work
underpin the IAEA’s mission: Safety and Security; Science and Technology; and Safeguards and
Verification. The IAEA has a wide range of programmes, including development of safety stan-
dards in regulatory language. IAEA Safety Series documents are often used as the principal
references for national regulatory policy. The most fundamental IAEA radiation protection
standard was the “International Basic Safety Standards for Protection against Ionizing Radiation
and for the Safety of Radiation Sources” [IAEA1996], which has been superseded by an interim
General Safety Requirements Part 3 report [IAEA2011]. IAEA publications can be found online
at: http://www-pub.iaea.org/books/.
3.4 Canadian Guidance

The regulatory body for nuclear power activities in Canada is the Canadian Nuclear Safety
Commission (CNSC). The regulatory system is designed to protect people and the environment
from licensed sources of anthropogenic radiation resulting from the use of nuclear energy and
materials. This is accomplished through a licensing process that requires the licensee to prove
that their operations are safe. The basis of the regulatory system is the principle that no tech-
nology is fail-proof and that therefore licensees must incorporate multiple layers of protection
(defence in depth) whenever radioactive materials are used. The CNSC also licenses the import,
export, and transportation of nuclear materials and other prescribed substances, equipment,
technology, and dual-use items. CNSC staff evaluates the performance of nuclear power plants
and radioactive material users, as well as participating in international activities for non-
proliferation of nuclear weapons.
The CNSC operates and enforces regulations under the Nuclear Safety and Control Act (NSCA,
1997). As the Canadian federal regulator, the CNSC executes licensing decisions made by the
Commission or its designates continually monitors licensees to ensure that they comply with
safety requirements that protect workers, the public, and the environment, and also upholds
Canada’s international commitments to the peaceful use of nuclear energy. Regulatory re-
quirements are codified in the NSCA, its associated regulations, licences, and directives pro-
vided by the CNSC. A number of regulations are referred to under the act and are listed below:
 Canadian Nuclear Safety Commission By-Laws (SOR/2000-212)
 Canadian Nuclear Safety Commission Cost Recovery Fees Regulations (SOR/2003-212)
 Canadian Nuclear Safety Commission Rules of Procedure (SOR/2000-211)
 Class I Nuclear Facilities Regulations (SOR/2000-204)
 Class II Nuclear Facilities and Prescribed Equipment Regulations (SOR/2000-205)
 Directive to the Canadian Nuclear Safety Commission Regarding the Health of Canadians
(SOR/2007-282)
 General Nuclear Safety and Control Regulations (SOR/2000-202)
 Nuclear Non-Proliferation Import and Export Control Regulations (SOR/2000-210)
 Nuclear Security Regulations (SOR/2000-209)
 Nuclear Substances and Radiation Devices Regulations (SOR/2000-207)
 Packaging and Transport of Nuclear Substances Regulations (SOR/2000-208)
 Radiation Protection Regulations (SOR/2000-203)
 Uranium Mines and Mills Regulations (SOR/2000-206)
Of significant interest are the Radiation Protection Regulations [NSCA2000], which set limits on
the amount of radiation that the public and nuclear energy workers (NEWs) may receive.
Canadian regulations are consistent with the most recent recommendations of the Interna-
tional Commission on Radiological Protection (currently adhering to guidance in ICRP60
[ICRP1990]). In Canada, standards and practices to protect people from radiation exposure are
also developed by the Federal-Provincial-Territorial Radiation Protection Committee (FPTRPC),
which provides a national forum on radiation-protection issues.
The CNSC offers instruction, assistance, and information on these requirements in the form of
regulatory documents, such as policies, standards, guides, and notices. Licensee compliance is
verified through inspections and reports.
In Canada, the current (2013) radiation-protection regulations are based on ICRP60 [ICRP1990].
However, the numerical values of the dose limits have not been changed in ICRP103
[ICRP2007], and therefore the basic values on which Canadian regulatory guidance is based will
continue for many years into the future. The operational values used by the CNSC are provided
in Table 12 with the inclusion of a balance-of-pregnancy dose limit for workers of 4 mSv. The
effective dose is calculated using Eq. (20) and compared to the dose limit for regulatory pur-
poses [NSCA2000]:
 mSv   I 
Etotal  E   5  RnP    20 mSv  , (20)
 WLM   ALI 
where
Etotal is the total effective dose from all sources (mSv)
E is the effective dose from external sources + committed effective dose measured di-
rectly or from excreta (mSv)
RnP is the exposure to radon progeny in working-level months (WLM)
I is the intake of radionuclides not already accounted for (Bq)
ALI is the annual limit on intake (Bq to give 20 mSv in 50 years).
3.5 Summary
Radiation-protection standards have evolved over the years to provide increased levels of
protection for workers and the public. Dose limits are established on the basis of scientific
knowledge which forms policy, best practices and guidance, and then national regulations.
The basic dose quantities used are absorbed dose, equivalent dose, and effective dose. Al-
though nuclear power plant operations often use obsolete units (such as Ci, Rad, and Rem) in
their plants, most of the nuclear industry uses SI units. Dose quantities for regulatory purposes,
as promulgated through the Nuclear Safety and Control Act, use SI units. Dose limits derived
from International Commission on Radiological Protection guidance are established in Canada,
and derived quantities are used in operations.
Background radiation is made up of natural and anthropogenic components. Dose limits are
established to provide protection above the ubiquitous background level, which has a wide
annual dose range world-wide. In Canada, the background dose is estimated to be approxi-
mately 3.6 mSv per year. Evidence from U.S. estimates indicates that increased use of advanced
medical imaging procedures such as computed tomography is resulting in a larger estimated
background dose; however, no such study has indicated that this is the case in Canada.
As research continues into large epidemiological studies and low-dose rate radiation effects,
dose-limit guidance may change in the future.
4 Radiation Instrumentation
Detection of ionizing radiation is vitally important to all aspects of the nuclear energy industry.
The ability to identify sources of radiation, quantities of radiation, and specific radioisotopes
enables the administration of comprehensive radiation protection, environmental monitoring,
and security programs.
Before discussing the mechanisms of how radiation interacts with detection materials, the
basics of radiation interactions with matter must be understood. These concepts were dis-
cussed in Chapter 3 (Nuclear Processes and Neutron Physics). This section considers radiation
instrumentation that may be pertinent to CANDU operations. The basics of detection, as well as
gas-filled, scintillation, and semiconductor detectors, are discussed. Dosimetry detectors,
especially as they relate to CANDU plant operations, are explored.
4.1 Basics of Detection

Ionizing radiation is tasteless, odourless, and colourless and cannot be detected by the human
senses. Therefore, devices to aid in detection and identification of ionizing radiation are re-
quired for safety purposes. Anything that responds to ionizing-radiation interactions can be
considered a radiation detector. In fact, the mechanisms of radiation interactions with tissue
and the human body, as discussed in Section 2, are fairly closely related to radiation detectors.
The primary differences are in the observable endpoints, where radiation detectors are de-
signed to provide an indication of radiation level through chemical/electric/electronic indica-
tions, whereas the endpoints in radiation biology tend to be deterministic or stochastic health
effects. That being said, the human body itself is a radiation detector of sorts. The first “radia-
tion dosimeters” used were based on skin reddening (or erythema). The variability in individual
human response to radiation (and of course the fact that skin reddening occurs only at unsafe
moderate to high radiation doses) makes this type of radiation detector impractical.
Radiation detectors work on one of two basic principles: (i) ionization and (ii) excitation.
Ionization
In an ionization-based detector, electrons bound to the atoms or molecules of the material
through which charged particles such as alpha and beta particles pass are released. The sepa-
rated electrons and ions (Figure 23) can be collected at two electrodes by imposing a potential
difference across the detector space; their presence is then measured as pulses or a current. An
example of generation of electron-ion pairs in a gas-filled counter (for example, a Geiger-Müller
tube) is depicted in Figure 24.
ion (atom with 1 e- less)

ionizing radiation
electron
+
Figure 23 Ionization process

radiation
anode (+v)
cathode (-v)
Figure 24 Ionization in a gas-filled tube
Excitation
In an excitation-based detector, part of the radiation energy is transferred to bound electrons
and raises them to an excited state in the atom or molecule. When the excited species returns
to its ground-state energy level, the excited atom or molecule may emit electromagnetic
energy in the ultraviolet to visible region (Figure 25). This light can be detected by a photomul-
tiplier tube (PMT), which generates secondary electrons using a photocathode and multiplies
them to generate a detectable signal. This process is depicted in the scintillation-type counter
system shown in Figure 26.
UV light
ionizing radiation
Figure 25 Excitation process
photocathode
UV light photomultiplier
signal out
e-
scintillator optical window
Figure 26 Excitation in a scintillator-type detector

Advances in radiation-detection instrumentation took place in the 1920s and have continued
ever since. A concise discussion of the history of radiation-detection instrumentation is pro-
vided by [Frame2005]. Important operational characteristics for all detector types include
resolving time and efficiency and are discussed below. An excellent reference on particle
counting is ICRU Report 52 [ICRU1994].
4.1.1 Resolving time

When ionizing radiation interacts with a detector, a signal in the detector material (whether it
be a gas, scintillator, or solid-state device) is produced that is related to the radiation event that
generated it. During the time it takes the system to process a pulse that is generated in the
detector, the detector may be unavailable to process additional events. For example, during the
time it takes for positive ions in a GM tube to reach the cathode, the tube is insensitive to any
radiation. During this time, if a second ionizing ray interacts with the detector, it will not be
observed because the tube cannot distinguish that there is another electron avalanche present;
the system observes only one large electron avalanche until it has been reset after detection. In
essence, the counter cannot produce pulses for more than one event because the counter is
“occupied” with the first event. This phenomenon is sometimes called coincidence, and as a
result, the observed counts are always lower than the true counts. The “resolving time” is often
referred to as “dead time”, because in essence the detector is “dead” and cannot detect any
other radiation in this time window. Note, however, that the resolving time is actually the sum
of the dead time and recovery time. In terms of definitions, dead time is the time required in a
detector before another unique pulse can form; recovery time is the time required after the
dead time until the pulse size is large enough to pass a discriminator; and resolving time is the
minimum time from the detection of one interaction until the next detection can occur. True
resolving times span a range from a few microseconds to 1000 microseconds, depending on the
detector. The loss of particles is important, especially when high count rates are involved and
the losses accumulate into large numbers.
For a counting rate, R, from a radioactive source, the presence of coincidence will mean that
the rate actually measured, r, will be less than the expected value (r < R). If the detector has a
dead time of T, then the true count rate is given by Eq. (21):
R  r  rRT . (21)
The counting rate can therefore be corrected for dead time using Eq. (22):
r
R . (22)
1  rT
The value of dead time, T, can be determined using a two-source method by measuring the
activity from two known sources r1 and r2. In theory, the measurement, r3, should be the simple
sum of the two sources as given by Eq. (23):
r1  r2  r3  b , (23)
where b is the background counting rate. If each of these counting rates is corrected for dead
time, then Eq. (23) becomes Eq. (24):
r1 r r3
 2  b. (24)
1  rT1 1  r2T 1  r 3T
Because the background count rate may be considered negligible for this measurement, Eq.
(24) can be expressed in the form of a quadratic equation:
r1r2 r3T 2  2 r1r2T  r1  r2  r3  0 . (25)

T should be on the order of microseconds, thereby making T2 negligible, enabling the simple
solution for the dead time, T, given by Eq. (26):
r1  r2  r3
T . (26)
2r1r2
The above expressions are useful in the limit of low interaction rates and are considered part of
what is known as the non-paralyzable model. If the dead time approaches 30% or more, the
above model fails, and the paralyzable model must be used (which is not developed here; the
reader is referred to [Knoll2010]. The significance of non-paralyzable versus paralyzable, from a
practical perspective, relates to ability to count radiation in extreme fields. An ideal detector
will have a linear response of count rate with interaction rate. A non-paralyzable detector will
have an increasing count rate with interaction rate (but not necessarily linear at high interac-
tion rate), but will tend to flatten out at high interaction rate. A paralyzable detector is one
which has an increasing response at lower interaction rate, a peak response, and then a de-
creasing response with higher interaction rate. Essentially, in a paralyzable system, an interac-
tion that occurs in the detector during the dead time will extend the dead time. The danger of a
paralyzable detector is that if the count rate goes down as the interaction rate goes up, a
person entering a progressively higher radiation field will think that he is moving into an area of
lower radiation. For example, a GM counter in a high radiation field will tend to become para-
lyzed and may severely under-predict the radiation hazard.
4.1.2 Energy response and efficiency

All radiation detectors respond to energy deposition differently, based upon the type of radia-
tion interacting with the detector and the energy of the radiation. For example, in a thin-
window GM tube, beta particles interact best with the filling gas, followed by alpha particles
and then gamma rays. This is due to the range of the respective particles in the gas in relation
to the anode and cathode of the GM detector.
No radiation detector counts all the particles which are emitted from a source for various
reasons. For example, resolving time affects the registered counts. In addition, many of the
particles do not strike the tube at all because they are emitted uniformly in all directions from
the source. The combination of effects that prevent the detector from counting all particles is
known as the efficiency. Efficiency of a radiation counting system is achieved by comparing the
measured count rate in the system to the disintegration rate (activity) of the source when the
activity is given in units of disintegrations per second (dps; 1 dps = 1 Bq), where the conversion
factor is 1 Ci = 3.7x1010 Bq.
The formula for determining the absolute efficiency is straightforward and is given as Eq. (27):
Cmeasured
% Efficiency    100% , (27)
A
where Cmeasured is the measured count rate and A is the calculated source activity. Note that the
time units of the numerator and denominator must be the same.
In fact, the absolute efficiency is a product of two separate phenomena, known as the geome-
tric efficiency and the intrinsic (or quantum) efficiency. The relationship for absolute efficiency,
ε, is given by Eq. (28):
Number detected
 100%
Number emitted
Number reaching detector Number detected
= × 100% . (28)
Number emitted Number reaching detector
 
Geometric efficiency Intrinsic efficiency
The intrinsic efficiency is determined by the energy of the particles, atomic number of the
detector, density of the detector, thickness of the detector, and other factors. A variety of
geometrical efficiency examples are depicted in Figure 27. In the left figure, the source is a
certain distance from the detector, and the geometrical efficiency will be low and based upon
the angle subtended by the source to the detector edge. In the centre figure, the geometrical
efficiency approaches optimal for a planar detector (often called a “2π geometry” because the
solid angle of a cone intersecting with the detector is approximately Ω=2π(1-cos(π/2))=2π). In
the right figure, the geometrical efficiency approaches unity because the source is surrounded
by detector material. This configuration is called a well detector (often called a “4π geometry”
because the solid angle of a cone intersecting with the detector is approximately Ω=2π(1-
cos(π))=4π). In practice, detector efficiencies are specified as a function of radioisotope (or
energy) for a given geometry.
FAR NEAR WELL

detector
detector
detector
source
source
source
Figure 27 Geometric efficiency examples (far planar, near planar, and well detectors)
For detectors that perform spectroscopy (measurement of incident radiation spectra), both
energy and efficiency calibration are generally required. As an example of how this calibration is
performed, consider a simple scintillation (sodium iodide) gamma spectroscopy system. When a
gamma ray (emitted during a change in an atom’s nucleus) interacts with a sodium iodide
crystal, NaI(Tl), the gamma ray will frequently give all its energy to an atomic electron through
the photoelectric effect (PE). This electron travels a short, erratic path in the crystal, converting
its energy into photons of light by colliding with many atoms in the crystal. The more energy
the gamma ray has, the more photons of light will be created. A photomultiplier tube (PMT)
converts each photon interaction into a small electrical current, and because the photons arrive
at the PMT at about the same time, the individual currents combine to produce a larger current
pulse. This pulse is converted into a voltage pulse with size proportional to the gamma-ray
energy. The voltage pulse is amplified and measured by an analog-to-digital conversion (ADC)
process. The result of this measurement is an integer between 1 and 1024 for a 10-bit ADC. One
is the measured value for a voltage pulse less than a hundredth of a volt, and 1024 is the
measured value for a pulse larger than approximately 8 volts (or the largest voltage pulse in the
ADC). Pulses between 0V and 8V are proportionately assigned an integer measured value
between 1 and 1024. This measure is called the channel number. The analog-to-digital conver-
sion process is performed, and the computer records the measurements as the number of
gamma rays observed for each integer measurement or channel number.
The spectrum is a visual display of the number of gamma rays as a function of the channel
number. In the case of known sources (location of the energy peak(s) is well known), then a
correlation of channel number with gamma-ray energy can be generated. The result is a graph
of the channel number (memory location) as a function of gamma-ray energy (photopeak). The
slope of this graph (energy/channel) is the energy calibration (a fitted equation may also be
used if the relationship is not linear over all energy values). An example of energy calibration is
shown in Figure 28.
2000
1500
Energy (keV)
1000
500
0 1000 2000 3000 4000

Channel Number
Figure 28 Example energy calibration graph
If the geometry is fixed and the activity of the calibration source is known, the peak area
(integral counts) can be correlated with activity to determine the efficiency for each photopeak,
as determined by Eq. (27). The efficiency can be plotted as a function of photopeak energy to
obtain an efficiency curve, as depicted in Figure 29. This curve enables determination of the
activity of an unknown source in a spectroscopic system.
0.040
0.035
0.030
0.025
Efficiency
0.020
0.015
0.010
0.005
0.000
0 500 1000 1500 2000
Energy (keV)
Figure 29 Example efficiency calibration graph
Energy calibration is essential for isotope identification, and efficiency is essential for activity
determination.
4.1.3 Counting statistics and decision levels

Statistics play a very important role in particle counting because atomic concentrations are on
the order of Avogadro’s number. Therefore, it is impossible to deal with atoms individually, and
statistics are used to assist in prediction of behaviour. Statistics uses two main quantities in
predicting behaviour, the mean and the standard deviation.
The mean ( x ) is computed as in Eq. (29):
N
xi
x  , (29)
i 1 N
where the number of measurements (xi) ranges from 1 to N.
The standard deviation of the data is computed as in Eq. (30):
N
x  x 
2
i
 i 1
. (30)
N 1
The standard deviation is a measure of the variation of the data without regard to position
above or below the mean (that is, an absolute value using the square). Because most meas-
urements tend to follow a Gaussian distribution about the mean (i.e., they are symmetrical),
the mean and standard deviation are widely used measures to predict behaviour. However,
particle counting is not symmetrical about a mean. Radioactive decay is a random process that
occurs very infrequently for low-activity samples and very frequently for high-activity samples.
Three statistical distributions useful in particle counting will be discussed below.
4.1.3.1 Binomial distribution

The binomial distribution is used when there are two possible outcomes of an event. The
probability of an outcome is a constant independently of the number of trials, and the selection
of either outcome does not affect the outcome of subsequent trials. Assume an experiment
where there are only two outcomes: (A) and (B). In a binomial distribution, p is defined as the
probability of obtaining one of the outcomes (A), and q is the probability that the other (B)
outcome occurs. Occasionally, statistical references use 1-p in place of q (which is how q is
calculated). The binomial distribution gives the probability, P(n), that n out of N objects are A, is
given by Eq. (31):
N! N!
P n  p n 1  p 
N n
p n q N n  . (31)
 N  n !n !  N  n !n !
For this distribution, the mean, m, is given by Eq. (32):
m  pN , (32)
and the standard deviation is calculated using Eq. (33):
  mq  Npq  Np 1  p  . (33)
To study radioactive decay, a distribution is required that represents a large number of counts
and a small probability of success (large N and small p). The Poisson distribution is appropriate
for this purpose.
4.1.3.2 Poisson distribution

The Poisson distribution is a special case of the binomial distribution given:
1. large number of counts, and
2. small probability of success.
The Poisson distribution, described by Eq. (34), gives the probability, P(n), that n out of N
objects are outcome (A):
mn  m
P  n  e , (34)
n!
where m is the mean of the distribution (m = Np). A primary feature of the Poisson distribution
is the ease of calculating the standard deviation using Eq. (35):
  m. (35)
For individual measurements, the standard deviation is the square root of the number of
counts, (N), as given by Eq. (36):
 N . (36)
For large values of m (m > 20), a more appropriate distribution, called the Gaussian distribution,
is used.
4.1.3.3 Gaussian distribution

The Gaussian (normal, or bell) distribution is used when the sample size is large. An important
feature of this distribution is that it uses continuous variables, unlike the Poisson and binomial
distributions which use discrete variables. The Gaussian distribution is described by Eq. (37):

 x  m 2
1
P n  e 2 2
. (37)
 2
For a large number of measurements, the data should follow a Gaussian distribution and yield
the mean and standard deviation presented by Eqs. (29) and (30).
4.1.3.4 Decision level

An important circumstance involving counting statistics is when low levels of radiation are
present (which is often the case, for example, in environmental measurements). When a
detector measurement is made at low counts, it must be determined whether the measure-
ment is indicative of anthropogenic radiation or due to fluctuations in the natural background
which is also being measured by the radiation-detection system.
The background can be estimated using two approaches. The first is a Poisson approach using
Eqs. (29) and (30) for the mean and standard deviation of the measurements and making a
judgement as to what multiple of the background measurement indicates a reading that is not
due to background radiation. For example, it may be decided that the background mean + 3
standard deviations of the mean is appropriate.
The second, more rigorous approach is to use detection limit statistics. A full discussion is
beyond the scope of this chapter, and the user is referred to [Currie1968] and [Chambless1992]
for broad discussions of detection limit statistics.
The decision criterion (sometimes called the detection limit, LD) is a value above some critical
level (LC) in which there is a quantifiable confidence that, if exceeded, you have actually meas-
ured radioactivity. The primary indicator is the critical level (LC), which is defined as the go/no-
go value above which a measurement is considered to represent radiation above a blank (or
background) sample. The critical level (LC) is defined by Eq. (38):
LC  2k b , (38)
where k is the Gaussian parameter related to area under the curve (sometimes termed “z”) and
σb is the standard deviation of the background radiation. Consider a normal distribution of
background counts, a 95% confidence level, and the fact that for a normal distribution, the 95%
confidence level occurs at k = 1.645; in this case, the critical level is defined as in Eq. (39):
LC  2 1.645   b  2.33 b . (39)
Therefore, the procedure to determine the critical level (the decision criterion) is:
1. Obtain numerous background measurements (for example, 20), spatially and temporally
separated, but NOT near contamination or radioactive sources.
2. Determine the mean and standard deviation of the measurements
3. Multiply the standard deviation by 2.33 and add it to the mean to obtain the decision
criterion.
For example, if the background-count rate mean is 30 cpm and the standard deviation is 5 cpm,
the decision criterion is 30 cpm+12 cpm = 42 cpm. If a measurement above 42 cpm is obtained,
it can be attributed to non-background radiation.
4.1.4 Particle counting versus spectroscopy

Generally speaking, gas-filled detectors are used for counting applications. In these, ionizing
radiation interacts within the detection material to create a pulse that is counted by an elec-
tronic counting system. Scintillation and semiconductor detectors can also be used strictly as
particle counters as well. In pulse mode, the signal from each interaction is processed indivi-
dually. In current mode, the electrical signals from individual interactions are averaged togeth-
er, forming a net current signal. In current mode, all information regarding individual interac-
tions is lost. If the amount of electrical charge collected from each interaction is proportional to
the energy deposited by that interaction, then the net current is proportional to the dose rate
in the detector material. This mode is used for detectors subjected to very high interaction
rates. The output from a particle-counting system will typically be counts per second (cps) or
counts per minute (cpm). It is also possible through calibration to convert the pulses counted to
dose rate or to integrated dose (Sv/h; Sv). A simple depiction of a particle-counting system is
shown in Figure 30. Radiation interacts in the detector and creates pulses. If the pulses are
above a discrimination threshold, they are passed to an integrator or counter circuit, and then if
being used with a digital display, through an analog-to-digital converter and digital display. In
addition, the count rate may be converted to dose rate using a calibrated radiation source and
mapping count rate against known dose rate. The important aspect of a particle-counting
system is that single values of count (rate) and/or dose (rate) are presented to the user as a
function of time. The user does not generally have any direct information about the radioiso-
tope or the activity being measured. To perform radioisotope identification and quantification,
spectroscopy is normally used.
Spectroscopy is strictly defined as the identification of a radioisotope using the energy of
emission of the decay particles. Spectrometry, on the other hand, is the quantification (i.e.,
activity determination) of a radioisotope using the intensity of the decay particle signal. For all
practical purposes, spectroscopy is used to denote identification and quantification.
analog
4 5 6 7
3
2 8
1 9
Integrator
threshold
or ADC
pulse
detector
123456
digital
Figure 30 Simple particle-counting system

A simple depiction of a spectroscopy system is provided in Figure 31. In the system, radiation is
emitted from the source and interacts with a material that responds to radiation with pulse
heights that vary with, and are proportional to, the incident radiation energy. The various pulse
heights, which are represented by a range of voltages, are processed in a multi-channel analyz-
er (MCA). The purpose of the multi-channel analyzer is to process the pulse heights into bins
that are determined by memory allocation units (typically called channels). For example, if the
pulses are amplified to have a maximum voltage of 10 volts and there are 1024 memory loca-
tions, then each memory location is 9.8 mV wide. The first channel is 0–9.8 mV, the second is
9.8–19.6 mV, and so on. It is normal to use a threshold discriminator to eliminate noise that
would otherwise be counted as real events, and event logging can be controlled by a lower-
level discriminator (LLD) and an upper-level discriminator (ULD). Events that fall within the
window will be processed by the multi-channel analyzer. Each pulse is counted within its
respective bin (analog-to-digital conversion) and as more counts accumulate, a spectrum is
produced which is displayed with counts on the y-axis and channel number on the x-axis. The
spectrum produced is characteristic of the radioisotope, and therefore identification is possible.
It can be readily seen that spectroscopy systems can also be used for particle counting. It is not
unusual for portable spectroscopy systems to use a scintillation detector for both dose rate and
spectral ID functions.
1024
.
SOURCE .
5
4
DETECTOR 1
pulse train
ULD
window
LLD
Threshold
Figure 31 Simple spectroscopy process

Spectrometer systems can be summarized by the following steps:
 When a gamma ray interacts with a detector, a pulse is produced. For gamma spectroscopy,
the pulses are produced when gamma rays interacting in the detector create electrons
through the photoelectric effect, Compton scattering, or pair production.
 The size of the pulse reflects the energy deposited in the detector.
 An analog-to-digital converter (ADC) sorts the pulses into bins according to size.
 The results of the ADC analysis are displayed. The display, a plot of number of pulses
(counts) versus pulse size (channel number), is referred to as a spectrum.
 If a large number of pulses have a similar size, they are sorted into adjacent channels and
appear on the spectrum as a peak. A peak represents a number of pulses of similar size.
In an ideal spectrometer system, a peak would be produced that corresponds to the photopeak
energy(ies) of the radioisotopes being measured, and the peak would be very narrow (ap-
proaching a delta function). In reality, a number of processes can occur within (a) the detector
and (b) the shielding that is often present around the system. These processes generally pro-
duce undesirable effects in the spectrum, and therefore attempts are made to minimize the
impact of these effects. The important aspect of that energy deposited in the detector (which is
what the spectrometer measures and displays) is simply the energy in minus the energy out
(Edep = Ein - Eout).
4.1.4.1 Detector Interactions

The primary interactions of importance in the detector are as follows:
 Photopeak. A photopeak is generated in the spectrum when gamma rays deposit all their
original energy in the detector through the photoelectric (PE) interaction. This is the prima-
ry signature in a gamma spectroscopy system and is used both to identify and to quantify
the radiation.
 X-Ray Escape Peak. An X-ray escape peak is generated in the spectrum when gamma rays
deposit all their original energy in the detector through the photoelectric effect (PE), except
for the energy of an iodine or germanium X-ray that leaves the detector.
 Compton Continuum. The Compton continuum is produced when gamma rays interact in
the detector through Compton scattering (CS) and then leave the detector.
 Double Escape Peak. The double escape peak is produced when gamma rays interact in the
detector through pair production (PP) and two 511-keV annihilation photons escape the de-
tector.
 Single Escape Peak. The single escape peak is produced when gamma rays interact in the
detector through pair production (PP) and one 511-keV annihilation photon escapes the de-
tector.
4.1.4.2 Shield interactions

If there is significant shielding around a spectroscopy system (for example, environmental
radioactivity gamma spectroscopy systems often use a scintillator or semiconductor detection
system in a lead shield, which is sometimes called a “castle”), then interaction occurring in the
shield can produced scattered or secondary radiation in the detector that will be logged in the
spectrum.
 Lead X-Rays. Lead X-rays are generated in the spectrum when gamma rays interact with a
lead shield through the photoelectric effect (PE) and the resulting X-rays produced in the
shield interact with the detector. The lead X-rays are around an energy level of 75 keV (73,
75, and 85 keV). If the shield were made of steel, iron X-rays might be seen, but these would
be of such low energy that they would probably go undetected.
 Backscatter Peak. A backscatter peak is generated in the spectrum when gamma rays
interact with the shield through Compton scattering (CS) and the 180o-scattered gamma
rays interact with the detector.
 Annihilation Peak. An annihilation peak is generated in the spectrum when gamma rays
interact with the shield through pair production (PP) and the resulting annihilation photons
at 511 keV interact with the detector.
 Summation Peak. A summation peak occurs when two gamma rays deposit all their energy
in the detector at the same time. The energy of this peak is the sum of the two photopeak
energies.
 Bremsstrahlung. Bremsstrahlung (“braking” radiation) is observed in the spectrum when
high-energy beta particles are produced and slow down in the sample.
4.2 Gas-Based Detectors

The various operating regions for gas-filled detectors are shown in Figure 32.
At low applied voltage, below the ion-saturation region, there is a region called the recombina-
tion region. Many electrons and ions produced in the gas recombine because the voltage
applied between cathode and anode is not large enough to collect all the electrons. This region
is not useful for counting radiation.
The next region is the ion-saturation region. The potential difference is sufficient to collect all
freed electrons. A detector working in the saturation region is called an ionization chamber, and
its output is proportional to the deposited radiation energy. Internal or thin-window ionization
chambers are used as alpha-particle and fission-fragment detectors.
The next important region is the proportional region. The applied voltage is large enough that
the electrons freed by the initial radiation are accelerated, so that they in turn ionize additional
atoms or molecules (secondary ionization) to free more electrons. This electron multiplication
generates an avalanche toward the anode for each primary electron that was freed. The
applied voltage domain is called the proportional region because each avalanche is character-
ized by the same electron multiplication at a given applied voltage. The output signal is directly
proportional to the deposited energy, although each pulse is many times larger than in the
ionization region. The limited proportional region is at slightly higher applied voltage, and the
proportionality of the output signal to the deposited energy at a given applied voltage no
longer applies. Amplification of the greater deposited energy reaches its limit while that for the
lesser deposited energy continues to increase. This region is usually avoided as a detection
region.
pulse
amplitude Geiger
Muller
region
2 MeV limited
prop.
1 MeV region
proportional
ion region
saturation
Applied voltage
Figure 32 Operating regions of gas-filled counters

The next usable region is the Geiger-Müller region, which is characterized by an applied voltage
high enough that any deposited energy produces sufficient secondary electrons to discharge
the entire counting gas. A linear amplifier is no longer needed because there is sufficient
electron production to generate a usable signal. In this region, it is not possible to distinguish
between small and large depositions of energy. At higher applied voltages, electrical discharge
occurs between the electrodes. This voltage region, called the continuous-discharge region, has
been used for some purposes, but generally is avoided because the discharge can disable the
detector.
Details about the two most common types of radiation detectors used in nuclear operations,
namely the proportional counter and the Geiger-Müller (GM) counter, are discussed in the
following sections.
4.2.1 Proportional counters

The proportional counter is a type of gas-filled detector introduced in the late 1940s. These
counters are similar to GM counters in that they rely on gas multiplication to generate a signal.
The primary difference between a GM and a proportional counter is that a GM counter pro-
duces a pulse when radiation interacts with it, irrespective of incident particle type or energy. A
proportional counter generates different-size pulses that are dependent on the incident parti-
cle energy. It is, therefore, possible to distinguish between alpha and beta particles using a
proportional counter, or between two different energies. Although proportional counters do
not have the resolution required to perform spectroscopy, they are extremely useful for dis-
criminating between alpha and beta particles.
The primary use of proportional counters in the nuclear industry is therefore in alpha-beta
counters (for example, swipe counters), where it is desirable to discriminate between alpha and
beta particles.
4.2.2 Geiger-Müller counters

Geiger-Müller (GM) counters were invented by H. Geiger and E.W. Müller in 1928 and are used
to detect radioactive particles (and β) and rays (γ and x). A GM tube usually consists of an
airtight metal cylinder closed at both ends and filled with a gas that is easily ionized (usually
neon, argon, or a halogen). One end consists of a “window” made of a thin material, mica,
which allows alpha particles to enter (these particles can be shielded easily). A wire, which runs
lengthwise down the centre of the tube, is positively charged with a relatively high voltage and
acts as an anode. The tube itself acts as the cathode. The anode and cathode are connected to
an electric circuit that maintains the high voltage between them.
When the radiation enters the GM tube, it will ionize some of the atoms in the gas. Because of
the large electric field created between the anode and the cathode, the resulting positive ions
and negative electrons accelerate toward the cathode and the anode. Electrons move or drift
through the gas at a speed of about 104 m/s, which is about 104 times faster than the positive
ions move. The electrons are collected a few microseconds after they are created, while the
positive ions take a few milliseconds to travel to the cathode. As the electrons travel toward the
anode, they ionize other atoms, producing a cascade of electrons called a gas multiplication or
avalanche. The multiplication factor is typically 106 to 108. The resulting discharge current
causes the voltage between the anode and cathode to drop. The counter (electric circuit)
detects this voltage drop and recognizes it as a signal of a particle’s presence. There are addi-
tional discharges triggered by UV photons liberated in the ionization process that start ava-
lanches away from the original ionization site. These short-lived discharges are called GM
discharges and do not affect performance. Once the avalanche of electrons is started, it must
be stopped (quenched) because the positive ions may still have enough energy to start a new
cascade. One early method was external quenching, which was done electronically by quickly
ramping down the voltage in the GM tube after a particle was detected. This meant that any
more electrons or positive ions created were not accelerated towards the anode or cathode.
The electrons and ions would recombine and no more signals would be produced. The modern
method is called internal quenching. A small concentration of a polyatomic gas (organic or

halogen) is added to the gas in the GM tube. The quenching gas is selected to have a lower
ionization potential (~10 eV) than the fill gas (26.4 eV). When the positive ions collide with the
molecules of the quenching gas gas, they are slowed or absorbed by giving their energy to the
quenching molecule. They break down the gas molecules in th the
e process (dissociation) instead of
ionizing the molecule. Any quenching molecule that may be accelerated to the cathode dissoci-
dissoc
ates upon impact, producing no signal. If organic molecules are used, GM tubes must be re- r
placed periodically because they perman
permanently
ently break down over time (about one billion counts).
GM tubes can also use a halogen molecule, which naturally recombines after breaking apart.
Different Geiger-Müller
Müller (GM) tubes have varying operating characteristics due to differences in
their fabrication. Consequently, each GM counter has a different high voltage that must be
applied to obtain optimal performance from the instrument. If a source of ionizing radiation is
positioned beneath a tube and the voltage of the GM tube is ramped up (slowly increased incr by
small intervals) from zero, the tube does not start counting right away. The tube must reach the
starting voltage where the electron “avalanche” can begin to produce a signal. As the voltage is
increased beyond that point, the counting rate increaincreases
ses quickly before it stabilizes. Where the
stabilization begins is a region commonly referred to as the knee, or threshold value. Past the
knee, further increases in voltage produce only small increases in the count rate. This region is
the plateau we are seeking. Determining the optimal operating voltage starts with identifying
the plateau. The end of the plateau is found when increasing the voltage produces a second
large rise in count rate. This last region is called the discharge region. To preserve theth life of the
tube, the operating voltage is selected near the middle middle, but towards the lower half of the
plateau (closer to the knee). If a GM tube operates near the discharge region and there is a
change in its performance, the tube will operate in a “con “continuous-discharge”
discharge” mode, which can
damage it. Figure 33 shows a typical plateau shape for a GM tube. Above approximately 1100
volts, the tube enters the continuous
continuous-discharge region.
Figure 33 GM counter sample plateau

For hand-held
held radiation detection instrumentation commonly used in nuclear power opera-
oper
tions, GM counters are the most common type of detectors
detectors; they can be found in radiation
survey instruments, contamination detectors (for example, pancake detectors),
detectors) and in personal
dosimeters.
4.3 Scintillation-Based Detectors

Scintillation detectors are the most commonly used detectors for field applications because
they do not require bulky cooling strategies. There are two stages in a scintillation detector
unit: the scintillator and the photomultiplier. The scintillation detection process is outlined as
follows:
1. Ionizing radiation deposits energy in the scintillator.
2. A portion of the absorbed energy causes electrons in the scintillator to move to a higher
energy level.
3. A portion of these electrons immediately fall back down to a lower energy level. As they
do, they emit a photon of light. Collectively, the photons of light form a flash, or scintil-
lation.
4. Scintillations are converted into an electronic pulse by a photomultiplier tube (PMT).
The brighter the scintillation, the larger is the pulse.
One particle of radiation (alpha, beta, gamma ray, X-ray, or neutron) interacting in the scintilla-
tor results in one scintillation (flash) of light and therefore one pulse. The greater the energy
transferred to the scintillator, the greater is the number of excited electrons and the greater is
the number of photons of light emitted in the scintillation. In other words, the greater the
deposition of energy, the brighter is the flash of light and the larger is the pulse.
4.3.1 Scintillators
Scintillators are categorized as fluors as opposed to phosphors. Fluorescence has a lifetime in
the excited state from 10-7 to 10-10 s, whereas phosphorescence has a lifetime in the excited
state of 10-3 s or longer. It is desirable to have a material with a short excited-state lifetime to
achieve a shorter resolving time. Scintillation materials can be classified as solid, liquid, or gas,
as detailed below.
 Gas scintillators: The radiation energy excites the electrons of the gas molecules, and light
photons are emitted when the electrons de-excite. Because these photons are typically in
the UV range, a PMT must be chosen that is UV-sensitive. Gas scintillators are extremely
fast, but their light output is poor (low efficiency). The most extensively investigated gas
scintillators use the noble gases xenon and helium.
 Liquid scintillators: Liquid scintillators consist of an organic scintillator dissolved in an
appropriate solvent. Normally, the sample to be counted is also dissolved in the scintillator.
The advantage of this arrangement is a high counting efficiency for beta particles, even for
betas of the lowest energies (e.g., 3H). In some cases, the liquid scintillator is used to meas-
ure external radiation sources (e.g., cosmic rays). In such cases, the liquid scintillator is
packaged and treated as if it were a solid crystal. One advantage of liquid scintillators is the
fact that extremely large detectors can be constructed. This is the basis of the commonly
found liquid scintillation counter (LSC).
 Solid scintillators: This is the most common type of material used for gamma spectroscopy.
Aside from their physical form, scintillators may also be classified chemically as organic or
inorganic:
 Inorganic: gas scintillators (e.g., He, Xe), inorganic crystals (e.g., NaI, ZnS, CsI, LiI, BGO), glass
scintillators.
 Organic: liquid scintillators (e.g., PPO, Bis-MSB, POPOP), organic crystals (e.g., anthracene,
stilbene), plastic scintillators.
Two extremely common solid scintillation materials used in the nuclear industry are ZnS and
NaI, which are discussed below.
Zinc Sulfide
Zinc sulfide is often doped (activated) with silver in the form of ZnS(Ag). It is typically used in
alpha detectors and for heavy-particle detection. In the past, ZnS was combined with 226Ra to
form radium paint, which was used to paint clock and aircraft dials so that they would “glow in
the dark”. ZnS is fairly opaque and has high scintillation efficiency (~40%) with a long decay
time.
Sodium Iodide
Sodium iodide is an alkali halide. Its use as a scintillator dates back to the late 1940s, and it is
often activated using thallium in the form NaI(Tl). Thallium is added in trace amounts, at a level
of approximately one atom in one thousand. NaI is used primarily for photon (X- or gamma)
radiation detection. NaI has a high light yield with a scintillation efficiency of approximately
12%. In other words, approximately 12% of the gamma-ray energy deposited in the crystal will
be emitted as light. The average energy of a photon emitted by NaI activated with thallium is 3
eV (which has a wavelength of approximately 415 nm, a wavelength readily detected by most
photomultiplier tubes). For a hypothetical 1-MeV gamma ray that deposits all its energy in the
NaI scintillator, one would expect 120 keV (12% of 1 MeV) to be emitted as light. With an
average photon energy of 3 eV, this would be equivalent to 40,000 photons of light. NaI crystals
can be grown to relatively large sizes for application in airborne/carborne spectroscopy systems
or portal monitors.
Scintillators produce photons of light in the UV to near-visible region and require photomultip-
lier tubes (PMT) as part of the detection process (Figure 26).
4.3.2 Photomultiplier tubes (PMT)

A photomultiplier tube is used to detect very weak light signals. It is a photoemissive device in
which the absorption of a photon results in the emission of an electron. It works by amplifying
electrons generated by a photocathode impinged upon by a UV or near-visible light source.
Photomultipliers acquire light through a glass or quartz window that covers the photosensitive
surface (photocathode), which then releases electrons that are multiplied by electrodes known
as metal channel dynodes. At the end of the dynode chain is an anode or collection electrode.
The current flowing from the anode to ground is directly proportional to the photoelectron flux
generated by the photocathode.
Electrons emitted by the photocathode are accelerated toward the dynode chain, which may
contain in excess of 14 elements. Focussing electrodes are usually present to ensure that
photoelectrons emitted near the edges of the photocathode will be likely to land on the first
dynode. Upon impacting the first dynode, a photoelectron will invoke the release of additional
electrons that are accelerated toward the next dynode. Note that photomultipliers produce a
signal even in the absence of light due to dark current arising from thermal emissions of elec-
trons from the photocathode, leakage current between dynodes, or background high-energy
radiation.
4.4 Semiconductor-Based Detectors

Electrons in a solid can occupy some energy levels, but not others. These levels are not single
discrete energies, but ranges of energies called bands. The highest energy band occupied by
electrons is referred to as the valence band. The conduction band is of even higher energy than
the valence band, but it is normally nearly void of electrons. The range of energies between the
valence and conduction bands (forbidden to electrons) is called the band gap (forbidden band).
When the valence band is full, the electrons in the band are essentially fixed in place, i.e., tied
to a particular site in the solid. Such a material will not conduct electricity unless electrons in
the valence band can be given the energy to reach the conduction band where they would be
free to move. If the valence band is not completely full, the electrons in it are free to move, and
the solid will conduct electricity.
Solids can be divided into the following three categories:
 Conductor. A material in which the valence band is partly, but not completely, full.
 Insulator. A material in which the valence is completely full and for which the band gap is
greater than 3–5 electron volts.
 Semiconductor. A material for which the valence band is completely full and for which the
band gap is less than 3–5 electron volts. Semiconductor materials will not normally conduct
electricity. However, the band gap is small enough that it is possible to lift enough electrons
up into the conduction band for electrical conduction to occur.
When alpha or beta particles enter a semiconductor detector, they create positive and negative
ion pairs (electron-hole pairs). Under the influence of an applied electric field, the positive
members of the ion pairs (called holes) and the negative members of the ion pairs (electrons)
move to the cathode and anode respectively. The result is an electronic pulse.
In the case of gamma detection, the gamma rays must first interact with the solid to produce
secondary electrons (through the photoelectric effect, pair production, or Compton scattering).
The secondary electrons then move through the detector and create electron-hole pairs which
are collected at the electrodes. A semiconductor detector is often considered to be the solid
equivalent of an ionization chamber.
Germanium is one of the most common semiconductor detector materials. A very common
semiconductor detector used for environmental investigation is the high-purity germanium
detector (HPGe). This detector needs to be cooled to near-liquid-nitrogen temperatures (~77°K)
to operate. This type of spectrometer is normally found in a health physics laboratory and is
used when high-resolution gamma spectroscopy is required. In general, for higher efficiency,
scintillation detectors are used because they tend to be much larger-volume detectors, whe-
reas semiconductor detectors are used when higher resolution is required. A typical commer-
cial HPGe spectroscopy system is shown in Figure 34. At the bottom of Figure 34, the details of
the opened lead castle can be seen, with the cylindrical HPGe detector element in the castle.
The castle is designed to hold cylindrical geometries. For environmental sample analysis, it is
very common to use a Marinelli beaker, which is a one-litre geometric design which covers the
HPGe detector element for maximum efficiency.
Figure 34 Commercial HPGe system (top) and detail of lead castle (bottom)
4.5 Portable Instrumentation

All nuclear facilities make use of various types of portable instrumentation. Portable or hand-
held instruments can be fabricated based on any detector technology, although some are more
suited than others to the mission. Broad categories of portable instrumentation include: (a)
general-purpose survey meter, (b) area contamination meter, (c) neutron meter, and (d)
portable spectrometer.
4.5.1 General-purpose survey meter

A general-purpose survey meter (GPSM) is almost always a gamma detection system using
scintillator, GM, or proportional detectors. This type of instrument is often called a gamma
survey meter and is used to detect and quantify ambient gamma-radiation fields. The instru-
ments have wide response to both incident photon energy and direction and provide readings
in counts per minute, counts per second, or both. If the instruments are calibrated, readings are
in dose rate (μ-, mSv, or Gy per hour) and if they have an integration function, in total dose.
4.5.2 Area contamination meter

An area contamination meter (ACM) is an instrument used to detect and quantify alpha or
beta/gamma surface contamination. It must be capable of detecting small quantities of radioac-
tive material, but is not capable of determining whether contamination is fixed to a surface or
loose. An ACM must be held very close to the surface to ensure accurate readings; however,
there are no suitable portable instruments for detecting small quantities of tritium (3H). A
common type of area contamination detector is the pancake probe (see Figure 35), which is a
relatively small-area thin-window GM tube that is very useful for frisking personnel for conta-
mination. ACMs are usually based on a GM or other gas-filled tube. A common alpha detector
uses a ZnS screen to convert alpha-particle interactions to UV photons, which are then detected
by a photomultiplier tube. In this case, no gas is required.
Figure 35 Portable contamination meter with pancake probe
4.5.3 Neutron meter

Portable neutron meters are commonly proportional counters which use neutron detection
reactions depending on the filling gas and generate charged particles. Two common reactions
(and hence, detection-tube types) are BF3- and 3He-based respectively according to the reac-
tions in Eqs. (40) and (41):
Li    2.79 MeV
7
10
B  n  7* (40)
Li    2.31 MeV  7 Li    0.478 MeV
3
He  n  3H  p  765 keV . (41)
These reactions are most likely to occur at low neutron energies, and therefore the counters
are wrapped in hydrogenous (neutron-moderating) material. Detector-response functions can
be altered using various techniques, such as perforated cadmium sleeves, so that a flat re-
sponse over a wide energy range of incident neutrons can be achieved. There are also a variety
of physical configurations, such as cylindrical, spherical, or even rectangular. Portable neutron
detectors have historically been called “neutron REM-meters”, “long counters”, “Bonner
spheres” (see Figure 36), “REM balls”, or by other names, depending on their origin and other
physical characteristics.
Figure 36 Neutron Bonner sphere
4.5.4 Portable gamma spectrometer

The possibility of portable gamma spectroscopy became a reality as electronics, processors,
microprocessors, and computers in general became smaller. Portable gamma spectrometers
typically use scintillation-type detectors with small photomultiplier tubes. Some “portable”
spectroscopy systems in the past have used semiconductor detectors; however, these are
actually quite bulky and heavy, and their true portability is doubtful. Although portable spec-
troscopy generally cannot give the same performance as a laboratory system, these instru-
ments can perform simplified radioisotope identification and typically offer count, count rate,
dose, and dose rate functions.
4.6 Specialized Detectors

A number of specialized detector systems are commonly used in nuclear facilities to process
personnel for potential internal or external contamination. These detection systems include
portal monitors, whole-body counters, hand/cuff/foot counters, and thyroid counters.
4.6.1 Portal monitors

Portal monitors are used to monitor people in a passive way as they move between areas. They
may be used as entrance and exit points from secure areas in nuclear plants or between clean
and dirty operational areas. A portal monitor typically is constructed as a tubular frame which
houses scintillation detectors (Figure 37). Portal monitors primarily respond to incident gamma
radiation that may arise from radioisotope contamination on personnel or equipment. If a
person is working in a potentially contaminated (dirty) area and some contamination is trans-
ferred to clothing, skin, or equipment, when that person passes through the portal monitor, the
gamma emissions from the contamination should be detected by the scintillators and trigger an
alarm on the electronic panel. This will elicit a response from health physics personnel to clean
the contamination and generate an incident report.
Scintillator
detectors
Alarm and
display
Figure 37 Portal monitor

Portal monitors are passive in that personnel only need to walk through the frame. Although
portal monitors cannot readily distinguish between external and internal contamination, some
are capable of determining radioisotopes and estimating activity (although this would tend to
be a rough estimate). A commercial portal monitor is shown in Figure 38.
Figure 38 Commercial portal monitor
4.6.2 Whole-body counters

Similar to portal monitors (and sometimes synonymous) in that they are often used at entry
and exit points, whole-body counters also detect radiation, although their primary mission is to
detect and quantify internal contamination (presumably after a person has been externally
decontaminated) through gamma emissions. Whole-body counters can use either scintillation
or semiconductor spectroscopy detectors. Primary differences from portal monitors are that
whole-body counters tend to have extensive shielding to reduce background radiation, they
require a dwell time to obtain the measurement, and they may be designed for either standing
or prone operation. A commercial whole-body counter is shown in Figure 39. Thin-window
detectors can be seen along the length of the body, the head, and the feet.
Figure 39 Commercial whole-body counter
4.6.3 Hand/cuff/foot counters

For radiation workers exiting contaminated areas, it is vital to determine that the extremities,
which have the highest probability of becoming contaminated, are adequately monitored. For
this purpose, there are specialized hand-, cuff-, and foot-contamination monitoring systems.
Generally, the person must stand on a detector and insert or place the hands or cuffs onto or
into a detection screen. The detectors may be scintillator or gas-flow proportional counters.
These types of counters are effective at detecting both alpha and beta radiation, which have
much shorter ranges than penetrating gamma radiation.
4.6.4 Thyroid and lung counters

Thyroid counters, as the name suggests, are specialized detectors used specifically for detecting
(principally iodine) radioisotopes in the thyroid (which may be present from fission-product
releases that migrate into the coolant system from failed fuel elements). The thyroid counter is
generally a scintillation-type detector aimed directly at the neck and thyroid area of the person
being monitored. A lung counter is used specifically to detect and quantify radioisotopes that
may have been inhaled. The detectors are arranged around the lung area, and due to the
geometrical constraints and the longer count times required (especially for actinides), the
person being monitored is generally sitting in a reclined seat, with the detectors near the lung
region. Detectors are usually scintillation- or semiconductor-based.
4.7 Dosimetry Techniques

This section will consider thermo-luminescent dosimeters (TLD, as dosimeter of record), elec-
tronic personal dosimeters (EPD), and liquid scintillation counting (LSC) for bioassay dose
estimation.
4.7.1 Thermo-luminescent dosimetry (TLD)

Thermo-luminescence (TL) is a thermally stimulated phenomenon; that is, material absorbs
energy from ionizing radiation or light. This energy is stored, and part of it is released in the
form of light when the material is heated. This phenomenon was first formally reported by
Robert Boyle in his address to the Royal Society of London on October 28, 1663, when he
described, “observation of a glimmering light when he heated a diamond in the dark”
[Uchrin1988]. In the 1950s, research suggested the possibility of applications of thermo-
luminescent properties, including radiation dosimetry, which led to development of more
sophisticated TL phosphors and commercially available TL readers, making thermo-luminescent
dosimetry (TLD) for ionizing radiation widely possible in the 1960s.
The main advantages of thermo-luminescent dosimeters (TLD) are their applicability over a
wide dose range (10-6–105 Gy), their usability for various types of radiation including mixed
fields, and their small physical size, which enables them to be readily worn as personal dosime-
ters. One of the fundamental disadvantages of TLDs is that the act of reading the dosimeter
effectively “zeros” the dosimeter (through a process called annealing), which means that there
is no permanent record of the dose.
The thermo-luminescence mechanism is highly complex and based on solid-state physics. TL
materials are insulators with a crystalline structure. In a perfect crystal lattice, the atomic
electronic levels are broadened into a series of continuous bands (the conduction band) sepa-
rated by a several electron volt-wide "forbidden" energy region from the highest filled band,
called the valence band. Usually, the conduction band is empty, and the valence band is filled.
Crystals always contain imperfections: thermal or intrinsic defects, extrinsic defects, or substitu-
tional impurities and radiation-induced defects [Uchrin1988]. The presence of lattice imperfec-
tions and impurities is essential for thermo-luminescent processes. The energy levels presented
by them are situated in the forbidden region and act as traps or recombination centres for
electrons and holes created in the crystal due to an excitation from interaction with ionizing
radiation.
Electrons and holes will remain in traps provided that they do not acquire sufficient energy to
escape. The number of trapped states is directly proportional to the number of ionizing-
radiation interactions with the material and hence the total dose. When the temperature of the
material is raised, the trapped charge carriers are given sufficient energy to escape from their
traps to the conduction band, where they recombine at a luminescence centre, and the excess
energy is radiated as visible or UV light, which is recorded by a photomultiplier tube (see
Section 4.3.2). The TLD reader unit provides a correspondence between the signal generated
from the measured light and the radiation exposure of the TL material.
Reading a TLD chip is relatively simple. The TL material is heated from the ambient temperature
up to 300°C–400°C, and the emitted light is collected and measured quantitatively. The TLD
reader consists of four components: (i) heating unit, (ii) light collection and detection system
(PMT), (iii) signal measuring system, and (iv) data recording system. The emission of light from
heated TLD material is often called a “glow”, and the resulting spectrum from heating the
material to release the traps is called a “glow curve”. A typical glow curve obtained from a
Harshaw model 3500 TLD reader for LiF TLD material irradiated to 500 μSv total dose from a 1
Ci 137Cs irradiator is depicted in Figure 40. The portion of the curve corresponding to the deliv-
ered dose is denoted by the region of interest marked with vertical lines.
Figure 40 TLD glow curve from irradiated LiF chip

A variety of physical and chemical forms of thermo-luminescent dosimeters are used for
different radiations and dose ranges. For example, physical forms of TLDs can include chips,
disks, powders, and rods; a subset of materials used as dosimeters include LiF, Li2B4O7, CaF2,
CaSO4, and Al2O3. TLDs are characterized by their dose-response properties, energy depend-
ence, sensitivity, and fading characteristics (fading is unintentional loss of stored signal due to
thermal or optical release of traps). The observed glow curve from reading TLDs may be af-
fected by chemo-luminescence (luminescence due to chemical reactions from material impuri-
ties) and tribo-luminescence (luminescence caused by mechanical effects in material prepara-
tion). These effects can be greatly reduced by inert gas (typically N2) flushing. A commercial TLD
reader and LiF chips are shown in Figure 41.
Figure 41 Commercial TLD reader (top) and tray of LiF chips (bottom)
The TLD is the dosimeter of record for nuclear energy workers (NEW) in Canada (National Dose
Registry, Radiation Protection Bureau, Health Canada), as it is in most countries world-wide. All
nuclear energy workers in Canada wear TLD dosimeters if they are working around sources of
radiation exposure and as directed by the health physics department.
4.7.2 Electronic personal dosimeter (EPD)

An electronic personal dosimeter (EPD) is a detector calibrated to provide dose rate and dose
readings in real time. The readings may be present on a display directly on the dosimeter,
stored for later retrieval, or transmitted to a base station. The primary advantage of an elec-
tronic personal dosimeter is that it can be set with alarm points to warn the user (through
audible and visual alarms) in real time if a dose or dose rate is being exceeded. This makes it a
very useful tool for ALARA adherence. Key features of EPDs include: small size, light weight, fast
response, continuous update and display, and timer/stay time functions. EPDs are generally
made from small GM tube or semiconductor detectors, and although the most common are
sensitive to X- and gamma radiation, some have neutron-detection capabilities.
Although EPDs do not provide a legal dose of record, they are often worn in conjunction with
TLDs by personnel working in nuclear-facility radiation environments, especially in non-
homogeneous fields or in areas where exposure may be in excess of action or regulatory limits.
4.7.3 Liquid scintillation counting (LSC)

Liquid scintillation counting (LSC) is a radioanalytical technique developed in the 1950s and
defined by incorporation of a radioisotope analyte into a scintillating liquid. It is a very sensitive
and widely used technique for detection and quantification of radioactivity and is applicable to
most forms of nuclear decay emissions (alpha and beta particles, electron capture, and gamma-
ray-emitting radionuclides).
Consider the case of a beta-emitting radioisotope in a liquid scintillation material. The beta
particle dissipates energy by collisions in the liquid, and the energy is absorbed by the medium
in three forms: heat, ionization, and excitation of the molecules in the solution. Excitation of
the solution molecules is the mechanism of the liquid scintillation technique. Facilitation of
efficient transfer of energy between beta particles and the solution is accomplished using a
solvent material and a scintillation solute solution (together called a “cocktail”). The scintillation
solute is a fluor, and excited solvent molecules can transfer energy to one another and also to
the solute. An excited solvent molecule creates an excited state in the solute, and as the excited
orbital electrons of the solute molecule return to the ground state, a photon of UV or near-
visible light is generated, which can be detected by a photomultiplier tube. The intensity of light
from the scintillation process is proportional to the initial energy of the beta particle.
The detection process for beta decay is summarized in Figure 42.
β UV PMT
Radionuclides Solvent Scintillator
chemical colour
quench quench
Figure 42 LSC detection process

The counting efficiency of the solvent-scintillator system is affected by many different factors
that may reduce detection efficiency. Two dominant factors are: (i) chemical quenching (some-
times called impurity quenching), which causes energy losses in the transfer from solvent to
solute, and (ii) colour quenching, which is the attenuation of UV light photons in the solution.
Both quench factors can be compensated for through calibration. Other effects that may
degrade LSC performance include: (a) thermionic effects, which are noise pulses that are
extraneous to the true signal, yet resemble electronic pulses resulting from nuclear decay
events; (b) photoluminescence, which results in activation of the cocktail or vial by ultraviolet
light, which can occur by exposure to sunlight (LSCs rely on light-tight counting chambers); and
(c) static electricity, which is a very common source of counting interference due to buildup and
subsequent discharge of static electricity on LSC vials.
The electronic process of light detection and pulse generation in an LSC analyzer is depicted in
Figure 43. In most scintillation counters, two photomultiplier tubes collect the total light
produced within the scintillation vial that either (a) falls directly onto the two photocathodes or
(b) is reflected onto each photocathode by a reflector centrally mounted between the two
PMTs. Radioactive decay events produce approximately 10 photons per keV of energy. Beta
decay yielding a multiplicity of photons will stimulate both PMTs at the same instant in time,
and the signal from each PMT is fed into a summing circuit which produces an output only if the
two signals occur simultaneously, which is called coincidence. Because electrical noise from the
PMTs is produced randomly over time, it occurs at a sufficiently low rate to be excluded by the
coincidence circuit (below the equivalent of 1 keV). The sample in the counting chamber and
the PMTs are surrounded by lead, typically about 5 cm in all directions, which generally reduces
the background radiation to low levels.
The output from the analog-to-digital converter is processed with a spectrum analyzer cali-
brated in keV, and regions between 0 and 2000 keV are used for sample analysis. For example,
the region of interest for 3H is 0–18.6 keV (3H maximum beta energy = 18.6 keV) and for 14C is
0–156 keV (14C maximum beta energy = 156 keV).
Figure 43 LSC electronics

Liquid scintillation counting is used in the nuclear industry for swipe analysis, environmental
sample analysis, and urine (bioassay) analysis. LSC is vitally important to detection of one
important radionuclide in CANDU nuclear power stations: tritium (3H). Tritium is extremely hard
to detect with field-portable instrumentation because the average energy of the beta particles
emitted is very low (5.6 keV). In nuclear plants, LSC is routinely used for determining tritium
intake through urine analysis. When tritium contamination is suspected on surfaces, quartz
fibre or paper swipes can be used to wipe down a surface and then be analyzed in an LSC.
Likewise, when performing environmental sample analysis for tritium, especially on water
samples, LSC is the analytical technique of choice for tritium. A commercial LSC (for urine
bioassay) is shown in Figure 44.
Figure 44 Commercial LSC counter (top) and detail of trays with vials (bottom)
4.8 Summary
Ionizing radiation cannot be detected by the human senses, and therefore instrumentation is
required. A variety of radiation detection instruments is used in nuclear facilities to support
routine operations such as dosimetry, radiation protection, and environmental monitoring.
Instrumentation is selected by health physics personnel based on its applicability to the mission
design. Radiation fields to be measured in nuclear facilities may be in the form of external
fields, water effluent, airborne gases, or particulates. Instrumentation is of critical importance
for the safe operation of nuclear facilities such as nuclear reactors and therefore is an integral
part of the radiation-safety management plan.
5 External Radiation Hazards and Shielding

External radiation hazards from beta, gamma, and neutron sources are discussed in this section,
with consideration of the ALARA principle. External radiation fields are always present within
nuclear reactors, and therefore consideration of expected doses and strategies to minimize
dose are required. Some historical perspective on external dosimetry and on shielding can be
found in [Poston2005] and [Shultis2005] respectively.
5.1 External Sources and Dosimetry

The concepts of exposure and dose were explored in Section 3.1. The following subsections
discuss the interactions of the primary radiations of concern to CANDU health physics—
specifically, light charged particles (such as beta particles and electrons), X- and gamma rays,
and neutrons. Note that under normal operating conditions, the primary radiation hazards to
personnel expected in a CANDU plant are from gamma, neutron, and beta radiation. Heavy
charged particles such as alpha particles are not an external hazard in CANDU plants because
they occur only as radioactive decay emissions and because their energy (typically on the order
of 4–5 MeV) is not enough to penetrate the outer layer of skin.
Beta radiation exposure is explored as surface contamination, hot particles on the skin, and a
submersion dose. Gamma-ray equations for point sources (specific gamma constant), line
sources, and volume sources are developed, and fast and thermal neutron-dose calculations
are introduced.
5.1.1 Beta radiation

The ways in which an external beta radiation dose can be delivered to human targets are
summarized in Figure 45. Generally speaking, it is possible to be submerged in a plume of beta-
emitting radioactive material and thereby be exposed in an isotropic field of beta particles (with
most radionuclides, there will be a corresponding gamma field that will require analysis). It is
possible to be exposed through a plane source of beta particles (for example, a spill of beta-
emitting radionuclides) where the target tissue is at some distance from the source, and it is
also possible to encounter a contact source of beta-emitting material (for example, by touching
a spill or a “hot” particle) that can generate a significant dose. Beta radiation can be signifi-
cantly attenuated by clothing, boots, gloves, and goggles. Detailed descriptions of charged
particle ranges and stopping powers can be found in Chapter 3.
Submersion
Contact
Stand-off
Contact
Figure 45 Beta dose delivery geometries
The intensity of beta radiation for depths less than the range of the beta particles can be
described by Eq. (42):
  t
  o e , (42)
where
φ is the beta intensity at depth t (particle flux
flux, energy flux, etc.)
φo is the initial intensity (same units as φ)
ray absorption coefficient (cm2/g)
 is the beta-ray
t is the depth(density thickness) in material (g/cm2).
The beta attenuation coefficients for tissue and air are given by Eq
Eqs. (43) and (44
44) respectively:
cm2
 , tissue  18.6  Emax  0.036 
1.37
(43)
g
cm2
 , air  16  Emax  0.036
1.4
, (44)
g
where Emax is the maximum beta
beta-ray energy in MeV.
ayer of skin provides some beta shielding and is typically represented by 0.007 g/cm2
The dead layer
density thickness. After the beta particles pass through the dead layer of skin, all beta energy is
deposited in living tissue (betas travel no more than approximately 1 cm in tissue), and the
energy deposited is essentially equal to the dose.
Beta doses from common industr
industrial scenarios are presented below.
5.1.1.1 Beta dose from a source on a surface

The dose calculated is from a surface contamination scenario with and without a stand-off
stand
distance of the tissue from the source. A contaminated surface may be represented by Figure
46. From a dose perspective, this may be considered as a “2π” geometry; half the betas travel
up towards the receptor, and half travel down into the sur
surface
face where the contamination is
located.
Figure 46 Beta surface contamination

The energy flux at the surface can be represented by Eq. (45):
  0.5Ca E , (45)
where
 is the energy flux at the surface (MeV cm-2 s-1)
Ē is the average energy per disintegration (MeV)
Ca is the areal activity (Bq cm-2).
The beta surface dose rate, considering the uncollided particles, may be determined using the
beta attenuation coefficient as in Eq. (46):
 MeV 
D       0.5Ca E   . (46)
 gs 
Rewriting Eq. (46) in standard SI dose units yields Eq. (47):
 MeV   Gy 
D   0.5Ca E   7
  2.88 10 Ca E   . (47)
 gs   h 
Of interest also is the contact dose to the skin (the scenario in which the tissue comes into
contact with a surface that has beta-emitting contamination). The dose rate to tissue on con-
tact is represented by Eq. (48):
g
Bq MeV J s   0.007 2
13
0.5  f   Ca 2
 E  1.6  10  3600     e cm
D  
cm transform MeV h
, (48)
0.001 J / g Gy
and the dose rate for tissue on contact with a contaminated surface is given by Eq. (49):
 Bq   MeV  0.007 tissue  mGy 
 
D contact  3.6 104  Ca  2   E      e
 cm   transform 

 h 
, (49)
which assumes that approximately 25% of the beta particles going down into the contaminated
surface are backscattered [Cember2009] towards the dose point (i.e., fβ=1.25) and the expo-
nential represents attenuation through 0.007 g/cm2 of tissue (the dead tissue-layer density
thickness).
For tissue at some distance d away from the contaminated surface, air attenuation must also be
taken into account, as in Eq. (50). Subsequent layers of attenuating material (such as gloves on
hands) may be taken into account using similar exponentials:
 Bq   MeV  0.007  tissue  mGy 
 
D at d  3.6 104  Ca  2   E      e
 cm   transform 
 d  air
e 
 h 
. (50)
5.1.1.2 Beta dose rate from contamination on skin (or hot particles)
For the case where the beta contaminant is in direct contact with external tissue with no
backing material, the beta dose rate is expressed by Eq. (51). Note that no backscatter (fb)
factor is required in this scenario because there is air “behind” the source:
g
Bq MeV J s    0.007 2
0.5  Ca 2
E 1.6 10 13  3600     e cm
D  
cm transform MeV h
0.001 J / g Gy . (51)
 mGy 
 
 2.88  104  Ca  E     e
   0.007

 h 

If the beta emitter is not in contact with the external tissue, other attenuating material can be
added using exponential attenuation factors with appropriate attenuation coefficients. Some
materials that can attenuate betas for skin doses in an industrial setting are listed in Table 13. If
the material is hydrocarbon-based, the tissue beta attenuation coefficient may be an appropri-
ate approximation in the exponential.
Table 13 Standard thickness of various beta attenuators (from [Cember2009])
Material Thickness (mm) Density (g/cm 3)

Lab Coat (Plastic) 0.1 0.036
Cotton Glove Liner 0.3 0.3
Surgeon's Glove 0.5 0.9
Outer Glove (thick) 0.45 1.1
Ribbed Outer Glove 0.55 0.9
Plastic Bootie 0.2 0.6
Rubber Shoe Cover 1.2 1.0
5.1.1.3 Beta dose from submersion in a plume

In an “infinite” cloud, the rate of energy emission is equal to the rate of energy absorption, and
the air dose is given by Eq. (52) for dry air:
k C  E  mGy 
D inf   4.45 107 C E  , (52)
 air  h 
where
k is a constant to yield units in mGy/h (5.76x10-7)
C is the air concentration of the radionuclide (Bq m-3)
E is the average beta energy (MeV)
ρair is the air density (1.293 kg/m3).
More interesting is the dose to tissue for a person submerged in an infinite cloud of beta-
emitting radionuclide. In this case, the expression given by Eq. (52) is modified by a factor of 1.1
for the approximation that tissue absorbs 10% more energy than air on a per-kg basis and by a
factor of 0.5 representing the fact that in the cloud, half the beta particles move inward (to-
wards tissue) and half move outward (away from tissue). Finally, a factor of e-μ * 0.007 compen-
sates for passage through the dead skin layer. Equation (52) therefore becomes Eq. (53):
Bq tps MeV J s
0.5 1.1 C 3
1  E 1.6 10 13  3600  e   0.007
D  
m Bq transform MeV h
J
kg kg . (53)
1.293 3 1
m Gy
Gy
 2.45  1010  C  E  e  0.007
h
Note that Eq. (53) can be modified for attenuation though clothing by multiplying by an expo-
nential factor e-μ * t, where μ represents the beta-attenuation coefficient for the material and t
is the thickness of the material.
5.1.2 Gamma radiation

Gamma rays are neutrally charged quanta of energy (photons). The principal interaction
mechanisms for photons, namely the photoelectric effect, Compton scattering, and pair pro-
duction were introduced in Chapter 3. In addition, photon cross sections and the Klein-Nishina
formula for scattering, as well as interaction cross sections, were introduced in Chapter 3.
Some ways in which external neutral particles such as gamma radiation deliver dose are de-
picted in Figure 47. The exposure can originate from point, area, volume, or distributed source
regions.
5.1.2.1 Point source - specific gamma constant (Γ)

A point source may be considered as small with respect to the target (vanishingly small com-
pared to the target). Consider the energy absorbed per unit mass of air at a specified distance
from a point source, as given by Eq. (54):
fi   / t   E i  MeV /   1.6 106  erg / MeV   3.7 1010  dps / Ci   3600  s / h    m 1   , (54)
 
4    1 m    kg / m3  87.7  erg / g / R 
2
where
fi is the fraction of transformations that yield a photon with energy Ei
μ is the linear energy absorption coefficient for dry air at density 1.293 kg/m3 (m-1).
Plane or volume
Point
Figure 47 Neutral-particle radiation dose-delivery geometries

By expressing the linear energy-absorption coefficient for air as a constant (which is a good
approximation over a large energy range of interest for most radionuclides, as shown in Figure
48), the terms in Eq. (54) can be reduced to a single conversion constant of 0.5, as shown in Eq.
(55):
erg Bq s
1.6  106  3.7  1010  0.0035 m 1  3600
h  0.5 R  m
2
MeV Ci , (55)
erg Ci  h  MeV
 4    1 m 2   1293 g  87.7 g
  m 3
R
which yields the specific gamma constant (Γ) in conventional units as Eq. (56) or in SI units as
Eq. (57). The importance of the specific gamma constant in conventional units is that, from an
operational perspective, it is much easier to memorize compared to the SI unit form:
R  m2
  0.5 fi  Ei (56)
i Ci  h
Sv  m 2
  1.24  107  f i  Ei . (57)
i MBq  h
-4
2x10
Attenuation
-4
10
Interaction coefficient (cm )
-1
Absorption
-5 -1
3.5x10 cm
-5
10
-1 0 1
10 10 10
Energy (MeV)
Figure 48 Linear interaction coefficients for dry air
Use of the specific gamma constant (Γ) to determine dose equivalent rate from a point source is
straightforward, as shown in Eq. (58):
 A  R or Sv 
H p  2  wr  , (58)
d  h 
where
H p is the dose equivalent rate (Rem or Sv per hour, depending upon which form of Γ is
used) at point p (distance d away from the source)
A is the source activity (Ci or MBq, depending upon which form of Γ is used)
d is the distance from source to receptor (m)
wr is the radiation weighting factor (=1 for gamma rays).
5.1.2.2 Line source

A line source configuration is one that is infinitely thin but of finite length, such as a pipe
carrying contaminated radioactive liquid. The geometry of such a system is described by Figure
49. The receptor is at point p at distance h from the line, at an arbitrary distance along the line.
l=l1+l2
dl
l1 l2
d1 h d2
l12 + h2=d12 l22 + h2=d22
θ
p
Figure 49 Line-source geometry
For a linear concentration of radioactivity Ci (Ci/m or MBq/m), the dose rate at point p can be
expressed as a function of the point source-specific gamma constant as defined in Section
5.1.2.1, using Eq. (59):
Cl  1 l1 l 
H p   tan  tan 1 2 
h  h h
or , (59)
 
H p 
Cl
h
where
H p is the dose rate (R/h or Sv/h, consistent with Γ)
l1, l2, and h are defined in Figure 49 (m)
θ is the opening angle (radians)
Γ is the specific gamma constant for a point source (conventional or SI units)
Ci is the linear activity (Ci m-1 or MBq m-1, consistent with Γ).
5.1.2.3 Planar source

A planar source in this case is represented as a disk source which is infinitely thin with area πr2.
This type of geometry could arise, for example, from a spill of radioactive material and is
represented by Figure 50.
dr
r
p
Figure 50 Planar source geometry

For an areal concentration of radioactivity Ci (Ci m-2 or MBq m-2), the dose rate at point p can be
expressed as a function of the point source-specific gamma constant as defined in Section
5.1.2.1, using Eq. (60):
 r 2  h2 
H p   Ca ln  2 , (60)
 h 
where
r and h are defined in Figure 50 (m)
Ca is the areal activity (Ci m-2 or MBq m-2, consistent with Γ).
5.1.2.4 Volume source

A volume source is one that has finite dimensions, such as a drum containing radioactive
material. The geometry of a cylindrical system used as an example is described by Figure 51.
The receptor is at point p on the central axis at distance h from the surface of the volume.
r
p
x
dx
h
Figure 51 Volume-source geometry
For a uniform volume concentration of radioactivity Cv (Ci m-3 or MBq m-3), the dose rate at
point p can be expressed as a function of the point source-specific gamma constant as defined
in Section 5.1.2.1, using Eq. (61):
 r 2  h2 

H p  
Cv

1 e   t
ln  2 
, (61)
 h 
where
r, t, and h are defined in Figure 51 (m)
Cv is the volume activity (Ci m-3 or MBq m-3, consistent with Γ)
μ is the linear absorption coefficient of the volume source material (cm-1).
It is possible to extend this volume treatment to other simple geometries in a similar manner.
5.1.3 Neutrons
Neutrons can be classified in a variety of ways. The most common way to classify neutrons is by
kinetic energy. Thermal neutrons and neutrons in thermal equilibrium with their environment
are distributed according to a Maxwell-Boltzmann distribution. Their most probable energy is
0.025 eV at 20°C. At higher energies from 0.01–0.1 MeV, neutrons may be classified as slow,
intermediate, or resonance. Fast neutrons have energies from a few MeV up to approximately
20 MeV, and relativistic neutrons have still higher energies. Neutrons are uncharged (neutral)
like photons and can travel appreciable distances without interacting. Neutrons are attenuated
exponentially under good geometry and do not interact appreciably with electron fields.
Neutrons collide with atomic nuclei in elastic and inelastic collisions. With elastic collisions, the
total energy is conserved, and the energy lost by the neutron is equal to the total energy of
recoil of the nucleus. With inelastic collisions, the nucleus absorbs some of the energy and is
left in an excited state, which upon de-excitation can yield neutrons, gammas, and light ions.
Fast neutrons undergo a series of primarily elastic scattering reactions, being slowed down in a
process called moderation. As neutron energy decreases, scattering continues, but the proba-
bility of capture by another nucleus increases. If the neutron reaches thermal energies, it will
randomly move around until absorbed by a nucleus.
5.1.3.1 Neutron reactions

A number of reactions important to neutron interactions are listed in Table 14. Some of these
reactions are important from a dose-to-tissue perspective, whereas some are important for
considering neutron detection.
Table 14 Important neutron reactions
Reaction Typical use

1
H(n,γ)2H Dose to tissue
3
He(n,p)3H Neutron proportional counters
6
Li(n,T)4He
Thermal neutron detection
10 7
B(n,α) Li
14
N(n,p)14C Dose to tissue
23
Na(n,γ)24Na Blood dosimetry
40
Ar(n,γ)41Ar Internal dosimetry
32
S(n,p)32P
113
Cd(n,γ)114Cd
Neutron spectroscopy
115 116m
In(n,γ) In
197
Au(n,γ)198Au
235
U(n,f) Thermal neutron detection
Besides the interactions presented in Table 14, neutron activation of materials can also be a
significant source of gamma dose in nuclear plants. Neutron activation is the production of a
radioactive isotope by absorption of a neutron and is governed by Eq. (62):
dN
 n N
dt
or , (62)
 N    n 1  e t
 A
where
A is the activity of the radioisotope produced through neutron activation (Bq)
 is the neutron flux (n cm-2 s-1)
 is the activation cross section (cm2)
 is the decay constant of the produced isotope (s-1)
N is the number of produced atoms
n is the number of target atoms.
Although the radioisotopes produced are generally a gamma radiation hazard, activation is an
important reaction with respect to this source of gamma (or beta) dose.
5.1.3.2 Neutron dosimetry

The absorbed dose from neutrons, from a biological dose perspective, involves primarily
neutron interaction with tissue elements H, C, O, and N and the dose resulting from these
interactions. Neutron interactions produce recoil protons and heavy charged particles with
short ranges that deposit their energy locally in tissue. Dosimetry estimates may be made
empirically be examining separately (i) fast neutrons and (ii) thermal neutrons.
I. Fast neutrons
The dose rate from fast neutrons in a material made up of a number of elements can be esti-
mated using Eq. (63):
 ( E )  E   Ni i fi

 Gy 
Dn  E   i
 
1
J
Gy  s , (63)
kg
where
(E) is the fast neutron flux (n cm-2s-1)
E is the neutron energy (Joules; 1.6E-19 Joules per eV)
Ni is the number of atoms per kg for the ith material element
i is the scattering cross section (cm2) for the ith material element
fi is the fractional energy transferred from neutrons for the ith material element, given
2M i
by f i  , where Mi is the atomic mass of the ith material element. fi
( M i  1) 2
represents the average fraction of neutron energy transferred in an elastic colli-
sion assuming isotropic scattering.
The fast neutron dose can be calculated as a function of different energy groups (corresponding
to the neutron energy spectrum) to determine the total fast neutron dose.
II. Thermal neutrons
The most important specific thermal neutron reactions for tissue dosimetry (hydrogen and
carbon) are described below.
Hydrogen capture - 1H(n,)2H
This is an example of a radiative capture process. The neutron is absorbed, followed by imme-
diate emission of a gamma photon, 01n  11H  12 H  00 , with photon energy 2.22 MeV. This
reaction is of significance when tissue is exposed to thermal neutrons. This reaction yields a
uniformly distributed gamma-emitting radioisotope, and therefore initially one must solve for
the specific activity (As) of this source using Eq. (64)
 Bq 
As   N H  H  , (64)
 kg 
where
 is the thermal neutron flux (n cm-2s-1)
NH is the number of hydrogen atoms per kg of tissue (5.98x1025 atoms/kg)

H is the absorption cross section for hydrogen (0.33x10-24 cm2).
Knowing the specific activity, it is possible to solve for the internal dose using approaches
presented in Section 6.
Nitrogen capture - 14N(n,p)14C
This reaction has a large thermal cross section of 1.78 barns and an energy release of 0.626
MeV in tissue. This is a significant contributor to dose when tissue is irradiated by neutrons
because the proton and the nucleus recoil have limited range and the energy is deposited
locally. The dose rate due to the energy released by charged particles for this reaction can be
determined by Eq. (65):
J
  N N   N  Q  (1.6 1013 )
MeV  Gy 

D np   , (65)
1
J
Gy  s 
kg
where
 is the thermal neutron flux (n cm-2s-1)
NN is the number of nitrogen atoms per kg of tissue (1.49x1024 atoms/kg)
N is the absorption cross section for nitrogen (1.75x10-24 cm2)
Q is the energy release in tissue (0.626 MeV).
5.2 External Hazards

A variety of potential external radiation hazards are associated with nuclear technology. When
a reactor is under power, a variety of areas inside containment can produce significant neutron
and gamma radiation fields. Generally speaking, these areas are well characterized, and admin-
istrative measures are put in place to minimize any hazard from these fields. When the reactor
is on power, there are hazards from fission neutrons, fission prompt gamma rays, and neutron-
capture gamma rays, as well as fission-product gamma rays, activation gamma rays, and photo-
neutrons. When the nuclear chain reaction is halted, the direct neutron component of the
hazard rapidly reduces to an insignificant proportion. The hazards that remain are therefore
fission-product gamma rays, activation gamma rays, and to a lesser extent, photoneutrons. In
an ideal situation, the fission products produced in the fuel would remain in the fuel cladding.
However, fuel failure does occur and contaminates water systems, which can lead to deposition
of fission products in piping and associated systems. In addition, activation products can mi-
grate away from their original locations due to corrosion processes.
The most common operational hazards from external radiation in a nuclear plant are due to:
(a) Fission-product activity in the core (present regardless of the power condition of the
core)
(b) Activation products in the core and surrounding material (including water)
(c) Fission products that migrate and deposit at various locations in the reactor
(d) Activation products that migrate and deposit around various reactor systems
(e) Contamination, in liquid or aerosol form, due to fission and activation products that be-
come inadvertently uncontained (which also may pose a significant external beta hazard
and internal alpha/beta hazard if inhaled).
The possible sources of the radiation that may contribute to external radiation dose while a
CANDU reactor is operating are depicted in Figure 52. The simplified drawing depicts the
biological shield (this shield also encompasses the thermal shield and reflector region), the core
(where the fission reaction takes place), and the primary heat transport (PHT) system that
circulates the hot pressurized water directly to the steam generators. The reactor operates as a
“closed primary circuit”, and therefore radioisotopes generated as a result of fission, activation,
or other nuclear processes may be circulating in the primary heat transport (PHT) system. The
primary circuit is physically separated from the secondary steam-generation circuit at the steam
generators. In theory, there should be no sources that migrate from the primary to the second-
ary side in a CANDU, and therefore no radioactivity at the turbines. Referring to Figure 52, the
potential sources may be described as follows (after [Goldstein1962]):
 Prompt fission neutrons – emitted within first few μs after fission
 Delayed fission neutrons – emitted from excited nuclei up to a few minutes after fission
 Activation neutrons – emitted from nuclear reaction products
 Photoneutrons – produced through threshold (γ,n) reactions
 Prompt-fission gammas – emitted in coincidence with fission (< μs)
 Short-period fission gammas – emitted by fission products within a few minutes (~10 min)
after fission
 Long-period fission gammas - emitted by fission products after ~10 minutes
 Capture gammas – emitted in (n,γ) reactions
 Inelastic-scatter gammas – emitted from excited nuclei after neutron inelastic scattering
 Reaction-product gammas – emitted from products of charged-particle reactions induced
by neutrons
 Activation-product gammas – emitted from radioactive products of nuclear reactions
 Annihilation photons – emitted from positron annihilation due to certain activation or
fission products
 Bremsstrahlung photons – emitted from decay electrons (beta particles) slowing down in
material.
Note that not all the listed sources are of equal importance, and often their importance is
dictated by their physical location in the reactor with respect to receptors (workers). In addi-
tion, the importance of sources is affected by the state of the reactor. On shutdown, some of
the sources disappear, and the relative importance of the remaining sources changes. There is
residual fission power in a CANDU core after shutdown, and therefore there are still prompt
fission gammas and neutrons, as well as capture gammas produced in the core. However, they
are produced at levels several orders of magnitude lower than during operation. Possible
external radiation sources ~10 minutes after shutdown are depicted in Figure 53.
Biological Shield
Neutrons
Photoneutrons
Core Secondary
Core side
Gamma rays
Capture Neutrons Gamma rays
Inelastic Prompt fission Prompt fission
Reaction Short period fission
Activation Capture
Annihilation Inelastic
PHT Gamma rays

Capture
Neutrons Inelastic
Delayed fission Activation
Activation Reaction
Photoneutrons Bremsstrahlung
Annihilation
Figure 52 Potential external exposure sources during reactor operation
Biological Shield
Neutrons
Photoneutrons
Core Secondary
Core side
Gamma rays
Activation Neutrons Gamma rays
Photoneutrons Long period fission
Activation
PHT
Gamma rays
Long period fission
Activation
Neutrons Bremsstrahlung
Photoneutrons Annihilation
Figure 53 Potential external exposure sources after shutdown (> 10 min)
Some examples of external hazards in a nuclear reactor environment include (but are not
limited to) the following:
 Mixed neutron/gamma fields from normal operation (under power or shut down) at various
locations around the reactor
 Contaminated coolant in a pipe (circulating or uncirculating)
 Contaminated resin columns
 Fission or activation products that have lodged or attached somewhere unexpectedly in a
piping system
 Leakage from a contaminated water system
 Outage activities that may involve opening piping or pump systems, welding and grinding
activities on potentially contaminated systems, etc.
Protection against these external hazards includes proper work planning (discussed in Section
7) along with use of the ALARA principle and shielding (discussed in Section 5.3).
5.3 External Protection (Radiation Shielding)

The essential principle of external radiation protection is embedded within the so-called ALARA
(as low as reasonably achievable) principle. The ALARA principle was formalized by the ICRP in
their 1977 recommendations [ICRP1977] and may be considered as the fundamental philoso-
phy of radiation protection that suggests that no practice dealing with radioactive material
should be adopted unless there is a net benefit arising from it. If radiation is to be used, all
efforts must be made to ensure that the dose is the lowest absolutely required for the given
task. The complete text of the ALARA principle clarifies the context of applicability and has been
paraphrased as [ICRP1977] making every reasonable effort to maintain exposures to radiation
as far below the dose limits as is practical, consistent with the purpose for which the activity is
undertaken, taking into account (i) the state of technology, (ii) the economics of improvements
in relation to the state of technology, (iii) the economics of improvements in relation to bene-
fits to public health and safety, (iv) other societal and socioeconomic considerations, and (v)
relationships to use of nuclear energy and radioactive materials in the public interest.
The typical radiation hazards for CANDU reactor operations are depicted in Figure 54. In applied
external radiation protection for CANDU reactors, it is normal to consider gamma, neutron, and
beta radiation. Alpha radiation, although a great concern for inhalation dose (considered in
Section 6), is of no concern as an external dose because the radiation cannot penetrate the
external dead layer of tissue and certainly not any clothing that a person may be wearing. Beta
radiation (discussed in Section 5.1.1) can be of concern externally if a worker is not protected
and is in contact with a source, submerged in a plume, or gets a “hot” particle on unprotected
skin. However, in this type of scenario, there are usually corresponding gamma emitters that
require shielding, and therefore the beta particles are generally well shielded as a result. For
other applications like nuclear medicine, low atomic number (low-Z) materials can provide
effective shielding from beta radiation (while simultaneously yielding less bremsstrahlung
radiation compared to high-Z shields). It is common to use PlexiglasTM shielding for beta-
emitting radioisotopes. Therefore, for normal shielding, the focus is on gamma and neutron
radiation.
Figure 54 External hazards

The reduction
uction of doses to ALARA for external radiation is primarily accomplished using three
principal concepts of radiation protection
protection: shielding, distance, and time. Simply stated, shield-
ing is placing something between the source and receptor that blocks (or attenuates)
attenuate radiation,
distance is keeping space between the source and the receptor, and time is spending as little
time around the source as required to reduce exposure. In addition, it is common to use the
effect of radioactive decay for short
short-lived
lived radioisotopes as part of the shielding strategy (for
example, delaying entry into an area after shutdown to allow time for radioactive decay).
Time is an obvious dose parameter because dose is accumulated over time and is delivered as a
function of time from an external source (dose rate). Therefore, less time = less dose. The role
of distance
istance is readily understood by considering the inverse
inverse-square
square law (which is also important
in shielding discussions)) depicted in Figure 55.
S
I
4 r 2
1
1 I
I 9
4
source emission I
S
surface area  4 r 2
r
2r
3r
Figure 55 Inverse-square law

The inverse-square
square law can be readily understood by rayray-tracing from a point source outwards
and considering the intensity per unit area at the intersection with an arbitrary sphere of radius
r. For every increase in radius, the intensity decrease
decreases by the square of the increased radius
because the intensity is propagating outwards across the expanding sphere. For example, twice
the radius leads to one-fourth the intensity, three times the radius, one-ninth the intensity, four
times the radius, one-sixteenth the intensity, and so on. Because dose is directly proportional to
radiation intensity, this demonstrates that dramatic decreases in dose can be achieved with
moderate changes in distance from the source, particularly close to the source.
As discussed earlier, shielding is the act of placing a material between the source of radiation
and the receptor, which has the effect of reducing the amount of radiation impinging on the
receptor. In radiation protection, shielding is often called biological shielding because its
primary purpose is to reduce the biological dose to humans. As such, shielding materials are
critically important to the effectiveness of shields. The following sections consider shielding for
gamma and neutron radiation.
5.3.1 Gamma shielding

In Section 5.1.2, the principles of gamma dosimetry were discussed. The relationship to deter-
mine intensity (fluence, flux, or dose) of radiation after passage through a shield is given by Eq.
(66):
I  I 0e  x , (66)
where
I is the intensity of the radiation after passing through the shield (fluence, flux, or dose
units)
Io is the intensity of the radiation before passing through the shield (same units as I)
μ is the linear attenuation coefficient for that material (cm-1)
x is the thickness of the shield (cm).
Taking into account both the distance from the source and receptor and the influence of the
shielding material, the intensity at a point r units away from a source shielded by a thickness x
of material is given by Eq. (67):
I 0e  x
I . (67)
4 r 2
The physical conditions that describe this type of attenuation are that (i) the radiation is a
narrow beam interacting with the shield and (ii) the shield is thin with respect to the mean free
path (1/μ) of the gamma radiation in the shield. This scenario is depicted in Figure 56. In this
case, the radiation, depicted by the dashed lines, scatters away from the receptor.
I0
x I 0e  x
4 r 2
Figure 56 Narrow-beam or thin-shield attenuation
5.3.1.1 Buildup
Although approximation of Eq. (66). may be valid in some cases, there are a number of situa-
tions in which the radiation source may be considered a broad beam with respect to the shield
or the shield may be thick with respect to the mean free path of radiation in the shield. This
case is depicted in Figure 57. It has been observed that in this scenario, some of the radiation
that would have scattered away from the receptor actually scatters towards the receptor and
contributes additional intensity to the receptor location that is not predicted by Eq. (66). In this
case, a factor is used to modify the standard attenuation equation to account for the scattering.
This factor is known as the buildup factor (B) and is dimensionless. The intensity equation
therefore becomes Eq. (68):
I 0 Be   x
I . (68)
4 r 2
r
I0 I 0 Be   x
x
4 r 2
Figure 57 Broad-beam or thick-shield attenuation
The buildup factor is both a function of energy and the shield material (μx, denoted as the
optical length, relaxation length, or number of mean free paths) and is given as Eq. (69):
I 0 ( E ,  x)  I s ( E ,  x) I ( E ,  x)
B( E ,  x)  1  s , (69)
I 0 ( E ,  x) I 0 ( E ,  x)
where
B is the buildup factor (dimensionless)
I0 is the non-collided intensity (fluence, dose, or dose rate) at the receptor
Is is the scattered (collided) radiation contribution at the receptor.
The buildup factor is the ratio of the radiation intensity, including both primary and scattered
radiation, at any point in a beam to the intensity of primary radiation only at that point. The
buildup factor accounts for scattering into an area due to poor (broad-beam) geometries and is
a function of shield material and shield thickness (x). Examining Eq. (69) leads to the conclusion
that B is always ≥ 1. Values for B have been experimentally determined; however, the bulk of
data available are from calculations, and many evaluations are presented in tables or in a
graphical format.
Two common ways of using the buildup factor for solving a shielding problem are: (i) knowing
the source and shield, calculate the intensity, and (ii) knowing the desired intensity at the
receptor point, determine the amount of shielding required.
Calculating intensity
The steps in calculating intensity (fluence, dose, or dose rate) are given as follows:
1. Look up B for:
 Material of interest (e.g., lead, iron, etc.)
 Energy of photons (e.g., 1 MeV)
 Relaxation length of shield (μx) – Relaxation length is the thickness of shield ma-
terial that will attenuate a narrow beam to 1/e of its original value.
2. Solve equation: I=I0*B*e-ux for fluence, dose, or dose rate.
Calculating required shield thickness
The calculation of the thickness of a shield required for a given desired fluence, dose, or dose
rate is difficult because the equation must be solved simultaneously for two variables, B and x,
which can be solved for only by iteration. The steps in calculating the required shielding thick-
ness are the following:
1. Define the target exposure I (fluence, dose, dose rate)
2. Solve the narrow-beam equation I=I0*e-ux to obtain an initial shielding thickness x. This
thickness will be too small.
3. Add one half-value of thickness to x (μ = (ln 2) / x1/2 → x1/2 = (ln 2) / μ)
 (new x) = (x + x1/2)
4. Look up B for material, energy, (new x)* μ.
5. Solve I=I0*B*e-ux with new values
 Check: Is the calculated value of I close to the target I?
6. If the calculated I is still too large, then add another x1/2 and recalculate (i.e., go to step
3).
Note that the shield may be placed anywhere between the source and the receptor; however,
when the shield is placed closer to the source, it provides the greatest solid-angle protection to
the receptor. Tabulated buildup factors can be found in a number of sources including, for
example, through the American Nuclear Society [ANS1991] and in work by Shultis and Faw
[Shultis2000].
A number of numerical techniques can be used to perform gamma-radiation transport calcula-
tions suitable for use in shielding estimates, for example, point-kernel techniques, discrete-
ordinate techniques, and Monte Carlo simulation. The need for advanced gamma-shielding
approaches arises from requirements to simulate complex geometries, multilayer shields,
complex materials, complex source terms, and differing cross-section evaluations.
5.3.1.2 Materials
In general, there are three categories of shield materials: (1) natural materials, (2) construction
materials, and (3) special materials. Natural materials include air (insofar as it can be a scatter-
ing medium), water, and soil, and often shielding design involves making judicious use of these
natural materials. Construction materials are a very advantageous design material from an
optimization standpoint because they can serve two purposes, as structural and as shield
materials. Construction materials may include concrete, steel, wood, plasterboard, and glass.
Special materials for gamma shielding are generally high-atomic-number materials such as
steel, iron, lead, tungsten, and uranium. The usefulness of a shield for any given application is
related to the mass-attenuation coefficient (directly related to the interaction cross section)
and the mass density of the material as a function of the photon energy spectrum. A plot of the
interaction cross sections of photons in lead is provided in Figure 58. It can be seen that the
photoelectric effect is the dominant interaction at energies below approximately 0.2 MeV,
Compton scattering is competing or dominant between approximately 0.2 and 4 MeV, and pair
production dominates above 4 MeV (further discussion of photon interactions is provided in
Chapter 3).
6
10
Compton scatter
Photoelectric
10
5 Pair production
Cross section (b/atom)
4
10
3
10
2
10
1
10
0
10
-3 -2 -1 0 1 2
10 10 10 10 10 10
Energy (MeV)
Figure 58 Photon interaction cross sections for lead

Photon total-interaction coefficients for uranium, lead, aluminum, and water are shown in
Figure 59. It can be seen that for high-Z materials such as uranium and lead, the interaction
mechanisms are very similar. For lower-Z materials such as aluminum and water, the interac-
tion coefficients are significantly lower at most energies. Note that the total interaction coeffi-
cients at moderate energies (0.5–3 MeV) are similar for all four materials and that therefore

their attenuating power comes primarily from their different mass densities because     x

is the argument of the exponential function when μ is the linear attenuation coefficient.
4
10
Interaction coefficient (cm /g)
3
10
2
Uranium
Lead
2
10
1
10
Water
0
10 Aluminum
-1
10
-2
10
-3 -2 -1 0 1
10 10 10 10 10
Photon Energy (MeV)
Figure 59 Photon interaction coefficients
5.3.2 Neutron shielding

Compared to the simplified gamma-shielding techniques, fast-neutron shielding is somewhat
more complicated. These complications arise from a number of considerations:
 Primary neutrons (usually fast) have a high scattering probability
 Inelastic scattered neutrons produce gamma photons
 Thermal neutrons are produced after slowing down
 Capture gamma photons from (n,γ) reactions arise mainly from thermal neutrons
 Secondary neutrons from (n,2n), fission, etc., can be generated
 Neutron production is usually accompanied by primary gamma photons
 Neutron absorption can activate shield material, which in turn emits delayed gamma
photons.
Taking all this into account, a simplified approach such as using buildup factors (similar to the
approach with photons) is not generally practical for neutron problems because [Shultis2000]:
 Neutrons scatter much more than photons, thereby making B very large
 There is a large buildup of low-energy neutrons because the absorption cross section (σa) is
small in the slowing-down energy region
 B depends strongly on the
o Isotopic (material) composition of the shield
o Physical geometry of the shield
o Incident neutron-energy spectrum.
Due to the difficulties outlined above, the buildup-factor approach will not be developed for
neutrons. However, in addition to the above, because neutrons have radiation-weighting
factors that are energy-dependent, the difference between dose conversion factors (response
functions) for air kerma, ambient dose, and dose equivalent can be significant, as shown in
Figure 60. This can lead to significant differences in dose estimation.
3
10
Response function (pGy cm or pSv cm )
2
2
10
Effective equivalent (AP)
1
10
2
Deep dose
Ambient dose
0
10
Tissue KERMA
-1
10
Air KERMA
-2
10
-3
10
-6 -5 -4 -3 -2 -1 0 1
10 10 10 10 10 10 10 10
Neutron Energy (MeV)
Figure 60 Response functions for neutrons
General considerations with respect to neutron shielding include:
 For biological shields, fast neutrons are of most concern because thermal neutrons are
readily absorbed
 Q (wR) varies with neutron energy and is greatest for 0.1 < E < 2 MeV
 Materials with large σa are used to absorb thermal neutrons (e.g., Cd, In, B, Li). However,
one must consider the resulting capture gamma radiation
 To stop fast neutrons, they can be thermalized by

o moderation (low-atomic-number materials) and
o inelastic scattering (e.g., by iron)
and then absorbed as thermal neutrons.
Most neutron-shielding analyzes require complex transport calculations. However, a few
simplified techniques are available. For example, point kernels can be used for fission neutrons
in hydrogenous and non-hydrogenous media, diffusion and removal theories for intermediate
energy neutrons, and point kernel techniques for capture gamma photons. In addition, tabu-
lated data for neutron shielding (for example, concrete shields) are often available.
5.3.2.1 Removal techniques

The removal cross section can be used to estimate the flux of fission neutrons in a hydrogenous
material (e.g., water) of distance r from source to receptor that is shielded by a non-
hydrogenous material of thickness t. In this case, the flux is given by Eq. (70) [Lamarsh2001]:
  r   S  G  r   e r t
where
, (70)
 rwater r
0.12  e
G r  
4 r 2
where
φ is the fission (> 1 MeV) neutron flux at r (n cm-2 s-1)
S is the source emission rate (n s-1)
Σr is the macroscopic removal cross section for the shield material (cm-1)
Σrwater is the macroscopic removal cross section for water (0.103 cm-1)
t is the thickness of the shield (cm)
r is the distance of the water from source to receptor (cm).
The macroscopic removal cross section can be expressed as a function of the microscopic cross
section, as in Eq. (71):
 r  N r
or , for mixtures . (71)
 r   Ni ri
i
Macroscopic and microscopic cross sections for a variety of elements and mixtures are provided
in Table 15 (adapted from [Shultis2005]).
Table 15 Removal cross sections for neutrons
Material Σr (cm-1) σr (b)
Aluminum 1.31
Hydrogen 1.00
Deuterium 0.92
Beryllium 0.132 1.07
Boron 1.07
Carbon 0.065 0.81
Oxygen 0.92
Sodium 0.032 1.26
Iron 0.168 1.98
Zirconium 0.101 2.36
Lead 0.118 3.53
Tungsten 3.36
Uranium 0.174 3.60
Zirconium 2.36
Water 0.103
Paraffin 80.50
Heavy water 0.092 2.76
Concrete
0.089
(6% water)
An approximation for the removal cross section (where Σr = μr) in terms of μr/ρ is given by Eq.
(72):
r 1.3 0.294  cm 
2
 0.206 A Z  . (72)
  g 
Functionally, the removal technique can be used to estimate the flux at the surface of a shield,
and then this flux can be used to estimate the flux at a position outside the shield. Knowing the
fast neutron flux at a receptor point outside the shield, the dose rate can be estimated using
Eq. (63). Note that this is a simplified approximation, and a more complex analysis (such as the
Monte Carlo technique) is required to determine the intensity accurately.
5.4 Summary
External radioisotope hazards associated with CANDU reactors include hazards directly associ-
ated with the nuclear chain reaction (neutron and gamma fields) or indirectly associated as is
the case with fission products from fuel (whether intact or defective) and activation products
(whether fixed or mobile). Dosimetry can be accomplished from first principles using source
activity and geometry or by using dose-conversion factors and fluence measurements.
External hazards in CANDU reactors involve gamma, neutron, and beta fields, and protection of
the worker from external hazards is accomplished by adhering to the principles of “as low as
reasonably achievable” dose and the concepts of maximum shielding (of source), maximum
distance (from source), and minimum time (around source) for any given work function.
6 Internal Radiation Hazards

Internal radiation hazards from alpha, beta, and gamma sources are discussed in this section.
Pathways for internal radiation exposure are periodically present within nuclear reactors, and
therefore consideration of expected doses and strategies to minimize dose are required. Some
historical perspective on internal dosimetry can be found in [Potter2005].
6.1 Internal Pathways and Dosimetry

When a material containing radioisotopes is inhaled, ingested, or absorbed through the skin or
a wound, this will lead to an internal dose, as depicted in Figure 61. For an ingestion pathway,
typically (1) contamination (for example, a radioactive spill) occurs. If the contamination is
touched (whether protected or unprotected), the radioisotope(s) can be either (2a) transferred
to the hands or (2b) transferred through a wound if present and directly to the blood. If the
hands are contaminated, the radioisotope(s) can be (3) transferred to food or the face directly,
from where they can be (4a) transported to the gastrointestinal (GI) tract. Alternately, for an
inhalation pathway, if a person is in proximity to airborne radioisotopes, the person may (4b)
take in the radiation directly from the plume into the respiratory tract. In both cases, radioiso-
tope(s) will be transported and (5) taken up into the blood and GI tract. From there, they can be
transported to target organ(s) and deposit decay energy, or in other words, a dose.
4b. Intake from

contaminated plume
3. Transfer of contamination to
food product, or hand-to-face
4a. Intake of directly
contamination from food
5. Uptake into blood product or orofacial 2b. If wound,
and GI tract contamination transfer into blood
2a. Transfer of
contamination to hands
1. Contamination
Figure 61 Internal contamination pathways

As was discussed in Section 3.4, the total dose assigned to an individual is the sum of all exter-
nal and internal exposures. In an industrial environment, the dominant pathway for internal
exposure is through inhalation of airborne radioisotopes.
Neutrons are generally not considered when discussing internal dose because the possibility of
internalizing a neutron emitter (for example, a spontaneous fission-neutron source) is extreme-
ly remote. However, in contrast to external dosimetry, alpha-emitting radioisotopes can be a
major contributor to the committed effective dose that a person might receive.
A number of factors affect internal dose, as outlined in Table 16. Note that not only does the
type and quantity of radioactive material contribute to dose, but also other factors such as the
chemical form of the material and its biological function in the human body.
Table 16 Parameters affecting internal dose
Physical Characteristics Biological Characteristics

Radionuclide Metabolic behaviour
Physical half-life Biological half-life
Chemical & physical form Chemical toxicity
Emitted radiation Tissue sensitivity
Intake route Age of individual
Duration of intake Individual health
Total intake Personal habits
The internal dose mechanism can be described as shown in Figure 62. Internal dose is facili-
tated by basic mechanisms for intake into the body (inhalation, ingestion, or absorption),
followed by uptake into extra-cellular fluids for circulation through the body, and finally deposi-
tion of radioisotopes into target organs and tissues where decay energy is deposited.
The dominant intake regimes related to transport through the body to excretion are outlined in
Figure 63. This is a simple representation of what is called a compartmental model. Dosimetry
can be accomplished by solving mathematically for the transport, deposition, and excretion of
radioisotopes through “compartments” in the body. Most internal-dosimetry computer codes
perform this type of calculation (for example, see [IMBA2010]). Compartmental modelling of
internal dose is beyond the scope of this chapter.
Inhalation into the respiratory system can be subdivided into three regions: (i) nasal-pharyngeal
(inhalable), (ii) tracheal-bronchial (thoracic), and (iii) pulmonary (respirable, or deep lung).
Particle deposition takes place in the three regions with an efficiency based on particle diame-
ter. Generally speaking, particles from submicron to ten microns are considered respirable,
whereas larger particles deposit in the upper airway regions. Although upper-airway deposition
can lead to dose, the respirable fraction is of primary significance for internal dosimetry.
Inhalation leads to deposition in one of the three regions of the respiratory tract, and if depo-
sited in the deep lung, particles can be absorbed into the blood for transport to target organ(s).
Inhalation can also lead to a fraction being removed mechanically (by coughing, etc.) from the
respiratory tract into the GI tract, where it may be treated as an ingestion and similarly ab-
sorbed into the blood for distribution to other organs before excretion. Excretion may occur
through urine, feces, sputum, sweat, tears, etc.; however, the dominant modes of biological
excretion are through urine or feces. Before radioisotopes are excreted, they will deposit
energy in the target tissue(s). The deposition of this energy into a mass of tissue is the defini-
tion of absorbed dose.
Intake Into Body
Extracellular
Uptake By
Fluid
Organ,
Deposition In
Tissue
Figure 62 Basic mechanisms for internal dose
Inhalation Ingestion
Respiratory
Tract Mechanical
Removal G-I
Tract
Absorption Absorption
Blood
(Cumulative Uptake)
Other Organs

Biological Removal
Figure 63 Intake to excretion pathways
The computation of internal dose requires three (3) critical pieces of information:
1. Rate of dose deposition (dose rate in tissue)
2. Elimination of radioactivity from tissue (decay and removal)
3. Total dose “committed” to tissue (committed dose).
In addition, simplified approximations can be made for short-range (high linear energy transfer)
particles such as alpha and beta radiation, compared to most penetrating (low linear energy
transfer) X- and gamma radiation. These approaches will be developed in Sections 6.1.1 and
6.1.2.
6.1.1 Alpha and beta internal emitters

The determination of absorbed dose from internal emitters follows directly from the definition
of dose (J/kg, or Gray). Assumptions of an infinitely large medium and a uniform distribution of
radioisotope in tissue are used. “Infinitely large” means that the target-tissue dimensions
exceed the range of the radiation. For particles with short ranges (alpha and most beta radia-
tion), the energy absorbed by the surroundings equals the energy emitted by the radioisotope.
These assumptions are captured by the concept of specific effective energy (SEE) given by Eq.
(73):
E  or    MeV 
SEE  or     , (73)
m  transform  kg 
where
E is the average energy of the emitted particle per disintegration (MeV per transform)
m is the mass of the target organ or tissue (kg).

The instantaneous dose rate at time=0 from intake of a short-range (charged particle) emitter
of activity Ct (Bq) can therefore be calculated from the SEE using Eq. (74):
 dps   MeV   J  s
 Ct Bq   1  SEE  E  1.6 ×10
-13
  86400 
 =  Bq   dis × kg   MeV   d  Gy 
D α /β  
 J / kg   d 
1 Gy  , (74)
 
 Gy 
= 1.382 ×10-8 × C t × SEE  
 d 
where
Ct is the radioisotope activity located in the target tissue (Bq)
SEE is the specific effective activity of the radioisotope contained with mass mt
(MeV/dis-kg).
Radioactivity is removed by the body through two mechanisms: (i) physical decay of the radio-
nuclide, and (ii) biological elimination of the radionuclide. To take into account both processes,
the concept of effective half-life is introduced. The effective half-life describes the residence
time of a radionuclide in a tissue or an organism and considers both radiological (physical)
decay (TR) and biological elimination (TB). If biological elimination follows first-order kinetics,
then it can be described by a loss-rate constant, B, which is related to half-life by the relation
B = (ln 2/ TB). The biological rate constant can be summed with the radiological rate constant,
R, to yield the effective elimination constant, E = B + R, which is related to half-life by E = (ln
2/ TE).
Biological (TB) and radiological (TR) half-lives together eliminate a radionuclide from the body
faster than either one alone. Combined, they become the effective half-life given by Eq. (75):
TB  TR
TE  . (75)
TB  TR
Effective half-lives are generally dependent upon the chemical form of the internal emitter.
Note that, although radiological half-lives are well-known quantities, biological half-lives are
specific to individuals, and therefore tabulated biological half-lives are approximations for some
reference person that may or may not represent any particular individual. With that caveat,
some representative examples of radiological, biological, and effective half-lives are given in
Table 17.
Table 17 Effective half-life

ha lf - lif e
Ra dioisot ope
r a diologic a l biologic a l e f f e c t ive
H- 3 ( t r it ium) 12 y 12 da ys 12 da ys
C- 14 5560 ye a r s 10 da ys 10 da ys
P - 32 14 da ys 257 da ys 14 da ys
S - 35 87 da ys 90 da ys 44 da ys
Co- 60 5 ye a r s 10 da ys 9.5 da ys
S r 90 28 ye a r s 50 ye a r s 18 ye a r s
I- 131 8 da ys 138 da ys 7.6 da ys
P o- 210 138 da ys 60 da ys 42 da ys
Ra - 226 1620 ye a r s 45 ye a r s 44 ye a r s
Using Eq. (74) as a basis, the estimated dose rate at some time t after intake of a radioisotope is
given by Eq. (76):
0.693
 =D e  e
- ×t
-λ e t TE
D t α/β =D α/β . (76)
Finally, the dose may be determined by integrating over time as in Eq. (77):
t
 e-λE t dt
D=D  0
 , (77)
1- e 
D α/β -λ E t
=
λE
where D is the dose (Gy), D /  is the instantaneous dose rate at time=0 (Gy/d), and E is in
units of d-1 (any time base will work as long as the dose rate and the effective elimination
constant have the same base unit). For long times (on the order of seven half-lives or more),
the dose approximation becomes Eq. (78):

D α/β
D= (Gy) . (78)
λE
For regulatory purposes, t = 50 years is normally used for a committed dose, and the absorbed
dose (or dose rate) can be converted to equivalent and effective quantities using the appropri-
ate weighting factors.
6.1.2 Gamma internal emitters

In the previous section, expressions were developed that are appropriate for short-range
particles such as alpha and beta particles. Long-range neutral particles such as gamma rays
require a different approach because they do not deposit all their energy locally. In other
words,, the absorbed dose cannot be calculated by assuming the tissue or organ to be infinitely
large because gammas are highly penetrating (unlike alphas or betas). Therefore,
Therefore only a fraction
of the energy carried by photons originating in the tissue containing the radioisotope will be
absorbed within that tissue. Therefore, an alternative approach to that used for charged
particles is required. For this scenario, the following parameters must be considered in the
calculation of absorbed dose:
 Energy per decay
 Total activity
 Mass of tissue or organ (target)
 Fraction of energy emitted which is absorbed in the target
 absorbed fraction, φ
o φ = (Energy absorbed by target)/(energy emitted by source)
o φ = 1 for both alpha and beta particles (therefore, this approach will also work
for these particles).
The concept of absorbed fraction is illustrated in Figure 64. The absorbed fraction can be
ΔE
calculated as  = .
E in
Figure 64 Absorbed fraction in mass

The specific absorbed fraction Φ is defined as the absorbed fraction φ divided by the mass of
the target tissue, as given by Eq. (79):
φ
Φ=
m
 kg  .
-1
(79)
Specific absorbed fractions are computed using mathematical phantom calculations (Monte
Carlo simulations) and are tabulated as a functio
functionn of target organ and photon energy for
sources in various tissues [MIRD1978]
[MIRD1978]. The generic equation for calculating absorbed dose rate
using the absorbed fraction concept is given by Eq. (80):
k  A   n i E ii
 =
D i
, (80)
m
where
D is the absorbed dose rate (Rad/s or Gy/s)
A is the activity in the source organ (Ci or Bq)
ni is the yield of particles at energy E emitted per disintegration (also denoted as fi or Yi)
E is the average energy per disintegration (MeV per transform)
φi is the absorbed fraction
m is the mass of target (kg)

k is a proportionality constant to produce the required dose-rate units. For example, if
the desired dose-rate units are Gy/s, the proportionality constant is k = (1.6E-13
J/MeV)(1 dps/Bq)/(1 J/kg/Gy) = 1.6x10-13.
The total absorbed dose can be determined using a cumulative activity A which is the integral
of the activity as a function of time. The time-dependent activity As(t) and the integrated
activity A in the source are given, assuming first-order kinetics, by Eq. (81):
As  t   As  0  e  E t
A  A  t  dt  A  0  e  E t dt  As  0  ,
 
(81)
0 s s 0 E
where As(0) is the activity in the source organ at time = 0. The units of A are activity-time (for
example, Ci-h or Bq-s, etc.). The total dose is then given by:
  n E
k A  i iii
D= . (82)
m
To calculate total dose in the body, consideration must be given to the source (S) and the target
(T). The absorbed fractions are usually denoted as  T  S  , and if there are multiple sources
or targets, this approach can be extended to sum all contributions. For example, for multiple
sources (S=1,2,…) contributing to a single target (T=1) of interest, the total-dose equation
becomes:
kA1  ni Eii 1  1 kA2  ni Eii 1  2 

D1  i
 i
 . (83)
m1 m1
An alternative method of computation, given by [MIRD1978] for the calculation of internal
dose, is given by the expressions in Eq. (84):
D i  s i  i
A
m
A
D   i  i
m . (84)
where

i  1.6  10 13  ni  E i
Note that consistent base units must be used in the evaluations to achieve the dose or dose-
rate units required. Also, as previously discussed, for alpha or beta particles, the absorbed
fractions will be equal to unity (φ=1).
6.2 Internal Hazards

A variety of internal hazards are associated with nuclear power plants and nuclear facilities in
general, and the dominant pathway to committed dose is through inhalation. The primary
hazards with respect to CANDU nuclear power stations under normal operating conditions are
tritium, carbon-14, radio-iodines, and short- or long-lived particulates, which are discussed
below.
6.2.1 Tritium
Tritium is an isotope of hydrogen with one proton and two neutrons. Tritium may be denoted
as 3H, H-3, or T for short. A tritium nucleus decays by emitting a single beta particle with a half-
life of 12.3 years. The maximum beta energy is 18 keV (average beta energy ~ 5.6 keV). Tritium
can contribute a large portion of the committed dose in a CANDU reactor and has been respon-
sible for 30% to 40% of the radiation dose received by nuclear station staff in the past [Burn-
ham1992]. Tritium, when taken into the body, distributes itself homogeneously throughout the
entire body.
Because a beta particle requires at least 70 keV to be able to penetrate the dead surface layer
of the skin, tritium is not a significant external hazard; however, it can be a significant internal
hazard in large quantities. Tritium is produced through neutron absorption in deuterium (
1
n  2 H  3H   ), which is abundant in the moderator and the primary heat-transport (PHT)
system of a CANDU reactor core. When neutrons irradiate heavy water (D2O), some of the
deuterium atoms (D) in the D2O absorb neutrons to become tritium atoms (T) and generate
TDO (tritiated heavy water). T2O will also be produced, but in insignificant amounts.
In the case of heavy water that spends many years in the reactor core (for example, moderator
water), after a few years, the tritium content of the D2O can be on the order of several TBq 3H
per kg D2O. For heavy water that spends most of its time outside the reactor core, for example,
the primary heat-transport system water, after a few years, the tritium content of the D2O can
be on the order of a few hundred GBq 3H per kg D2O. A moderator leak can produce significant
committed dose rates if unprotected. For example, for 2 TBq/kg of moderator water that leaks
into the vault, an upper estimate of the committed dose rate is approximately 3 Sv/h, or 50
mSv/m [Burnham1992]. Workers around the moderator system must therefore be aware of the
enormous tritium risk that can exist there if D2O leaks are present, and therefore evaluation of
the concentration of tritium in air is required before entry. People working with moderator
water or performing maintenance on moderator D2O systems must wear water-resistant
(plastic) suits and breathe supplied air.
Because the PHT system operates at high temperature and pressure (300°C / 10 MPa), leaks are
inevitable in a CANDU reactor. Each fuelling machine also spills about 1 L of D2O during each
channel visit, the result being that the atmosphere of the fuelling-machine vaults and the boiler
room can be expected to have levels of tritiated water vapour up to 50 Sv/h under normal
conditions [Burnham1992]. Design improvements to the fuelling-machine (FM) snout assembly
to catch D2O that escapes during re-fuelling operations have been made to reduce tritium-in-air
levels in the FM vaults and boiler room [Aydogdu2013]. Any spill or leak of water in any of the
CANDU reactor water-handling systems should be assumed to be radioactive, and workers
must be aware that high localized levels of exposure may be possible.
Due to the relatively long half-life of tritium, it should be noted that the risk is present whether
or not the reactor is on power and remains for a long time after shutdown.
6.2.2 Radio-iodines
Several radioactive isotopes of iodine (I-131, -132, -133, -134, -135) are produced in the fuel as
fission products. These are volatile and can easily escape from defective fuel. All radioiodines
decay by beta emission with associated gamma emission. Radioiodine is routinely expected to
be circulating in the PHT system, and therefore any leakage from the PHT system gives rise to
an airborne radioiodine hazard. Radioiodines, when taken into the body, migrate rapidly to, and
are taken up by, the thyroid.
Radio-iodines can exist as elemental iodine (I2 molecules), as organic iodine (frequently as
methyl iodide (CH3I) from ion-exchange resins in the PHT purification circuits), and as volatile
hypoiodous acid (HOI). If a leak of PHT steam occurs, the radioiodines enter the air and are
picked up by dust particles to exist in particulate form [Burnham1992].
Both tritium and radio-iodines are produced when the reactor operates, and production of both
stops when the reactor shuts down. However the radio-iodines, with half-lives on the order of
hours to days, decay away over a period of weeks, whereas the tritium does not. Therefore, the
radio-iodines are a transient hazard, whereas the tritium is a persistent hazard in a CANDU
reactor. There are gaseous fission-product and delayed-neutron monitoring systems in CANDU
reactors to detect and locate failed fuel promptly. The FM can then remove the failed fuel from
the channel to reduce radio-iodine levels in the coolant. The heavy-water purification system
also removes radio-iodines from the coolant. Note that the fuel-failure rate in CANDU reactors
is low (<< 0.1%), and therefore, out of approximately 5000 bundles discharged from the core
annually, only a few bundles have a single fuel-element defect [Aydogdu2013].
6.2.3 Particulates
Radioactive particulates can be present in CANDU reactors as (i) fission products, (ii) activation
products, or (iii) other radioactive volatiles attached to dust particles. Radiation hazard from
particulates arises when airborne particulates are inhaled or when “hot” particles come into
close proximity to or contact with tissue. There are two general categories of particulates:
short-lived and long-lived.
Short-lived particulates, principally fission-product noble gas progeny 88Rb and 138Cs, are not a
significant internal hazard due to their short half-lives (18 and 32 min respectively [Burn-
ham1992]).
Long-lived particulates such as 141,144Ce, 140Ba, 140La, 134,137Cs, 131I, 103,106Ru, 95Zr/95Nb, 90Sr, 65Zn,
60
Co, 59Fe, 54Mn, 51Cr, and 14C are all beta emitters, and most also emit gamma radiation. These
particulates may be found at various locations in a CANDU reactor, essentially anywhere that a
leak can occur or maintenance is being done.
6.2.4 Carbon-14
Carbon-14 decays through beta emission with a maximum energy of 156 keV (average ~50 keV)
and has a 5730-year half-life. Carbon-14 is generated in CANDU reactors by three processes:
1
n  17O  14C   moderator and PHT water
1
n  14 N  14C  p air
1
n  13C  14C   .
About 1985, during a pressure-tube change outage at the Pickering NGS, 14C was found to be
widely distributed within the plant. Because of the large volume of moderator (and therefore
relatively large quantity of 17O), generation of 14C in CANDU reactors is about an order of
magnitude greater than in other pressurized-water reactors. 14C was also generated in large
quantities from 14N in the N2-filled annulus gas spaces of the pre-1985 Pickering A reactors and
is also formed in CANDU reactors whenever there is air in-leakage into the core (for example,
from the flux detectors). Carbon-14 was primarily a problem in the Pickering A Unit 1 & 2
CANDU station due to the annulus-filling gas and the lack of awareness of the extent of the
condition. 14C particulates were released when pressure tubes were replaced at Pickering A in
the 1980s and is still present as residuals in the annulus space and in waste from the pressure-
tube refurbishment. Although 14C is always present in CANDU reactors, it is less of a hazard in
other CANDU reactors than in Pickering A. Since the annulus gas was changed from N2 to CO2 in
the Pickering A reactors, 14C formation in the annulus gas has been reduced significantly. The
production of 14C in the calandria vault air applies to the Pickering A reactors because the vault
atmosphere is air. The calandria is inside a water-filled tank in Bruce A, B, and Darlington
reactors and inside a water-filled concrete vault in CANDU 6 and Pickering B reactors [Ay-
dogdu2013]. Note, however, that due to its long half-life, 14C can be a persistent problem for
operation, shutdown, and subsequent decommissioning and disposal activities.
6.3 Internal Protection

The two primary techniques to prevent internal radiation from entering the body are (1) to
block portals of entry into the body (control of the worker), and (2) to interrupt the transmis-
sion of radioactivity from the source to the worker (control of the source). Both techniques are
standard industrial hygiene practice and are discussed below.
6.3.1 Control of the source

In an industrial setting, the airborne hazard is the most significant. Therefore, it is important to
understand and characterize aerosols. Airborne particles are often classified by size and manner
of production. Fumes are airborne solid particles formed by vapour condensation with a diame-
ter range of 0.001–1.0 μm. Mists are suspended liquid droplets formed by condensation and
can be of any diameter (generally in a range of 0.01–1.0 μm). Dusts are solid particles formed
by mechanical action, with a diameter range of 0.1–30.0 μm. Smoke is a product of combustion
with a diameter range from 0.1 to 1.0 μm. Inertials are large particles with diameters > 50 μm.
In CANDU nuclear plants, particulates are made up of fission products and activation products.
Gases are made up of noble fission gases which escape from fuel defects, such as 133Xe, 135Xe,
and 85Kr, as well as activation products such as 41Ar. However, these noble gases are more of an
external than an internal hazard. Vapours can be present in the form of radio-iodines and
tritium and are readily absorbed by the human body.
The primary way to control radioactive emissions from the source is by ventilation. There are
two ways to provide ventilation: (1) general exhaust ventilation (GEV), which is the removal of
contaminants by movement of the entire air mass into, around, and out of the workplace
[ANSI2007], and (2) local exhaust ventilation (LEV) by systems that remove air at the point
where the hazard is generated [ANSI2006,2003] [CSA2004]. The selection of a given ventilation
solution depends upon multiple factors. General guidance principles are the following: GEV for
non-toxic contaminants, multiple sources, widely distributed sources, and non-contaminated
dilution air; LEV for moderately to highly toxic contaminants, one or a few sources, and when
there is risk of direct worker exposure. Nuclear facilities use combinations of GEV and LEV with
highly monitored and filtered air discharge systems.
6.3.2 Control of the worker

The main method for protecting the worker is the proper use of personal protective equipment
(PPE), and for internal dose mitigation, the main protective equipment is respiratory protection
appropriately used, although other protective clothing such as suits, gloves, and eye protection
will help prevent skin absorption and contamination transfer.
In areas where a worker may be in contact with (or submerged in) radioactive material, plastic
or water-resistant protective outer garments are suitable protection from beta emitters (alphas
are also blocked). Gamma and neutron radiation will not be attenuated greatly by these gar-
ments, hence the need for dose control and shielding (ALARA) for these hazards. A worker
performing a radiological survey around contaminated components is shown in Figure 65. The
worker is shown wearing a TyvekTM polyethylene suite with hood, double latex gloves, safety
glasses, and a half-face respirator (P100 toxic particulate filter).
Respirators fall into three broad categories: (i) escape, (ii) filtering, and (iii) supplied air. Escape
respirators are for emergency purposes and are designed for one use or for one year from
placement. Filtering and supplied-air respirators are fit-for-purpose assigned and depend on
the nature of the contaminant, the nature of the work, and the level of protection required.
The expression for respirator performance is defined by the ratio of the contaminant concen-
tration outside the mask (Cout) to the concentration inside the mask (Cin), which is called the
protection factor (PF) and is given by Eq. (85):
Cout
PF = . (85)
Cin
Protection factors are estimated by the manufacturer; however, an accurate protection factor
can be determined only after performing a fit test on the respirator [ANSI2001]. Guidance for
proper selection, use, and care of respirators is provided by the Canadian Standards Association
[CSA2012].
Figure 65 Worker in PPE

Respirators can also be categorized as (i) air-purifying (for particulates only) and (ii) atmos-
phere-supplying (for particulates, gases, and vapours), which can be further subdivided into (ii-
a) air-line respirators and (ii-b) self-contained breathing apparatus (SCBA). Combination-type
respirators also exist which incorporate both air purification and atmosphere supply. Respira-
tors are also classified by modes of regulator operation as follows: negative pressure (NP),
positive pressure (PP), continuous flow (CF), demand (D), pressure demand (PD), and recirculat-
ing pressure demand (RP). The protection factors afforded can range from approximately 5–
1000 for air-purifying respirators and 5–10,000+ for atmosphere-supplying respirators, with the
largest protection factors being for self-contained breathing apparatus. Although manufactur-
ers state nominal protection factors for respirators, the actual PF for any usage is determined
through an individual fit test and check. For internal dosimetry calculations, manufacturer-listed
protection factors (assuming that the respirator has been fitted and used properly by the
worker) are often sufficient. Air-purifying respirator cartridges are colour-coded by contami-
nant type and designated by usage type and efficiency, as shown in Table 18.
Table 18 Air-purifying cartridge designations
Contaminant type Use Efficiency (%)
N – Solid and water-based, particu- 99.97

White – acid gases
lates only (Not oil resistant) (denoted as 100)
R – Any particulate; one shift only

Black – organic vapours 99
for oily particulates (oil Resistant)
Green – ammonia gas P – Any particulate (oil Proof) 95
Yellow – acid gases and

organic vapours
Purple – toxic particulates
For example, the worker depicted in Figure 65 is wearing P100 toxic-particulate cartridges on a
half-face air-purifying respirator. The P100 is the most common filter type for work in low-level-
contaminated particulate environments. In many environments found in CANDU plants, the
contamination is in the form of contaminated water, aerosols, or vapour, and therefore atmos-
phere-supplying respirators would be required. Air-line supply may be preferable when work
will take longer than an SCBA tank full of air will allow, or when complete body coverage is
required (such as might be the case in a tritiated water-vapour environment). SCBA may be
preferable when increased mobility of the worker is required or when a higher inhalation
protection factor is needed. As previously stated, the choice of respirator solution must fit the
objectives of the task.
In addition to personal protective gear, another strategy for protection from radio-iodine is to
use a thyroid blocker in the form of stable iodine (potassium iodide (KI)). The mechanism of KI is
to saturate the thyroid with non-radioactive iodine so that when radioiodine is inhaled, it will
be rejected by the thyroid and excreted, primarily through urine. The use of KI is not recom-
mended for routine operations and is therefore only a strategy for emergency response.
Charcoal-based cartridges for air-purifying respirators provide radioiodine protection in the
absence of atmosphere-supplying respirators.
6.4 Summary
A variety of internal radioisotope hazards are associated with CANDU reactors. Dosimetry
methods are well established, either using dose conversion factors on activity estimates (de-
termined from bioassay measurements) or from first principles and compartmental models.
Internal hazards in CANDU reactors involve alpha, beta, and gamma emitters and include 3H
(the principal internal hazard), 14C, radio-iodines, and particulates that can be present from
leaks of fission products (defective fuel) and activation products (directly or through corrosion).
Protection of the worker from internal hazards is accomplished by ventilation systems with
filter trains, heavy-water vapour-recovery systems, atmospheric separation and access control
features, control of sources of contamination, and judicious selection of personal protective
equipment. Respiratory protection can afford a worker a significant protection factor when
used appropriately.
7 Radiation Safety Management

Management of radiation safety programs and actions is essential to the overall effectiveness
of CANDU plant operations. Management includes reporting doses and keeping records of dose
history for workers, generation of work plans for radiation areas, assignment of personal
protective equipment, contamination control principles, waste management, and emergency
preparedness. General principles of radiation-protection management can be found in
[Miller1992] and details specific to CANDU operations in [Burnham1992].
7.1 Dose Records

Dose records are required by CNSC regulations [NSCA2000] and are maintained by the National
Dose Registry, Radiation Protection Bureau, Health Canada. Routine dose monitoring is per-
formed for nuclear energy workers (NEW) as defined under the CNSC Radiation Protection
Regulations [NSCA2000], and the dose limits for workers are as defined in Table 12. Nuclear
energy workers (NEW) are personnel who have been trained to work around radiation and
radioactive materials, who have been educated about the risks (including fetal risks) involved in
working around radiation and radioactive materials, who are monitored for dose if they are
reasonably expected to be exposed to more than 5 mSv in a one-year dosimetry period and
who have been informed in writing that they are designated as nuclear energy workers. Porta-
bility of dose records is ensured by coding the records to individual workers’ social insurance
numbers (SIN). Dose monitoring has a variety of objectives [Miller1992], as listed below:
 Compliance with regulatory limits
 Optimization of radiation protection program
 Assessing adequacy of protective controls
 Detecting changes in work practices
 Transparency in hazard assessment to management and workers.
Dose monitoring is required whenever radiation work is to be performed which can result in
significant exposure or contamination. Different organizations can determine what exactly
constitutes radiation work. The following are representative of radiation work [Burnham1992]:
 Work with radioactive sources and materials
 Work in external fields greater than 10 μSv/h
 Work in airborne concentrations greater than 10 μSv/h committed effective dose
 Entry into rubber areas (areas where loose contamination is expected and a temporary
surface is used in the area with a clean-dirty boundary)
 Any activity in which one expects to receive a dose greater than 0.2 mSv.
Monitoring periods (time between reading dosimeters or bioassays) for CANDU stations tend to
be much shorter than for other users (for example, university or medical). A typical monitoring
period for a CANDU station may be two weeks or shorter, whereas other users typically submit
dosimetry records every three months. Dosimetry estimates may be performed daily depending
on the magnitude of the anticipated hazard of the radiation work being performed.
The detection of radiation fields or potential contamination is facilitated by air monitors, area
monitors, portal monitors, and hand-held instrument readings. The principal personal expo-
sures in CANDU stations being monitored for are:
 External dose – thermo-luminescent dosimetry (TLD)
o Whole-body gamma (deep and shallow)
o Whole-body neutron (could be from instrument reading instead)
 Extremity dose – extremity pack or finger (ring) TLD
o typically for near-field and contact beta/gamma exposures
 Internal dose - bioassay
o Tritium
o Other (14C).
The results of dosimetry estimates at any given time can be used to optimize worker exposure
to ALARA levels, or if needed, to remove a worker from further exposure. Yearly reports are
provided to NEWs which show their historical dose of record, dose received every year during
employment, quarterly dose in the past year, and lifetime occupational dose. These data are
maintained locally and are available through the National Dose Registry [NDR2013] upon
request.
7.2 Radiation Work Planning

Germane to the effectiveness of work planning is that the persons developing and executing
the plans be properly trained. Personnel working in CANDU stations require training to various
levels depending on the jobs they will be required to perform. In CANDU operations, workers
are categorized by colour scheme as described in the following subsection.
7.2.1 Radiation-worker colour scheme

 Green-qualified workers are personnel with sufficient experience and formal training to give
them the qualifications to manage themselves and to supervise the radioactive protection
of other workers.
 Yellow-qualified workers are trained to manage their own radiation-protection require-
ments, but not the protection of others.
 Orange-qualified workers have minimal training and radiation work experience, but
may perform radioactive work under the direct or indirect supervision of a Green-
qualified person. They have sufficient qualifications to go into the various areas of
the plant that are not posted with radiation hazard signage.
 Red-“qualified” persons have no formal radiation work qualifications; however, they can
perform work in radiation areas only under the direct supervision of a Green-qualified per-
son. The work has hazard limits that must be adhered to.
Whereas Green- and Yellow-qualified workers are typically designated as nuclear energy
workers (NEWs), Orange- and Red-qualified workers typically are not designated as NEWs.
7.2.2 Zoning
In addition to categorization of workers, CANDU plants are also categorized by location and
potential for radioactive contamination. Although any given CANDU station may have its own
zone categories, a typical scheme is presented below [Burnham1992].
 Unzoned – areas with absolutely no contamination, generally not within the boundary of a
reactor, but within the station boundary (for example, a switchyard).
 Zone 1 – clean areas where absolutely no contamination is permitted; no radiation training
is required to be in these areas (for example, administrative offices).
 Zone 2 – areas that contain no radioactive systems and are normally free of contamination
(although the potential exists for contamination due to ventilation and personnel traffic (for
example, a turbine hall).
 Zone 3 – areas that contain radioactive systems that may act as sources of contamination
(for example, a reactor building).
Contamination monitoring stations must be used before crossing the boundary from Zone 3 to
Zone 2 or from Zone 2 to Zone 1. A two-zone system is to be introduced in new CANDU reac-
tors: a radiation-controlled area (RCA) and non-RCA spaces [Aydogdu2013]. Zone 3 becomes
RCA, and Zones 1 and 2 become non-RCA areas. Work plans must take zoning into account
because this impacts the level of qualification required for workers and the PPE or dosimetry
that will be assigned.
7.2.3 Work planning

Planning for work in a radiation area goes far beyond categorization of workers. Work planning
involves identifying hazards and planning work while acknowledging these hazards. Work
planning specifications can be subdivided into approximately 11 steps [Burnham1992]:
1. Identify hazard(s)
2. Control hazard(s), which can be further subdivided into
a. Elimination of hazard
b. Minimization of hazard
c. Install physical barriers
d. Use warning devices
e. Minimize human-error potential
f. Establish procedures commensurate with the hazards
g. Train and supervise personnel
h. Accept the hazard and work with it
3. Simplify tasks
4. Verify instruments and assign as appropriate
5. Post signage appropriate to the hazard
6. Identify and issue dosimetry
7. Identify and issue appropriate PPE
8. Identify and assign personnel appropriate to task(s)
9. Perform pre-task briefings and clarify assignments
10. Monitor progress during work
11. Follow up after work.
A radiation-exposure permit (REP) is generated by a qualified person (a responsible health
physicist, for example) consistent with the hazard and adhering to the “as low as reasonably
achievable” (ALARA) tenet. The work plan takes into account all personnel qualifications and
assigns responsibilities accordingly.
7.3 Personal Protective Equipment

 Personal protective equipment is needed for activities in Zones 2 and 3 both from an
industrial hygiene and occupational health and safety perspective and a radiation-
protection perspective. Normal industrial protective equipment may be required for
work in Zone 1 and unzoned areas as well. In Section 6.3, control of the source (i.e.,
ventilation) and control of the worker (i.e., respirators) were discussed for airborne
hazards. Additional routine personal protective equipment includes active work-
zone clothing (in CANDU operations, historically called “browns” because the outer
garment was a brown-coloured coverall). In fact, personnel working in active areas
have a complete change of clothing consisting of coveralls, undergarments, socks,
and safety shoes. The important provision is that the active-area clothing remains in
the potentially contaminated areas and no personal clothing is allowed in Zones 2 or
3. Besides appropriate respiratory protection commensurate with the task (from
cartridge respirator to plastic suit with air-line supply), a worker may also require a
hard hat, safety glasses, hearing protection, lab coat, disposable coveralls, appropri-
ate gloves, safety shoes, booties, or any combination of the above.
7.4 Contamination Control

The most effective way to control contamination is at the source, specifically by preventing it
from occurring in the first place. However, there are inevitable instances where radioactive
contamination will be generated in a CANDU reactor, for example from fuelling-machine visits,
primary heat-transport (PHT) system leaks, and shutdown maintenance. Contamination can be
present in solid, liquid, and aerosol forms, and may be found near the source or at some
distance from the source. Controls are established in CANDU plants in the form of permanent
control zones with defined access-control points (Section 7.2.2), and temporary control zones
for ad hoc decontamination activities (sometimes called “rubber areas”).
Personal contamination protection is achieved by proper identification and use of protective
clothing (described in Section 7.3) and proper respiratory protection (described in Section 6.3).
Spread of contamination is minimized by using proper work planning techniques (described in
Section 7.2).
If personnel or equipment become contaminated (as determined by radiation-detection
equipment), then decontamination must be performed. Decontamination of equipment can be
performed using techniques similar to normal cleaning (washing and scrubbing), although more
attention must be given to the waste stream generated (collection and disposal of cleanup
waste).
Radiation-detection instrumentation is essential for identifying contamination and determining
the efficacy of decontamination. Three distinct limits must be considered with respect to
contamination monitoring and decontamination activities: (i) detection limit, (ii) action level,
and (iii) regulatory limit. The detection limit is a function of the instrument used to detect
radioactive material. The action (or reference) level is established by the licensee (station
operator) and is typically some fraction of the regulatory limit. The purpose of this level is to
identify potential problems through a comprehensive quality-assurance program and rectify
them before contamination approaches the regulatory limit. The regulatory limit is set by a
governing body (in Canada, the CNSC) and is a limit which should not be exceeded in normal
operations. The relative relationships, with four different measurement cases, can be seen in
Figure 66.
regulatory limit
action level
Level
detection limit
background
1 2 3 4 5
Figure 66 Relationship between background, action level, detection limit, and regulatory limit
In Case 1, the measurement lies at the background level and therefore is considered back-
ground. Note that a measurement cannot, by definition, be less than background, although the
background level may vary from one place to another. In Case 2, the measurement lies be-
tween the background level and the detection limit of the instrument. In this case, with a given
confidence, the measurement is below the detection limit of the contamination-survey instru-
ment, and generally no further action is taken. In Case 3, the measurement lies above the
detection limit, and radiation has been positively detected with a selected confidence level.
This measurement is useful for identifying the presence of radioactive material and may assist
in decontamination. In Case 4, the reading has exceeded an action level (as put in place by the
station operator), and generally decontamination may be required to mitigate the hazard
(although it is not strictly required by the regulations). The location of the detection limit and
the action level may be exchanged. In Case 5, the measurement lies above the regulatory limit.
In this case, decontamination procedures must be carried out to bring the levels of measured
radioactivity to levels below the regulatory limit and as low as reasonably achievable (typically
below the action level).
7.4.1.1 Waste management

Radioactive waste is generated in all CANDU plants during routine operations. Waste can
originate from contamination sources (Section 7.4) and may include cleanup and laboratory
wastes (paper, rags, glassware, etc.), contaminated tools or components, filters, discarded
components that have been activated, and ion-exchange resins from process systems. These
types of waste are typically called stored wastes. A second source often categorized as waste is
spent (or used) fuel. This material is not really waste as such because there is a strong possibil-
ity of reuse (or reprocessing) of the fuel in the future. The International Atomic Energy Agency’s
philosophy for long-term storage is that future generations should be afforded the same degree
of protection as the current population. Spent nuclear fuel will not be considered further in this
chapter.
There are two general approaches to radioactive waste management [IAEA1995]:
1. Dilute and disperse
a. Aerosol: discharge as gases or other fine particulates
b. Liquid: discharge into marine or fresh-water environments
2. Concentrate and contain
a. Solid: storage (delay and decay).
Dilution and dispersion occur routinely under the CNSC licence conditions for the CANDU
operator. Solid-waste management may be considered in three practical categories for CANDU
operations: (i) sources (which may include activated materials or “hot” particles, (ii) loose
contamination, and (iii) tritium waste, as discussed in the next subsection.
7.4.1.2 Sources
Sources, including fixed surface contamination, will normally be collected by removing the
entire item to solid-waste storage. Efforts to remove sources of radioactive material from the
equipment or item to which they belong (such as removing a hot particle from a section of
pipe) may not be undertaken depending on the hazard analysis. Measures to ensure the
integrity of the source or fixed contamination must be implemented to prevent the spread of
contamination while the material is being moved.
7.4.1.3 Loose contamination

For loose contamination, it is first essential to establish contamination-control measures. An
appropriate method for removing the contamination will be used. Typically this will include
brushing, sweeping, wiping, cleaning, or some combination of these (as discussed in Section
7.4)
In some instances, decontamination may not be practical (e.g., in contaminated piping). Dis-
posal of the entire item (reduced to the smallest practical volume) will then be undertaken.
Decontamination efforts must continue until measurement readings are below the detection
limit of the instruments, below the action level, and below the regulatory limit. Upon comple-
tion of decontamination activities, a thorough survey of the decontaminated location is per-
formed.
7.4.1.4 Special case: tritium contamination

Tritium is a major component of the contamination associated with CANDU plants. Tritium
contamination is characterized by the high mobility of tritium, resulting in ready cross-
contamination, absorption, and permeation in many materials. Tritium outgassed from surfaces
is primarily in the form of tritiated water. The properties of tritium make tritium decontamina-
tion more complex than decontamination of other radioisotopes. Surface tritium will generally
be found as:
 Removable tritium – tritium near the surface that is easily removed by light washing.
 Fixed tritium – located deep within the surface of the material and not amenable to surface
decontamination techniques; and
 Transferable tritium – tritium that is near the surface and acts as a reservoir of tritium to
replenish the removable tritium if that layer is significantly diluted.
Accordingly, decontamination only of the removable layer will likely result in reappearance of
the tritium after a short time (less than one month) as removable tritium. Repetitive decon-
tamination of the surface layer will eventually reduce the inventory and hence the eventual
level of removable tritium on a surface.
The most commonly used methods for tritium decontamination are washing, vacuuming,
purging, thermal desorption, and isotopic exchange by adding water or hydrogen gas in the
purge gas. These are generally adequate, but are time-consuming. More sophisticated tech-
niques such as chemical or electrochemical etching and plasma can potentially provide high
decontamination factors. Baking of the tritium in a contaminated surface is an alternate
method for decontamination, but is not normally used.
Tritium contamination of walls and floors, particularly concrete, poses a chronic outgassing
problem. Concrete decontamination is very difficult and slow. Covering the surfaces with non-
porous covers may be an alternative approach to decontamination. Walls inside a CANDU
reactor building are painted with reinforced epoxy-resin paint for ease of decontamination and
to reduce tritium contamination.
7.5 Emergency Planning

Besides routine planning functions, health physics/radiation protection also has duties related
to emergency planning. Emergencies are classified by the IAEA on the INES Nuclear and Radio-
logical Event Scale [IAEA2009], which outlines methods for assessing releases, estimating doses,
and relating these to the INES scale. Emergency planning for different categories of events is
critical for protection of workers, the public, and the environment.
Nuclear emergency plans consist of on-site emergency contingency planning as well as off-site
planning. In any type of emergency, the top-level management has the ultimate responsibility
for emergency response. From a practical perspective, the duty shift supervisor at a CANDU
reactor will be responsible for the initial response to an emergency, whether or not it involves
radioactive material (for example, emergencies may involve any combination of radioactive
material, fire, chemical, medical, and security issues). Emergency response involves activating
an emergency response team that is specially trained for a variety of emergencies. If there is an
emergency that involves, or may involve, the release of radioactive material from the confines
of the station site, an off-site emergency plan will be activated which interfaces with provincial
and federal assets. To ensure that emergency plans are relevant to the current plant conditions
and that the various assets are prepared for an emergency, periodic training is conducted and
emergency-response exercises performed.
7.6 Summary
Radiation-protection program management for CANDU operations is essential for both person-
nel safety and reactor safety functions. The health physics office will be involved in or responsi-
ble for various management aspects including licensing, training, personnel monitoring and
bioassay, work planning and review, surveys and equipment maintenance, inventory control,
waste management, emergency preparedness, quality assurance, and audits. Well-defined and
well-executed management plans at CANDU plants ensure that the stations operate in the most
safe and reliable configuration possible.
8 Radiation and the Environment

CANDU plants routinely release radioactivity to the environment. This section considers re-
leases due to routine CANDU operations and discusses the philosophy behind protection of the
environment from these releases. Some historical perspective on environmental radioactivity
can be found in [Moeller2005], and guidance on environmental radiological assessment can be
found in [Till2008] and [Faw1999].
8.1 Radiological Environmental Monitoring Program (REMP)

There are three general reasons to sample the environment for radioactivity:
1. Establish a baseline level of radionuclides;
2. Show compliance with regulations for some activity; and
3. Track radionuclides in the event of a suspected unplanned release.
There is no essential difference in the sampling strategies for any of the above. In all cases, one
is trying to determine a level of radionuclides , biological dose, or both and to define a course of
action based upon measurements.
The main differences are the priorities given to various aspects of the sampling:
 For a baseline study, the priority is to determine the levels and the statistical variations to a
high degree of confidence;
 For routine environmental sampling, the priority is to determine the lowest minimum
detectable activity levels and compare them with what the expected releases might be and
what the regulatory criteria demand; and
 For emergency response, the priority is to determine the trends and rough magnitude of
the release to provide protective-action guidance to the public.
For routine and baseline sampling, a number of biota that are in the pathway between radia-
tion source and receptor must be sampled. These biota are identified during the planning stage
of the environmental monitoring program. Likewise, in a post-accident environment, a number
of parameters require monitoring, but their importance has time dependencies because some
pathways from radioactive source to human exposure have a longer residency time than
others.
The design of an environmental monitoring program and requisite sampling strategies are
individually dependent on purposes, goals, and specific locations. However, some general
guiding principles for environmental monitoring exist and are presented in the following
sections.
8.1.1 Design of a monitoring program

The design of an effective radiological environmental monitoring program requires a thorough
understanding of the specific location and the surrounding environment, especially with regard
to environmental pathways which lead to human populations.
An environmental radiation monitoring program is normally used to:
 assess independently the actual or potential dose to an average member of a Critical Group
resulting from normal operation of a nuclear facility; and to
 provide an independent assessment of the effectiveness of the source and effluent control
and effluent monitoring programs.
A Critical Group is made up of individuals in a population which because of habits of lifestyle
and diet may be more exposed to radiation than others. Protection of members of the Critical
Group ensures that the greater majority of the exposed population will be protected. For
example, one such Critical Group is that of one-year-old infants around land-based nuclear
facilities where dairy farming is important.
The program can also serve to:
 Validate, as far as possible, the assumptions made in any assessment of derived release (or
emission) limits (DRL/DEL) for the facility; and to
 Provide assurance to the regulator and the public that human health and the environment
are being monitored and protected.
Normally an operator (licensee) is regulated to develop and implement an approved radiologi-
cal environmental monitoring program. Elements of an environmental monitoring program can
be found in an emergency response program, although their management and deployment will
be different.
The primary objective of a long-term radiological environmental monitoring program is to
define a regime of both continuous and periodic sampling and analysis to assess those radionu-
clides which are found in the various identified and sampled pathways. The accumulation of
these data over a period of years provides a baseline against which all subsequent variations
can be compared and assessed.
8.1.2 CANDU facilities

CANDU facilities have conducted individual radiological environmental monitoring programs for
more than 20 years. The rationale for sample collection and the selection of sampling media are
discussed below.
There are three significant comparable aspects in the radiological environmental monitoring
programs at Canadian utilities which have nuclear reactors. The three aspects are:
(i) source emissions monitoring and control of source emissions;
(ii) derived release (or emission) limits (DRL/DEL) specific to each facility, to define
the dose impact to humans of identified radionuclides released into specific
pathways; and
(iii) radiological environmental monitoring programs.
These are discussed in the following sections.
8.1.2.1 Source emissions

Source emissions occur into two major pathways from all Canadian facilities: air and water.
Continuous monitoring of all airborne radioactive emissions at points of release for the gaseous
effluent monitoring (GEM) program alerts station operators to radionuclide releases that may
exceed release setpoints defined by derived release (or emission) limit (DRL/DEL) calculations.
Radioactive liquid effluent releases from holding tanks are analyzed before pump-out to the
environment and are monitored during pump-out to ensure that releases do not exceed
setpoints defined (by the DEL calculations) for monthly releases in the liquid effluent monitor-
ing (LEM) program. Such in-plant source-emissions monitoring is the first line of assessment of
what is being released on a continuous or interim basis from the facility and forms the basis for
control of emissions where possible, for plant shutdown, or for notification of the relevant
authorities (for example, advising emergency measures organizations that the emissions exceed
the DELs and could lead to exceeding the regulatory dose limits).
Source emissions are controlled at CANDU stations as follows [Aydogdu2013]:
 Airborne and waterborne releases are controlled by limiting releases to 1% of the weekly
DRL (see Section 8.1.2.2) for airborne releases and 1% of the monthly DRLs for waterborne
releases (these are internal targets); see Section 8.2.
 The gaseous-effluent monitoring system consists of a gaseous-effluent monitor (GEM)
which continuously monitors and measures I-131, noble gases, and particulate emissions,
and a tritium cartridge and a 14C sampler which continuously monitor tritium and 14C emis-
sions (measurements are made in the laboratory on a daily or weekly basis).
 The reactor-building (RB) ventilation system has filter trains that contain charcoal and HEPA
filters to reduce iodine and particulate releases. The off-gas management system in a
CANDU reactor is used to reduce noble gas emissions.
 CANDU reactors have liquid-waste decontamination facilities (except Pt. Lepreau Generat-
ing Station) to decontaminate liquid waste in high-activity tanks (if necessary) before dis-
charge.
8.1.2.2 Derived release (or emission) limits

Assessment of source radiation releases is gauged against derived release (sometimes called
emission) limits (DRL/DEL) calculated for each radionuclide in each source pathway for each
facility. The DRL for any radionuclide is derived following a process defined by the Canadian
Standards Association in publication N288.1-08 [CSA2008], and guidance has also been pro-
vided by the CANDU Owners Group [COG2008]. The DRL is defined as the activity of a particular
radionuclide or a group of radionuclides (i.e., noble gases, particulates) that, if released over
the course of a year into defined environments, will cause the general public dose limit to be
reached. DRLs are calculated in five categories (tritium, I-131, 14C, noble gases, and particulates)
for airborne releases and in three categories (tritium, 14C, and gross /) for waterborne re-
leases. This assessment is the means by which human risk from radiation releases from any
defined source can be estimated.
8.1.2.3 Radiological environmental assessment

Once releases from a facility have taken place, their impact is assessed by means of a radiologi-
cal environmental monitoring program. This program is based on site-specific criteria concern-
ing release pathways leading to humans; those humans likely to be most exposed because of
specific habits (one or more critical groups); sample types in these pathways; and collection
frequency of these sample types. Usually, collection frequency and sample type are based on
analyzing those samples that are most likely to provide evidence of specific radionuclides that
would be indicative of releases from the facility into the environment. There are constraints on
detection because of dispersion (dilution) from the point of release. In most airborne pathways,
a dispersion factor of about one million is applied up to the first kilometre. In water effluents,
the dispersion factor is a function of cooling-water flow and tidal mixing and may range from a
factor of ten to thousands. In addition, radionuclides dispersed into the environment may
become concentrated into various biota, making their detection relatively simple. For example,
many marine biota concentrate various metals (including radioisotopes of those metals) from
seawater. Much of this information is summarized in Canadian Standards Association document
N288.4-10, which defines the desirable features of an environmental radiation-monitoring
program, and to which Canadian nuclear utilities generally conform.
Table 19 indicates some significant sampling media and their approximate frequency of collec-
tion in the two major pathways (air and water) by the various provinces, all of which operate
CANDU nuclear reactors. The most significant data outside plant boundaries are obtained from
samples gathered close to the point of emission in each of the two major pathways (air and
water). The air pathway leads the most directly to humans. Air is monitored by means of
continuous sampling for airborne radiation particulates, radioiodines, noble gases, tritium, and
carbon-14. Sampling is continuous, and measurements by the GEM of 131I, noble gases, and
particulate emissions are also continuous, but measurements of tritium and 14C releases are
not; tritium cartridges and 14C samples are taken to the laboratory for tritium and 14C meas-
urements on a regular basis (daily or weekly). Samples which are not amenable to continuous
collection (soils, vegetation, vegetables, fruits, milk, and potable water) are taken as grab
samples on a monthly or quarterly basis or as the season permits. The effectiveness of air-
pathway monitoring relies upon continuous and repetitive sampling of the given media. In the
liquid effluent pathway, the primary sample is the water itself. For water which is not used for
human consumption, such as seawater, grab samples may be taken quarterly from several
locations. In the case of potable lake water from which communities draw their drinking water,
for example, water from Lakes Ontario and Huron, daily samples (sometimes several samples
per day) may be taken, primarily to monitor for tritium. Less frequently, samples of raw lake
water, fish, aquatic plants and sediments may be collected, usually on a quarterly (or less
frequent) basis.
Table 19 Example of samples collected by Canadian utilities [Cole1997]
Group New Brunswick Ontario Quebec
Air pathway samples Air particulates (m) Air particulates (m) Air particulates
(m)
(continuous sampling)
Air iodine (m) Air iodine (m) Air iodine (6m)
Air tritium (m) Air tritium (m) Air tritium (m)

(grab samples)
Air 14C (m) Air 14C (m) Air 14C (m)
Soil (q) Forage (s) Vegetation (y,s)
Fruit (s) Fruit (s) Soil (y)
Vegetation (s) Vegetables (s) Vegetables (s)
Vegetables (s) Drinking water (m) Drinking water (q)
Milk (q) Milk (w)w Milk (m)
Well water (q) Well water (m) Farm produce (q,s)
Surface water (q) Greenhouse vegetables (q)
Water pathway sam- Seawater (q) Lake water (y) River water (m)
ples
Mud (q) Aquaculture fish (m) Sediment (q)
Beach sediment Water treatment plants Fish (m)

(q) (d)
Seafood (s) Sediment (y) Mollusks (m)
Lake fish (y) Aquatic plants (a)
Other Rain water (m) Rain water (m) Rain water (m)
Berries (s) Honey (s) Maple syrup (s)
Lichen (q) Farm samples (w,s) TLD (c )
TLD (c ) TLD (c )
N.B.: collection frequency is provided in parenthesis. Key: d (daily), w (weekly), m (monthly), y

(yearly), c (continuous), s (seasonal), q (quarterly).
Table 20 indicates some of the more significant radionuclides that are characteristic of potential
nuclear-reactor releases into the environment because of leakage of either fission or activation
products. Selection of key indicator radionuclides (e.g., 131I, 137Cs, 95Zr, 95Nb, 144Ce, 59Fe, 54Mn,
60
Co, etc.) for analysis is sufficient to define whether or not such leakage can be detected and
avoids the need to address unusual and difficult-to-analyze radionuclides such as 90Sr (which is
a pure beta emitter and is more difficult to identify). If the indicator nuclides are not detected,
then it can be assumed that an isotope such as 90Sr is unlikely to be present.
Table 20 Some key fission and activation radionuclides expected from a reactor release
Argon 41 Iodine 131 Krypton 89 Thorium 228
Carbon 14 Iodine 132 Manganese 54 Thorium 234
Cerium 144 Iodine 133 Nickel 63 Tungsten 187
Cesium 134 Iodine 135 Niobium 95 Xenon 131m
Cesium 137 Iron 55 Nitrogen 13 Xenon 133m
Chromium 51 Iron 59 Ruthenium 103 Xenon 133
Cobalt 58 Krypton 83m Ruthenium 106 Xenon 135m
Cobalt 60 Krypton 85m Strontium 89 Xenon 135
Hafnium 181 Krypton 85 Strontium 90 Xenon 138
Helium 3 Krypton 87 Tantalum 182 Zinc 65
Hydrogen 3 Krypton 88 Thorium 227 Zirconium 95
Special Cases: 131I and 137Cs

Care must be taken when interpreting results for at least two of the more significant indicator
radionuclides (131I, 137Cs) because they can enter the environment from releases other than
from nuclear reactors. Iodine-131 is widely used in medical programs at larger hospitals for
thyroid diagnosis and therapy. It is readily detectable from time to time in samples taken from
the areas surrounding these hospitals, but its existence is not generally publicized. Routine
releases of iodine-131 from such medical usage, even from one patient, is generally many
thousands of times larger than typical total releases that might be associated with reactor
operations. Cesium-137 and 90Sr are associated with hardwood ash. Fallout from the major
bomb tests conducted in the 1950s and 1960s resulted in deposition of fission radionuclides in
north-eastern North American forests. These radionuclides have relatively long half-lives, are
still concentrated in hardwoods in this region, and can still be detected in tree lichens. Careless
or inadvertent disposal of wood ash can lead to the detection of 137Cs in many sample types,
including milk and soils.
8.1.3 Environmental transfer model

Routine environmental monitoring requires a comprehensive environmental-transfer model
which describes the pathways from source to critical group. The principal radionuclides of
investigation will be dependent on routine release criteria. For example, in CANDU reactors, 3H,
14
C, and noble gases are important environmental radionuclides for monitoring.
By way of comparison, the principal radionuclide releases to the environment under accidental
conditions would be, in order of decreasing importance in terms of dose:
 Radioactive particulates, in particular 137Cs, are often released in the form of cesium iodide
(CsI) following severe fuel failure and loss of containment. In very severe fuel-melt events,
other radioisotopes such as actinides may also be released. However, this is unlikely, and as
demonstrated following the Chernobyl accident, actinides would not be the most significant
dose contributor; 137Cs and 90Sr would be the most significant dose contributors;
 Radioactive volatiles (mainly radio-iodines) which, along with radioactive noble gases, are
the first radioisotopes released from the fuel at the onset of fuel failure; and
 Noble-gas radionuclides (mostly radio-isotopes of krypton and xenon), which may be
released following failure of the reactor fuel cladding, even in the absence of reactor fuel
melting.
Whether planning for routine monitoring or emergency management, the modelling analysis
for pathways is similar. See Chapter 13 for a more in-depth discussion.
8.1.3.1 Modelling
A formal pathways analysis model is required to determine the significant pathways and the
types of samples and measurements required for all locations around the source of release. In
addition, baseline meteorological data showing long-term weather patterns, especially with
respect to prevailing wind directions at different seasons, may or may not indicate that a
specific geographical direction should be emphasized for sample collection at each site.
8.1.3.2 Pathways
In pathways analysis, environmental components are analyzed to determine:
 ability to accumulate environmental contaminants;
 relative residence time of contaminants in the medium;
 importance as a food source to humans; and
 suitability for inclusion in a radiation monitoring program.
A generalized environmental transfer model is depicted in Figure 67.
Immersion
Inhalation
Vegetated Soil External

(c)
Ingestion
Atmosphere Forage and Crops
(b) (d)
Ingestion
Animal Produce
(e)
Source Dose
(a) (f)
Ingestion
Aquatic Animals
(c)
Ingestion
Aquatic Plants
Sea Water (c)
(b)
Sediment External
(c)
Immersion
Figure 67 General environmental transfer model

The labels (a) through (f) for each component in Figure 67 may be considered as “levels of
desirability” for a given sample from an environmental component. For instance, the most
desirable sample for an environmental radiation-monitoring program comes from near the
source, whereas dosimetry information at the end of the chain provides inconclusive informa-
tion about the source of the radiation. The most useful information is, in general, obtained
from components (b) through (e). Atmospheric-pathway samples are most useful if obtained
from a continuous air-sampling regime because of the temporal variability of these samples.
The most useful secondary indicator for the air pathway can be seen to be vegetated soil. For
seawater samples, seawater, aquatic animals, plants, and sediments all provide useful data.
Note that the environmental-transfer model will change depending on the location where the
sampling is taking place. For example, if the sampling locations are not near sea water (as is the
case in mainland Ontario), then the sea-water pathway may be replaced by a lake-water
pathway. The main point is that the target group must be adequately identified and modelled in
the pathways analysis.
8.1.3.3 Indicator species

It is important that food species which constitute critical pathways to humans, including terres-
trial and marine organisms, are identified. The uptake through food chains is extremely com-
plex because of the:
 Physical and chemical behaviour of radionuclides;

 Food chain being considered;
 Bio-accumulation factors in different species and foods;
 Seasonal variability; and
 Distance from the source.
The indicator species will vary depending on the individual pathways analysis performed. Some
typical indicator species include:
 Cow’s milk;
 Beef;
 Vegetables and fruits;
 Soil (vegetation);
 Marine plants; and
 Seafood.
The indicator species will typically be determined by close examination of the critical groups.
8.1.3.4 Seasonality
Time of year (for release and sampling) can have a significant impact on the way in which
radionuclides move through the environment, particularly in their use in the various food
chains. Although this feature is generally not a primary consideration when performing emer-
gency environmental analysis, it can be extremely important for establishing baseline levels of
radionuclides in the environment.
8.1.4 Practical considerations for environmental sampling

When performing sample acquisition during an emergency (such as an unplanned reactor
release), care must be taken to prevent over-exposure to radiation. Dosimetry and adherence
to dose restrictions will be monitored by the health physics department. A calibrated and
functioning field-survey instrument must be used to monitor dose rate and total dose. There
will likely be airborne hazards that are not immediately identified and that have implications for
personal protective equipment. Generally, the greatest hazards will often be non-radiological
hazards (as discussed below).
During routine environmental monitoring, radiation poses no threat to the person obtaining the
samples because the work is being undertaken at background levels of radiation. However,
there are many non-radiological threats that can pose a hazard to the sampling team. These
hazards are generally of the same magnitude of hazard that a person may be exposed to daily.
The following aspects must be considered before performing any environmental survey:
 Is the area to be sampled clear of obvious threats, such as
o Wild animals
o Loose terrain
o Electrical hazards
o Slippery surfaces?
 Has the owner of the property been contacted, and has approval to sample been obtained?
Hazards include:
o Agitated land owners
o Guard dogs
o Hunters.
 Is the sampling team prepared personally to be out-of-doors?
o Anti-allergens (EpiPens®)
o Appropriate footwear
o Appropriate clothing for time of year, terrain, and weather.
The above is not an exhaustive list, but provides examples of some of the considerations
required to ensure safe environmental sampling.
8.2 CANDU Releases

CANDU reactor operators monitor both emissions of radionuclides and quantities in environ-
mental samples. The results from these radiological environmental monitoring programs
(REMPs) can be found on utility Web sites, and summaries are available for anyone to review.
Detailed analyzes and reports are maintained by the health physics office and are made availa-
ble to the regulator (CNSC). General emissions monitored and reported by the stations are
available in these documents; a representative dataset for a CANDU is provided in Table 21
(adapted from [OPG2012]) along with the order-of-magnitude DRLs and the percentage activity
of each DRL . Note that the emissions may vary from year to year, but generally are within an
order of magnitude of those presented in Table 21. In addition, note that the emissions are
typically much less than 1% of the derived release (or emission) limit for each radionuclide or
group of radionuclides.
Environmental samples provide a basis for calculation of dose to critical-group representative
persons. The doses estimated from environmental samples from CANDU plants have, up to the
time of writing this document, never approached the CNSC public dose limit of 1 mSv/y and
have generally been approximately three orders (0.1%) of magnitude below this limit.
Table 21 Order-of-magnitude CANDU radionuclide emissions
Emission Magnitude of Magnitude of Activity % of DRL

activity (Bq) DRL
Air
HTO 1014 1017 0.10
Noble gas 1013 1017 0.01
I-131 108 1012 0.01
7 11
Particulates 10 10 0.01
14 12 15
C 10 10 0.01
Water
HTO 1014 1018 0.01
Gross β/γ 1010 1014 0.01
14
C 109 1015 0.0001
It is clear from Table 21 that tritium and noble gases are the greatest contributors to the
emitted activity from CANDU reactors. They are correspondingly the major contributors to
public dose as well (see, for example, [OPG2012]). However, as was previously discussed, the
dose attributable to these radionuclides is approximately 1 μSv/y, which is an extremely small
fraction of the average person’s dose attributable to background radiation exposure.
8.3 Protection of the Environment

The radiological environmental monitoring program is designed to [OPG2012]:
i. demonstrate that radioactive materials are properly controlled at the plant, in-
dependently of effluent monitoring results;
ii. enable estimation of annual doses to the public from operation of the nuclear
facility;
iii. provide data to be used for calculating derived release (or emission) limits and
public doses.
There has always been an underlying assumption that if adequate protection of the public is
achieved, protection of the environment and environmental-ecosystem components is de facto
afforded. For example, in the 1990 recommendations of the ICRP [ICRP1990], it was stated that:
“The Commission concerns itself with mankind’s environment only with regard to the
transfer of radionuclides through the environment, since this directly affects the radio-
logical protection of man. The Commission believes that the standards of environmental
control needed to protect man to the degree currently thought desirable will ensure
that other species are not put at risk.”
The underlying assumptions related to ICRP 60 are: (i) the environment is protected through
the protection of humankind, (ii) reproductive capacity is the relevant endpoint, and (iii) the
appropriate level of protection is to avoid endangering the existence of species or creating
ecological imbalances. In the 1990 recommendations, the ICRP has not explicitly stated that the
environment should be protected.
In the 2007 recommendations of the ICRP [ICRP2007], this concept is also addressed, and
human protection may be considered as an indicator of environmental protection. However,
the guidance suggests that it must be demonstrated directly that the environment is not
affected by radiological releases. Therefore, protection of the environment shifts the focus
from human to non-human biota.
Dose calculations require reference values to describe the anatomical and physiological charac-
teristics of an exposed individual and values for tissues and organs which define a reference
individual. A reference individual is not intended to describe an “average” individual, but serves
to create a standard and a point of reference for dose-estimation procedures. The concept of a
“reference man” is one of the cornerstones of radiological protection [ICRP2003b].
The ICRP has provided guidance on assessment of radiological impact for non-human biota
[ICRP2003a]. To this end, the concept of reference animals and plants (RAPs) was introduced to
provide a common reference for establishing environmental protection [ICRP2008]. ICRP
initiatives include defining a reference set of dosimetric models and environmental geometries
that are applied to reference animals and plants (RAP). This enables assessment of the likely
consequences for individuals, the population, or the local environment. The common approach
in ICRP 103 [ICRP2007] is delineated in Figure 68.
Practices
Environmental radionuclide concentration(s)
Reference Man Reference Animals and

Plants
Dose limits, Derived Consideration

Dose constraints Levels
Decision-Making regarding public health and environment

for the same environmental situation
Figure 68 Common radiological assessment approach for human and non-human biota
Whereas assessment of risk for human exposure involves comparison against dose limits and
constraints, non-human biota assessment involves comparison against derived consideration
levels. The concept of derived consideration levels is depicted in Figure 69.
RAISED CONCERN
> Several times the natural
background
Natural background for a

reference organism
< Natural background
LITTLE CONCERN
Figure 69 Derived consideration levels
In Figure 69, estimated values of dose to the reference organism that are less than the natural
background are of little concern, whereas levels that are several times greater than the natural
background do raise concern. The derived consideration level is a band of dose rate for a type
of RAP within which there is some chance of deleterious radiation effects.
A reference animal or plant (reference organism) is a hypothetical entity with the assumed
basic characteristics of a specific animal or plant, described to the taxonomic level of family,
with precisely defined anatomical, physiological, and life-history properties [ICRP2007]. RAPs
are used for relating exposure to dose, and dose to effect, for that type of organism. The ICRP
reference animals and plants are presented in Table 22, along with their respective wildlife
groups and environments.
Table 22 Reference animals and plants (RAP)
Organism Wildlife group Terrestrial Freshwater Marine
Deer Large terrestrial animals X
Rodent Small terrestrial animals X
Duck Aquatic birds X X
Frog Amphibians X X
Trout Fresh-water fish X X
Flatfish Marine fish X
Bee Terrestrial insects X
Crab Marine crustaceans X X
Earthworm Terrestrial annelids X X X
Pine tree Large terrestrial plants X
Wild grass Small terrestrial plants X X
Brown seaweed Seaweeds X
It is highly likely that the future of radiological environmental monitoring programs will involve
incorporating explicit consideration of RAPs to demonstrate protection of both human and non-
human biota.
8.4 Summary
CANDU reactors release routine quantities of radioactive material to the environment as
permitted under their licensed operating conditions. Radiological environmental monitoring
programs are designed to demonstrate that there is a minimal impact of CANDU nuclear
reactors as related to human exposure and that the radiological emissions are below derived
release limits established by the federal regulatory body (CNSC). Over the operating lifetime of
CANDU reactors, the releases have had minimal impact on the environment, and the dose
calculated for critical groups have been generally been orders of magnitude below the limits set
by the Canadian Nuclear Safety Commission. The future of environmental protection around
nuclear facilities in Canada may involve explicit consideration of reference animals and plants
and require demonstration that a high degree of protection is afforded to non-human biota.

Chapter 12 is related to the following chapters in the text:
Chapter 3 – Nuclear Processes and Neutron Physics
This chapter presents background data regarding particles, electromagnetic radiation,
and the interactions of ionizing radiation with matter. This background information is
germane to understanding the concepts developed in Sections 2 through 6.
Chapter 13 – Reactor Safety Design and Safety Analysis
This chapter discusses radiation hazards, risk, and risk models, which are related to Sec-
tions 2.3 and 2.4.
Chapter 16 – Regulatory Requirements and Licensing
This chapter is related moderately to Sections 7 and 8.
Chapter 19 – Interim Fuel Storage and Disposal
This chapter is related to Sections 7 and 8.
10 Problems
Note: Solutions to the problems in this chapter will require reference to other texts, data,
materials and/or Web resources.
1. Calculate the individual electron densities and the ratio of the electron densities of a 1.5-cm-
thick piece of aluminum to that of the equivalent density thickness of a piece of lead.
2. How much energy does an alpha particle require to penetrate the minimal protective epi-
dermal layer of skin (thickness ~7 mg/cm2)?
3. How much energy does a beta particle require to penetrate the minimal protective epider-
mal layer of skin (thickness ~7 mg/cm2)?
4. Some important reactions in nuclear fuel have implications with respect to nuclear non-
proliferation. Two important isotopes are Pu-239 and U-233, which are both fissionable.
a. Starting with a neutron absorption in U-238, write the decay relationship to the end-
point of Pu-239.
b. Assuming that you can have neutron absorptions in both Pu-239 and Pu-240, write
down the decay relationships to the endpoint of U-233.
5. A beam of photons is incident on a slab thickness of aluminum shield on one side and exits
on the other. Assume that the beam consists of photons at only two energies: 500 keV and
1000 keV, and that the incident intensities of the two energies are equal. Your goal in designing
the shield is that there be 25% intensity of the 500-keV beam component compared to the
1000-keV component on the exit side of the shield. What thickness of shield is required?
6. A primary gamma ray from the decay of K-40 is scattered twice: first, through an angle of 25
degrees, and then through an angle of 130 degrees.
a. What is the energy of the photon after the second scattering? Show intermediate en-
ergy.
b. What would the energy of the photon be if you reversed the order of scattering (i.e.,
first 130, then 25 degrees)? Show intermediate energy.
7. You are designing a slab shield for the outer region of a small light-water-moderated reactor.
Assume that the neutron flux is 100% thermal inthe outer reactor region. What thickness of
natural cadmium shield is required so that no more than 1% of the thermal flux penetrates the
shield? Assume σcapture = 2450 barns.
8. Assuming that the specific heat of the body is 1 cal g-1 oC-1,
a. Plot the approximate temperature rise as a function of whole-body doses ranging
from 100 mGy to 10 Gy;
b. What temperature rise in the body corresponds to an LD50 dose of 4 Gy? Comment
on this temperature rise.
c. What whole-body absorbed radiation dose corresponds to a 3 mL sip of 60oC coffee?
Show calculations for both males and females.
9. Pocket ionization chambers are built in the form of an electrical capacitor. From previous
physics courses, you know that capacitance is equal to a change in charge over a change in
voltage, as follows:
Q
C
V ,
where the capacitance, C, is in Farads (F), the voltage, V, is in volts (V) and the charge, Q, is in
coulombs (C). You have a chamber that has a sensitive cylindrical volume with a diameter of
6.35 cm and a length of 63.5 cm. The standard density of the air in the chamber is 0.001293
g/cm3. The chamber is calibrated so that its quartz fibre has full deflection when there is 100 μC
per kg of air. The capacitance of the chamber is 1.2 nF.
a. If it takes 300 volts to charge the chamber, what is the voltage in the chamber at full-
scale deflection?
b. To what exposure (in Roentgen) does full-scale deflection correspond? To what ab-
sorbed dose in air and in tissue does it correspond?
10. You have 10 GBq of each of Co-60, Cs-137, I-131, and Ir-192 (which you can assume to be
point sources of radiation).
a. Compute the exposure rates, in mGy/h, at 25 cm, 50 cm, and 100 cm away from the
sources (present your results in a table for full value: rows = isotope; columns = dis-
tance)
b. For each isotope at the 1-m distance, how long would it take to get to the public dose
limit of 1 mSv/y (assume 1 Sv = 1 Gy)?
11. A standard-size male has an injection of 150 mCi of I-131. Assuming that 90% of this injec-
tion goes to, and is retained in, the thyroid, compute the initial dose rate (Gy/h) to the thyroid
from the beta radiation.
12. You are working in a nuclear plant, and the area monitors have detected a puff release of
CO2 that has escaped from maintenance being performed on a sealed pressure tube. The only
radionuclide identified in the plume is 14C. You are wearing full SCBA (self-contained breathing
apparatus), so that you have no possibility of getting an inhalation dose. However, it is later
found that you were submerged in the plume for a total of 25 minutes. Using the assumptions
listed below, determine your dose from this exposure.
Assumptions:
14
 C activity detected on 1000-L air sample by area monitors = 55 kBq
 Air at STP
 Average material covering the worker has the standard thickness of a thick outer
glove (assume that the beta attenuation coefficient for tissue will approximate
this material).
13. You work in a plant that manufactures sources for radioisotope thermoelectric generators
(RTGs). The source is fabricated by pressing and sintering powdered strontium titanate
(90SrTiO3). A spill of the powdered material occurred in the plant after a shielded vessel which
transports the raw material from shipping and receiving to the handling area fell over and
broke open. While the person who was handling the material was trying to find a manager, one
maintenance worker (Worker A), who was not wearing personal protective gear, worked near
the spill (which was spread out evenly on the floor). You are the health physicist charged with
determining the potential exposure to the worker. You interview the worker and co-workers
and take measurements near the spill. You gather the following data to use in your analysis:
 Sr-90 contamination measured: 2 MBq cm-2 (assume uniform)
 Approximate distance of Worker A’s hands from the spill: 1 foot
 Approximate time Worker A’s hands spent in the vicinity of the spill: 12 minutes
a. Calculate the dose rate to Worker A’s hands using the above data.
During your interview, you also determine that Worker A came into brief contact with
the spill and contaminated her hands. She washed her hands immediately after coming
in contact with the spill (two hours ago). You estimate that casual hand washing is only
65% effective in removing this type of contamination. You immediately assist her to de-
contaminate her hands with some strong detergent and a scrub pad.
b. Calculate the dose to Worker A’s hands from this contact exposure.
14. A solution is being prepared using 0.835 GBq of Ba-133. The solution spread into a roughly
circular area of diameter 50 cm. What is the maximum dose equivalent rate (SI units) 50 cm
above the spill?
15. You have the following sources with associated activities:
Bq Source
5.00E+09 Co-60
7.50E+09 Ir-192
1.00E+09 Na-22
5.00E+10 Tc-99m
7.50E+10 Au-198
1.00E+11 Ra-226
a. Compute the exposure rates, in mGy/h, at 50 cm away from these sources (present
your results in a table for full value: rows = isotope; columns = distance)
b. For each isotope at the 50-cm distance, how long would it take to get to the public
dose limit of 1 mSv/y (assume 1 Sv = 1 Gy)?
16. You are asked to decommission an old radioactive-waste storage room, and inside you find
a large lead pig (sphere). To dispose of the source, you need to know both the radioisotope and
the activity. Unfortunately, there is no documentation for this source, and there are no mark-
ings on the pig. You take a dose-rate measurement at about 4 feet away from the surface of the
lead sphere and find that it is 100 μGy/h. You assume that there is a point gamma source at the
centre of the sphere, and using a ruler, you determine the diameter of the sphere to be about
20 cm. Using your portable GR-135 gamma spectrometer, you determine the isotope to be Co-
60 by identification of the 1.17 and 1.33 MeV peaks. What do you estimate the activity of this
source to be (SI units)? (Hint: do not neglect buildup).
17. You are responsible for health physics at a nuclear plant. You have just been told that there
are measurable levels of 14C in air in the form of carbon dioxide. This has previously not been a
hazard, and therefore there are no derived limits for this radioisotope. Estimate
a. the allowable intake limit (Bq), and
b. the derived air concentration (Bq/m3) for 14CO2 using the stochastic limit for nuclear
energy workers at your plant (20 mSv/annum). Note: Assume that after inhalation, the
14
CO2 is highly soluble and transfers directly into the blood.
You have been told that two workers, who did not know there was an airborne 14CO2
hazard, performed their duties in a 2-DAC environment. Worker “A” was wearing a half-
face air-purifying respirator with P100 (purple) cartridges while performing his duties for
three hours in the morning and four hours in the afternoon. A welder (Worker “B”) was
wearing an air-line, continuous-flow, half-mask respirator while performing her duties
for five total hours.
c. Estimate the committed effective dose from this hazard for both Worker A and
Worker B (assume no external hazard).
18. A standard-size male working in a nuclear plant breathes in 100 mCi of Kr-87. Assuming that
75% of this intake is retained in the lung mass, compute the initial dose rate (Gy/h) to the lungs
from the beta radiation (N.B.: only use betas with a probability of emission greater than 1%).
19. Calculate the allowable limit on intake for tritium (HTO vapour) in Bq using the stochastic
limit for nuclear energy workers of 20 mSv per annum. Assume that the source is whole-body
and the target is whole-body.
20. The DAC value for tritiated water is typically adjusted to allow for the fact that a person
working in a tritiated atmosphere will absorb half as much tritium through the skin as through
inhalation (factor of 1.5). For example, if a person inhales 100 Bq of tritium, another 50 Bq will
have been absorbed through the skin (i.e., 150 Bq total). Using the above, what is the DAC for
tritium (Bq/m3)?
21. A worker is performing his duties in a tritiated atmosphere of 1 DAC for eight hours wearing
a positive-pressure (PP) air-purifying full-facepiece respirator. What is the worker’s committed
dose from his exposure to this environment?
22. You are the health physicist at a nuclear power station. The chemistry manager has asked
you to review a purchase requisition that has been initiated for a nitrogen-16 calibration
source. The source generates N-16 through an (α,p) reaction involving curium-244 (spontane-
ous fission radioisotope) and carbon-13. The source gamma-emission strength is 2.2x106 γ/s,
and the neutron-emission strength is 2.0x105 n/s. Assume a gamma energy of 6.1 MeV and an
average neutron energy of 2.5 MeV. Data provided are:
 Neutron flux-to-dose equivalent (at 2.5 MeV);
20 n/cm2-s = 25 μSv/h.
a. Write the stoichiometric relationship for the N-16 production reaction.
b. Calculate the total gamma dose-equivalent rate at 1 foot (μSv/h), assuming 100%
emission rate from the principal gamma peak.
c. Calculate the total neutron dose-equivalent rate at 1 foot (mRem/h).
Both lead and polyethylene are available to shield the source.
d. Which shielding arrangement (from the source outward) listed below would be ex-
pected to yield the lowest overall dose rate? Explain your answer.
i. Lead only
ii. Polyethylene followed by lead
iii. Polyethylene only
iv. Lead followed by polyethylene
v. No shielding is necessary because the 12-inch air gap will sufficiently scatter or at-
tenuate the neutrons.
e. If the source is surrounded by 3.93 cm of lead, by what percentage will the gamma
dose rate at 1 foot be reduced?
23. The radioisotope 140La is a common fission product that requires shielding. For La-140, find
the:
 Physical half-life (h)
 Decay scheme, and
 Average beta energy (keV).
a. Calculate the specific gamma constant for La-140 using the four (4) most dominant
gamma lines.
24. List the (strongest transition) capture gamma-ray energy (MeV) and capture cross section
(barns) for the following isotopes: H-1, B-10, Cd-113, Gd-157. Answer the following questions:
a. Why are cadmium and gadolinium used by control and safety systems in nuclear reac-
tors? Where are they used?
b. With respect to B-10, discuss the principal thermal-neutron capture reactions and
discuss the importance of the (n,α) reaction for shielding. What other practical use does
the 10B(n,α) reaction have?
25. Derive the working rule for specific gamma constant in classical units (i.e.,   0.5 EiYi ).
i
Show all work and units (i.e., be explicit about where the factor of 0.5 comes from).
26. To account for the number of atoms in a material, the density of atoms in the material must
be deduced. This is necessary because the interaction cross section for a material depends on
the number of atoms that may potentially interact with the transiting radiation per unit path
length. The terminology “mixing cross sections” refers to the proportion of the cross section
assigned to each element in a material. For example, water has two atoms of hydrogen and one
atom of oxygen per molecule in water, and both hydrogen and oxygen have interaction cross
sections that are a function of energy and radiation type. The number density of a single-
element material is typically expressed as follows:
  Av
N
MW ,
where
N is the atom density (atoms/cm3)
ρ is the mass density (g/cm3)
Av is Avogadro’s number (6.02E23 atoms/mole)
MW is the molecular weight of the element (atoms/mole).
For compounds or mixtures, the number density of the ith element in the mixture can be
calculated using the weight fraction wi of the ith element in the mixture. The weight fraction can
be determined as:
ni Ai
wi 
MW ,
where
ni is the number of atoms of element i in the mixture;
Ai is the molecular mass of element i in the mixture (g/mole); and
MW is the molecule mass of the mixture (g/mole).
The number density of the ith constituent in the mixture is therefore given as:
  Av
N i  wi
MWi ,
where Ni is the number density (atoms/cm3) of the ith constituent in the mixture (atoms/cm3). It
is common to express the number density in terms of (atoms/b-cm), and therefore:
  Av  atom  24  cm 2    Av 1024  atom 

Ni  wi   10    wi  
MWi  cm3   barn  MWi  b  cm  .
Note that atom density is the same thing as number density. The macroscopic cross section for
the ith element is:
 1 
i  Ni   i  
 cm  .
This is related to material mixing in that the macroscopic cross section for the mixture is pro-
portional to the individual number densities as:
 1 
   Ni  i  
i  cm  .
Note that depending on the nomenclature used, μ = Σ.
In addition, many computer-based codes require, for material definition, the elemental compo-
sitions in terms of weight fraction or number (atomic) density, and therefore it is very useful to
know how to compute these (because this is how the computer-based code will use the cross
sections!).
a. Show that the weight fraction of oxygen is 0.888 and that of hydrogen is 0.112 in wa-
ter.
b. Calculate the atom fractions of oxygen and hydrogen in water.
c. Calculate the number densities (atoms/b-cm) of oxygen and hydrogen in water.
27. Using the following absorption and scatter cross sections, determine the macroscopic
absorption, scattering, and total cross sections for H2O. What is the mean free path of neutrons
in water based on your calculation?
28. Define and/or briefly explain the significance of the following:
a. ICRP
b. IAEA
c. NCRP
d. WHO
e. MARSSIM
f. MARLAP
g. RESRAD
h. Barriers between outer and inner body

i. Basis for different types of environmental monitoring programs and the concept of an
environmental baseline study
j. Three general types of environmental samples
k. Classical method to determine whether a radioactive release has occurred from a nu-
clear plant accident or a weapon detonation
l. Work disciplines that make up environmental health physics.
29. It is usual to evaluate airborne derived release limits (or derived emission limits) by first
evaluating the all-relevant-compartment values in units of (Sv a-1)/(Bq m-3) and multiplying the
sum by the airborne dispersion factor (P01, in s m-3) to obtain [X9/X0] in units of (Sv a-1)/(Bq s-1).
The DRL is then evaluated using the annual dose limit. The applicable transfer factors are given
below:
Transfer factor Description Value Units
P(i)19 Dose by inhalation 1.5E-11 Sv Bq-1
P(e)19 Dose by immersion Nil
P49, P59 Dose by ingestion 1.6E-9 Sv Bq-1
P13, P34 Air to soil to forage, crops No pathway
P39 External dose from ground Nil
P14 Air to forage 2.75E+3 m3 kg-1
Air to crops 3.75E+2 m3 kg-1
P45 Forage to milk 0.15 kg kg-1
Forage to meat 0.64 kg kg-1
P15 Cow inhalation to milk 4.2 m3 kg-1
Cow inhalation to meat 5.1 m3 kg-1
P01 Source-Atmosphere 2.0E-7 s m-3
Data are provided as follows:

DRL Data
Parameter Adult Infant

Vegetables consumed, fraction grown locally 0.3 0.1
Meat consumed, fraction locally produced 1.0 0.1
Milk consumed, fraction locally produced 1.0 1.0

All vegetables, kg per year 200 84
All meat, kg per year 70 24
All milk, kg per year 170 220
Breathing rate, m3 per year 8400 1400
Determine the DRL for 14C in the airborne pathway, assuming an infant critical group at 1 km
from the nuclear station and all produce and animals at 1 km (i.e., all data apply for 1 km).
30. A nuclear power plant located on Lake Ontario operates an environmental monitoring
program which includes annual fish sampling. At the laboratory, the fish are analyzed for the
following radionuclides:
 Gamma emitters
 Sr-90/Y-90
 Carbon-14
 Tritium.
One year, some fish samples (lake trout) are found to have the following radionuclide concen-
trations:
 Cs-137: 3.33 ± 0.19 Bq/kg (wet weight)
14
 C: 250 ± 30 Bq/kgC.
 Tritium: 241 ± 6 Bq/L (free water)
 Sr-90/Y-90: Less than minimum detectable amount.
Relevant data are provided as follows:
Dose factors (Sv/Bq)
E 10y E 15y E Adult

Tritium 2.4E-11 1.8E-11 1.8E-11
14
C 8.5E-10 5.6E-10 5.6E-10
Cs-137 1E-8 1.3E-8 1.3E-8
SR-90 5.2E-8 2.6E-8 1.6E-8
Y-90 9E-9 5.4E-9 4.2E-9
Background concentrations:
Tritium in Lake Ontario water: 1.0 Bq/L
Carbon-14 in living material: 24.0 Bq/kgC.
Cs-137 in trout from Lake Ontario: 10. Bq/kg (wet weight)
Station cooling water discharge rate: 10,000 m3/min
Dilution factor: 30
Cs-137 bioaccumulation factor for Lake Ontario: 10,000
Assume that fish are 90% water
Assume that fish are 45% carbon
Lake Ontario fish consumption by Boy Scouts: 10 kg/y

Calculate the committed dose attributable to radioactivity from the nuclear power plant to Boy
Scouts from a nearby camp eating the fish (assume that any radionuclide concentration above
background is due to releases from the nuclear power plant).
11 Further Reading
F. H. Attix, Introduction to Radiological Physics and Radiation Dosimetry. Toronto, ON: Wiley-
Interscience, 1986.
V. Baryakhtar, V. Kukhar, I. Los, V. Poyarkov, V. Kholosha, and V. Shestopalov, Comprehensive
Risk Assessment of the Consequences of the Chernobyl Accident. Science and Technology
Centre in Ukraine – Ukrainian Radiation Training Centre. Project No. 369, Kiev, Ukraine,
1998.
R. Bailey, H. Clark, J. Ferris, S. Krause, and R. Strong, Chemistry of the Environment, 2nd Edition.
New York: Academic Press, 2002.
E. Baum, H. Knox, T. Miller, Nuclides and Isotopes, 16th ed. Schenectady, NY: Lockheed Martin,
Knolls Atomic Power Laboratory, 2002.
J. J. Bevelacqua, Health Physics in the 21st Century. New York: Wiley-VCH, 2008.
A. Brodsky, Review of Radiation Risks and Uranium Toxicity with Application to Decisions
Associated with Decommissioning Clean-Up Criteria. Hebron, CT: RSA Publications, 1996.
Chernobyl Forum, Chernobyl’s Legacy: Health, Environmental, and Socio-Economic Impacts and
Recommendations to the Governments of Belarus, the Russian Federation, and Ukraine.
The Chernobyl Forum: 2003-2005, 2nd revised edition, 2005.
CIBA, Health Impacts of Large Releases of Radionuclides. Toronto: John Wiley, CIBA Foundation
Symposium 203, 1997.
J. Cooper, K. Randall, and R. Sokhi, Radioactive Releases in the Environment—Impact and
Assessment. Etobicoke, ON: John Wiley, 2003.
M. Eisenbud and T. Gesell, Environmental Radioactivity from Natural, Industrial, and Military
Sources, 4th Edition. Toronto, ON: Academic Press, 1997.
M. Gomez, Radiation Hazards in Mining: Control, Measurement, and Medical Aspects. New
York: Society of Mining Engineers, American Institute of Mining, Metallurgical, and Pe-
troleum Engineers, 1981.
E. M. A. Hussein, Radiation Mechanics—Principles and Practice. New York: Elsevier, 2007.
IAEA. Protection of the Environment from the Effects of Ionizing Radiation—A Report for Discus-
sion. Vienna, Austria: International Atomic Energy Agency, IAEA-TECDOC-1091, 1999.
IAEA, Classification of Radioactive Waste. Vienna, Austria: International Atomic Energy Agency,
General Safety Guide GSG-1, 2009.
Journal of Environmental Radioactivity, Elsevier, ISSN: 0265-931X.
R. L. Kathren, Radioactivity in the Environment—Sources, Distribution and Surveillance. London,
U.K.: Harwood, 1984.
J. Lehr, M. Hyman, T. Gass, and W. Seevers, Handbook of Complex Environmental Remediation
Problems. Toronto, ON: McGraw-Hill, 2002.
K. H. Lieser, Nuclear and Radiochemistry—Fundamentals and Applications, 2nd ed. Toronto, ON:
Wiley-VCH, 2001.
J. Louvar and B. Louvar, Health and Environmental Risk Analysis. Toronto, ON: Prentice-Hall,
1998.
J. Magill and J. Galy, Radioactivity, Radionuclides, Radiation. New York: Springer, 2005.
J. E. Martin and C. Lee, Principles of Radiological Health and Safety. Toronto, ON: Wiley-
Interscience, 2003.
D. Moeller, Environmental Health, 3rd Edition. Cambridge, MA: Harvard University Press, 2005.
T. Moore and D. Dietrich, “Chernobyl and its Legacy”, EPRI Journal, 5:21 (1987).
G. Paić, Ionizing Radiation: Protection and Dosimetry. Boca Raton, FL: CRC Press, 1988.
J. Turner, Atoms, Radiation, and Radiation Protection, 2nd Edition. Toronto, ON: Wiley Inter-
science, 1995.
R. Tykva and D. Berg (eds.), Man-Made and Natural Radioactivity in Environmental Pollution
and Radiochronology. Dordrecht, The Netherlands: Kluwer, 2004.
UNSCEAR, Sources, Effects, and Risks of Ionizing Radiation. New York: United Nations Scientific
Committee on the Effects of Atomic Radiation (UNSCEAR), Report to the General As-
sembly, with scientific annexes, 1988.
UNSCEAR, Sources and Effects of Ionizing Radiation, Volumes I and II. New York: United Nations,
2000.
V. Valković, Radioactivity in the Environment. Amsterdam, The Netherlands: Elsevier Science,
2000.
X. G. Xu and F. K. Eckerman, Handbook of Anatomical Models for Radiation Dosimetry. Boca
Raton, FL: CRC Press, 2010.
12 References
[ANS1991] ANS, Gamma Ray Attenuation Coefficients and Buildup Factors for Engineering
Materials. La Grange IL: ANSI/ANS-6.4.3-1991.
[ANSI2001] ANSI, American National Standard – Respirator Fit Testing Methods. Fairfax, VA:
American National Standards Institute, ANSI/AIHA Z88.10-2001.
[ANSI2003] ANSI, American National Standard – Laboratory Ventilation. Fairfax, VA: American
National Standards Institute, ANSI/AIHA Z9.5-2003.
[ANSI2006] ANSI, American National Standard – Fundamentals Governing the Design and
Operation of Local Exhaust Ventilation Systems. Fairfax, VA: American National Stan-
dards Institute, ANSI/AIHA Z9.2-2006.
[ANSI2007] ANSI, American National Standard – Recirculation of Air from Industrial Process
Exhaust Systems. Fairfax, VA: American National Standards Institute, ANSI/AIHA Z9.7-
2007.
[Aydogdu2013] K. Aydogdu, Personal communication via email with Kam Aydogdu, February 20,
2013.
[BEIRVII2006] BEIR VII, Health Risks from Exposure to Low Levels of Ionizing Radiation: Phase 2.
Washington DC: National Research Council of the National Academies, National Acad-
emies Press, 2006.
[Bentur1991] Y. Bentur, N. Horlatsch, and G. Koren, “Exposure to Ionizing Radiation during
Pregnancy: Perception of Teratogenic Risk and Outcome”, Teratology 43: 109-112
(1991).
[Bergonie1906] J. Bergonie and L. Tribondeau, “De quelques resultats de la radiothérapie et
essai de fixation d’une technique rationnelle”, Comptes Rendus des Séances de l'Acadé-
mie des Sciences 143: 983-985 (1906).
[Bond1996] V. P. Bond, L. Wielopolski, and G. Shani. “Current Misinterpretations of the Linear
No-Threshold Hypothesis”, Health Phys. 70: 877-882 (1996).
[Brenner2003] D. J. Brenner, R. Doll, D. T. Goodhead, E. J. Hall, C. E. Land, J. B. Little, J. H. Lubin,
D. L. Preston, R. J. Preston, J. S. Puskin, E. Ron, R. K. Sachs, J. M. Samet, R. B. Setlow, and
M. Zaider, “Cancer Risks Attributable to Low Doses of Ionizing Radiation: Assessing What
We Really Know”, PNAS 100: 13761-13766 (2003).
[Burnham1992] J. U. Burnham, Radiation Protection, Rev. 3. Point Lepreau Generating Station,
NB: New Brunswick Power Corporation, 1992.
[Cember2009] H. Cember and T. Johnson, Introduction to Health Physics, 4th Edition. Toronto,
ON: McGraw-Hill, 2009.
[Chambless1992] D. A. Chambless, S. S. Dubose, and E. L. Sensintaffar, “Detection Limit Con-
cepts: Foundations, Myths, and Utilization”, Health Phys. 63(3): 338-340 (1992).
[Clarke2005] R. Clarke and J. Valentine, “A History of the International Commission on Radio-
logical Protection”, Health Phys. 88(6): 201-216 (2005).
[Clement2009] C. Clement, Radiological Protection Standards. Seminar at University of Ontario
Institute of Technology, March 17, 2009.
[CBSC2004] CNSC, Keeping Radiation Exposures and Doses “As Low as Reasonably Achievable
(ALARA)”. Ottawa, ON: Canadian Nuclear Safety Commission, Regulatory Guide G-129,
Revision 1, 2004.
[COG2008] COG, “Derived Release Limits Guidance”. In: CANDU Owners Group Document COG-
06-3090-R2-1, Hart, D. (ed.), Toronto, ON, 2008.
[Cole1997] D. Cole and E. Waller, Environmental Radionuclide Baseline Study. Department of
National Defence, Maritime Command Headquarters, prepared by SAIC Canada, 1997.
[CSA2004] CSA, Fume Hoods and Associated Exhaust Systems. Mississauga, ON: Standards
Council of Canada, Canadian Standards Association, CAN/CSA-Z316.5-04, 2004.
[CSA2008] CSA, Guidelines for Calculating Derived Release Limits for Radioactive Material in
Airborne and Liquid Effluents for Normal Operation of Nuclear Facilities. Mississauga,
ON: Standards Council of Canada, Canadian Standards Association, CAN/CSA-N288.1-08,
2008.
[CSA2010] CSA, Environmental Monitoring Programs at Class I Nuclear Facilities and Uranium
Mines and Mills. Mississauga, ON: Standards Council of Canada, Canadian Standards As-
sociation, CAN/CSA-N288.4-10, 2010.
[CSA2012] CSA, Selection, Use, and Care of Respirators. Mississauga, ON: Standards Council of
Canada, Canadian Standards Association, CAN/CSA-Z94.4-11, 2012.
[Currie1968] L. A. Currie, “Limits for Qualitative Detection and Quantitative Determination:
Application to Radiochemistry”. Anal. Chem. 40: 586-593 (1968).
[Deichmann1986] W. B. Deichmann, D. Henschler, B. Holmstedt, and G. Kell, “What Is There
that is Not Poison? A Study of the Third Defense by Paracelsus”. Arch. Toxicol. 58: 207-
213 (1986).
[DoseResponse2013] Dose-Response, Dose-Response, an International Journal – Assessing the
Nature, Mechanisms, and Implications of Dose-Response Relationships. ISSN 1559-3258,
http://www.dose-response.com/, 2013 .
[Faw1999] R. E. Faw and J. K. Shultis, Radiological Assessment: Sources and Doses. La Grange, IL:
American Nuclear Society, 1999.
[Frame2005] P. W. Frame, “A History of Radiation Detection Instrumentation”, Health Phys.
88(6): 97-121 (2005).
[Goans1997] R. E. Goans, E. C. Holloway, M. E. Berger, and R. C. Ricks, “Early Dose Assessment
Following Severe Radiation Accidents”, Health Phys. 72: 513-518 (1997).
[Goldstein1962] H. Goldstein, “Shielding”, Vol. III, Part B. In: Reactor Handbook, 2nd Edition, E. P.
Blizzard and L. S. Abbott (eds.), New York: Interscience, 1962.
[Grasty2004] R. L. Grasty and J. R. LaMarre, “The Annual Effective Dose from Natural Sources of
Ionizing Radiation in Canada”, Rad. Prot. Dos. 108(3): 251-226 (2004).
[Gusev2001] I. A. Gusev, A. K. Guskova, and F. A. Mettler, Medical Management of Radiation
Accidents, 2nd Edition. New York: CRC Press, 2001.
[IAEA1995] IAEA, The Principles of Radioactive Waste Management. Vienna, Austria: Interna-
tional Atomic Energy Agency, IAEA Safety Series No. 111-F, 1995.
[IAEA1996] IAEA, International Basic Safety Standards for Protection against Ionizing Radiation
and for the Safety of Radiation Sources. Vienna, Austria: International Atomic Energy
Agency, IAEA Safety Series No. 115, 1996.
[IAEA2009] IAEA, INES – The International Nuclear and Radiological Event Scale User’s Manual,
2008 Edition. Vienna, Austria: International Atomic Energy Agency, IAEA-INES-2009.
[IAEA2011] IAEA, Radiation Protection and Safety of Radiation Sources: International Basic Safety
Standards. Vienna, Austria: International Atomic Energy Agency, IAEA Safety Standards
Series No. GSR, Part 3 (Interim), 2011.
[ICRP1977] ICRP, Recommendations of the International Commission on Radiological Protection.
International Commission on Radiological Protection, ICRP Publication 26; Annals of the
ICRP 1(3), 1977.
[ICRP1987] ICRP, Lung Cancer Risk from Indoor Exposures to Radon Daughters. International
Commission on Radiological Protection, ICRP Publication 50; Annals of the ICRP 17(1),
1987.
New York: International Commission on Radiological Protection, ICRP Publication 60,
Permagon, 1990.
[ICRP1994] ICRP, “Dose Coefficients for Intakes of Radionuclides by Workers”, ICRP Publication
68; Annals of the ICRP 24(4) (1994).
[ICRP1996a]. ICRP, “Age-Dependent Doses to Members of the Public from Intake of Radionu-
clides: Part 5: Compilation of Ingestion and Inhalation Dose Coefficients”, ICRP Publica-
tion 72; Annals of the ICRP 26(1) (1996a).
[ICRP1996b] ICRP, “Conversion Coefficients for Use in Radiological Protection against External
Radiation”, ICRP Publication 74; Annals of the ICRP 26(3/4) (1996b).
[ICRP2003a] ICRP, A Framework for Assessing the Impact of Ionizing Radiation on Non-Human
Species. Toronto, ON: International Commission on Radiological Protection, ICRP Publi-
cation 91, Elsevier, 2003a.
[ICRP2003b] ICRP, Basic Anatomical and Physiological Data for Use in Radiological Protection.
Toronto, ON: International Commission on Radiological Protection, ICRP Publication 89,
Elsevier, 2003b.
Elsevier, 2007.
[ICRP2008] ICRP, Environmental Protection: Concept and Use of Reference Animals and Plants.
Elsevier, 2008.
[ICRU1994] ICRU, Particle Counting in Radioactivity Measurements. Bethesda, MD: International
Commission on Radiation Units and Measurements, ICRU Report 52, 1994.
[IMBA2010] IMBA, IMBA Professional Plus Computer Code for Internal Dosimetry Calculations –
Version 4.1.3 HPA. Chilton, UK: http://www.imbaprofessional.com, 2010.
[Inkret1995] W. C. Inkret, C. B. Meinhold, and J. C. Tascher, “A Brief History of Radiation Protec-
tion Standards”, Los Alamos Science 23: 116-123 (1995).
[Jones2005] C. G. Jones, “A Review of the History of U.S. Radiation Protection Regulations,
Recommendations, and Standards”, Health Phys. 88(6): 181-200 (2005).
[Knoll2010] G. F. Knoll, Radiation Detection and Measurement, 4th Edition. New York: Wiley,
2010.
[Lamarsh2001] J. R. Lamarsh and A. J. Baratta, Introduction to Nuclear Engineering, 3rd Edition.
Upper Saddle River, NJ: Prentice-Hall, 2001.
[Lindell1996] B. Lindell, “The History of Radiation Protection”, Rad. Prot. Dos. 68(1/2): 83-95
(1996).
[Mattson2008] M. P. Mattson, “Hormesis Defined”, Aging Res. Rev. 7(1): 1-7 (2008).
[Miller1992] K. Miller (ed.), Handbook of Management of Radiation Protection Programs, 2nd

Edition. Boca Raton, FL: CRC Press, 1992.
[MIRD1978] MIRD, “Estimates of Specific Absorbed Fractions for Photon Sources Uniformly
Distributed in Various Organs of a Heterogeneous Phantom”. In: Society of Nuclear
Medicine, Medical Internal Radiation Dose Committee, MIRD Pamphlet No. 5, W. S.
Snyder, M. R. Ford, and G. G. Warner (eds.), New York, 1978.
[Moeller2005] D. W. Moeller, “Environmental Health Physics: 50 Years of Progress”, Health
Phys. 88(6): 160-180 (2005).
[Mosby1990] C. V. Mosby, Mosby’s Medical, Nursing, and Allied Health Dictionary, 3rd Edition.
Toronto, ON: C. V. Mosby, 1990.
[NCRP1987] NCRP, Radiation Exposure of the U.S. Population from Consumer Products and
Miscellaneous Sources. Bethesda, MD: National Council on Radiation Protection and
Measurements, Report No. 95, 1987.
[NCRP2001] NCRP, Management of Terrorist Events Involving Radioactive Material. Bethesda,
MD: National Council on Radiation Protection and Measurements, NCRP Report No. 138,
2001.
[NCRP2009] NCRP, Ionizing Radiation Exposure of the Population of the United States. Be-
thesda, MD: National Council on Radiation Protection and Measurements, Report No.
160, 2009.
[NDR2013] NDR, National Dose Registry, Radiation Protection Bureau, Health Canada.
http://www.hc-sc.gc.ca/ewh-semt/occup-travail/radiation/regist/index-eng.php, 2013.
[NSCA1997] NSCA, Nuclear Safety and Control Act. Government of Canada, S.C. 1997, c.9.
Available at http://laws-lois.justice.gc.ca/eng/acts/N-28.3/. 1997.
[NSCA2000] NSCA, Radiation Protection Regulations. Government of Canada, SOR/2000-203.
Available at http://laws-lois.justice.gc.ca/eng/regulations/SOR-2000-203/index.html.
2000.
[OPG2012] OPG, 2011 Results of Radiological Environmental Monitoring Programs. Ontario
Power Generation, N-REP-03481-10010. Downloadable at:
http://www.opg.com/news/reports/ index.asp. 2012.
[ORNL2013] ORNL, RadToolbox. Downloadable from the Centre for Biokinetic and Dosimetric
Research, http://ordose.ornl.gov/downloads.html. 2013.
[Poston2005] J. W. Poston, “External Dosimetry and Personnel Monitoring”, Health Phys. 88(6):
41-48 (2005).
[Potter2005] C. A. Potter, “Internal Dosimetry: A Review”, Health Phys. 88(6): 49-62 (2005).
[Preston2005] R. J. Preston, “Radiation Biology: Concepts for Radiation Protection”, Health
Phys. 88(6): 29-40 (2005).
[Scott2008] B. R. Scott, “It’s Time for a New Low-Dose Radiation Risk-Assessment Paradigm—
One that Acknowledges Hormesis”, Dose-Response, 6: 333-352 (2008).
[Scott2012] B. R. Scott, V. R. Bruce, K. M. Gott, J. Wilder, and T. March, “Small γ-Ray Doses
Prevent Rather than Increase Lung Tumors in Mice”, Dose-Response 10: 527-540 (2012).
[Shultis2000] J. K. Shultis and R. E. Faw, Radiation Shielding. La Grange Park, IL: American
Nuclear Society, 2000.
[Shultis2005] J. K. Shultis and R. E. Faw, “Radiation Shielding Technology”, Health Phys. 88(6):
71-96 (2005).
[Siegel2012] J. A. Siegel and M. G. Stabin, “RADAR Commentary: Use of Linear No-Threshold
Hypothesis in Radiation Protection Regulation in the United States”, Health Phys. 102:
90-99 (2012).
[Till2008] J. Till and H. Grogan, Radiological Risk Assessment and Environmental Analysis.
Toronto, ON: Oxford University Press, 2008.
[Uchrin1988] G. Uchrin and M. Ranogajec-Komor, “Thermoluminescent Dosimetry”. In: Ionizing
Radiation: Protection and Dosimetry, Paić, G., (ed.), Boca Raton, FL: CRC Press, 1988.
[UNSCEAR2000] UNSCEAR, United Nations Scientific Committee of the Effects of Atomic Radia-
tion – Report to the General Assembly. New York: United Nations, 2000.
[UNSCEAR2008] UNSCEAR, Sources and Effects of Ionizing Radiation. New York: United Nations
Scientific Committee on the Effects of Atomic Radiation, Report to the General Assembly
(A/63/46), 2008.
[Walker2000] J. S. Walker, Permissible Dose – A History of Radiation Protection in the Twentieth
Century. Berkeley, CA: University of California Press, ISBN 0-520-22328-4, 2000.
[Waller2013] E. J. Waller, “Sources of Radiation in the Environment, Including Natural Radia-
tion, Naturally Occurring Radioactive Materials (NORM), Technically Enhanced Materi-
als, Weapons Tests, and Nuclear Accidents”. In: Encyclopedia of Sustainability Science
and Technology, ISBN 978-0-387-89469-0, New York: Springer, 2013.
13 Acknowledgements
ments during the development of this Chapter. Their feedback has much improved it. Of
course the responsibility for any errors or omissions lies entirely with the authors.
Kam Aydogdu
Esam Hussein
1
CHAPTER 13
Reactor Safety Design and Safety Analysis
prepared by
Dr. Victor G. Snell
Summary:
The chapter covers safety design and safety analysis of nuclear reactors. Topics include concepts
of risk, probability tools and techniques, safety criteria, design basis accidents, risk assessment,
safety analysis, safety-system design, general safety policy and principles, and future trends. It
makes heavy use of case studies of actual accidents both in the text and in the exercises.
Table of Contents
1 Introduction ............................................................................................................................ 6
1.1 Overview ............................................................................................................................. 6
1.3 Risk ...................................................................................................................................... 8
1.4 Hazards from a Nuclear Power Plant ................................................................................ 10
1.5 Types of Radiation in a Nuclear Power Plant .................................................................... 12
1.6 Effects of Radiation ........................................................................................................... 12
1.7 Sources of Radiation ......................................................................................................... 14
1.8 Risk .................................................................................................................................... 15
1.9 Problems ........................................................................................................................... 17
2 Design Basis Accidents .......................................................................................................... 18
2.1 Top-Down Approach ......................................................................................................... 19
2.2 Bottom-Up Approach........................................................................................................ 21
2.3 Probabilistic Safety Analysis.............................................................................................. 22
2.4 Experience......................................................................................................................... 22
2.5 Canadian Approach to DBAs ............................................................................................. 22
2.6 Other Design Basis Events................................................................................................. 25
2.7 Problems ........................................................................................................................... 25
3 Experience............................................................................................................................. 31
3.1 Criticality Accidents and Power Excursions....................................................................... 32
3.2 Loss of Cooling / Heat Removal ........................................................................................ 45
3.3 Problems ........................................................................................................................... 53
4 Safety Goals and Risk Assessment ........................................................................................ 54
4.1 Safety Goals....................................................................................................................... 54
4.2 Risk Assessment ................................................................................................................ 58
4.3 Problems ........................................................................................................................... 74
5 Mitigating systems ................................................................................................................ 76
5.1 Defence-in-Depth.............................................................................................................. 76
5.2 Shutdown Systems ............................................................................................................ 77
©UNENE, all rights reserved. For educational use only, no assumed liability. Reactor Safety Design and Safety Analysis – October 2015
5.3 Heat-Removal Systems...................................................................................................... 86

5.4 Emergency Core Cooling System....................................................................................... 91
5.5 Containment ..................................................................................................................... 94
5.6 Monitoring ........................................................................................................................ 98
5.7 Problems ........................................................................................................................... 98
6 Safety Analysis – Accident Phenomenology ....................................................................... 100
6.1 Accidents by Phenomena................................................................................................ 100
6.2 Margins ........................................................................................................................... 101
6.3 Major Computer Analysis Tools Required for DBAs........................................................ 103
6.4 Code Validation and R&D................................................................................................ 105
6.5 Selection of Initial Conditions ......................................................................................... 111
6.6 Accident Walk-Through: Large LOCA .............................................................................. 115
6.7 Accident Walk-Through: Small LOCA .............................................................................. 124
6.8 Accident Walk-Through: Single-Channel Event............................................................... 125
6.9 Accident Walk-Through: Loss of Reactivity Control........................................................ 126
6.10 Accident Walk-Through: Loss of Forced Circulation ....................................................... 127
6.11 Accident Walk-Through: Loss of Secondary-Side Heat Removal .................................... 127
6.12 Accident Walk-Through: Fuel-Handling Accident ........................................................... 128
6.13 Accident Walk-Through: Loss of Moderator Inventory or Heat Removal....................... 129
6.14 Severe Accidents ............................................................................................................. 129
6.15 Problems ......................................................................................................................... 134
7 Safety Analysis – Mathematical Models ............................................................................. 135
7.1 Reactor Physics................................................................................................................ 135
7.2 Decay Power.................................................................................................................... 136
7.3 Fuel.................................................................................................................................. 137
7.4 Heat-Transport System.................................................................................................... 141
7.5 Fuel Channels.................................................................................................................. 143
7.6 Moderator....................................................................................................................... 145
7.7 Containment ................................................................................................................... 145
7.8 Fission Products, Atmospheric Dispersion, and Dose..................................................... 146
7.9 Problems ......................................................................................................................... 149
8 Safety of Operation ............................................................................................................. 150
8.1 Safety Culture.................................................................................................................. 150
8.2 International Nuclear and Radiological Event Scale ....................................................... 152
8.3 Safety Aspects of Future Designs.................................................................................... 154
8.4 Problems ......................................................................................................................... 158
9 Review................................................................................................................................. 159
10 References........................................................................................................................... 159
11 Further Reading .................................................................................................................. 168
12 Glossary............................................................................................................................... 169
13 Appendix 1 – Basic Rules of Boolean Algebra..................................................................... 171
13.1 Operators ........................................................................................................................ 171
13.2 Basic Principles................................................................................................................ 172
Reactor Safety Design and Safety Analysis 3
13.3 Theorems ........................................................................................................................ 172

13.4 Combining Probabilities.................................................................................................. 174
14 Appendix 2 – Common-Cause Failures – An Example ........................................................ 177
15 Appendix 3 – Why a Reactor Cannot Explode Like an Atomic Bomb ................................. 177
16 Acknowledgments............................................................................................................... 178
List of Figures
Figure 1 Chapter concept map........................................................................................................ 7
Figure 2 Risk optimization............................................................................................................. 10
Figure 3 Examples of radiation dose............................................................................................. 12
Figure 4 Fuel element cross section showing location of fission products .................................. 14
Figure 5 Simplified top-down approach........................................................................................ 20
Figure 6 Simplified bottom-up example ....................................................................................... 21
Figure 7 Criticality experiment...................................................................................................... 26
Figure 8 SES-10.............................................................................................................................. 27
Figure 9 ZED-2 cutaway................................................................................................................. 29
Figure 10 ZED-2 top view .............................................................................................................. 30
Figure 11 Learning and forgetting................................................................................................. 31
Figure 12 SL-1 cutaway ................................................................................................................. 34
Figure 13 NRX fuel cross section................................................................................................... 36
Figure 14 NRX elevation................................................................................................................ 37
Figure 15 RBMK schematic diagram ............................................................................................. 41
Figure 16 RBMK reactor ................................................................................................................ 43
Figure 17 Reverse shutdown in RBMK ......................................................................................... 44
Figure 18 RBMK building cross section ......................................................................................... 45
Figure 19 TMI schematic............................................................................................................... 46
Figure 20 TMI core end state ........................................................................................................ 47
Figure 21 Fukushima Dai-ichi before earthquake ......................................................................... 49
Figure 22 BWR Mark 1 containment............................................................................................. 50
Figure 23 Design basis versus actual flood level........................................................................... 51
Figure 24 Typical λ(t) versus time ................................................................................................. 64
Figure 25 Reliability for constant λ................................................................................................ 66
Figure 26 Availability with repair .................................................................................................. 67
Figure 27 Simple pumped system ................................................................................................. 69
Figure 28 Sample fault tree with labels ........................................................................................ 70
Figure 29 Simple event tree .......................................................................................................... 72
Figure 30 Contributors to SCDF for CANDU 6 ............................................................................... 74
Figure 31 Simple fault-tree exercise ............................................................................................. 75
Figure 32 Defence-in-depth: barriers............................................................................................ 76
Figure 33 CANDU shutdown systems............................................................................................ 79
Figure 34 Xenon transient after shutdown and start-up .............................................................. 81
Figure 35 2/3 logic......................................................................................................................... 85
Figure 36 SDS2 testing .................................................................................................................. 86
Figure 37 Bundle power after shutdown ...................................................................................... 88

Figure 38 Shutdown cooling system ............................................................................................. 89
Figure 39 Moderator and shield cooling....................................................................................... 90
Figure 40 ECC layout ..................................................................................................................... 92
Figure 41 Three phases of ECC...................................................................................................... 93
Figure 42 Single-unit containment................................................................................................ 95
Figure 43 Vacuum containment concept ...................................................................................... 96
Figure 44 Margins ....................................................................................................................... 102
Figure 45 Safety analysis codes................................................................................................... 105
Figure 46 Code validation process .............................................................................................. 106
Figure 47 RD-14M schematic ...................................................................................................... 107
Figure 48 Contact boiling tests.................................................................................................... 108
Figure 49 Heat-transfer regimes on outside of calandria tube................................................... 109
Figure 50 Large-scale containment facility ................................................................................. 110
Figure 51 Large-scale vented combustion facility....................................................................... 110
Figure 52 Containment test facility............................................................................................. 111
Figure 53 HTS layout ................................................................................................................... 116
Figure 54 Core flow vs. break size for a group of channels ........................................................ 117
Figure 55 Typical LBLOCA timescale............................................................................................ 118
Figure 56 Containment pressure transient for 100% ROH LOCA ................................................ 119
Figure 57 Response of PWR and CANDU to reactivity increase ................................................. 122
Figure 58 Importance of reactivity effects.................................................................................. 123
Figure 59 Sources of water near the fuel.................................................................................... 130
Figure 60 Channel collapse in severe core-damage accident ..................................................... 131
Figure 61 Core collapse ............................................................................................................... 131
Figure 62 Mechanistic model of channel collapse...................................................................... 132
Figure 63 Debris heating transient.............................................................................................. 133
Figure 64 Heat flux on calandria wall.......................................................................................... 133
Figure 65 Temperature distribution across a fuel pin ................................................................. 139
Figure 66 Node/link structure..................................................................................................... 142
Figure 67 Heat transfer to pressure tube.................................................................................... 143
Figure 68 Atmospheric dispersion .............................................................................................. 148
Figure 69 AND gate ..................................................................................................................... 171
Figure 70 OR gate........................................................................................................................ 171
Figure 71 NOT gate...................................................................................................................... 171
Figure 72 Probability of both of two events ............................................................................... 175
Figure 73 Probability of either of two events ............................................................................. 176
List of Tables
Table 1 Single / dual failure limits ................................................................................................. 23
Table 2 Consultative document C-6 limits .................................................................................... 24
Table 3 Dose limits ........................................................................................................................ 24
Table 4 SL-1 chronology ................................................................................................................ 35
Table 5 Bayes’ theorem example .................................................................................................. 62
Table 6 Reliability terms and relationships ................................................................................... 65
Table 7 Levels of defence in depth................................................................................................ 77
Table 8 Typical trip signals for CANDU .......................................................................................... 83
Table 9 Grouping and separation example ................................................................................... 85
Table 10 Operating pressure of decay heat-removal systems ...................................................... 91
Table 11 Some conservative assumptions and parameters........................................................ 112
Table 12 Reactivity effects in PWR and CANDU.......................................................................... 123
Table 13 Examples of reactivity response to accidents for PWR and CANDU ............................ 124
Table 14 Typical core inventory of volatile fission products..................................................... 147
Table 15 Stages of organizational decline ................................................................................... 152
Table 16 INES event scale............................................................................................................ 153
Table 17 Categories of passive safety characteristics ................................................................. 156
1 Introduction
The purpose of this chapter is to describe the safety characteristics of nuclear reactors and how
their safety performance is predicted and verified.
1.1 Overview
Figure 1 summarizes the chapter concepts and their logical relationship. We describe each box
in turn.
Section 0 (this section) defines risk. The type of risk depends on the activity. Most people want
their activities and surroundings to be “safe”. However, this is a meaningless ideal, and impos-
sible to achieve in absolute terms, because every activity imposes some risk. Fortunately, risk
can be quantified. Then society can set acceptable levels of risk in a reasonably objective
manner.
The risk from nuclear reactors comes from accidental release of radioactive material. Most of
the radioactive material is in the reactor fuel. Therefore, one can postulate accidents which
might allow radioactivity to escape and then design systems (called mitigating systems or safety
systems) to prevent or control such accidents. These design basis accidents1 (DBA) are covered
in Section 2. As well, much knowledge about the risk from nuclear reactors comes from actual
experience in both small research and large power reactors. Case histories of the most impor-
tant events which influenced the development of nuclear power reactor safety are covered in
Section 3. Lessons learned from these were extracted in the form of specific deterministic
requirements, which described the accidents that had to be designed for and the assumptions
used in showing the safety systems were effective. These accidents also became design basis
accidents.
This approach limits risk, but does not lend itself to quantifying risk because the deterministic
requirements and the design basis accidents were chosen “conservatively”, i.e., to be worse
than what would happen in reality, and with little regard to frequency. In addition, descriptions
of these accidents were not complete. A parallel approach is to start off by setting numerical
risk targets for the plant as a whole (safety goals). Possible accidents are identified and classi-
fied using a frequency-based approach. This probabilistic approach is covered in Section 4. It
leads to another list of accidents which overlaps with, but is different from, the list of design
basis accidents.
1
We bow to common practice in using this term. The term was not used in Canada until recently, but has been
used for a long time in the United States. It originally implied that as long as the plant was designed to withstand
“design basis” accidents, all would be well. The accidents at Three Mile Island and Chernobyl showed the weakness
of this concept. At Fukushima, the design basis was wrong (and was known to be – see [IAEA, 2011] p. 75). This
means that design basis accidents do not define a strong boundary between possible and “incredible”. “Beyond
design basis” accidents are now of much interest—indeed, although infrequent, they are the only ones which
could have significant consequences.
Introduction
Section 1
Experience Safety Goals
Section 3 Section 4
Deterministic Requirements Probabilistic Requirements
Section 4
Safety Mitigating Probabilistic

Design Basis Credible
Analysis Systems Safety Analysis
Accidents Accidents
Sections 6 & 7 Section 5
Section 2
Plant safety
as designed
Section 8 Safety Culture

Good Operating
Plant safety Practice
as operated
Future Trends
Review Section 9
Figure 1 Chapter concept map

The mitigating systems must be able to handle both design basis accidents and accidents
identified by the probabilistic approach. They are described in Section 5.
The methods to confirm that the mitigating systems are effective are:
 Probabilistic safety analysis (PSA), which drives design requirements for, and verifies the
reliability of, normal and mitigating systems. It also ensures to the extent practical that
an accident which requires a mitigating system does not also impair it (Section 4). It
uses Boolean algebra to determine the frequencies of initiating events, given the fre-
quency of component failures, and mathematically combines these with the reliability of
mitigating systems to determine the frequency of severe accidents, thereby showing
that the safety goals have been achieved.
 Deterministic safety analysis, or simply safety analysis, which drives design requirements
for, and verifies the performance of, mitigating systems. It uses mathematical models of
all the key systems in a plant to predict (ultimately) how much radioactive material will
escape from the fuel and where it will be transported. It shows that the deterministic
requirements have been met. Section 6 describes the phenomenology of accidents, and
Section 7 summarizes the mathematical tools used to predict how they evolve.
Ultimately, the safety of a plant, however well-designed, depends on the people who run it.
Section 8 covers safety aspects of operation, including safety culture. A brief summary of
innovative designs which promise to deliver increased safety wraps up Section 8.
Section 9 summarizes the key points of the Chapter. Section 10 gives the references used in this
Chapter, and Section 11 lists other sources of information on reactor safety. A Glossary in
Section 12 is a ready reference for the abbreviations used in this Chapter. Each subsection
includes problems for further self-study and the occasional worked example.

The goal of this Chapter is for the student to develop:
 An understanding of the approach to nuclear reactor safety concepts and safety de-
sign
 Familiarity with the hazards involved in nuclear power reactors
 Knowledge of the concepts of risk, risk quantification, and risk optimization
 An understanding of the root causes of key real-world accidents, leading to “respect
for the reactor core”
 The ability to develop systematically a list of credible accidents for a nuclear power
plant design
 The ability to understand and create simple fault trees and event sequence diagrams
 Familiarity at an overview level with the physical and mathematical models used in
safety analysis.
It is assumed that the reader is generally familiar with the material in prior chapters.
1.3 Risk
Risk involves three key ideas:
 all technologies involve risk,
 every endeavour involves a risk/benefit trade-off, and
 risk can be quantified and reduced to an acceptably low level.
Nuclear reactors, hydro dams, and fossil-fuel electrical generating stations are all inherently
dangerous. The nature of the hazard in each case is quite different. Hazards can be sudden
(acute) or delayed. For hydro dams, an acute hazard is rupture of the dam, causing massive
floods downstream. A delayed hazard is build-up of toxic mercury in the water behind the dam
due to leaching from the rocks. For natural-gas plants, there is a local acute hazard due to
explosion and a global delayed hazard due to climate change from the release of combustion
products (greenhouse gases) to the atmosphere. Coal plants are likewise a major source of
greenhouse gases and in addition can cause respiratory disease from the combustion and
release of toxic chemicals in the coal; see [Inhaber, 1978] and [Rogers, 2004] for Canadian
examples. Some coal plants emit more radioactive material to atmosphere in normal operation
than a nuclear power plant. For a nuclear power plant, the hazard of most interest is the
release of radioactive material in accidents. Unlike a coal plant, an inadvertent rise in power in
a nuclear plant (if it is not stopped) can both drive the release of radioactive material out of the
fuel and rapidly cause damage to the reactor and its containment structure.
We “accept” hazards of technologies when they have a benefit which is perceived to offset the
risk. Sometimes this decision is made on an individual basis: you may go sky-diving (an activity
so objectively risky that you cannot get insurance coverage for it) because you believe that the
unique thrill is worth the risk. You accept the hazards of electrical shock and fires for the
convenience of using electric lights and appliances. Nothing that we do on a day-to-day basis is
as risky as hurtling down a narrow strip of levelled ground at 100 km/h in a thin metal container
containing 60 litres of explosive liquid towards someone else in a similar device, using a painted
strip as a guide to avoid collision. Yet almost all people believe the benefits of driving are worth
the risk.
Sometimes the decision is made on a societal basis: if you live in a city, you cannot easily choose
to accept or reject risks such as being hit by a car (even if you choose not to drive one), breath-
ing polluted air, or getting mugged. Activities which pose an involuntary risk are often regulated
by law. In our three examples, the respective regulatory devices would be traffic laws, emission
controls, and the criminal laws.
The benefits of nuclear power include production of clean electricity. In Ontario, over half the
electricity comes from nuclear power; in countries such as France, as much as 80%. Other
benefits of nuclear technology are medical and industrial applications, insect control, environ-
mental protection, and scientific research. Canada is the source of much of the world’s produc-
tion of medical isotopes, largely originating from the NRU reactor at Chalk River.
This chapter is concerned with risk to humans. Many technologies also pose risk to other living
organisms. For nuclear technology, in general, if radiation risk to humans is acceptable, the risk
to other living things will also be acceptable because they are less susceptible to radiation (e.g.,
they do not live as long (and therefore do not develop cancer as easily) or are inherently more
resistant to radiation damage (e.g., insects)). Radioactive elements and compounds can be
concentrated as they move up the food chain, and therefore these pathways must be modelled
to provide a scientific basis to show that humans are limiting this risk.
Safety can be thought of as the complement to risk; however, usually it is risk that is quantified,
and we shall focus on risk in this Chapter.
Because risk cannot be eliminated, it must ideally be optimized. This means a cost/benefit
analysis, although such an analysis often looks only at the risk of a technology, without factoring
in benefits. At a risk optimum, the additional resources used to provide additional risk reduc-
tion would come at a disproportionate cost, and any resources removed from risk reduction
would cause a disproportionate increase in risk. This situation is shown in Figure 2. This opti-
mum is never achieved in practice. This is partly because the risk from nuclear power is per-
ceived to be greater than that of other technologies, even if the numerical risk is the same, due
to social factors such as unfamiliarity with the technology and its association with atomic
weapons and cancer. Hence, regulation of nuclear power often includes a penalty on the
allowed risk, called risk aversion. Such a topic is beyond the scope of this Chapter; see [Slovic,
1987] or [Siddall, 1981] for examples. However, we still need an objective means of quantifying
risk in terms of frequency and consequences. We start with the latter—what are the conse-
quences of a nuclear accident?
Figure 2 Risk optimization
1.4 Hazards from a Nuclear Power Plant

Most of this chapter covers radiological hazards. However, we will start by systematically listing
possible hazards to make sure not to miss one. A hazard in a broad sense can be physical,
chemical, biological, or radiological.
Nuclear power plants do not pose a physical hazard due to the nuclear process—there is no risk
of off-site injury due to a nuclear explosion (the explosion at Chernobyl was a steam explosion
and that at Fukushima a hydrogen one—both with far less energy than a nuclear explosion
would have generated).
A simple explanation for this is the following: an atomic (or nuclear) bomb works by making a
mass of fissile material supercritical and holding it together long enough to reach very large
energies. The hard part is holding it together, which requires three things:
1. banging two sub-critical masses together very fast, so that the supercritical mass formed
does not disintegrate due to heating as the pieces approach each other, and
2. ensuring that the source of neutrons that initiates the explosion is located at the centre
and is triggered at the right time, and
3. using pure fissile material—U235 or Pu239—so that the mass goes critical on fast neu-
trons. Fast neutrons have very short lifetimes. The basic time unit that bomb designers
use is a “shake”, or 10-8 seconds. It takes only about 50 chain-reaction generations of
neutrons to produce enormous nuclear energies in the few shakes before the mass
blows apart and the chain reaction stops.

Most power reactors, however, slow down the neutrons to thermal energies, and thermal
neutrons have lifetimes of milliseconds. (In fact, as covered in Chapter 4, a power plant is
critical on delayed thermal neutrons, with lifetimes of the order of tenths of seconds to several
seconds.) This means that if you somehow made a power reactor (e.g., a CANDU) supercritical,
the energy doubling time would be the order of hundreds of milliseconds. This is slow enough
to stop with mechanical or hydraulic devices, but if these failed, the thermal energy build-up
would destroy the fuel and the reactor geometry before the power rose above perhaps ten
times normal, ending the chain reaction. The result is not minor (as at Chernobyl), but is not a
nuclear bomb. For further detail, assuming you have read the physics Chapter, see Appendix 3 –
Why a Reactor Cannot Explode Like an Atomic Bomb.
Most people do not think of a nuclear power plant as posing a chemical hazard, but thermal
power plants need a large supply of cooling water. Approximately one-half to two-thirds of the
energy produced by any thermal power plant is wasted because of the second law of thermo-
dynamics; the waste energy is rejected to a lake, river, sea, or the atmosphere. For many sites
near large bodies of water, most of the cooling water is used in once-through mode in the main
condenser, and in many plants (fossil-fuel as well as nuclear), it is chlorinated to avoid growth of
biological material such as zebra mussels in the plant equipment. It follows that such plants
have relatively large tanks of chlorine somewhere on-site. The consequences of rupture of
these tanks can be severe off-site contamination (as in the Mississauga train derailment in 1979
[Liverman, 1979], [OMSC, 1981]).
There is no biological hazard associated with a nuclear plant because these plants do not
contain or produce bacteria2 or viruses.
A summary of radiological hazards is provided in Chapter 15. Key points will be summarized
here. The effects can be somatic—affecting the living individual who is exposed to radiation—
or genetic—appearing in the yet-to-be conceived offspring of that person or in later genera-
tions.
 Somatic effects can occur:
o soon after the exposure (acute, or prompt, or early, or non-stochastic effects—
these terms all mean the same thing) or
o later (delayed or latent or stochastic effects). The word stochastic means random
and reflects the fact that if many individuals are exposed to a moderately “high”
dose of radiation (above about 0.2 Sv each), one can predict the number of such
individuals who will one day get cancer as a result of the exposure, but one can-
not predict which individuals will be affected.
o in the fetus of a woman who is exposed while pregnant (teratogenic).
 Genetic or hereditary effects have been observed in animals, but not in people.
2
There could be a small biological hazard if the plant uses cooling towers and does not keep them clean, in which
case they could become a source of bacterial growth.
[UNSCEAR, 2001].
1.5 Types of Radiation in a Nuclear Power Plant

The radiation in a nuclear power plant comes from fission fragments and from the activation of
non-radioactive material. It therefore includes:
 alpha rays, or helium nuclei,
 beta rays, or electrons, and
 gamma rays, or photons.
Neutrons are not normally a concern to the public in reactor accidents because they slow down
very rapidly in the reactor structures; however, they can be a concern to workers if local shield-
ing is inadequate, or in non-reactor facilities in case of inadvertent criticality.
1.6 Effects of Radiation

The biological effect of radiation is measured in Sieverts (Sv), a unit which combines the energy
deposited in tissue and the effectiveness of that particular form of radiation in causing damage
to cells. Chapter 15 provides more detail on this topic. Figure 3 gives a perspective on the
range of actual doses, from normal activities to severe accidents. Sources: [CNSC, 2011],
[Talbot, 2003], [Lewis, 1999], [TEPCO, 2011], [UNSCEAR, 2000].
Figure 3 Examples of radiation dose

Below about 0.1 Sv per person in a large human population, there are no observable stochastic
effects. Therefore, even the most exposed staff members at Fukushima during the accident are
not expected to show symptoms of exposure. One Sv individual dose marks the onset of acute
symptoms, and above 5 Sv, as received by the firefighters at Chernobyl, severe illness and death
may ensue.
1.6.1 Linear Hypothesis

As discussed in Chapter 15, the linear hypothesis is used to extrapolate the observed effects on
human populations exposed to high doses of radiation (as in Japanese atomic bomb survivors)
to low doses. For large doses in the stochastic range, the linear hypothesis states that [ICRP,
1990]:
100 person-Sv will produce about 5 fatal cancers in the exposed (general) population.
Sample Problem
Rank the magnitude of the following risks to a group of people (express the answers numerically
and explain your reasoning):
a. A collective dose of 1000 person-Sv given to 1,000,000 people
b. A collective dose of 1000 person-Sv given to 100,000 people
c. A collective dose of 1000 person-Sv given to 100 people
d. A collective dose of 1000 person-Sv given to 5000 people.
Answer:
According to the linear hypothesis, a collective dose of 1000 person-Sv should result in 50 fatal
cancers. However, one must be careful of the range of applicability. For each case in turn:
a) The average dose is 0.001 Sv, which is so small that there is no evidence that the linear
hypothesis applies. Incidentally, this is also less than the annual dose in North America
from natural background radiation.
b) The average dose is 0.01 Sv, and therefore the linear hypothesis is likely inapplicable.
c) The average dose is 10 Sv, which would cause prompt injury or death, and therefore the
linear hypothesis is inapplicable.
d) The average dose is 0.2 Sv, and therefore the linear hypothesis may be used and would
predict 50 fatal cancers over time in the exposed population.
1.6.2 Hormesis
The difficulty with the linear hypothesis is that the effects of radiation on humans at low doses
(below 0.1 Sv) are so small (or even beneficial—the hormesis hypothesis) that the models are
very difficult to validate. At the moment, for better or worse, the linear hypothesis is used
internationally as the basis for setting regulations for radiation protection. However, the
Fukushima accident has led to a rethinking of overly conservative regulations because there is a
real risk in evacuating people unnecessarily, which may be greater in some cases than the risk of
radiation exposure. Moreover, ICRP has taken great pains to distinguish a priori theoretical risks
from a posteriori predictions of “real” health effects resulting from low radiation doses to many
people; see, e.g., [Gonzales, 2013], from which the following is quoted: “Following exposure to
low radiation doses below about 100 mSv, an increase of cancer has not been convincingly or
consistently observed in epidemiological or experimental studies and will probably never be
observed because of overwhelming statistical and biasing factors.”
For more details on hormesis, see, e.g., [Cuttler, 2009]. As an example [Chen, 2007], steel
containing Co60 was used in the construction of 1700 apartments in Taiwan, resulting in an
average dose of 0.4Sv to the 10,000 occupants. What was observed was a significant decrease,
rather than an increase, in cancer deaths.
1.7 Sources of Radiation

Because this chapter will now focus on the radiological hazards of nuclear power plants, we
need to know where the radioactive material is normally and how it can escape. There are
several repositories of radioactive material (we use CANDU here as an example):
 Most of the radioactive material (fission products) is in the fuel in the core.
 Large quantities of long-lived radioactive isotopes are in the spent fuel, located either in
wet pools or in dry shielded concrete containers.
 Tritium (T) is produced in the heavy-water moderator and
coolant (about twenty times more in the former) through Figure 4 Fuel element cross
activation of deuterium (D) by neutrons: section showing location of
fission products
12 yr

D ( n, p )  n  T ( 2n, p )  He(n,2 p )  e .
(1)
Tritium is radioactive, with a half-life of about 12 years, and decays
to helium with emission of an electron. It is hazardous if inhaled or
ingested or if it comes in contact with skin, but little shielding is
needed to protect a person: beta particles can be stopped by a
sheet of plastic.
 Carbon-14 is produced by neutron bombardment of dis-
solved nitrogen in the moderator.
In terms of public safety, the fission products in the fuel are by far
the most significant hazard due to their quantity and their potential
to be made mobile or volatile. In normal operation, the radioactive material in the fuel consists
of:
 fission products trapped within the ceramic UO2, and
 fission product gases in bubbles or interlinked spaces within the fuel ceramic, or free be-
tween the fuel and the sheath.
These exist in the ratio of about 9:1 for the highest-powered fuel element in a CANDU reactor,
as shown in [Ionescu, 2009].
Therefore, accidents which cause perforation of the fuel sheath (but which do not damage the
UO2) have the potential to release only somewhat less than 10% of the gaseous fission prod-
ucts. Sheaths can be damaged mechanically (in fuel-handling accidents) or by overheating: if
the sheath overheats from its normal temperature of 300°C to about 600°C–800°C, it will
plastically deform because of the pressure of the fission product gases it contains, and eventu-
ally rupture. To drive out the remaining gaseous fission products and the solid fission products
such as cesium and strontium, the fuel temperature has to be raised to close to the melting
point (2840°C) or the fuel itself must be heavily oxidized by direct exposure to air or steam.
Therefore, accidents which release significant amounts of radioactive material are initiated by:
 overheating the fuel in the core due to a mismatch between power and cooling
 leaks or pipe breaks in the coolant or moderator
 mechanical damage to the fuel
 overheating of the spent fuel in storage due to a power/cooling mismatch.
All accident analyzes reduce to these categories of failures.
It follows that the fundamental safety functions that must be carried out after an accident are:
 Control the fission reaction (and shut down the reactor)
 Cool the fuel (remove the decay heat)
 Contain any release of radioactive material
 Monitor the state of the plant.
In summary, control / cool / contain / monitor is the essence of safety design.
1.8 Risk
Safety concerns are ultimately expressed in terms of risk. The risk of a system, which must be
specified (e.g., risk of a component failure, of an activity, of a nuclear reactor, of the nuclear fuel
cycle) is customarily defined as:
Risk   fi ci
i , (2)
where fi is the expected frequency of event i and ci is its consequence; the summation is over all
events. Like the risk of betting money, this summation makes sense only for a large number of
systems over a long period of time; it does not apply to a single machine (or a single bet), for
which the event outcome is binary (e.g., for one bet, you either win or lose). In nuclear safety,
probabilistic risk analysis can be used to quantify the risk posed by a nuclear power plant; more
typically, it calculates risk indicators such as severe core-damage frequency or large off-site
release frequency. Such indicators can identify potential plant improvements to reduce risk
where practical.
Sample problem
The frequency of a (fully-contained) core-damage accident in a certain 1000 MWe nuclear
power plant is 10-7 per year. If your insurance conglomerate were asked to insure the plant,
what premium would you have to charge to break even over the long term?
Answer:
We need to use a common measure of comparison, in this case average cost / year. A contained
core melt would require removal of the damaged core, decontamination, and replacement of
the reactor. This could take ten years. In addition, the electricity formerly generated by the
plant would have to be replaced. Since the accident is stated to be contained, we do not add
costs for off-site decontamination, evacuation, relocation, nor health effects – for an accident
resulting in a significant release, these would have to be considered. This means that the
average cost / year is (using rough ballpark values—the problem can easily be made very
complex if greater sophistication and accuracy is needed):
{[$5 × 109 for decommissioning / decontamination + $5 × 109 for rebuilding the plant] +
[106 kW × $0.06 /kW-hour × 24 hours / day × 365 days / year × 10 years]} × [10-7 / year]
= [$1000 + $526] / year or ~$1500 /year.
Note the importance of the replacement electricity cost.
Because accidents cannot be prevented in any significant human endeavour, the broad goals of
reactor safety analysis are to:
 Show that the frequency and consequences of accidents are within acceptable limits
and/or
 Show that the frequency of an accident is too small to consider.
Acceptable limits are defined (broadly) with respect to the event frequency. For example,
frequent occurrences (minor faults such as loss of electrical power) should not stress the
system, damage fuel, or invoke protective systems. Very infrequent events, like a large loss of
coolant, are permitted to push the physical systems into plastic deformation or damage fuel, but
not to allow radioactive release beyond a prescribed limit. Severe accidents may damage the
core, but should not fail the containment building. This approach may, but does not necessarily,
address the direct economic costs of a severe accident, which can be huge, even if the public
safety impact is minimal, e.g., Fukushima.
This framework implies that:
1. We have to know what the possible accidents are.
2. We have to be able to predict their frequency.
3. We have to be able to predict their consequences.
Safety analysis is carried out at several stages in the plant life cycle:
 During preliminary design (design assist), to ensure that the design concepts meet safety
requirements.
 During final design, to confirm that safety requirements are met and to include the re-
sults as part of the applications for the licences required to construct and operate the
plant (see Chapter 16).
 For an operating plant, to incorporate the effects of any changes in the plant, in funda-
mental knowledge, in operating experience, and in safety analysis methodology.
Three safety-analysis methods are used during some or all of these stages. These are comple-
mentary, not mutually exclusive, and in practice all three are used:
1. Rule - e.g., use the ASME code for pressure-vessel design. It is implied that following the
code or the standard reduces the likelihood of failure of the material to a very low level.
This is largely based on long experience, testing, and more recently, analysis.
2. Deterministic safety analysis - i.e., assess a prescribed list of failures which are selected
based on past experience and judgement. Sometimes these are called “design basis ac-
cidents”, as discussed in Section Error! Reference source not found.. Each accident se-
quence is chosen to be severe enough that the consequences of a “real” accident should
be less; this means that only a subset of possible accidents is analyzed. For example, the
Emergency Core Cooling (ECC) system is sized to provide enough flow to refill the core
after a break in the heat-transport system that is up to twice the flow area of the largest
pipe and on a time scale that prevents excessive fuel damage. The consequences of
these stylized accidents are predicted and compared against analysis limits. Such analy-
sis limits can be loosely based on frequency.
3. Probabilistic safety analysis - i.e., assess the frequency and consequences of failures,
optimizing to deal with the high-risk contributors. PSAs therefore proceed using the fol-
lowing methodology:
a. define risk-based criteria,
b. generate a set of accidents to consider,
c. predict the frequency and consequences of each event,
d. show that the criteria are met.
Much of the rest of this Chapter covers these three methods.
1.9 Problems
1. International bodies set limits for the dose an individual should receive from all man-
made sources. A number of issues exist behind this framework. Discuss the following four
questions and draw reasoned conclusions:
a. How should exposure from radiation used for medical purposes be controlled
(i.e., what factors should determine whether or not, and how much, radiation
should be used)?
b. Should large power reactors have the same dose limits as small research reactors
such as the McMaster Nuclear Reactor (which also produces medical isotopes)?
Why?
c. You are a safety expert and have been asked to approve a smoke detector. As-
sume that the smoke detectors give a whole-body dose [USNRC, 2001] of 10-5
mSv per year to each of the 35,000,000 people in Canada. What would your de-
cision be, and why? What factors would you look at in normal operation of the
detector? What is the most severe accidental exposure that can happen with a
smoke detector, and how would you assess its acceptability on a risk basis?
d. What dose would you accept for voluntary lifesaving (i.e., your colleague is
trapped in a very high radiation field and you are asked to go in and save him)?
Give reasons.
2. A nuclear designer is trying to optimize his design. He knows of an accident with a
frequency of 10-7 per year which leads to a contained core melt and causes the following
effects:
a. Permanent damage to the plant (i.e., cannot be recovered)

b. Evacuation of nearby people (5,000) for three days
c. No prompt fatalities
d. A collective dose to the closest population of 100 Sv.
He can reduce the frequency (but not the consequences) of this accident by a factor of ten by
putting in an extra heat-removal system costing M$10 in capital costs and an extra $100,000 per
year in maintenance and operating costs. How would you make this decision in a quantitative
way? Hint: Consider expressing all accident consequences in terms of dollars. Then calculate
the total average annual costs in each scenario. Some organizations assign a cost per Sv of
dose—find out typical values. This problem will take you some time.
3. A massive spontaneous failure (i.e., due to a material flaw, not as a result of a core melt)
in an LWR pressure vessel would simultaneously breach all the physical barriers which prevent
radioactive material from escaping: the fuel, the primary-coolant pressure boundary, and the
containment. Research and describe the approach taken by LWR designers to show that this is
“incredible”. Does this issue apply in any way to CANDU?
4. A nuclear regulator is considering a high-level safety goal for new nuclear power plants
in Canada. He proposes two requirements:
a. The risk to an individual close to the nuclear power plant of dying immediately
from an accident must be less than 10-6 per year.
b. The risk to an individual close to the nuclear power plant of getting cancer from
an accident must be less than 10-5 per year.
Two nuclear power plants apply for a licence. Each has done an accident analysis, and the
results are as follows:
1. For plant 1, no significant releases occur for any accident above a frequency of
10-7 per year. However, there is a core melt at that frequency which fails con-
tainment and gives a dose of 10 Sv to each individual in the nearby population.
2. For plant 2, two accidents are the major contributors to risk. One causes severe
fuel damage, but prevents core melt. It occurs at a frequency of 10-4 per year
and gives a dose of 0.25 Sv to a number of individuals in the nearby population.
The other is a core melt, but it is contained—it occurs at a frequency of 10-6 per
year and gives a dose of 1 Sv to a number of individuals in the nearby population.
Determine numerically whether these plants meet either or both safety goals or neither one.
Hint: consider converting average dose to risk.
2 Design Basis Accidents

In Section 1.8, the concept of stylized accidents (design basis accidents) that could be used in
designing parts of the plant was introduced. This section shows how to go about defining these.
More specifically, design basis accidents are the set of accidents for which the designer makes
explicit provision (defence), while remembering that more severe or peculiar accidents can
occur and ensuring that his/her design has some capability to deal with them.
Unfortunately, there is no way of identifying possible accidents beforehand (in any large-scale
engineering field) that is guaranteed to be complete. The history of any technology is replete
with unpleasant surprises, especially at the beginning—just think of the Hindenburg disaster,
the Flixborough cyclohexane explosion [Venart, 2004], and of course the Titanic. Technolo-
gies—if we are fortunate—have their accidents early on, when the scale is small and the lessons
learned can be applied to commercial applications.
The best way to obtain a “nearly complete” list of accidents, apart from experience, is to use
more than one technique. In the next few subsections, we describe several such techniques.
2.1 Top-Down Approach

One technique is called the “top-down” approach, which starts by specifying an undesired
outcome and then asks what the direct and immediate causes of that outcome could be. Then
one looks at each cause in turn and asks what are its direct and immediate causes, and so on.
After a few such cycles, one has generated a list of accidents, from which design basis accidents
can be selected.
For a power reactor, the top event could be taken as “unwanted movement of radioactive
materials”. Because most radioactive materials are in the fuel, the coolant, the moderator, or
the spent-fuel bay, we then ask how in each of these cases they could become mobile, i.e.,
airborne or waterborne. Radioactive material could be released from the fuel by overheating or
mechanical damage; from the coolant and moderator, by pipe breaks or overheating (which
later then releases the liquid or steam through relief valves); and from the spent-fuel bay, also
by overheating. Fuel overheating in the core (power-cooling mismatch) can be caused by loss of
heat removal from the coolant (loss of heat sink), loss of the coolant itself, coolant flow impair-
ment, or a loss of reactor reactivity control which causes the power to rise.
Figure 5 illustrates the event-generation sequence graphically. Events in ovals represent possi-
ble end-points of the chain, where further detailed decomposition would not add much more
information, i.e., the ovals are potential design basis accidents. Once we have reached this
reasonable level of resolution, we can sort the events into design basis accidents and beyond
design basis accidents (BDBAs), using expected likelihood, and possibly consequences, as
criteria. Note that the figure is both highly simplified and incomplete, with a number of unde-
veloped branches.
Figure 5 Simplified top-down approach
2.2 Bottom-Up Approach

Another approach is to look at each component of the nuclear power plant and ask what is the
consequence if this component fails. If systems exist which are supposed to protect the reactor
against such a failure, what are they, and what happens if they also fail? Eventually, if the power
plant is well designed, one gets to a very low frequency and draws the boundary between
design basis and beyond design basis accidents again. This is called a “bottom-up approach”.
Another term is FMEA, which stands for “failure modes and effects analysis”, although this
approach tends to be more limited because it stops after the first failure.
Sample Problem: Do a bottom-up analysis of the feed-water system in a nuclear power plant.
(The reader is assumed to be familiar with CANDU design—if not, please consult the earlier
Chapters in this book.)
Answer: The feed-water system can fail in a number of ways: power to the pumps may fail, or a
feed-water line may break, or the feed-water valve(s) may inadvertently close. Taking each
feed-water system component in turn and assuming that it fails, we then identify the systems
which are there to protect against such failures—e.g., the auxiliary feed-water system, the
shutdown cooling system, and the emergency water system (these are described in Section 5).
We then look at the failure of each of these mitigating systems. From these results, using
judgment to determine frequencies, we can select likely candidates for DBAs, shown in ovals in
Figure 6. The un-terminated arrows represent further development of the diagram.
Figure 6 Simplified bottom-up example

A combination of the top-down and the bottom-up methods will give a large number of acci-
dents which cover most possible events. However, the treatment of frequency is not very
precise.
2.3 Probabilistic Safety Analysis

Probabilistic safety analysis (PSA) is a rigorous tool for identifying accidents and assigning
frequencies to them. PSA uses a top-down approach to generate the failure frequencies of
operating systems and the failure probabilities of safety systems (fault trees). It then combines
failure of each operating system with successive failures of the required safety systems to obtain
a long list of accidents and their frequencies (event trees). One can also select design basis
accidents from this list if they have not already been picked up by other techniques. This topic
is covered in more detail in Section 4.
2.4 Experience
Design basis accidents may also be added because of experience.
For example, Canadian practice requires that each design basis accident be analyzed assuming
complete failure of one shutdown system, no matter how low the frequency. The reason goes
back to 1952, when the core of the NRX research reactor in Chalk River, Ontario, was damaged
in an accident ([Lewis, 1953] [Hurst, 1953]) in which the power increased and the shutdown
system was impaired. One of the causes was a complex control/shutdown system design; the
shut-off rods were hydraulically driven, and their performance was sensitive to dirt in the
system. The accident resulted in a large subsequent emphasis on shutdown-system reliability,
testability, and robustness, to the extent that, even though the shutdown systems in CANDU
bear no resemblance to those in NRX (lessons having since been learned about shutdown-
system design), the CANDU reactor had to be designed to survive an accident even if the
shutdown system failed. Although one could show that the most severe accident without
shutdown (a large LOCA) would not release enough energy to break containment, the designers
decided (eventually) to add another, fully independent shutdown system, so that even if one
shutdown system failed completely, the other could be credited.
The accident at Fukushima in 2011 [JNTI, 2011] is a recent example of a poor choice of the
parameters for a design basis accident. Although the plant had a Design Basis Flood level, it was
clearly inadequate, and since the accident, all operating plants have had to change their provi-
sions for both Design Basis Floods and Beyond Design Basis Floods. We shall cover key historical
events in more detail in Section 3.
2.5 Canadian Approach to DBAs

We do not have space to cover the entire historical development of the Canadian approach to
DBAs. See [Snell, 1978] for an overview. Instead, we shall describe the basis for selection of
DBAs for currently operating CANDUs and planned future plants.
2.5.1 Siting guide

In 1972, D. G. Hurst and F. C. Boyd of the Atomic Energy Control Board (AECB) — the name at
the time of the nuclear regulator (now CNSC) — laid the ground rules for the deterministic
licensing guidelines under which all large operating CANDU plants up to but excluding Darling-
ton have been licensed [Hurst, 1972]. The spectrum of possible design basis accidents was
collapsed into two broad categories: single failures, or the failure of any one process system in
the plant, and dual failures, a much less likely event defined as a single failure coupled with the
unavailability of either a shutdown system, or containment, or the emergency core cooling
system: these constituting the so-called special safety systems. For each category, a frequency
and a consequence limit was chosen that had to be satisfied. In addition, to deal with the siting
of a reactor (Pickering A) next to a major population centre (Toronto), population dose limits
were defined for each category of accident.
The limits were as follows (sometimes this is called the Siting Guide):
Table 1 Single / dual failure limits

Accident Maximum Fre- Individual Dose Limit Population Dose
quency Limit
Single Failure 1 per 3 years 0.005 Sv 102 Sv
0.03 Sv thyroid 102 Sv thyroid
Dual Failure 1 per 3000 years 0.25 Sv 104 Sv
For example, loss of reactivity control, loss of Class IV power, and a loss-of-coolant accident are
all single failures; loss of coolant plus failure of the ECC, or loss of coolant with failure of the
containment isolation dampers to close, are dual failures. In this framework, both single and
dual failures are design basis accidents.
Safety-system demand unavailability can be inferred from the frequency limits in this table:
because a dual failure is a single failure plus unavailability of a safety system and must occur no
more often than one in 3000 years, each safety system must fail no more often than 1 in 1000
times, and therefore the demand unavailability is 0.001.
This approach had a number of deficiencies, such as lumping events with widely differing
frequencies into the same class—e.g., a large pipe break or a loss-of-coolant accident (LOCA)
and loss of off-site power both fell into the single-failure category. Combinations of higher-
frequency events were also not addressed.
2.5.2 Consultative Document C-6

To address some of the deficiencies in the single-dual failure methodology for design-basis
accidents, the AECB issued document C-6 in June 1980 [AECB, 1980]. This retained the concept
of classes of events, five in this case. As with the Siting Guide, the classes roughly grouped
events based on frequency, but the assignment of events to classes was done by AECB staff
based on their beliefs about the likelihood of the event. Design basis accidents included, for
example:
Class 1: failure of reactivity or pressure control; failure of normal electrical power; loss of feed-
water flow; loss of service-water flow; loss of instrument air; and a number of other events that
one might expect to occur occasionally.
Class 2: feeder-pipe failure; pressure-tube failure; channel-flow blockage; pump-seal failure;

other events that would not be expected to occur more than once (if that) in a plant lifetime.
Class 3: large LOCA; earthquakes and other events that are rare and could damage the fuel or
portions of the plant.
Class 4: Class 1 events + unavailability of a special safety system
Class 5: Class 2 or 3 events + unavailability of a special safety system, e.g., LOCA plus ECC
impairment.
Dose limits were defined for individual members of the public only (see Table 2).
Table 2 Consultative document C-6 limits

Event Expected Frequency Whole-Body Dose (Sv) Thyroid Dose (Sv)
Class per reactor-y [Charak,
1995]3
1 > 10-2 0.0005 0.005
2 10-2 to 10-3 0.005 0.05
3 10-3 to 10-4 0.03 0.3
4 10-4 to 10-5 0.1 1
5 < 10-5 0.25 2.5
2.5.3 RD-337
As the nuclear industry has become more international and more competitive, the Canadian
Nuclear Safety Commission (CNSC) has understood the need to align its requirements, espe-
cially for new builds, more closely with international ones and has therefore developed top-
level design requirements [CNSC, 2008] which align more closely with IAEA standards [IAEA,
2000] and are less technology-specific. Events are divided into three classes: anticipated
operational occurrences (AOOs), which are expected to occur at least once in the plant lifetime;
design basis accidents; and beyond design basis accidents (BDBAs), including event sequences
that may lead to a severe accident. The limits are from [CNSC, 2008] and [CNSC, 2008a].
Table 3 Dose limits
Event Frequency Dose Limit (Sv)
Anticipated Operational Occurrence  10-2 / reactor-year 0.0005
Design Basis Accident 10-2 to 10-5 / reactor-year 0.020
3
Expected frequency ranges are not part of C-6; they were used by Ontario Hydro in the licensing of Darlington to
classify events not listed in C-6.
There are no dose limits for BDBAs, but there are numerical safety goals and specific system
design requirements; see Section 4.
2.6 Other Design Basis Events

A complete list of design basis accidents includes external hazards: earthquakes, fires, torna-
does, tsunamis, floods, etc. The magnitude and frequency of each hazard are site-dependent.
The unique aspect of these hazards is that they can affect more than one system at the same
time.
Design basis accidents also include man-made hazards, both internal (operator error, sabotage)
and external (explosions from nearby industrial or transportation facilities, terrorism). Defining
what should be the “design basis” and “beyond design basis” for malevolent acts is the respon-
sibility of the national government.
2.7 Problems
The reader should have read and understood Chapters 3, 4, and 5 on reactor physics before
proceeding.
1. A laboratory experiment has been set up to determine the critical mass of enriched
uranium. Two hemispheres of U235 metal are supported in an unshielded facility by one scien-
tist. Other scientists are in the same room, observing. A screwdriver is used to slowly push one
hemisphere closer to the other, while a neutron detector measures the increase in neutron flux
as they approach each other (Figure 7). (This scenario is modelled on, but is not quite the same
as, the Lewis Slotin accident in 1946 [LANL, 2000]). Develop a safety approach using the con-
cept of design basis accidents as follows:
a) Use both “top-down” and “bottom-up” approaches to define a set of accidents. Specifi-
cally: What is the “top event” that is to be avoided? What could cause the accidents?
b) How fast do the accidents occur (i.e., what physical process determines the time scale)?
What inherently limits the consequences (why don’t you get a nuclear bomb)?
c) Compare the nature of the hazard to the scientists with that to the public.
d) How could the consequence of an accident be prevented or mitigated:
 Without any further equipment, i.e., just after it has occurred?
 With equipment installed beforehand?
Figure 7 Criticality experiment

2. Calculate the “risk” in Sv/year to an individual at the site boundary from a (not very
good) reactor designed and operated so it exactly meets the dose limits in:
a) The two classes of accidents in the single/dual failure approach
b) Event classes 1 to 5 inclusive from Consultative Document C-6
c) The AOO and DBA limits from RD-337.
What conclusions can you draw (if you think the comparison is not meaningful, explain). What
relative contribution do the more severe accidents make to the risk?
3. Consider the small reactor for urban district heating shown in Figure 8 [Snell, 1989]. It is
intended to be located in urban areas in buildings such as hospitals or universities [Currie,
1985]. Salient safety-related characteristics are:
 pool reactor, natural circulation, atmospheric pressure
 double-walled pool (350,000 litres) with a purification system (small pump and ion-
exchange resins, outside the pool)
 10 MW(th) output
 forced-flow secondary side, with a heat exchanger immersed in the pool
 tertiary heat exchanger connected to the heating grid (why?)
 negative reactivity feedback from fuel temperature, coolant temperature, and cool-
ant void (e.g., an increase in coolant temperature decreases the reactivity)
 active reactor-control devices (rods) with limits on rate (a few mk/hour, compared
to, say, CANDU, which can go up to several mk/minute) and worth (no rod in excess
of a couple of mk).
 low fuel temperatures, so that there are no free fission products in the fuel
 two shutdown systems—one passive system actuated by a signal (drops the control
rods) and one fully passive system (rods within the core which are thermally acti-
vated: the absorber material inside the rods, normally above the core, melts and fall
into the core on high coolant outlet temperature)

 a confinement boundary (not shown in the figure) covering the pool top; however,
the building is conventional
 no emergency core cooling system (why?)
 a licensed operator is not required to be in the control room. Any upset sounds an
alarm which notifies a local attendant (who can shut the reactor down, but not re-
start it). Licensed operators can remotely monitor the reactor, but not control it.
Develop a set of design basis accidents for this reactor. It is important to show how you did this,
not whether you get the same answer as the designer did (there is not really enough informa-
tion given in the exercise to get the “right” answer—it’s your thinking process that counts). If
you are getting design basis accidents which seem inconsistent with an urban location, how
could they be made impossible?
Figure 8 SES-10
4. Consider a low-energy research reactor used to determine fundamental physics parame-

ters. It consists of a vertical cylindrical heavy-water tank in which are suspended fuel assem-
blies (Figure 9 and Figure 10). Salient safety-related characteristics are (somewhat simplified):
 pool reactor, natural circulation, atmospheric pressure
 nominal zero energy (a few watts), no engineered heat-removal systems
 low fuel temperatures, with very few fission products in the fuel
 fuel rods suspended from hangars which can be arranged manually into different lat-
tice pitches and geometries. Fuel rods are stored beside the pool.
 capability to use fuel with a large range of enrichment ratios (but not highly irradi-
ated fuel)
 provision for insertion of a few channels consisting of fuel inside a pressure tube
containing electrically heated coolant at high pressure and high temperature inside a
calandria tube (but still nominally ~zero nuclear power)
 control through moderator level (pump-up and drain); pump-up speed limited by
pump capacity
 manual start-up and shutdown
 three redundant dump valves which open to trigger a heavy-water dump on high
neutron power or high log rate
 no emergency core-cooling system and no containment. A cover provides shielding
of operators when the reactor is critical.
a) Develop a set of design basis accidents for this reactor. It is important to show how you did
this, not whether you get the same answer as the designer did (there is not really enough
information given in the exercise to get the “right” answer, it’s your thinking process that
counts). Start from a long list developed using at least two of the techniques discussed in
this Chapter and then suggest which accidents you would consider too rare to design
against, and why. Provide details, e.g., it is not enough to say “increase in power”; list all the
ways that this could occur.
b) If you wanted to reduce the risk from this reactor (based on your list of design basis acci-
dents and a judgment about probability), what design changes would you make first?
Figure 9 ZED-2 cutaway

[AECL, 2011]
Figure 10 ZED-2 top view
3 Experience
Current power reactor safety design has been powerfully influenced by accidents that hap-
pened during the development of the technology. In this Section, we describe a few of them—
space constraints do not permit an exhaustive study.
Real accidents almost never follow the simple assumptions in design basis safety analysis. Most
real accidents tend to be very complex, have more than one contributing factor, have a high
component of human error and extraordinary human recovery, and often leave one wondering,
“How did all these things happen together?”
The answer is, of course, that if they didn’t happen together, there would be no accident. It is a
poorly designed nuclear power plant where a single failure or error causes a catastrophic
accident. Most accidents are the end-point of a chain of events, prevention of any one of which
would have stopped the accident from happening. We will discuss this later in Section 5.1
(Defence-in-depth).
Figure 11 Learning and forgetting

The fact that such chains have occurred provides powerful lessons learned for future designs
and current operation.
In general, when any new technology is introduced, its characteristics are not fully known, and
there is an early peak in the accident rate. As people learn from mistakes in design and opera-
tion, the accident rate declines, as shown in Figure 11. See also [Ott, 1981] for a more detailed
discussion. The decline does not, however, account for organizational forgetting, and if nothing
is done to preserve corporate or industrial memory, the rate will start to rise again. One of the
purposes of case studies is to ensure that past mistakes are not repeated.
What you are expected to obtain from this section is the ability to dissect an accident. What
was its nature? What were the root causes? What were the lessons learned for design? For
operation? The exercises at the end of the chapter give you some additional actual incidents to
assess.
You should have read and understood Chapters 3, 4, and 5 on reactor physics before proceed-
ing.
Each of the case studies described below gives a brief description of the facility, the event, and
lessons learned. These are all of necessity very abbreviated. In particular, of the dozens of
lessons learned, only the few most important are listed. The reader should consult the refer-
ences for more detail.
3.1 Criticality Accidents and Power Excursions

The earliest accidents in nuclear reactors were criticality accidents. This is not surprising
because most early reactors were research reactors, which were small enough that if the
reactor was shut down, the decay heat could be removed by the water and by structures
surrounding the core for long periods of time. Loss of heat sink was not a major issue. There-
fore, the main safety concern was avoiding unwanted criticality and shutting down quickly if it
occurred. This concern was emphasized by the large worth of some of the reactivity-control
devices: because reactors were small, space for control rods was limited, and refuelling was
infrequent, the control rods had to compensate for burnup and tended to have high reactivity
worth.
A number of reactivity accidents have taken place when the reactor was supposed to be shut
down. An initially sub-critical reactor is not as safe as it sounds. The power is fairly constant
and proportional to the number of source neutrons divided by the absolute value of the (nega-
tive) reactivity [Cameron, 1996]:
Se
Nf  , (3)

where:
Nf is the number of neutrons,
Se is the source term (neutrons produced per unit of time, either from spontaneous fis-
sion, fission product decay, photoneutrons or an artificial neutron source), and
ρ is the reactivity, negative in this case.
However, the reactor is not under active “control”. Detectors are insensitive, and indications
that the reactor might be near-critical are not very obvious. In some cases, the reason for the
shutdown is to do maintenance on the control or shutdown systems, which means that they are
not available. Then safety depends on ensuring an adequate shutdown margin (net negative
reactivity) by procedural means, i.e., inserting a number of absorber rods, or removing fuel, or
adding a liquid absorber to the moderator or coolant, or removing the neutron reflector sur-
rounding the core. Safety also depends on maintaining the shutdown margin over time. This is
the reason for the guaranteed shutdown state in CANDUs: to have a large enough negative
margin that the reactor is resistant to operator mistakes or equipment failures. It is also good
practice, as we shall see from the case studies, to have an additional means of shutdown always
poised, even during shutdown.
You may want to think about whether it is safer to refuel a reactor while it is critical or during a
shutdown.
3.1.1 SL-1
3.1.1.1 Description
This summary uses material from [USAEC, 1962], [USAEC, 1962a], [USAEC, 1962b], [Stacy, 2000],
and [Thompson, 1965].
The SL-1 (Stationary Low Power Reactor No. 1) was a natural-recirculation pressurized boiling-
water military reactor with a thermal power of 3 MW. It was located at the Atomic Energy
Commission National Reactor Testing Station in Idaho Falls, Idaho. Figure 12 shows a cutaway of
the reactor vessel. At the time of the accident, on January 3, 1961, there were 40 fuel assem-
blies and 5 control rods in the core.
3.1.1.2 The event

Three operators were engaged in reassembling one of the control-rod drive mechanisms while
the reactor was shut down. Part of the assembly process required raising the shaft of the rod a
limited amount by hand. This was done by unclamping the rod and raising it; the rod was then
re-clamped, and the only remaining task was to unclamp and lower the rod. For some unknown
reason, the rod was instead raised rapidly, making the reactor super-prompt critical. The
pressure vessel jumped up nine feet, with the chain reaction being compensated for (probably)
initially by the negative feedback due to heating of the fuel and moderator and then by the
melting and vaporization of the fuel itself.
Two operators were standing on the lid at the moment of the accident. One was thrown to one
side as the vessel rose. The other was impaled on a shield plug ejected from the top of the
reactor and was carried up to the roof of the reactor room, where he remained suspended. The
third man was killed by radiation and flying debris. Table 4 gives a chronology of the events.
Figure 12 SL-1 cutaway
Table 4 SL-1 chronology

Time Event
-500 msec Central control-rod withdrawal starts.
-120 msec Reactor goes critical with rod at 16.7 in. (40.6 cm).
Central rod at 20 in. (50.8 cm), period = 3.9 ±0.5 msec, (2.4 ± 0.3)% Δk.
0 Peak power burst (1.9 ± 0.4) × 104 MW.
~2 msec Prompt nuclear energy release ends; total nuclear energy of excursion =
(133±10) MW-sec [+(24 ±10) MW-sec in metal-water reaction].
20% of plate area destroyed; centre 16 elements 50% melted; central shroud
and control blade ejected from core.
Water column above core accelerated by average pressure (500 psi or 35 atm)
to a velocity of 160 ft/sec (49 m/sec).
34 msec Water slams against lid of vessel. Maximum pressure ≈10,000 psi (~700 atm).
Head shielding ejected. Plugs ejected with velocity of 85 ft/sec (26 m/sec) or
less. Vessel rises, shearing connecting pipes. Guide tubes collapse. Nozzles
and vessel expand.
160 msec First plug hits ceiling.
Two-thirds of water expelled. 5%–10% of fission products expelled.
Vessel hits ceiling. Total kinetic energy involved ~1% of total energy released.
Insulation ripped from vessel.
2000–4000 Vessel comes to rest in support cylinder.

msec
3.1.1.3 Lessons learned

The philosophy of reactivity design is vitally important. The fuel in SL-1 had boron strips at-
tached to the fuel; these were supposed to burn up with the fuel to keep the core reactivity
within the control range of the control rods. These strips had burned up faster than planned,
and in addition, pieces of boron had fallen off the fuel during operation. The net result was that
the core became more reactive to the point that it could be made critical with four rods fully in
the core and the central one partially withdrawn. The criticality position of the central rod at
the time of the accident was 16.7 inches withdrawn; the extra 4 inches that the rod was with-
drawn above this value added more than four times the delayed neutron fraction β, giving a
reactor period of 4 msec. Clearly, once the transient had started, no operator action and no
shutdown system could stop it in time.

Many lessons were learned from SL-1; here we summarize the major ones concerning reactivity
control:
1. It should be impossible for a reactor to be made critical by withdrawing the single most
effective rod. Conversely, it should always be possible to shut down a reactor with the
single most effective rod stuck in its outermost position. This so-called “single-rod rule”
has been followed ever since in safe reactor design.
2. It should not be possible to withdraw a high-worth rod quickly. It is still unknown why
the operator withdrew the rod. In any case, the lesson is that if a reactor requires high-
worth rods, there must be mechanical means to prevent their rapid withdrawal.
3. In a reactor where large amounts of reactivity can be added in short times, there must
be inherently fast negative reactivity feedback, for example from fuel temperature, so
that the transient is stopped short of reactor damage. This lesson was later applied in
commercial light-water reactors.
Note that the total worth of reactivity devices in CANDU is quite low because on-power refuel-
ling means that the movable control devices do not have to compensate for burnup. Moreover,
the dispersed CANDU core enables the reactivity worth to be spread among many individual
reactivity devices.
3.1.2 NRX
3.1.2.1 Description
This summary uses material from [Lewis, 1953], [Hurst, 1953], [Larson, 1961], and [Cross, 1980].
NRX was Canada’s first large research
reactor, built in 1947 with a thermal power
of 30MW. The moderator was heavy water
contained in a cylindrical calandria. Cool-
ing was once-through, light-water, taken
from and returning to the Ottawa River.
Passing through the calandria were vertical
tubes open to the air at top and bottom.
Each fuel rod, made from metallic uranium
sheathed in aluminum, had its own cooling
jacket (Figure 13); each rod, with its cooling
jacket, was located in one of the vertical
tubes (Figure 14). A stream of air passed Figure 13 NRX fuel cross section
between each fuel rod and its calandria
tube. Twelve boron shut-off rods passed through 12 of these tubes. Start-up was achieved by
removing half these rods, after which the reactor was controlled by varying moderator level.
The shut-off rods could be raised by compressed air and were then held up by an electromag-
net. They could be driven back in using high-pressure air, although they would fall more slowly
under gravity if released.
Figure 14 NRX elevation

The 12 rods were divided into six groups, or banks, of rods, ranging from one to four rods per
bank. Normally a “safeguard bank” was held in reserve: that is, bank 1 containing four rods was
the first to be removed during a start-up and was designed to overcome any reactivity added if
the reactor should go inadvertently critical on removal of any of the remaining banks of rods.
Removal of the safeguard bank was prohibited by limit switches on each rod unless all other
rods were down; however, the switches had been giving trouble and had been disabled.
To ensure fast start-up after a shutdown (and avoid xenon poison-out), it was possible to drive
up the first four rods, worth 30 mk, in a few seconds.
There were four pushbuttons on the control panel. Pushbutton 4 charged air to the heads of
the shut-off rods; release of this air drove the rods down. Pushbutton 3 temporarily increased
the current in the electromagnet and ensured that the rods were properly seated. Pushbutton
1 raised the safeguard bank, and Pushbutton 2 raised the remaining banks in automatic se-
quence. Normal operation was to press Pushbutton 3 along with 4, 1, and 2. If 3 was not
pressed with 4, air might leak from the head system; if 3 was not pressed with 1 and 2, the shut-
off rods might not be drawn fully home, and the safety circuits would prevent start-up from
proceeding.
The void reactivity was positive and large; that is, removal of all the light-water coolant would
result in an increase in reactivity of 25 mk.
3.1.2.2 The event

At the time of the accident, the reactivity of certain fuel rods was being measured at low power.
The cooling to these rods was reduced, and one was being cooled only by air.
The following sequence of events is taken from [Cross, 1980]:
“The accident occurred during a start-up procedure. Just as the first group of shut-off
rods was about to be removed, an operator in the basement of the building (who had
nothing to do with the start-up) mistakenly turned some air-valves which caused several
shut-off rods to rise. This was immediately shown by the indicator lights in the control
room. The reactor supervisor phoned the operator to stop and went down to the base-
ment himself to make sure that the valves were properly reset. When this was done the
rods should have gone down into the reactor. In fact, they went down only partway, but
far enough that the lights in the control room indicated that they were down.”
This failure to reinsert is not explained in any of the published reports. Cross continues:
“The supervisor in the basement phoned the control room and told his assistant to press
two numbered buttons. He gave the wrong number for one of the buttons, and when it
was pressed, instead of resetting the air pressure as was intended, it raised four more
shut-off rods. If the first group of shut-off rods had been down, as their lights indicated
they were, raising four rods was a reasonable thing to do, so the mistake was not recog-
nized.”
Specifically, the supervisor asked the assistant to press buttons 4 and 1. This would charge the
air to the heads and raises the safeguard bank. He had meant to say 4 and 3, which charges the
air to the heads and seats the rods. However, because button 3 was not pressed, the air to the
heads brought up by button 4 leaked away, and the safeguard bank (which appeared to have
been up already) was raised. The supervisor realized his mistake, but the assistant in the
control room had to leave the phone to use both hands to push the two buttons.
From [Cross, 1980]:
“It was soon apparent from instruments in the control room that the reactor was above
critical and the power level was rising. This was surprising, but not alarming, since the
reactor could easily be turned off by dropping the shut-off rods just raised. However,
when after 20 seconds, the button was pressed to do this, only one of the four rods ac-
tually went down. The power level continued to climb and, after some discussion in the
control room, it was decided to dump the moderator into a storage tank. Within less
than 30 seconds, the power-level metres were back on scale and the power dropped
rapidly to zero.”
The removal of the safeguard bank made the reactor supercritical by about 6 mk and started a
power rise to about 100 kW. The lone shut-off rod started to fall in at this point, but the power
kept rising to about 17 MW. Boiling then occurred in some of the temporarily cooled rods,
expelling the light water and increasing the reactivity by another 2.5 mk. Power continued to
rise to 60–90 MW, when it was stopped by dumping the moderator.
The power surge melted a number of fuel rods and caused a number of calandria tubes to fail.
Eventually, in a major operation, the reactor calandria was removed and buried; the building
was decontaminated, and the reactor was replaced.

1. Some negative reactivity should always be held in reserve (safeguard bank). Before a
reactor is made critical, the safeguard bank is removed and poised, so that if anything goes
wrong, it can be quickly reinserted. In CANDU, both shutdown systems 1 and 2 are poised
before the reactor is allowed to go critical using the control devices.
2. The shut-off rods must be of a simple design. After NRX, designers mistrusted shut-off rods
as a safety shutdown mechanism; hence, NPD and Douglas Point used moderator dump as
their emergency shutdown system and had no safety rods. Pickering had a few safety rods
which acted in conjunction with the moderator dump. Only by the time of Bruce A were
shut-off rods restored as a primary shutdown mechanism. They were, of course, very dif-
ferent from those in NRX: they consisted of a rod on a cable, which was wound around a
pulley and held in place by an electromagnetic clutch. On a shutdown signal, current to the
clutch was cut off, and a spring assisted the rod to drop by gravity. It was, and is, a simple
and reliable design with large clearances.
3. The control and shutdown systems should be separated and independent, with the latter
devices being used only for shutdown and held outside the reactor during critical operation.
4. An accident, even with a shutdown failure, should not be a public-health disaster. This
meant that either the reactor containment had to withstand the energy released, or (much
later on) two shutdown systems would have to be provided so that failure to shut down af-
ter an accident would be a very low-probability event.
3.1.3 Chernobyl
This summary uses material from [USSR, 1986], [USDOE, 1986], [Howieson, 1987], [Chan, 1987],
[Rogers, 1987], and [INSAG, 1992].
3.1.3.1 Description
Chernobyl unit 4 was of the RBMK type (Реактор Большой Мощности Канальный, or High-
Power Channel-Type Reactor) and the most recent of the 1,000 MW(e) series. It was a graphite-
moderated, boiling-light-water-cooled, vertical-pressure-tube design using enriched (2% U235)
UO2 fuel with on-power refuelling. It used a direct cycle to produce electricity from twin
turbines (Figure 15, from [INSAG, 1992]). The reactor cutaway is shown in Figure 16, from
[Semonov, 1983].
There are two independent primary coolant circuits, each containing about 830 fuel channels,
two steam separators, and four pumps (with one normally on stand-by). Refuelling is done
during operation from the top of the core. A containment structure encloses the inlet piping in
the lower portion of the reactor and provides pressure relief to a water pool located beneath
the reactor (Figure 18). Control and shutdown are performed by movable absorbing rods in
lattice positions. Emergency core cooling is provided for pipe breaks through a system consist-
ing of a combination of pressurized-water accumulators and electric pumps.
The reactor has a void coefficient which varies from negative to positive according to the
operating state. It is limited in normal operation through control of operating conditions.
The containment is a pressure-suppression type (see Section 5) and is compartmentalized
(Figure 18). All inlet pipes are enclosed in leak-tight compartments which connect to a bubbler
pond or suppression pool below the reactor. Steam discharged from a pipe break is directed to
the bubbler pond, where it condenses. Each channel is accessible (for refuelling) from the top
through a large cover plate, to which its extension is welded (Figure 16). The (small) outlet
pipes above the core are in an enclosure designed for a rupture of up to two channels or their
outlet pipes, because rupture of more than one was believed to be highly improbable. As noted
in [INSAG, 1992], “Simultaneous failure of a greater number of fuel channels would generate a
pressure high enough to fail the containment function by lifting the cover plate, in the process
severing the remainder of the fuel channels”.
3.1.3.2 The event

The accident resulted from a test to show that after a main pump trip, the small amounts of
steam generated from the low residual reactor power could provide enough electricity through
the turbine generators to extend the rundown time of the main coolant pumps. The test was to
be done while the reactor was critical. However, shortly after pump trip, the reactor power
started to increase rapidly, could not be terminated, and caused the violent destruction of the
reactor.
The test was supposed to have been done at a starting power of 700 MW(th). However, it was
initiated at a power level of 200 MW(th). The reasons are somewhat lengthy, but during the
transfer from the local power control system to the main range automatic power, an unplanned
drop to zero neutron power occurred and lasted for four to five minutes. The operator, instead
of aborting the test, then attempted to recover power, but because of xenon decay, had to raise
the control rods so that most of them were out of the core. In the end, the reactor never
reached the planned initial power level.
Key: 1: graphite reactor core; 2: fuel channel; 3: in-core instrumentation tube; 4: feedwater
pipe; 5: steam separator drum; 6: downcomer; 7: intake header; 8: main circulating pump; 9:
bypass; 10: high-pressure header; 11: stop valve; 12: distribution group header; 13: steam
header; 14: steam dump valve; 15: accident localization system; 16: ECCS water reserve tank;
17: pressure controller; 18: turbogenerator; 19: condenser; 20: moisture separator/reheater;
21: condensate pump; 22: preheater; 23: deaerator; 24: emergency electric feedwater pump;
25: electric feedwater pump; 26: mixing preheater; 27: condensate collecting tank; 28: moisture
separator/reheater condensate pump; 29: level controller; 30: ECCS hydraulic accumulator; 31:
ECCS pump; 32: ECCS header; 33: ECCS fast-acting valve; 34: leak limiter.
Figure 15 RBMK schematic diagram
In the immediate aftermath of the accident, the Soviets [USSR, 1986] placed much blame on the
operators for not following test procedures and performing the test in an unauthorized reactor
configuration. The initial cause of the power rise was stated to be the increase in void reactivity
as coolant boiling increased following pump trip. It was implied that because most of the dual
control/safety rods were withdrawn, the reactor trip was ineffective in stopping the power rise.
Later analysis ([Chan, 1987] and [INSAG, 1992]) indicated that a major cause of the accident was
the unexpected behaviour of the shut-off rods, which actually caused an increase in reactor
power rather than shutting the reactor down. This surprising design aspect requires more
detailed explanation. The absorber rods have graphite displacers or followers attached to their
lower ends which are designed to increase the absorber reactivity worth. As they are inserted,
the absorber rods move into the high-flux region in the centre of the core, which was previously
occupied by the graphite, so that the absorber-rod effectiveness is enhanced (see Figure 17). If
no graphite were present, the rod would displace water—also an absorber—in which case the
change in reactivity with insertion would not be as great.
However, during the accident, most of the absorbers were well removed from the core, which
was an abnormal configuration. The flux was peaked at the top and bottom, where most of the
reactor power was being generated. Therefore, when absorber insertion first started, the water
in the high-flux region at the bottom of the core was first displaced by the graphite follower,
leading to a reactivity increase. Therefore, operating the plant in an abnormal condition re-
sulted in an unusually large hold-up of void reactivity, which combined with a deficient shut-
down-system design, led to the large power excursion and the resulting core damage. [INSAG,
1992] indicates that this deficiency was known by designers before the event.
The power rise failed a number of fuel channels; the resulting steam release blew off the top
cover and severed the remaining fuel channels, effectively exposing the core to the environ-
ment and by-passing the containment.
The core damage rendered it subcritical, although to make sure this was the case, the Soviets
used helicopters to dump boron into the open reactor structure. The graphite moderator, when
exposed to air and heated by the fuel fragments, began to burn and did so for several days until
it was smothered by a combination of sand dropped onto the core pit and liquid nitrogen
pumped in from underneath. The core melted and flowed into the rooms below the reactor
vault, although this was not realized until some years afterwards, when the reactor cavity was
inspected remotely and found to be empty. A temporary cover (the sarcophagus) was rapidly
built over the damaged structure to prevent radioactive material from escaping and is currently
being replaced by a more permanent cover.
Figure 16 RBMK reactor
In terms of health effects ([UNSCEAR,

2000], [UNSCEAR, 2008]), 28 prompt
deaths occurred among the reactor
operators and the firefighters called
in to extinguish fires on the turbine-
hall roof, arising from the ejected
fuel. Some increase in leukemia has
been observed among the members
of the clean-up crew. Among the
public, approximately 6,000 excess
thyroid cancers have been observed
in children, almost all of which are
curable. No other cancers have been
observed at statistically significant
levels in the population.

1. The effectiveness of shutdown
systems should never be impaired
by the operating state of the reac-
tor. The shutdown systems
should be fast enough and inde-
pendent enough to overcome any power Figure 17 Reverse shutdown in RBMK
transient, even from an abnormal operat-
ing state.
2. The concept of “safety culture” was used to address inadequacies in the processes of
design, operations, and regulation in the USSR, which contributed to the accident. We shall
cover this in Section 8. Briefly [INSAG, 1991], “Safety culture is that assembly of characteris-
tics and attitudes in organizations and individuals which establishes that, as an overriding
priority, nuclear plant safety issues receive the attention warranted by their significance.”
3. Some discussion took place about the role of the containment (Figure 18). It is possible that
a complete containment envelope as used in Western PWRs and CANDUs, even if not de-
signed for a reactor runaway, would have mitigated the release of radioactive material, but
this has not been demonstrated. At the very least, one can conclude that the upper con-
finement design basis (rupture of two channels) was not robust in terms of defence-in-
depth.
Figure 18 RBMK building cross section
3.2 Loss of Cooling / Heat Removal

Shutting down a power reactor does not eliminate the hazard; the decay heat in the fuel, if not
removed, will eventually cause the fuel to melt. We now look at two such events: Three Mile
Island and Fukushima Dai-ichi.
3.2.1 Three Mile Island

This summary uses material from [Kemeny, 1979], [Yaremy, 1979], [Rogovin, 1980], and [Brooks,
1980].
3.2.1.1 Description
The Three Mile Island reactor was a conventional pressurized-water reactor; a flow-sheet is
shown in Figure 19. The reactor core is inside a large pressure vessel; unlike most CANDUs and
BWRs, the coolant is highly sub-cooled and kept so by a pressurizer connected to the pressure
vessel. A pressure relief valve on top of the pressurizer provides overpressure protection for the
reactor coolant system (RCS).
Figure 19 TMI schematic
3.2.2 The event

On March 28, 1979, a maintenance error resulted in loss of feed water to the steam generators
when the reactor was on power. Because there was no heat sink for the primary coolant, the
coolant pressure began to rise, and the pressurizer relief valve opened to limit the pressure, as
it was supposed to do, following which the reactor tripped (shut down automatically). As the
pressure fell, the emergency core cooling system activated to make up for lost water, again as
designed. The pressurizer relief valve should have re-closed, but did not. This led to a small and
continuing loss of coolant through the open valve, which went unrecognized for hours. The
result was that the loss of inventory uncovered the core, leading to a partial core melt, hydro-
gen production from oxidation of the fuel sheaths at very high temperatures, a hydrogen burn
inside containment, and a partial melt-through of the pressure-vessel wall at the bottom.
However, the wall remained intact. Finally, the failed valve was recognized and then isolated by
closing a blocking valve, the main pumps were turned back on to re-pressurize the system, and
high-pressure make-up water was added to refill it. Figure 20 shows the end state of the core
[Smithsonian, 2004].
A stuck-open pressure-relief valve (PRV) is

uncommon, but not rare. What converted
this operating transient into a partial core
melt was that the operators did not realize
that the valve had stuck open, partly because
the valve indicator in the control room was
derived from the valve actuating signal, not
the actual valve-stem position, and partly
because the pressurizer level was rising due
to the loss of coolant from the top—the
opposite of what one would expect for loss of
coolant elsewhere in the RCS. The operators
had been trained when operating a “solid”
system (no boiling) to use pressurizer level as
a measure of circuit inventory, and they came
to the erroneous conclusion that there was
too much water in the system. If water is
added to a sub-cooled system, the pressure
can rise very fast indeed when the steam Figure 20 TMI core end state
space in the pressurizer is filled up, and the
operators were trying to avoid this. Their actions over the next several hours were complex and
confused, clearly indicating that they did not understand what was happening and did not have
the training or the tools to do so.
The containment did its job. Small amounts of noble gases and iodine were released through a
leakage recovery system (which could not be isolated because that would also stop primary
pump seal cooling) and by venting the make-up tank to control its pressure. It is a credit to the
containment design concept that a partial core melt resulted only in very small releases.
In addition, this was a “wet” accident, i.e., the radioactive material had the opportunity to
interact with quantities of water in the reactor vessel and containment. This is in contrast to
Chernobyl, where the moderator caught fire. Fission products such as iodine and cesium
combine when released from overheated fuel to form cesium iodide (an aerosol). However,
when CsI meets water, it dissociates:
CsI  Cs  I  ,
and is thereafter very difficult to remove (compare this to trying to remove chlorine after
dissolving table salt (NaCl) in water). TMI taught that in terms of public safety, “wetter is
better”.
The lifetime health effects on the surrounding population were predicted to be effectively zero.

1. A number of obvious improvements were needed to valve position indicators, event log-
ging, and alarm prioritization (the printers were swamped by the number of alarms,
running far behind the event), as well as provision of diagnostic aids to the operators.
2. The prescriptive approach to reactor safety in the United States, which was based on
regulatory-driven lists of design basis accidents, had focused attention on too narrow a
range of accidents, not just because it omitted severe accidents, but also because it fo-
cused on bounding design basis accidents and treated superficially the supposedly less-
severe DBAs such as a small LOCA.
3. PSA became far more important and more widely used to define and therefore prevent
possible severe accidents.
4. Sharing of prior experience with other plants was poor. TMI led to the establishment of
the Institute of Nuclear Power Operations (INPO) for industry-wide cooperation in op-
erations and training.
3.2.3 Fukushima Dai-ichi

Much of this material is taken from official reports or presentations by the Japanese Govern-
ment [Japan, 2011], the Tokyo Electric Power Company (TEPCO) [TEPCO, 2011a], the Japan
Nuclear and Industrial Safety Agency (NISA) [NISA, 2011], the International Atomic Energy
Agency (IAEA) [IAEA, 2011], and the United States Nuclear Regulatory Commission (USNRC)
[USNRC, 2011]. Vast amounts of detailed material exist on this event—it is probably the most
photographed accident in history—and many details are still unclear or unknown, e.g., the exact
state of the cores in Units 1, 2, and 3 and the containment status.
3.2.3.1 Description
There are six nuclear power reactors at the Fukushima I (Dai-Ichi) site. All are boiling-water
reactors (BWRs). There are four more BWR reactors nearby at Fukushima II (Dai-Ini). Figure 21
(from [USNRC, 2011]) shows the overall station layout of Fukushima Dai-Ichi.
Mark 1 type BWRs have an “inverted light-bulb” containment (the drywell) connected to a
toroidal suppression pool or chamber (the wetwell), as can be seen in Figure 22. The purpose
of the suppression pool is to condense any steam released inside the containment in an acci-
dent, such as a loss-of-coolant accident, and also to act as a temporary or permanent heat sink
for decay power in case the main heat-removal mechanisms are unavailable. If the suppression
chamber (S/C) is to act as a long-term heat sink, it must be cooled by an external circuit, e.g., for
Unit 1 (which we shall use as the example), the isolation condenser.
Figure 21 Fukushima Dai-ichi before earthquake

The containment vessel is surrounded by a reactor building, which, among other things, con-
tains and supports the elevated spent-fuel pool. The reactor building is not designed to with-
stand significant pressure.
The IAEA [IAEA, 2011] notes that:
“In Unit 1, the Isolation Condenser (IC) is designed to operate through gravity driven
natural circulation of coolant from the reactor pressure vessel (RPV) through a heat ex-
changer immersed in a large tank of water in the reactor building at an elevation above
the core. The Unit 1 IC was designed to have a decay heat-removal capacity of about
eight hours. A valve must be manipulated to bring the IC into service.”
Figure 22 BWR Mark 1 containment
3.2.3.2 The event

At 14:46 on Friday, March 11, 2011, a magnitude 9 earthquake took place off the eastern coast
of Japan. The earthquake occurred at a depth of 24 km and initiated a tsunami which was
estimated to be 14 m high when it hit Fukushima approximately 40 minutes later, followed by
multiple additional waves. The measured on-site accelerations due to the earthquake were
greater than the design basis for Fukushima I, but less than the design basis for Fukushima II.
Although the earthquake caused significant on-site damage and loss of all off-site electrical
power, in both stations, all operating units were automatically shut down, and the emergency
diesel generators started and functioned until the tsunami hit.
The design basis tsunami for Fukushima I was O.P. (base level) +5.7m; the actual height was O.P.
+14–15 metres. Figure 23 shows a cross section of the station and the flood level.
Figure 23 Design basis versus actual flood level

The emergency diesel generators at Fukushima I (which are cooled by sea-water) were all
disabled by the flood; because off-site AC power was already lost, the result was a total loss of
all AC power at Fukushima I, units 1 through 5; hence, all motor-operated pumps became
inoperable. In addition, lighting in the main control room was lost. The sea-water system (used
to reject heat to the ultimate heat sink) was likewise inoperable. Unit 6 had one functioning air-
cooled diesel generator, which was cross-tied after the flood, with some effort, to Unit 5 and
was instrumental in achieving cold shutdown in those units with no fuel damage. In addition,
because the emergency batteries could no longer be charged, control and instrumentation
power in Units 1–4 was eventually also lost, crippling the ability to monitor the reactor state or
to operate motorized valves.
Taking Unit 1 as an example (the sequence of events in other units was similar, but the timing
was somewhat prolonged), the isolation condenser was initially used to remove decay heat.
However, the isolation condenser stopped operating fairly soon thereafter. With no core heat
sink, the water in the vessel began to boil away, with the steam condensing in (and heating up)
the suppression pool and the level in the reactor vessel falling, eventually exposing fuel. As the
fuel overheated in a steam environment, the zirconium cladding oxidized and produced hydro-
gen.
The challenges were then:
 to restore water over the fuel
 to prevent containment failure due to over-pressure
 to control and release the hydrogen.
About 15 hours after the earthquake (March 12 at 05:46), the operators managed to inject fresh
water using fire pumps, followed about 13 hours later (March 12 ~19:00–20:00) by direct sea-
water injection. To control containment pressure, the primary containment vessel (PCV) was
apparently vented on March 12 at 10:17. This appears to have allowed the hydrogen in con-
tainment to migrate to and accumulate in the reactor building; it exploded at 15:36 on March
12, destroying much of the superstructure of the building.
It appears from analyzes done with severe accident codes that the fuel in Units 1, 2, and 3
melted a few hours after being uncovered and may have penetrated the reactor vessel and the
primary containment. (This could also be deduced from fairly simple bounding calculations.)
The current state and location of the core and the integrity of the reactor vessel and the pri-
mary containment are not known with certainty, although damage to both is expected.
A further issue was the condition of the fuel in the spent-fuel pools. Normally, the decay heat
generated by the fuel bundles in the pools is removed by active systems, which of course were
lost when all AC power was lost. In the absence of heat removal, or if water is lost due to
leakage, heating of the spent fuel is slower than heating of the core, but if not addressed, will
eventually lead to uncovering of the fuel and fuel damage. In Fukushima, the spent-fuel bays
are located high in the reactor building to facilitate transfer of fuel assemblies from the core.
Spent-fuel cooling was made more difficult by the elevation (and inaccessibility) of the spent-
fuel bays, and concerns were voiced about their structural integrity following the earthquake.
In the early phases after the accident, water was added manually to the bays (initially from
helicopters and later from cranes); stable cooling has since been restored to all pools. Inspec-
tion of the bays shows debris from the damaged reactor buildings; TEPCO has stated, “Most
spent fuels estimated to be undamaged” based on the radioactive contents of the pool water
and remote inspection. As of December 2013, fuel in the Unit 4 pool is now being removed.
Fukushima II was not flooded nearly as severely as Fukushima I and successfully achieved cold
shutdown of all units.
The safety goal is to achieve cold stable shutdown, with temperatures below 100°C and any
releases of radioactive materials from containment under control.
The injection of sea-water, while presumably stopping further core melting, led to a large
accumulation of highly radioactive water in the basement of the buildings because cooling was
essentially “once-through”. Some of this water intermittently leaked to the sea and to ground.
Once-through cooling was later converted to closed-loop cooling and continuously decontami-
nated using on-line filtration. In addition, to prevent further release of airborne radioactive
material from debris or leakage to ground or sea from rain, the reactor buildings are being
covered by weatherproof covers.
Finally the large amounts of highly contaminated water on site have to be treated and securely
stored.

Although it is still early after the accident to formulate complete lessons learned, it seems
reasonable that they will include the following issues:
1. Ensure completeness of design basis accidents, especially external events, and in par-
ticular those with a cliff-edge effect like flooding (dramatic worsening of consequences
for a small change in the severity of the event).
2. Ensure that resources are available in the long term after a severe event to control, cool,
and maintain: for example, portable generators with pre-designed hookups to key
equipment, flexible hoses that can be run to sources of water.
3. Ensure prolonged core cooling in the absence of off-site power and failure of on-site die-
sel generators.
4. Ensure prolonged spent-fuel bay cooling or make-up and monitoring after a prolonged
loss of power.
5. Prevent loss of monitoring (e.g., design the station emergency batteries to last for a sig-
nificant length of time, protect them against external events, and provide recharging ca-
pability). Maintain on-site and off-site communications.
5. Provide facilities on site to deal with physical destruction and high radiation fields (e.g.,
bulldozers to clear debris).
6. Review lessons learned about institutional behaviour, e.g., the role of the regulator. See
Chapter 16 for a case study.
7. Examine the impact on accident evolution of having multiple units at one site.
3.3 Problems
1. Browns Ferry Fire, 1975:
a. Locate (as a minimum) the following source material (try the USNRC web site):
Fire at Browns Ferry Nuclear Plant; Tennessee Valley Authority; March 22, 1975 -
Final Report of Preliminary Investigating Committee, May 7, 1975.
b. Assess the Browns Ferry fire in terms of threats to the ability to control, cool,
contain, and monitor.
c. Indicate lessons learned. Describe how you derived these lessons learned. Do
not just copy official lessons learned.
d. Comment on the robustness of the design and indicate whether design or oper-
ating changes should have been considered (and why).
2. Tokai-Mura, 1999:
a. Locate (as a minimum) the following material: International Atomic Energy
Agency, Report on the Preliminary Fact Finding Mission Following the Accident at
the Nuclear Fuel Processing Facility in Tokaimura, Japan; IAEA report, 1999.
b. Assess the accident in terms of threats to the ability to control, cool, contain, and
monitor.
c. Indicate lessons learned. Describe how you derived these lessons learned. Do
not just copy official lessons learned.
d. Comment on the robustness of the facility design and operation, and indicate
whether design or operating changes should have been considered (and why).
4 Safety Goals and Risk Assessment

In this section, we expand on the concept of risk introduced in Section 1.8. We start off by
defining numerical safety goals, which are a means of quantifying risk, and then present some
probabilistic safety assessment tools that can be used to calculate whether or not the plant
meets the safety goals.
4.1 Safety Goals

A safety goal partially answers the question “how safe is safe enough?” In this sub-section, we
will develop a worked example of how a safety goal might be derived and then compare it to
safety goals adopted by other organizations. We shall also point out some of the pitfalls of
using a safety goal as the only safety criterion. The intent of this section is that the reader
should be able to understand how safety goals are derived, what their limitations are, and how
they are used in design.
To enable meaningful decisions, a safety goal should be expressed in quantitative terms. A
safety goal such as “Make the reactor as safe as possible” will mean different things to different
people and provides no guidance to the designer or operator. A goal such as “The reactor must
never have a severe accident” is probably physically impossible and sets up expectations that
cannot be met.
It is not easy to define a safety goal. Here is a possible starting point for a safety goal which has
numerical requirements (and which is often used as the basis for regulatory decision-making):
“The annual risk of death to the most exposed member of the public due to accidents in a
reactor should be small in comparison to his/her total risk of premature death.”
Some of the aspects of this safety goal are:
 It uses a common risk measure (annual risk of death). Although this is objective, it may
not be all that relevant to nuclear accidents: experience has shown that significant pub-
lic dose from a severe accident develops slowly, giving time to move people out of
harm’s way. The results of real events have shown few public health effects.
 It compares the risk from a nuclear reactor to all other risks of premature death. This is
not at all obvious. One could set up the comparison to all other risks of electricity gen-
eration, or all energy generation, or all (average) industrial activity. In fact, the first two
would be much more logical because they compare risks of producing the same product
(energy).
 No relative benefits are mentioned.
 The safety goal (in this example) is for the most exposed individual. It does not consider
social effects such as radiation exposure to a large number of individuals, evacuation,
land contamination, and effects on the environment such as on animals and plants. The
assumption in this safety goal is that protection of the most exposed individual member
of the public also provides sufficient protection for society at large and for the environ-
ment.
 The goal refers to members of the public, not workers in the plant. It is generally ac-
cepted that people will implicitly accept a somewhat higher risk to life and limb if it
comes as part of their job, i.e., if a direct benefit is obtained. Attempts have been made
(e.g., in the United Kingdom, see [HSE, 2006]) to set (nuclear) safety goals for plant
workers in nuclear power plants, but the risk to such workers is dominated by conven-
tional industrial risk, which itself turns out to be less than in comparable non-nuclear in-
dustries.
 The goal refers to the risk of nuclear power in isolation. Just as there is a risk to having
nuclear power, there is also a risk to not having it, because the electricity would then
have to be generated from other sources with greater (or lesser) risk.
The goal does not distinguish between prompt and delayed risk of death. We can break down
the safety goal proposed above into two sub-goals:
“The annual risk of prompt death to the most exposed member of the public due to ac-
cidents in a reactor should be small in comparison to his/her total annual risk of prompt
death due to all accidents”,
and
“The annual risk of fatal cancer to the most exposed member of the public due to acci-
dents in a reactor should be small in comparison to his/her total annual risk of fatal can-
cer due to all causes.”
4.1.1 Risk of prompt death

In Canada in 2009, accidents were the fifth leading cause of death over the whole population, at
a rate of 30.4 deaths for every 100,000 people [StatsCan, 2009]. This means that the average
person’s risk of death from an accident is ~3 × 10-4 per year (note that the rate for males is 50%
higher than that for females). We could then say that the risk from a nuclear power plant of
premature death to an individual should be small compared to 3 × 10-4 per year, say by a factor
of 100 (this may be too conservative, but we will use it for illustration), or 3 × 10-6 per year.
Because the only way of causing prompt fatalities to the public in a nuclear accident is through a
core melt and failure of containment, and even that will likely not cause prompt public fatalities
(given the experience of Chernobyl and Fukushima), this suggests that our safety goal should
be:
“The likelihood of a large release from a nuclear power plant in an accident should be
less than 3 per 106 reactor years”.
(Note that to risk prompt deaths, a large release by itself is not sufficient; it must also be sud-
den, otherwise people can be evacuated beforehand.) These changes have yielded a goal that
can be used in design: it is relatively straightforward using PSA to calculate the likelihood of a
large rapid release (core melt plus early failure of containment), to see (if the goal is exceeded)
where the dominant contributors are, and to fix them if needed to meet the safety goal.
4.1.2 Risk of delayed death

Cancer fatalities can be considered in a similar fashion. In the same year in Canada (2009),
malignant neoplasms were the leading cause of death over the whole population, at a rate of
211 deaths per 100,000 people. Therefore, the average person’s risk of dying from cancer is 2.1
× 10-3 per year (or about 16% over a 75-year lifetime). Recall from Chapter 1 that 100 person-Sv
will produce about five fatal cancers in the exposed population, or a risk of 5 × 10-2 fatal cancers
per Sv. Therefore, using the linear dose-effect hypothesis, 211 cancer deaths per 100,000
people would be induced by a collective dose D calculated as follows:
D(Sv) × 5 × 10-2 (cancers / Sv) = 211 cancer fatalities, (4)
which means that D = 4220 Sv, or an average individual dose of 42 mSv. However, the linear
dose-effect hypothesis is not applicable to such low doses. Regardless, if we divide by 100 again,
then the maximum time-averaged individual dose from accidents should be less than 0.4 mSv
per year, averaged over a group of people, or about 25% of natural background radiation in
Toronto (1.6 mSv / year). The paradox with this safety goal is that it would make nuclear power
safer than natural background radiation.
This is not as useful a safety goal as the previous one because it does not tell us anything about
the frequency distribution of accidents. However, it can likewise be validated by summing all
the events in a PSA which cause a release of radioactive material.
At a more basic level, summing low doses over large numbers of people is incorrect, and a
safety goal derived this way is fundamentally flawed.
4.1.3 Safety goals in Canada

The CNSC has recently published safety goals, for new reactors built in Canada. The rationale is
similar to our sample case. Specifically, CNSC states [CNSC, 2008]:
“A limit is placed on the societal risks posed by nuclear power plant operation. For this
purpose, the following two qualitative safety goals have been established:
1. Individual members of the public are provided a level of protection from the conse-
quences of nuclear power plant operation such that there is no significant additional risk
to the life and health of individuals; and
2. Societal risks to life and health from nuclear power plant operation are comparable to
or less than the risks of generating electricity by viable competing technologies, and
should not significantly add to other societal risks.”
These are developed into design goals:
1. “Core Damage Frequency: The sum of frequencies of all event sequences that can lead
to significant core degradation is less than 10-5 per reactor year
2. Low Release Frequency: The sum of frequencies of all event sequences that can lead to a
release to the environment of more than 1015 Becquerel of iodine-131 is less than 10-5
per reactor year. A greater release may require temporary evacuation of the local popu-
lation.
3. High Release Frequency: The sum of frequencies of all event sequences that can lead to
a release to the environment of more than 1014 Becquerel of cesium-137 is less than 10-6
per reactor year. A greater release may require long term relocation of the local popula-
tion.”
This is largely consistent with international practice for new reactors, with one new concept: the
low release frequency. It is intended to address those accident scenarios which may result in
small but significant releases. These accidents may require emergency measures such as
sheltering or short-term evacuation of an area around the plant, and the low release frequency
sets a limit on these [Rzentkowski, 2013].
An approach to safety based only on a safety goal has both benefits and limitations. The
greatest benefit is that risk-based decisions require greater levels of protection on those areas
of greatest risk, and conversely, i.e., they optimize safety resources. Some limitations are:
 To determine compliance with a risk target, all significant events have to be identified
and summed. This might be challenging in the early phases of a design.
 Safety goals are meaningful only for events for which frequencies and consequences are
reasonably calculable. In practice, this includes most “internal” events for which actual
data exist, for which the failure combinations can be calculated, or for which a reason-
able extrapolation from the historical record can be made. However, if the design has
innovative features, with little operating experience, it may be difficult to support the
reliability values and hard to spot the cross-links4. Passive safety systems pose a par-
ticular challenge in this regard because they can be difficult to test, and therefore it is
hard to build up a reliability database.
 Not all (rare) events can be assigned a frequency and consequence with confidence, for
example:
o massive structural failure
o massive failure of pressure vessels
o very low-frequency, high-consequence external events such as earthquakes be-
yond historical record-keeping
o sabotage, terrorism, and war.
The approach to the first two is usually to design to accepted engineering codes and standards.
Then one can infer from experience the likelihood of sudden failure of structures and compo-
nents so designed. Massive failure of a LWR pressure vessel would be a catastrophic event
because it would lead to an immediate release of fission products and probably would damage
containment at the same time. Calculations have been done to show that such massive failures
are less frequent than 10-8 per year, but such low frequencies must be treated with some
skepticism. Rare events can happen by unanticipated sneak paths; for example, a precursor to a
pressure-vessel boundary failure was the undetected almost-through-wall erosion of the
reactor pressure vessel in the Davis-Besse plant [USNRC, 2002]. Similarly, historical records
make it possible to define the intensity of earthquakes down to about the ten-thousand-year
return frequency, which is taken as the design basis earthquake (DBE) in Canada. More severe
but rarer earthquakes are hard to characterize. One can carry out a “seismic margin” analysis to
calculate the likelihood of surviving an earthquake somewhat more severe than the DBE; much
beyond that, about all one can say is that the effects of damage to the nuclear plant would be
4
A cross-link is a failure that affects more than one system or component, e.g. a common supply of instrument air.
small compared to the havoc wreaked by such an earthquake on the rest of society, as was the
case for Fukushima. Finally for events resulting from hostile human actions, the approach has
generally been to design according to rules (e.g., the plant’s inherent defences plus the local
security force should be able to delay an attack of x people armed with y type of weapons for z
minutes); x, y, and z are indeed chosen based on reasonableness (likelihood), but the historical
database of hostile acts against nuclear power plants is sparse. In any case, the defences being
built into new plants for severe accidents are also helpful against malevolent acts.
For these reasons, those regulators that have safety goals use them in addition to whatever
deterministic criteria they have developed. This combination is called a risk-informed approach,
in which the quantified risk provides a powerful rationality check.
4.2 Risk Assessment

This section summarizes the probabilistic safety assessment concepts and tools used to demon-
strate numerically that the plant meets its safety goals, as well as for many other purposes. For
basic references, see [McCormick, 1981] and [USNRC, 1981].
4.2.1 The basics

First, you may want to refresh your memory by reading Appendix 1 – Basic Rules of Boolean
Algebra and the rules about the probability of multiple events. The following text assumes you
already know, or have read, all this material.
Sample Problems:
a) A nuclear reactor has two fully independent shutdown systems. The probability that a
shutdown system will fail on demand is 1 in 1000. What is the probability that they will both
fail in an accident?
Answer:
Let A1 be the event where shutdown system 1 fails and A2 be the event where shutdown system
2 fails. These are stated to be independent events, so:
P(A1 A2) = P(A1) P(A2) = 10-3 × 10-3 = 10-6.
(Sometimes it is hard to ensure that redundant systems are completely independent).
b) As above, but what is the probability that either will fail in an accident?
Answer:
We can use the rare events approximation because P(A1) and P(A2) are <<1.
Here, we want P(A1 + A2):
P(A1 + A2)  P(A1) + P(A2) = 10-3 + 10-3 = 2 × 10-3.
c) Now consider two independent diesel generators, each with a probability of failure to start
of 15%. What is the probability that either will fail to start?
Answer:
We cannot really use the rare events approximation here, so:

P(A1 + A2) = P(A1) + P(A2) - P (A1A2) = 0.15 + 0.15 – (0.15)*(0.15) = 0.28.
4.2.2 Bayes equation

The Bayes equation is typically used in nuclear power applications to predict the likelihood of a
rare event given the history of few (or no) such events. We will prove it first and then give some
examples which make it clearer.
Given an event or hypothesis, B, and An mutually exclusive events or hypotheses (n=1, 2....N):
P( An B)  P( An ) P( B | An )  P( B) P( An B)
, (5)
 P( B | An ) 
 P( An | B)  P( An )  . (6)
 P( B ) 
Because the events An are mutually exclusive,
N
 P( A
n 1
n | B)  1 . (7)
Multiplying by P(B):
N N
P ( B )   P ( B ) P ( An | B )   P ( An ) P ( B | An )
n 1 n 1
. (8)
Substituting (8) into (6)

P( An ) P( B | An )
P ( An | B )  N
. (9)
 P ( A ) P( B | A
m 1
m m )
This is Bayes’ theorem. Therefore, if we know P(B|An), we can calculate P(An|B). A couple of
examples will make this clearer.
Sample Problems
Pipe Inspection
Suppose that you are radiographing a Class I pipe for a defect. You know from past experience
that the likelihood of a defect is one per 100,000 radiographs. You also know that the likelihood
that the instrument indicates a defect when there is no defect (false positive) is 1%, and the
likelihood that the instrument indicates a defect when a defect in fact exists is 99%. Your test
indicates a defect. What is the probability that the pipe actually has a defect when the instru-
ment says it does?
Answer:
Apply Bayes’ theorem to two events:
A: pipe has a defect, therefore P(A) = 0.00001

B: instrument says that pipe has a defect, therefore P(B)=0.011
B|A: instrument says pipe has a defect when it has a defect, therefore P(B|A) = 0.99
What we want is P(A|B), the probability that the pipe actually has a defect when the instrument
says it has one.
Using Bayes’ theorem:
P(A|B) = P(B|A)[P(A)/P(B)]
= 0.99 × 0.00001 / 0.01
= 0.00099
because the denominator in Equation (9) is just P(B), which in this case is known. The result
seems counterintuitive and suggests that the test is not very good in detecting defects, despite
the instrument’s high accuracy rate. However, the fact that the defect is so rare (we need about
a hundred thousand samples before we have a good chance of seeing a real positive) magnifies
the small false positive rate, so that most positive tests are false positives.
This is quite important in medical tests; even a very accurate test for a rare cancer will often
give far more false positives than real ones.
4.2.2.1 Failure rate estimation when no failures have occurred

We can use Bayes’ equation to glean information from non-events as well [Kaplan, 1979].
Large LOCA
There have been 14,800 reactor years of experience without a large break LOCA. What is an
upper estimate of the frequency?
Answer
Let B = 14,800 reactor years of experience without a large break LOCA (LBLOCA).
What we do now is take (say) six cases, in each of which we hypothesize the value of the LLOCA
frequency. We then use Bayes’ theorem to test how good our hypotheses are (i.e., calculate the
probability that each hypotheses is correct). We label our hypotheses A1 to A6.
1
This is a slight simplification using the fact that P(A) is small. Actually
P(B) = P(B|A) P(A) + P(B|Ᾱ) P(Ᾱ)
= 0.99 × .00001 + 0.01 × 0.99999
= 0.0100098
or approximately 0.01 as stated.
In words, the first term is the 99% chance of detecting the defect in the one pipe in 100,000 that has the defect;
the second term is the 1% chance of indicating a false positive in the remaining 99,999 pipes out of 100,000.
A1 = LBLOCA probability = 10-2 / year

If A1 were true, then:
P(B|A1) = (1 – 10-2)14800 = 2.5 × 10-65.
Because we can assume that each reactor-year of experience is independent, the probability of
each reactor-year without a LBLOCA is 1 – 10-2, and P(B|A1) is just the intersection of 14,800
events.
Likewise, we find that:
P(B|A2) = (1-10-3)14800 = 3.7 × 10-7
P(B|A3) = (1-10-4)14800 = 0.2276
P(B|A4) = (1-10-5)14800 = 0.8624
P(B|A5) = (1-10-6)14800 = 0.9853
P(B|A6) = (1-10-7)14800 = 0.9985
If we knew P(A1),...,P(A6), we could calculate P(An|B), or the probability that our statement An is
actually true. But we don’t know. Instead, we can make an assumption, called an a priori
probability: let’s just assume that P(An) = 1/N = 1/6, which is the case labelled “uniform prior” in
Table 5, i.e., that each hypothesis is equally likely. Then, using Bayes’ theorem, we find that
P(A1|B) = 8.14 × 10-66 , i.e., A1 is not very likely. From Table 5, we see that A2 is more likely than
A1, but that A4 to A6 are even more likely. If we use these results to postulate a more likely
series of P(An), labelled “non-uniform prior” in the table, we can see that P(An|B) is further
adjusted, and we can conclude that the LLOCA frequency is significantly less than 10-3. The
practical application of this is in assigning a reasonable upper limit to a frequency in a probabil-
istic safety analysis, when no such events have actually occurred and all we have is the number
of reactor-years of experience.
Note that one of the criticisms of Bayes’ Theorem when used this way is that the answer
depends on the appropriateness of the initial hypotheses. If few data are available and you put
in strange hypotheses, you get back strange answers. Note also that if no events have occurred,
Bayes’ Theorem does not give information about the lower limit of the frequency.
Table 5 Bayes’ theorem example

n 1 2 3 4 5 6
An 10-2 / year 10-3 / year 10-4 / year 10-5 / year 10-6 / year 10-7 / year
P(B|An) 2.5 x 10-65 3.7 x 10-7 0.2276 0.8624 0.9853 0.9985
Uniform Prior
P(An) 0.1667 0.1667 0.1667 0.1667 0.1667 0.1667
P(An|B) 8.14 x 10-66 1.20 x 10-7 0.0741 0.281 0.321 0.325
Non-Uniform Prior
P(An) 0.001 0.01 0.1 0.4 0.4 0.089
P(An|B) 2.27 x 10-66 3.36 x 10-9 0.0251 0.358 0.394 0.0808
4.2.3 Probability distributions

We will now develop the theory behind failure probabilities, failure rates, and reliability.
Let p(x)dx be the probability that an event occurs in an interval x to x+dx—the probability
density function. Let P(X) be the cumulative probability that the event occurs somewhere
between xmin and X. Then
X
P( X )  
xmin
p( x)dx  P( x  X ) , (10)
where p(x) is the probability density function. If p(x) is a constant, po, then P(X) = po(X-xmin), as
expected.
There are two types of systems:
1) Those that operate on demand (i.e., safety systems)
2) Those that operate continuously (i.e., process systems).
4.2.3.1 Demand Systems

We define:
Dn = nth demand
P(Dn) = probability of success on demand n
_
P(Dn ) = probability of failure on demand n
Wn = case where system works for each demand up to and including demand n
What is the probability that it works for n-1 demands and fails on demand n?
P(Wn1 )  P( D1D2 ...Dn 1 )

, (11)
P ( DnWn 1 )  P ( Dn | Wn 1 ) P (Wn 1 ) . (12)
Therefore,
P( D1D2 ...Dn 1Dn )  P( Dn | Wn 1 ) P (Wn 1 )
 P( Dn | D1D2 ...Dn 1 )  P( Dn 1 | D1 D2 ...Dn 2 )  ...  P( D2 | D1 )  P( D1 ) . (13)
If all demands are alike and independent, this reduces to:
P ( D1 D2 ...Dn 1 Dn )  P ( D )[1  P ( D )]n 1 . (14)
4.2.3.2 Failure dynamics

Failures can be time-dependent.
Let: f(t)dt = probability of failure at time t in the time interval dt
F (t )  cumulative failure probability
t
. (15)
  f ( )d
0
Assuming that the device eventually fails, the reliability, R(t) is defined as
R(t )  1  F (t )
 t
  f ( )d   f ( )d . (16)
0 0

  f ( )d
t
Therefore,
dR(t ) dF (t )
f (t )    . (17)
dt dt
Let λ(t) dt = probability of failure at time t given successful operation up to time t (i.e., the
conditional failure rate); then:
f (t )dt   (t )dtR(t )
or , (18)
dR(t )
f (t )   (t ) R(t )  
dt
and hence
dR(t )
  (t ) R(t ) . (19)
dt
Solving and using the boundary condition R(0)=1:

t

  ( ) d
R(t )  e 0
. (20)
If λ is constant, (i.e., random failures):
R(t )  e t . (21)
A typical λ(t) is shown in [McCormick, 1981], page 26—the “bathtub” curve (Figure 24).
Figure 24 Typical λ(t) versus time
Table 6 summarizes the relationships among these quantities.

Table 6 Reliability terms and relationships
Description Symbol = = =
Hazard rate λ(t) 1 dR f (t ) f (t )

R dt 1  F (t ) R(t )
Reliability R(t) 
1-F(t) t

 f ( )d
  ( ) d
t
e 0
Cumulative failure probability F(t) t

1-R(t) t

 f ( )d
  ( ) d
0
1 e 0
Failure probability density f(t) dF (t ) dR(t ) λ(t) R(t)


dt dt
4.2.3.2.1 Mean time to failure (MTTF)

This is simply the time-weighted average over all time of the failure probability density:

  f ( )d 
MTTF= 0

   f ( )d
 f ( )d
0
0
(22)

since  f ( )d  1
0
(a component will eventually fail).
If λ is constant, then

1
MTTF    e   d  . (23)
0

4.2.3.2.2 Availability, A(t)

Availability is the proportion of the time that a system is in a functioning condition. If no repairs
have been made, it is the same as reliability. If a system has been repaired, then its availability
will satisfy:
R(t) ≤ A(t) ≤1.
Assume random failures. This implies that λ = constant, and therefore
R(t) = e-λt = reliability,
as illustrated in Figure 25. Eventually, the reliability goes to zero for long times.
Conversely, the failure probability is
F(t) ≡ 1 - R(t) = 1 - e-λt,
as illustrated in Figure 25, and goes asymptotically to one for long times.
Let repair occur at some time interval, τ. This resets the failure probability to zero each time the
component is repaired, assuming that the repair is perfect and restores the component to its
original condition. Then F(t) is a sawtooth curve, as illustrated in Figure 26.
If τ << λ, then
 2t 2
F (t )  1  (1  t   ...)  t for t< in any interval , (24)
2
with t measured after the time of last repair.
Hence:

F , (25)
2
i.e., the failure probability is half the test interval times the failure rate.
Figure 25 Reliability for constant λ
Figure 26 Availability with repair

For constant λ, but τ not << λ, the following applies:
 
  e    1
F   F (t )dt   (1  e  t )dt  . (26)
0 0

A common task is to design a system (composed of components that have a known failure rate)
to meet some target unavailability (1-A) or Ᾱ. Given a design, the repair interval is the remain-
ing variable. A frequent repair cycle (low τ) gives a low Ᾱ, but such frequent repair may be
impractical due to excessive cost, radiation exposure, or wearout (i.e., repair does not restore
the equipment to an as-built state). In such a situation, alternative designs would have to be
considered.
Often repair may not be required to return F to zero. It may be sufficient simply to test the
components to ensure that they are available. This is usually the case for “demand” systems.
Sample Problem
Consider the case of a single shut-off rod (SOR) for a reactor. Given a failure rate based on
previous experience of λ = 0.002/year and a required unavailability of ≤10-3, what is the re-
quired test period, τ?
Answer:

A  0.001 . (27)
2
To meet the Ᾱ target of 10-3, we solve for τ:
103
  1 year . (28)
0.001/ yr
4.2.3.3 System Unavailability

Of course, a real shutdown system has many more rods, say, N in total. Let En be the event
where all rods except for n drop successfully. The system as a whole is designed to meet its
requirements for the cases of E0, E1, E2—i.e., if zero, one, or two rods fail to drop. Hence, the
system unavailability Ᾱ is the sum of all the failure probabilities:
N 2
A   P( Ek )  1  P( Ek ) . (29)
k 3 k 0
The conversion to the second term makes the calculation much easier because there are fewer
terms to sum. Assuming that the rods fail independently and that the failure rate is λ, then the
probability of a given rod failing on average is:

 F   p for conciseness (30)
2
as before. The success probability is 1-p. In general, the probability of event Ek is
N!
P( Ek )  (1  p )n  k p k . (31)
( N  k )!k !
N!
The factor gives the number of possible ways for the event to happen, the factor
( N  k )!k !
(1-p)n-k is the probability that N-k rods all successfully drop, and the factor pk is the probability
that k all fail to drop.
4.2.3.4 Fault trees

A more systematic way of calculating system unavailability is through fault trees. One starts at
the top with the event of interest, usually a system failure. Then one determines each and
every immediate cause of such an outcome. If either of several immediate causes is sufficient
to cause the top event, then they are joined by an OR gate (please review Appendix 1 – Basic
Rules of Boolean Algebra—for a refresher on probability theory).
Conversely, if all of several immediate causes must occur to cause the “top” event, then they
are joined by an AND gate. Both AND and OR gates can have more than two inputs.
In more detail, the steps in drawing and then calculating a fault tree are:
 Start (at the top) from the undesired event, e.g., loss of feed water.
 Carefully define the boundaries of the system to be evaluated.
 Using the principle of immediate cause, draw the tree downwards using AND, OR,
etc. Boolean logic.
 Stop at basic events for which we can no longer decompose the event or when we
arrive at a point where we know the probability of failure, e.g., failure of a relay.
 Assign probabilities to each event in the tree.
 Calculate the Boolean algebra to determine the numerical probability of failure of
the top event.
The last step is not conceptually difficult, but the amount of arithmetic can be huge. It is almost
always done using special-purpose codes for all but the simplest of fault trees.
One test of a properly drawn fault tree is that causality cannot pass through an OR gate—if it
does, an AND gate is missing. For an OR gate, the input faults are never the causes of the
output fault—they are identical to the output, but more specifically defined as to cause. For an
AND gate, there is a causal relationship between the inputs and the output—the input faults
collectively represent the causes of the output fault.
There are additional types of gates which we will not cover here. The U.S. Nuclear Regulatory
Commission Fault Tree Handbook [USNRC, 1981] is a superb source for learning fault-tree
analysis.
Sample Problem
Figure 27 Simple pumped system

Consider a simple (and rather badly designed) pumped system as shown in Figure 27. Each
pump can supply 100% of the necessary flow. The demand failure probabilities for each com-
ponent are:
Pump (P): 0.01
Valve (V): 0.01
Signal (S): 0.001
What is the unavailability of the system?
Answer
The first step is to define the system boundaries carefully. For this example, we will exclude the
water tank and the reactor. That means that we will not consider failure modes for the tank
(e.g., water freezing, tank drained, line opening blocked); in a real application, these would
need to be covered. We also have to define the failure criterion: less than 100% flow to the
reactor. We can then draw the fault tree. It may help to think the problem through in words
first:
Immediate cause of insufficient flow to reactor = no flow through valve V3
Immediate cause of no flow through valve V3 = (no flow through valves V1 OR no flow
through valve V2) OR valve V3 closed
Immediate cause of no flow through valve V2 = pump P2 fails OR valve V2 fails to open
Immediate cause of valve V2 fails to open = valve V2 mechanically fails OR signal S fails,
etc.
This results in the fault tree shown in Figure 28. Note that there are many ways of drawing this;
for example, one could recognize that the failure of the signal affects all components and
therefore could be in an “OR” gate just under the top event. It is always safer, however, to
follow the principle of immediate cause.
Having drawn the fault tree, we now assign labels to each basic event (V1, V2, V3, S, P1, P2), as
shown in Figure 28. We can also assign intermediate labels (A, B, C) to help in the calculations.
Now we can write out the Boolean algebra, starting from the bottom. It is important NOT to
assign numerical values immediately to the basic events and work out the final answer before
doing the Boolean algebra. In many cases, this will lead to double counting because terms such
as S+S will be counted twice. Remember that in Boolean logic, S+S=S, not 2S, and S●S=S, not S2.
Figure 28 Sample fault tree with labels

Let us calculate the chain up to point C first:

C=(S+P1)+(S+V1).
Similarly,
D=(S+P2)+(S+V2),
B = C●D = ((S+P1)+(S+V1))●((S+P2)+(S+V2)),
A = S+V3.
Top event = A+B
= (S+V3) + [((S+P1)+(S+V1)) ● ((S+P2)+(S+V2))]
= S+V3+(S+P1+V1)●(S+P2+V2) because S+S=S
= S+V3+S●(S+P2+V2)+ P1●(S+P2+V2)+V1●(S+P2+V2)
= S+V3+P1●P2+P1●V2+P2●V1+V1●V2 -- using S●S=S, S+S=S and S● (S+x)=S.
We can see immediately (as is obvious from the diagram) that S and V3 are single points of
failure for this poorly designed system.
Now we can plug in the numbers:
Top event = 0.001+0.01+4(0.01)2 = 0.0114.
4.2.3.5 Event trees

The companion to a fault tree is an event tree. While a fault tree answers the question “what is
the unavailability or failure rate of this system?”, an event tree answers the question “what
happens after the system fails?” The end result of an event tree is that the plant ends up in one
of three main states: stable, core damage, or large release (in practice, these states are often
subdivided).
Unlike fault trees, which are written from the top to the bottom of the page, event trees are
written from left to right. Here are the steps:
 Start (at left) from the undesired initiating event (e.g., loss of feed water, loss of
coolant).
 List the relevant mitigating systems along the top from left to right, in approximate
order of actuation.
 At each branch where a mitigating system is called on, assign the probability of suc-
cess or failure.
 Develop each branch logically to a stable end state, core damage, or a large release.
 Calculate the frequency of each end-point of the branches, starting from the fre-
quency of the initiating event and multiplying by the probabilities of success or fail-
ure at each branch.
 Sum these to obtain the frequencies for core damage and large release.
This is a very simplified explanation, and in practice the methods are much more complex—e.g.,
they have to account for a failure in the fault tree (example, loss of instrument air) also occur-
ring in the event tree, or a failure of a mitigating system (e.g., loss of service water) which is
common to the failure of other mitigating systems. This topic is well beyond the scope of this
section. For more on common-cause failures, see [USNRC, 1981].
A simple example, which should be self-explanatory, is shown in Figure 29, where P1, P2, P3,
etc. are the probabilities of success for each mitigating system. Note that if the failure prob-
abilities (1 – Pi) are small, the success probabilities can be assumed to be equal to one when
calculating each success branch. So, for example, the frequency of core damage in the se-
quence just below the box labelled “Stable” is:
Small LOCA frequency x (1-P3) x P4,
and one could set P4=1 in calculating this branch, as a good approximation. Note also that even
if ECC fails, the moderator will maintain pressure tube integrity and prevent fuel melting by
removing heat from the fuel channel, as described in Section 5.3.2.3.
Figure 29 Simple event tree
4.2.3.6 Dormant vs. active systems

So far, we have focused on systems that are normally dormant and are required to operate on
demand. Safety systems generally fall into this category. However, some systems, like the
Emergency Core Cooling System (ECCS), are required to activate on demand and to continue to
function for some defined mission time. The normal response of the ECCS to a heat-transport
system (HTS) break (loss-of-coolant accident or LOCA) is for the ECCS to detect the event and
initiate the injection of high-pressure (HP) cooling water (strictly speaking, the water-injection
function of ECC is called ECI, or Emergency Coolant Injection, because ECC also has other
functions such as steam generator cooldown and loop isolation1). Then, after the HTS has been
depressurized, medium-pressure and finally low-pressure water is injected. The HP water is
supplied, for example, from a water tank (accumulator) pressurized by huge gas cylinders.
Medium-pressure cooling water can be supplied from a water tank by ECC pumps, and low-
pressure water is retrieved from the sumps in the reactor-building basement, cooled, and
pumped back into the HTS. For CANDU reactors, an ECC mission time of one to three months
has been set2. The ECCS is consequently divided into two separate fault trees for the purposes
of analysis: dormant ECC and long-term ECC (designated DECC and LTECC respectively). The
DECC fault tree focuses on failure to detect the LOCA event, failure to initiate high-pressure (HP)
cooling water, failure to distribute the flow, and failure to provide medium- and low-pressure
water. The LTECC fault tree focuses on failure to provide long-term low-pressure cooling due to
pump failure, valve failure, flow blockage, loss of electrical power, or loss of coolant supply.
4.2.4 Typical results

We have now summarized fault trees and event trees. A PSA is the integrated collection of
these across an entire plant. There are three recognized levels of PSA, which can be defined as
follows (see, e.g., [IAEA, 2010]):
Level 1 – identifies the sequences of events that can lead to core damage and calculates the
core-damage frequency.
Level 2 – uses the information from Level 1 to calculate the magnitude and frequency of re-
leases to the environment of radioactive material.
Level 3 – uses the information from Level 2 to calculate health effects and other consequences
such as contamination of food and land.
In practice, Level 1 is used to assess the design and operation of the plant, Level 2 is used to
assess the effectiveness of containment, and Level 3 is used in off-site emergency planning.
As an example, we summarize the results of a Level 1 PSA done for an operating CANDU 6 plant
[Santamaura, 1998]. The summed severe core-damage frequency (SCDF) was predicted to be
6.1 × 10-6 per year. We have already discussed the underlying reasons that this number is low:
redundant shutdown systems mean that accidents with failure to shut down occur at very low
frequency, and the moderator, if available, is effective in preventing or arresting severe core
damage. Moreover a severe core-damage accident (after shutdown) at high pressure is not
possible because failure of a limited number of pressure tubes converts the event inherently to
a low-pressure accident. Note that [Santamaura, 1998] covered internal events only; typically,
inclusion of external events doubles this value. Also note that accidents in shutdown states
1
Such a distinction is made in Canada, but most other places just use ECC.
2
The mission time is defined here as the time beyond which the decay heat can be removed from the fuel to the
moderator without any water in the fuel channel, so as to prevent any further fuel failures due to overheating. It is
a calculated number.
were excluded.
The contributors to the SCDF are shown in Figure 30. Note that it is the high-frequency initiat-
ing events, including many AOOs, that dominate the SCDF; the bounding accidents such as large
LOCAs contribute very little and hence are much less risk-significant.
Figure 30 Contributors to SCDF for CANDU 6
4.3 Problems
1. What can you say quantitatively about the probability of a pressure-tube/calandria tube
failure in a CANDU, based on that fact that only one has occurred? You will need to calculate the
number of reactor-years of experience with CANDU operation. Use Bayes’ Theorem.
2. Two 100% electrical motor-driven auxiliary feed-water pumps are located in the turbine
building and depend on recirculated cooling water and two independent power supplies. They
have the flow capability to mitigate steam- and feed water-line breaks. Identify common mode
and common-cause failures. The utility wants to extend the life of the plant and is planning
refurbishment. What mitigation features or modifications would you include to reduce or
eliminate these failures?
3. Develop a set of high-level safety goals for a military-use nuclear submarine. They can
be, but do not have to be, numerical. The most important part of your answer is to explain and
justify it, not whether or not it matches someone else’s “official” goals. Consider any differ-
ences due to docked versus at-sea and peacetime versus war.
4. Small reactors could be used in remote northern communities for heating and electricity
production. They would replace very expensive diesel generators, the fuel for which has to be
flown in, whereas the reactor could be designed to be refueled once in twenty years. Small
reactors have also been used for powering unattended remote military installations. Propose
safety goals for each case, with reasons.

5. Assuming that the small LOCA frequency is 10-2 per year, work out the frequency of core
damage in Figure 29. Use P1=P2=P3=10-3.
6. Simple fault tree
a. A system consists of two redundant trains as shown in Figure 31 (i.e., 2 × 100%,
connected in parallel). If the probability that each component P1, P2, V1, V2
works successfully is 0.9, what is the reliability of the system? Write out the Boo-
lean algebra in full. To achieve a reliability of 0.99, what would the reliability of
each component (assumed to be all the same) have to be? List any assumptions.
b. Now do the same calculation without the cross-tie.
c. The cross-tie looks like an obvious thing to do. Think of some engineering disad-
vantages to the cross-tie. Hint: what might happen if one pump fails to start?
How would you fix the design? What impact would your fix have on reliability?
Figure 31 Simple fault-tree exercise
5 Mitigating systems
In previous sections, we have referred to the four safety functions required in a nuclear reactor:
 shut down the reactor
 remove decay heat
 contain any radioactive material
 monitor the state of the plant.
In this section, we shall describe the major systems that perform these functions. We shall
concentrate on CANDU for our examples, although other reactor types have similar systems.
First, however, we need to ask how much redundancy and independence we need for each
safety function. This concept is called defence-in-depth and is a fundamental underpinning of
nuclear safety.
5.1 Defence-in-Depth
The concept of defence-in-depth was developed for military purposes. At one level, it consists
of a series of physical barriers, so that breaching any one barrier (or even more) does not lead
to disaster. The Krak Castle in Syria is one such example, where an (ancient) intruder would
have to breach the moat, the outer wall, the inner wall, and the keep to capture the ruler. More
generally, defence in depth requires that a defender deploy his resources, such as fortifications,
field works, and military units both at and well behind the front line, so that a breach in the
front line does not lead directly to defeat.
Figure 32 Defence-in-depth: barriers

The analogy for a nuclear power plant involves the physical barriers to release of radioactive
material from the fuel, namely the fuel matrix itself (which we have seen is highly resistant to
release of fission products except at melting temperatures), the fuel sheath (which prevents the
gaseous fission products in the fuel-sheath gap from escaping), the heat-transport system, the
containment, and the exclusion zone (which provides dilution rather than a physical barrier);
see Figure 32.
A complementary approach to defence in depth is to define a series of five successive levels:
 Prevent abnormal operation and system failures.
 Control abnormal operation and detect further failures.

 Control accidents within the design basis.
 Control severe accidents.
 Apply off-site emergency response.
These are objectives rather than physical barriers. Note that the last is largely independent of
reactor design. There are various formulations of these concepts; see [INSAG, 1996], from
which the following summary Table 7 is extracted.
Table 7 Levels of defence in depth
Levels of defence Objective Essential means
in depth
Level 1 Prevention of abnormal operation and Conservative design and high
failures quality in construction and
operation
Level 2 Control of abnormal operation and detec- Control, limiting, and protec-
tion of failures tion systems and other
surveillance features
Level 3 Control of accidents within the design Engineered safety features
basis and accident procedures
Level 4 Control of severe plant conditions, includ- Complementary measures
ing prevention of accident progression and and accident management
mitigation of the consequences of severe
accidents
Level 5 Mitigation of radiological consequences of Off-site emergency response
significant releases of radioactive materi-
als
Defence-in-depth is fundamental to the achievement of nuclear safety; try Problem 1 in this
section.
5.2 Shutdown Systems

Typically, a reactor has one or more systems for each of the four safety functions (control, cool,
contain, monitor). Defence-in-depth influences the amount of redundancy provided. For each
system, its performance is specified by design requirements. These in turn arise from listing and
evaluating the challenges that the system must overcome, which make up the design basis.
Shutdown is one of the most important safety functions in a reactor because it reduces the
amount of energy that has to be removed from the fuel after an accident. It is usually accom-
plished through rapid insertion of a neutron-absorbing material into the core. Another way is to
remove from the core material which is essential to the chain reaction, e.g., the moderator.
More radical engineered concepts are possible in principle, such as removing fuel or changing
the core geometry, but they are not in widespread use for fast shutdown.
The design bases of the shutdown systems are typically set by those accidents developed using
the various approaches in Section 2, e.g., large loss-of-coolant accident, loss of reactivity
control, loss of Class IV power, etc. Each of these accidents contributes to answering the
following questions, and the limiting value in each case becomes a design requirement for the
shutdown system:
 How is negative reactivity inserted into the core?
 How fast does the system have to act once it receives a signal?
 How much reactivity depth must it have (how many negative milli-k?)
 How reliable must it be?
 What are the analysis limits that describe effectiveness?
 What sorts of signals are available and practical to trigger the shutdown system for
each accident?
 What sort of environment must the shutdown system be designed to withstand?
 How do we ensure that a fault does not affect both the control system and a shut-
down system? Or both shutdown systems?
 How do we know that the systems will work as designed?
 How does the operator know that the system has been required and that it has
worked?
We shall cover these topics in turn. Many of these questions are common to other safety
systems, and therefore we shall explore them in detail the first time and simply refer to them
later.
5.2.1 Mechanical design

The most basic part of shutdown design is inserting a neutron absorber into the core. Because
modern power reactors are large, a single absorbing device is generally not sufficient. Almost
all reactor types use some form of absorber rod, multiples of which are inserted vertically into
the core. In most reactors in the world, the rods do double duty, being driven in and out of the
core for control purposes, and being driven or dropped in rapidly for shutdown purposes.
CANDU, however, separates the control rods from the shut-off rods, as one of the lessons
learned from the NRX accident.
The first CANDU shutdown system was moderator dump: large valves at the base of the calan-
dria would open, and the moderator would drain out of the calandria vessel under the action of
gravity, somewhat like the ZED-2 design. The system would be re-poised by closing the valves
and pumping the moderator back into the calandria. NPD, Douglas Point, and Pickering A used
this system. Pickering A also used a few shut-off rods which fell in by gravity. The moderator
dump system was (and is) highly reliable, but is somewhat slow compared to shut-off rods,
especially for larger cores, and in addition removes a source of water surrounding the fuel
channels, which could be used in an emergency (we will discuss this later).
Modern CANDUs have two separate shutdown systems: rods and poison injection (Figure 33,
from [AECL, 2005]). Specifically, Shutdown System 1 for CANDU consists of 26 or 28 shut-off
rods, normally suspended above the core and released on a signal. They act in the moderator,
between the rows of pressure tubes, as can be seen in Figure 33. They fall in by gravity and are
spring-loaded, which accelerates them over the first few feet of travel. Mechanically, the rod is
suspended on a cable running over a pulley, which is released by a clutch and wound back up by
a motor. The rod itself falls into a perforated guide tube within the moderator; the purpose of
the guide tube is to make sure that the rod falls straight in and does not tip over or snag.
Figure 33 CANDU shutdown systems

Shutdown System 2 consists of a liquid neutron absorber (gadolinium nitrate) which is injected
directly into the moderator water through perforated metal tubes. The liquid is accelerated by
gas pressure from a common helium tank which pressurizes one poison tank per nozzle, as
shown in Figure 33. The system is actuated by opening the fast-acting valves connecting the
helium tank to the poison tanks. A disadvantage of this system (in case of a spurious trip) is that
because it injects poison into the moderator water, the poison must all be removed chemically
(by ion exchange) before the reactor can start up again, a process that takes almost two days.
Note that there are no normally closed valves between the poison tanks and the moderator for
reliability reasons. As a consequence of this, however, poison gradually diffuses toward the
core down the pipe and must be back-flushed from time to time to drive the diffusion front
away from the moderator.
5.2.2 Speed
The required speed is set by the fastest accident. In CANDU, this is the large loss-of-coolant
accident, where the break is assumed to grow to full size instantaneously—a very pessimistic
assumption and inconsistent with what we know of pipe dynamics [Tregoning, 2008]. Core
voiding for such a case inserts positive reactivity at a rate of approximately 4 mk/sec. One of
the main safety requirements during this phase is to prevent melting of the central part of any
fuel pin due to overpower, because significant amounts of molten fuel could risk failure of the
nearby pressure tube. As long as the net positive reactivity is kept below approximately 6 mk,
depending on the design and the assumptions, the energy is not sufficient to melt the centre of
the fuel. This then suggests that the shutdown system has to start to bite (insert negative
reactivity) in approximately one second, and that the initial reactivity insertion rate has to
overcome the 6 mk already inserted, as well as turning the transient over by 1.5 seconds—in
other words, tens of (negative) mk/sec.
5.2.3 Reactivity depth

Reactivity depth means the total negative reactivity inserted once the shutdown system has
fully operated. For shut-off rods, this occurs when the rods have been fully inserted; for poison
injection, when the poison has been fully injected and mixed with the moderator. The reactivity
depth requirement is set by the accident which inserts the greatest positive reactivity. For
CANDU, this is not the large LOCA, but a small LOCA, specifically a pressure-tube break followed
by an assumed calandria-tube rupture. Why? When a reactor is operating at full power, nega-
tive reactivity load is present due to xenon, a neutron absorber which is formed from the decay
of iodine, which in turn is formed from the fission product tellurium as follows:
135
Te ⇒<1min ⇒ 135I ⇒6.7h ⇒ 135Xe ⇒9.2h ⇒ 135Cs ⇒ 135Ba (stable),
where the times represent half-lives. The xenon load in a CANDU at full power is approximately
-25 mk and has to be compensated for by positive reactivity from the fuel to keep the reactor
critical. When the reactor is suddenly shut down, the xenon decays more slowly than the
iodine, so that the absorption due to xenon initially increases to about -40 mk, then decreases
as the iodine decays away (Figure 34).
Figure 34 Xenon transient after shutdown and start-up

After a long shutdown, the xenon load is small, so that poison is added to the moderator to
compensate for the absent xenon load (because the normal control system does not have the
required negative range). As the reactor comes back to power again, the iodine and therefore
the xenon gradually build up, and the poison can be chemically removed from the moderator
(or a burnable poison is used, i.e., one which is made into a less absorptive species over time by
neutron absorption, typically gadolinium). If a pressure-tube break with assumed failure of the
surrounding calandria tube occurs during start-up, then the “poisoned” heavy-water moderator
can be replaced by non-poisoned heavy-water coolant. The amount of positive reactivity that
can be added in, say, the first 15 minutes is dependent on:
 coolant void in the fuel channels, from the fraction of coolant that is lost through the
break.
 fuel temperature, due to cooldown of the fuel after trip.
 “clean” coolant displacing “poisoned” moderator.
 increase in moderator temperature (albeit small and slow) due to mixing of the hot
discharging coolant with cold moderator.
 decay after shutdown of any xenon that has been formed during on-power opera-
tion.
Offsetting this is the negative reactivity due to the shutdown systems and eventually the very
large negative reactivity due to injection of emergency-coolant light water. The last effect is not
credited in licensing analysis for CANDU (although in LWR licensing analysis, the reactivity effect
of borated ECC must be credited [Westinghouse, 2011a]). For CANDU, the analysis of the
reactivity balance is usually done 15 minutes after the first clear signal of a pressure-tube break,
when the operator can be assumed to supplement the shutdown-system reactivity (e.g., by
manual poison addition). One then designs the shutdown depth of the shutdown system to
achieve an adequate shutdown margin (the net result of the reactivity balance) using conserva-
tive assumptions. Physically, one achieves greater shutdown depth by adding more rods (new
designs), by putting rods preferentially in the high-flux region of the core, etc. There is a practi-
cal space limit, however, on the number and location of rods. The typical shutdown depth for
existing CANDU shutdown systems is -70 mk for the shut-off rods and < -200 mk for poison
injection. Hence, shutdown margins for the rods are smaller, -5 to -10 mk for existing CANDUs,
under the most pessimistic conditions (e.g., assuming that the two most effective rods are
unavailable).
5.2.4 Availability
Demand unavailability is expressed in dimensionless units, although it is also written as
hours/year or years/year. The required unavailability is set primarily by the safety goals applied
to the reactor by the regulator, and for current CANDUs, it is at most one failure per 1000
demands (10-3 per demand) for safety systems. Experience has shown that the shutdown
systems normally meet approximately 10-4 per demand. Unavailability values much lower than
this are possible, but are usually not credited for a single system because of:
 the number of tests that would be needed to establish such an unavailability;

 the suspicion that unknown cross-link effects and common-cause failures impose a
lower limit on the unavailability of a single system.
If each shutdown system has an unavailability of 10-3 per demand, then one is tempted to say
that the unavailability of both systems together is 10-6 per demand. This is true only if the
systems are sufficiently independent and diverse that they are not subject to common-cause
failures. See Section 13.
Recent CANDUs have used software for both SDS1 and SDS2. Software can be more reliable
than relay logic, but its failure modes can be subtle, so that while testing in operation is impor-
tant, much more emphasis must be placed on software design to avoid subtle and common-
cause failures. Methods include a rigorous development process, diversity in development tools
and platforms, simple programming logic, and strict independence between the software
engineers responsible for designing each shutdown system. In CANDU, the control-system
software is completely independent of and separate from the shutdown-systems software, and
in current CANDUs, they run on independent hardware platforms.
5.2.5 Analysis limits

Analysis limits are used to judge the effectiveness of shutdown. Trip signals are chosen to
reduce the challenge to the fuel and to downstream safety systems—in other words, to prevent
or postpone consequences. For example, for accidents which are expected to happen perhaps
once or more in the plant lifetime, i.e., anticipated operational occurrences or AOOs, the
shutdown systems should act early enough to prevent fuel-sheath failure, thus avoiding any
challenge to the HTS and containment as well as the cost of clean-up. This category includes,
among others, loss of Class IV power, a very small LOCA, or a failure in the reactivity-control
system. In turn, prevention of fuel-sheath failure generates a number of secondary criteria,
which we will cover in Section 6. For loss of heat sink accidents, a subsidiary requirement is that
the shutdown systems should act soon enough to give the operator adequate time to bring in a
backup heat sink—typically 15 to 30 minutes, although new designs aim for at least eight hours
before operator action is required. For rare accidents such as a large LOCA, one should limit
(but not necessarily prevent) fuel damage (to reduce the challenge to containment) and ensure
that the fuel is not made so brittle from the Zircaloy-steam reaction that it forms debris and
cannot be cooled by ECC; see Section 6. Limiting fuel damage in a large LOCA is a task shared
between shutdown and ECC: the job of the shutdown system is to deliver the fuel to the ECC in
reasonable condition so that ECC can remove decay heat from a known fuel geometry. Finally,
in any accident, shutdown systems should act early enough so that there is no risk to the
pressure boundary (or at least no risk in addition to that from the initiating event).
5.2.6 Signals
A shutdown system must detect an accident soon enough that the analysis limits are met.
Some of the commonly used trip signals are listed in Table 8 (this list is illustrative rather than
complete).
Manual trip is credited if the time scales are long—typically fifteen minutes from the first clear
signal of the accident in the main control room—and if there is no practical alternative.
Table 8 Typical trip signals for CANDU

Accident Symptoms Typical Trip Signals
Loss of reactor Reactor power rises High neutron flux
power control Reactor power rises rapidly High log rate of neutron flux
Heat-transport system pressure High heat-transport system pressure
rises
Loss of forced Coolant flow drops Low heat-transport system flow
circulation Low core pressure drop
HTS pressure rises High heat-transport system pressure
Reactor power rises High neutron flux

Medium to Reactor power rises High neutron flux
large loss of Reactor power rises rapidly High log rate of neutron flux
coolant Containment pressure rises High containment pressure
Coolant flow drops Low heat-transport system flow
Coolant pressure drops Low heat-transport system pressure
Small loss of Pressurizer level drops Low pressurizer level

coolant Coolant flow drops Low heat-transport system flow
Containment pressure rises High containment pressure
Moderator level rises (in-core High moderator level
break)
Coolant pressure drops Low heat-transport system pressure
Loss of feed HTS pressure rises High heat-transport system pressure
water Feed-water flow drops Low feed-water flow
Steam generator level falls Low steam generator level
Large steam HTS pressure falls Low HTS pressure
main failure Pressurizer level falls Low pressurizer level
Steam generator level falls Low steam generator level
Feed-water pressure falls Low feed-water pressure
Reactor power decreases
Containment pressure rises5 High containment pressure
5.2.7 Operating environment

A safety system must be able to function in, or be protected from, the conditions caused by an
5
For plants where portions of the steam mains are within containment
accident. If it is not possible to meet this requirement, then a redundant system is needed. For
example, a major fire in or near the main control room would require shutdown (because it
could affect the control computers) and at the same time would possibly damage some of the
components of Shutdown System 1 so that it would not respond (although it would very likely
fail safe)—and therefore SDS2 (which does not depend on the integrity of the main control
room) is used as a back-up to perform this safety function.
The shutdown mechanisms in CANDU act mostly within the moderator, which protects them
from some of the effects of accidents—but not all. For example, we must design so that:
 no high-energy pipes are within striking distance of the reactivity mechanisms deck
on top of the reactor, where the shut-off rod clutches and pulleys are located.
 an in-core break cannot disable shut-off rods to the extent that the system does not
meet its reactivity depth requirements (it is not possible to protect all the guide
tubes and rods from pipe whip and jet forces from an in-core break).
 shutdown-system cables and instruments are separated to the extent practical so
that a local fire will not incapacitate both shutdown systems.
 steam, high temperatures, water, and high radiation fields from an accident must not
prevent a shutdown system from firing when needed (once it has fired, however,
such protection is no longer needed).
 shaking due to an earthquake must not prevent the shutdown system from actuating
nor slow it down so severely that it cannot meet its analysis limits.
5.2.8 Common-cause failures

We gave an example earlier where a single cause (fire) could disable more than one system.
This is a serious challenge to the protection provided by seemingly independent systems.
Known common-cause failures can be “designed out”, but to cover the possibility that we
cannot anticipate them all, a two-group philosophy is followed in CANDU. In summary, this
philosophy is:
 for each failure, ensure that there are at least two ways of performing the required
safety function.
 separate these two ways geometrically (so that they are not subject to local damag-
ing hazards such as fire, turbine missiles, or aircraft crash).
 use diverse equipment and diverse means of operation.
 protect equipment against the environmental results of the failure.
 qualify or protect at least one of the two systems against plant-wide external events
such as tornadoes and earthquakes.
The practical application involves many trade-offs: diversity is not always possible, and there is
only so much space within which systems can be separated. Whereas ideally one would like to
route control system cables separate from SDS1 cables and SDS2 cables, it would be almost
impossible from a plant layout point of view, and therefore grouping of control system and SDS1
cables is allowed, but they must be separated from SDS2. However, all three logic channels (see
Section 5.2.9) within any given group must be separated, so that a local cable fire cannot
generally disable all three channels of any one system.

If a compromise in separation must be made, then one must show that one or more of the
following applies:
 there is no credible hazard in the area
 another system outside the area will mitigate the event
 the system or component is protected by a barrier
 the system or component is fail-safe
 the component is designed to withstand the hazard.
Table 9 shows an example of how grouping and separation were implemented on CANDU 6.
Table 9 Grouping and separation example
Safety Function Group 1 Group 2
Shutdown Reactor Control System Shutdown System 2
Shutdown System 1
Heat Removal From Fuel Heat Transport System Emergency Water System
Steam & Feedwater Systems
Shutdown Cooling System
ECC
Moderator
Contain Radioactive Material Reactor building air coolers Containment & containment
subsystems
Monitoring & Control Main Control Centre Secondary Control Area
Most safety systems need services such as air, water, and power. These must also be grouped
and separated.
5.2.9 Testing
Section 4.2.3.3 discussed the link between testing and avail-
ability. It is not practical to test safety systems fully during on-
power operation; instead, they are designed to have their
components tested in such a way as to preclude actual system
operation. Specifically, the testing of the logic is separated
from that of the final mechanism, and the mechanism is
tested in stages, not all at once.
All safety systems in CANDU have three logic channels, with
two out of three being sufficient to initiate the trip or safety- Figure 35 2/3 logic
system action (Figure 35). The requirement that two channels
must both vote for a trip reduces the likelihood of spurious
trips due to a single component failure. On the other hand, the reactor will still trip if required
even if one channel is unavailable (failed unsafe). Finally, a single channel can be tested without
tripping the reactor.
The shut-off rods are designed to be partially dropped individually in a test. In other words,
shortly after the clutch has released and the rod begins to fall, the clutch can be re-energized
and the rod “caught” before it enters the core. This proves that the rod is not stuck in the first
part of its travel. Typically, each rod is partially dropped about once a week.
For Shutdown System 2, Figure 36 shows one way of testing each logic channel and each valve
without firing the system, while any two channels which trip will fire the system. Testing any
single logic channel opens two valves, but does not allow (much) poison to leak into the mod-
erator.
Performance testing is usually done by
measuring the speed of actuation of one or
more components. Therefore, on a partial
rod drop, one can determine how long it
takes the rod to reach the point when it is
caught, or one can test the valve opening
times on SDS2. Before a scheduled shut-
down, a full actuation test of the shutdown
system measures the rod drop versus time
and the actual power rundown.
5.2.10 Human interface

An operator must: Figure 36 SDS2 testing
 Be notified that the shutdown system has tripped

 Be able to confirm that it has actuated correctly
 Have procedures to follow in case it has not.
Normally, notification consists of an alarm window on the SDS panel, showing (for each chan-
nel) that a trip parameter has passed its trip set-point and a window showing that the SDS has
fired. The latter is not sufficient to establish that the system has indeed worked, and an opera-
tor will have supplementary information available such as shut-off rod position; he will also
check the neutron power measurement to ensure that it is consistent with a shut-down reactor.
Finally, should these latter measurements suggest that the system has not fired, he will have
and will follow backup procedures such as manually tripping the shutdown system, manually
firing the second shutdown system, or dropping the mechanical control absorbers (manual
stepback).
We have now covered all the high-level design requirements and design bases of the shutdown
systems. Our in-depth example can be extended to the systems responsible for heat removal,
containment, and monitoring, but we will cover these more briefly.
5.3 Heat-Removal Systems

The high-level design requirements of the heat-removal systems are determined by the answers
to the following:
 How much heat must they remove (full power, decay heat, decay heat after x min-
utes, etc.)?
 Where are they connected (primary side, secondary side, etc.)?
 How are they initiated?
 Under what conditions can they operate (pressure, temperature)?
 What is their reliability?
5.3.1 Heat-removal capability

Heat removal at high power is performed by the primary coolant flowing through the fuel
channels and transporting the fuel heat to the steam generators. Coolant is transferred across
the steam-generator tubes to the main steam and feed-water systems. The case of a sudden
loss of heat removal from the secondary side (e.g., turbine trip with loss of condenser vacuum)
must be provided for, so that a capability for 100% steam dump to atmosphere is provided. This
is accomplished by banks of main steam-safety valves (MSSVs) on the main steam lines. Full
steam flow is needed only until shortly after the reactor has tripped.
For economic reasons, another subsystem is provided which can dump up to 60% steam directly
from the steam generators to the condenser. This is used for poison prevention—that is, if the
turbine trips, but the operator thinks that he can restart it reasonably soon, instead of shutting
the reactor down, he can set back the reactor power to a level just sufficient to prevent a
poison-out due to xenon build-up. The heat must still be removed, however, and therefore the
design makes it possible to dump the steam directly to the condenser, by-passing the turbine.
Because (unlike steam dump to atmosphere) the secondary water is recycled from the con-
denser back to the steam generators, it is possible to do this for considerable periods of time.
This is a big advantage in the case of a prolonged loss of electrical grid power; if the reactor can
avoid a trip when the grid is lost (e.g., through stepback to 60% power) and is kept running, it
remains ready to supply power as soon as grid stability is restored.
This capability was demonstrated after the power blackout of Thursday, August 14, 2003. When
the grid failed at 16:10, all operating reactors in Ontario tripped. The four Bruce Power units
and Darlington unit 3 were quickly put into poison-prevent mode, operating at about 60% of full
reactor power and by-passing steam to the condensers. These reactors were able to start
providing power within a few hours as grid restoration began. For details, see [Rogers, 2004a].
The maximum decay heat immediately after shutdown is approximately 6%. However, we can
sometimes design to remove less if the heat sink can be brought on-line later. Figure 37 shows
the variation of decay heat with time after shutdown.
The desired end-point of an accident is a stable cold shutdown—i.e., the reactor is well sub-
critical, the decay heat generated by the fuel is being removed, no further release of radioactive
material from the fuel is taking place, and the reactor is depressurized and “cold” (i.e., ~100°C
or less). Therefore, in addition to removing the decay heat from the fuel as it is generated,
there must be some additional heat-removal capability to cool the reactor down.
5.3.2 Location of heat removal

Heat can be removed from the secondary cooling system and from the primary cooling system.
Figure 37 Bundle power after shutdown
5.3.2.1 Secondary-side heat removal

The secondary cooling system is used under the following conditions:
 If all components and systems on the secondary side are working, the main feed-water
pumps provide water to the steam generators. The condenser removes heat from the
steam and provides a continuous source of cooled water to the feed-water pumps. This
relies on the availability of Class IV electrical power to run the main feed-water pumps.
 Alternatively, the task can be undertaken by one or more auxiliary feed-water pumps,
sized to remove decay heat. These feed-water pumps are typically powered by Class III
electrical power, or directly by a diesel engine, or by a steam turbine connected to the
steam generators. It is not necessary to size them for 6% power because they are not
needed until some of the water already in the steam generators has boiled away. This
takes about half an hour, so that the heat-removal capability of the auxiliary feed-water
pumps is typically about 4%. Note that if Class IV power is unavailable, the steam cannot
be condensed, and the heat is removed by steaming to atmosphere from the steam
generators, either through the atmospheric steam discharge valves (ASDVs), with a ca-
pacity of about 10% of full-power steam flow, or through the main steam-safety valves
(MSSVs), with a capacity of about 115% of full-power steam flow. After an hour or more,
the water in the feed-water train will be used up, and the operator will have to establish
another heat sink.
 In some CANDUs, the dousing tank located in the top portion of the containment can
supply a long-term source of water by gravity to the steam generators. In others, an ele-
vated tank outside containment (the boiler emergency cooling system, or BECS) per-
forms this function.

 Most recent CANDUs have a seismically qualified source of water (e.g., a large pond) for
use after an earthquake. This emergency water system (EWS) has its own seismically
qualified power and pumps and can supply water independently to the steam genera-
tors for about three days.
5.3.2.2 Primary-side heat removal

Because many of the options for secondary-side heat removal are valid only for a limited period
of time after an accident, CANDUs also have a primary-side system to remove decay heat, the
shutdown cooling system. It is a closed system connected to the reactor headers with its own
pumps and heat exchangers (Figure 38). It can be brought on-line immediately after shutdown
at high temperature and pressure.
Figure 38 Shutdown cooling system

The emergency core cooling system can be viewed as a decay heat-removal system for the
special case of a break in the primary cooling system piping. We shall cover this system later.
5.3.2.3 Moderator and shield tank

The moderator surrounding the fuel channels can be used in a severe accident (LOCA with loss
of ECC) to remove decay heat. This heat-removal pathway is efficient enough to prevent fuel
melting, but will not prevent extensive fuel damage and distortion of the fuel channels (Figure
39).
The shield tank surrounding the calandria (see also Figure 39) has its own heat-removal system
(pumps and heat exchangers) and can be used in a severe core-damage accident, e.g., if a LOCA
plus loss of ECC plus loss of moderator heat removal all occur. In current CANDUs, the shield
tank does not have enough heat-removal capability (0.3%) to match that being generated by the
fuel; in addition, the causes for the failure of ECC and moderator cooling may also have disabled
the heat-removal capability of the shield
tank (e.g., loss of electrical power, loss of
service water). However, several current
CANDUs and all new ones, have installed
gravity-driven makeup to the shield tank
and/or the calandria to provide a much
longer-term heat-removal capability.
5.3.3 Initiation of decay heat removal

Decay heat-removal systems can be either
automatic or manual, based on when
they are needed. Typically, if they are not
needed for 15–30 minutes, they can be
manually initiated (e.g., shutdown cooling
system, EWS); if they are needed sooner,
they must be automatic (e.g., ECC,
auxiliary feed-water system).
5.3.4 Operating pressure

A key design choice is the operating
pressure of the heat-removal system.
Table 10 summarizes the advantages and Figure 39 Moderator and shield cooling
disadvantages of high vs. low pressure. In
many cases, the pressure is determined
by the nature of the design (e.g., the
moderator).
Table 10 Operating pressure of decay heat-removal systems

Operating Advantages Disadvantages
Pressure
High Can be brought in at any stage of an More stringent requirements on code
accident class of piping and components
Components tend to be smaller due Need to ensure that it is tolerant if
to more efficient heat removal brought on-line when system is at low
(larger ΔT) pressure (e.g., risk of pump cavitation)
More easily automated
Low Can be simpler and cheaper Requires previous depressurization of
Can be made more passive the system (i.e., depends on another
system)
5.3.5 Reliability
Unlike the case of shutdown systems, the mission time of decay heat-removal systems can
extend to days or months, and therefore we need to determine both the demand availability
(reliability to start) and the running reliability once the system has started. A typical active
decay heat-removal system consists of pumps, which require electrical power, and heat ex-
changers, which require a source of cooling water, which in turn requires electrical power.
Unavailabilities in the range of 10-2 to 10-3 to start and 10-1 to 10-2 over a specified mission time
are typical. Therefore, redundancy in decay heat-removal systems is necessary.
5.4 Emergency Core Cooling System

The functions of the ECCS are to refill the core and to remove decay heat after a LOCA. We now
discuss high-level design requirements.
5.4.1 Performance requirements

Sudden large primary pipe breaks have never occurred in a Western nuclear reactor; the prob-
ability is less than 10-5 per reactor year. A modern CANDU reactor has between 360 and 480
fuel channels and therefore 720 and 960 feeder pipes. The probability of a small pipe break is
about 10-2 per reactor year based on experience, which is an issue not only for safety but also
for plant investment protection. Therefore, the ECCS has safety requirements for both large and
small breaks, and in addition, investment protection requirements for small breaks only.
The ECC safety requirements for all breaks are to:
 meet public dose limits by limiting the extent of fuel damage,
 prevent pressure-tube failure,
 ensure that the fuel in the fuel channels retains a coolable geometry,
and for small breaks, in addition, to:
 prevent sheath failure.
5.4.2 Location of water injection

The CANDU ECC injects light water into the reactor inlet and outlet headers in both heat-
transport system loops (for reactors which have two loops): eight headers in all in CANDU 6
(Figure 40). Because each channel is connected to two headers and the headers are above the
core, this provides a water pathway to every channel in the core. The disadvantages of this
scheme are:
 water injected near or at the break discharges without removing heat from the fuel
(depending on the break size and location). The design must provide sufficient pres-
sure and water volume that flow of water out of the injection point at or near the
break is accounted for.
 the (feeder) pipe to each channel is fairly small in diameter and contains a large
amount of stored heat, and the injection water may have to flow countercurrent to
the steam exiting the channel as it refills.
Light-water injection also ensures long-term shutdown in CANDU, but is not credited with this
function in safety analysis.
Figure 40 ECC layout
5.4.3 Injection pressure and flow rate

The current design (using CANDU 6 as a model) is a three-stage ECC:
 A high-pressure initial injection phase limits fuel overheating for small breaks and forces
early cooling for large breaks (to limit pressure-tube deformation and early fuel damage).
Typical injection pressure is 5.3 MPa. In CANDU 6, high-pressure injection comes from two
water tanks which are pressurized by gas at the time of a LOCA signal. In some multi-unit
CANDU plants, this high-pressure phase is supplied by electrically driven pumps because the
reliability of Class IV power tends to be very high (it can be obtained from other operating
units as well as the grid).
 Medium-pressure injection takes over when the high-pressure water tanks are nearly empty.
It uses medium-pressure (~1 MPa) pumps and draws cold water from the dousing tank or a
similar reservoir. It pumps this water into the same locations (all headers) as the high-
pressure ECC. The medium-pressure phase ensures that enough water has collected in the
basement of the containment building before the next (recovery) phase starts.
 In recovery injection, the medium-pressure ECC pumps are switched over to take water from
the sump in the basement. They pump this water through dedicated ECC heat exchangers
before returning it to the heat-transport system. This phase is the long-term heat sink.
Figure 41 shows the three phases schematically for a large LOCA.
Figure 41 Three phases of ECC
5.4.4 Other Functions

There are two further functions that ECC must perform: loop isolation and crash cooldown.
Loop isolation simply closes the valves connecting the two heat-transport system loops (for
some CANDUs which have two loops). It limits (most of) a LOCA to one loop. In addition, for a
LOCA with loss of ECC injection, loop isolation (which is an independent signal) limits the
possible source of hydrogen.
Crash cooldown is more significant. If a break is small, less than, say, a feeder failure, it is not
able to depressurize the heat-transport system down to ECC pressure or to keep it below ECC
pressure once injection begins. Crash cooldown opens all the MSSVs on all the steam lines and
blows down the steam-generator secondary side to near atmospheric pressure over about 15
minutes. Because the steam generators are still a heat sink for the primary coolant in a small
LOCA, this forces the primary-side pressure down over the same time scale. Therefore, it
ensures that ECC is not blocked by the heat-transport pressure “hanging up” at secondary-side
pressure and that the unfailed loop will also be refilled by ECC. Some CANDUs (Darlington) use
high-pressure pumps for a small LOCA and are not as dependent on crash cooldown for this
purpose.
5.4.5 Initiation Signals

Clearly, ECC initiation must be automated. The basic signal is low heat-transport system pres-
sure. By itself, this is not unique to a LOCA, and a spurious injection is costly due to heavy-water
downgrading, and therefore it is conditioned (ANDed) by one or more of high building pressure,
sustained low heat-transport system pressure, and high moderator level (the last is for an in-
core break). A separate signal isolates the loops (in some designs) on low pressure. Crash
cooldown is part of the ECC signal; however, because of its importance, recent designs have two
independent crash cooldown signals.
5.4.6 Reliability
As a safety system, the ECC must meet a demand unavailability of 10-3 or less. A running
unreliability of 10-2 over the three-month mission time has been used as a design target.
Availability will be better because fuel will heat up more slowly in the longer term if interrup-
tions in ECC occur; hence, there may be time to repair the fault in ECC before further fuel
damage occurs.
5.5 Containment
The important aspects of containment are the following:
1. What is the design pressure?
2. What is the leak rate at design pressure?
3. How is pressure controlled? How is heat removed?
4. How is containment isolation ensured?
5. What is the containment reliability?
6. What other functions must containment perform?
5.5.1 Design pressure and leak rate

Containment is an envelope around those systems containing or potentially containing signifi-
cant amounts of fission products and is designed to prevent their escape to the environment.
Of course any structure will leak, and the leak rate will increase with internal pressure. Typi-
cally, containment surrounds at least the reactor core and the primary heat-transport system.
The design pressure is chosen to be greater than the maximum pressure reached in any accident
for which a predictable degree of containment leak-tightness is a requirement. To determine

the design pressure, all design basis accidents which release significant radioactive material into
containment are analyzed, and the peak pressure reached in any one, plus some margin, is
chosen to be the design pressure. The important aspect of design pressure is that the leak rate
at the design pressure is known. In addition, for severe accidents [CNSC, 2008]:
“Containment maintains its role as a leak-tight barrier for a period that allows sufficient
time for the implementation of off-site emergency procedures following the onset of
core damage. Containment also prevents uncontrolled releases of radioactivity after this
period.”
This requirement may also affect the design pressure.
Leak rate is generally not selected independently in advance, but results from the construction
technique. If very low leak rates are required (≤0.2% / day at design pressure), a steel liner is
used.
Figure 42 Single-unit containment

For CANDU 6, for example (Figure 42), the design pressure is 124 kPa(g), and the leak rate at the
design pressure is 0.5% of the contained volume per day. Vacuum containments have lower
design pressures (because the vacuum building terminates the pressure rise), and (after the
initial short-term overpressure), leak rates are negative (inward) for several days; see Section
5.5.2. For the Darlington containment, for example, the design pressure is +96 kPa(g). The
normal operating pressure for the vacuum building is -96 kPa(g) [Huterer, 1984]. Typically, the
design pressure is set by the large LOCA because this both causes high short-term pressure and
has the potential to release fission products into the containment. The leakage rate at design
pressure is confirmed by proof testing before the plant is operational and by periodic testing
thereafter.
Should the pressure exceed the design pressure, the building will not explode (typical safety
margins on massive failure of the building are a factor of about three over the design pressure;
see [Rizkalla, 1984], [Rizkalla, 1986]). However, the leak rate is harder to predict because the
leak area may increase. In particular the leak-tightness of any penetrations and seals above
design pressure is dependent on their detailed design and will need to be determined on a case-
by-case basis. In any event, for an epoxy-lined containment, leakage would increase through
penetrations and cracks before the internal pressure reached failure pressure, and therefore it
would be very unlikely for the building to fail catastrophically.
5.5.2 Pressure control and heat removal

Without some means of removing heat, the containment pressure in an accident such as a pipe
break will rise rapidly as the broken system discharges steam and empties, then more slowly as
the decay heat is steamed into containment. It is possible to build a high-pressure containment
to withstand this, at least for some time. To date, CANDU containments have had some means
of short-term pressure suppression and/or some means of long-term heat removal to provide
for these two phases of an accident: the addition of pressure suppression enables a lower
containment design pressure.
Figure 43 Vacuum containment concept

Large nuclear stations in Ontario use multi-unit containment in which parts of the containment
envelope are shared among four or eight units. The individual reactor containment buildings
are all connected to a common vacuum building kept at very low pressure. Inside the vacuum
building is an elevated water tank; when a LOCA occurs, the vacuum valves open (self-actuated
on pressure differential), thereby connecting the vacuum building to the reactor building(s), and
the contents of the water tank are sprayed over the vacuum-building volume (Figure 43, from
[Morison, 1987]). The sprays are likewise self-actuated on the pressure differential caused by
the LOCA. The water sprays condense the steam and reduce the internal pressure. The con-
tainment pressure quickly goes sub-atmospheric and remains so for several days after an
accident, meaning that leakage is inward, not outward.
The single-unit CANDU 6 also uses pressure suppression (Figure 42). The sprays and the ele-
vated water tank are located in the reactor building. There are six spray arms, each with spray
valves, arranged as two valves in series on each spray arm (to avoid a spurious douse). The
valves on three spray arms are pneumatically operated, and for diversity, the valves on the
other three are electrically operated. The containment building is made of pre-stressed, post-
tensioned concrete with an epoxy liner for leak-tightness. Dousing cycles on and off for small
breaks, but for large breaks, it remains on until the dousing water is used up and is all on the
basement floor of the reactor building. Dousing does not control pressure in the long term
because it is used up in the early part of an accident.
In the longer term, heat can be removed by containment air coolers. These require both water
and electrical power. Alternatively, low-flow sprays can also be used, with heat being removed
from the spray water by heat exchangers and the cooled water pumped back up to the spray
arms.
5.5.3 Isolation
In normal operation, containment is not a sealed system. Many fluid lines penetrate the build-
ing (e.g., steam lines, feed-water lines, service water lines, instrument lines). For CANDUs, the
building is normally ventilated for atmospheric temperature control, especially because person-
nel access to parts of the building is possible and required during operation. Some penetra-
tions, particularly ventilation, could be pathways for release of radioactive material if an acci-
dent should occur, and therefore, on an accident signal, these are automatically isolated.
Steam, feed-water, and service-water lines are not isolated immediately in CANDU practice
because continuing to use a running system is more reliable than stopping it and starting up
another one. For LWRs, the opposite approach is taken due to a different philosophy of con-
tainment isolation [USNRC, 2013] and partly due to the reactivity increase on a steam-line break
in LWRs. However, main steam isolating valves (MSIVs) are used in recent CANDUs to prevent
leakage to the environment through a pre-existing steam-generator tube failure after an acci-
dent, but they are manually operated and slow. Feed water and service water are not normally
isolated.
5.5.4 Reliability
As a special safety system, the containment must meet a demand unavailability of 10-3
years/year or less. Containment leak-tightness is tested every few years by pressurizing the
building and measuring the leak rate. This is an invasive and expensive test, and if the leak rate
exceeds the requirement, one must assume that a containment impairment has existed for half
the interval between tests. Therefore, on-line leakage monitoring systems are being deployed
in existing reactors. The containment isolation system is likewise tested during operation to
prove that its unavailability target is not being exceeded.
5.5.5 Other functions

Containment also acts as a barrier to protect reactor systems from external events (tornadoes,
turbine missiles, aircraft crashes, and malevolent acts). These may impose additional design
requirements on the structure.
Hydrogen can build up in containment after an accident. After a LOCA, hydrogen is formed
slowly by radiolysis of the water circulating through the core. A severe accident such as a LOCA
plus loss of emergency core cooling can also produce hydrogen early on because of the chemi-
cal reaction between the hot fuel sheaths and the steam in the fuel channels (cf. Sections 3.2.2
and 3.2.3.2). The containment building promotes some mixing of hydrogen due to natural
circulation. Air cooler fans provide forced mixing. In addition, igniters are placed in various
rooms to burn any local hydrogen concentration before it can detonate. Passive autocatalytic
recombiners can be used for long-term hydrogen control; they do not need electrical power or
controls. They present a catalyst bed to the containment atmosphere, on which the hydrogen
recombines with oxygen. The heat of reaction causes a convection flow through the device,
which helps mix the containment atmosphere.
5.6 Monitoring
For most accidents, the plant state is monitored from the main control room (MCR), and the
safety functions of shutdown, heat removal, and containment can be performed from there.
Some accidents, however, can render the MCR uninhabitable or inoperable: for example,
earthquakes, fire in the MCR, hostile takeover, aircraft strikes, and high radiation fields. A
secondary control area (SCA) is provided for such eventualities; the operators relocate to the
SCA and can perform the required safety functions from there.
5.7 Problems
1. Select an accident from the case studies discussed in Section 3 and analyze it in terms of
the five levels of defence in depth (both barriers and objectives). Which aspects were present?
Which were missing?
2. Consider the ZED-2 reactor described in Section 2.7. What elements of defence in depth
are present in this design? What elements are missing? Why might it be acceptable to have
missing or weak levels of defence-in-depth?
3. Using the SLOWPOKE 10MW heating reactor described in Section 2.7, describe the
possible shutdown-system requirements in terms of design, rate, depth, signals, margins,
environment, and independence. Give reasons—do not simply copy existing material.
4. Using the ZED-2 zero-power research reactor described in Section 2.7, describe the
possible shutdown-system (moderator dump) requirements in terms of design, rate, depth,
signals, margins, environment, and independence. Give reasons—do not simply copy existing
material.
5. Look up (e.g., from the USNRC web site) the design of either the EPR or AP1000 decay
heat-removal systems. Summarize them and discuss advantages and disadvantages compared
to the CANDU decay heat-removal systems (choose one CANDU plant for your comparison). If
you do not have access to CANDU information, simply compare EPR and AP1000 decay heat
removal systems.
6. What are the options for heat removal from containment after a severe accident (core
damage)? What are the pros and cons of each of these options? Feel free to look up what
choices have been made by modern designs.
6 Safety Analysis – Accident Phenomenology

This Section provides a high-level discussion of accident phenomena in CANDUs and how the
results are judged. The technical capabilities that computer codes used in CANDU safety
analysis should possess are then described. Severe accident phenomena and analysis are
summarized, followed by a discussion on uncertainty analysis.
6.1 Accidents by Phenomena

In Section 1.4, we described the hazards posed by a nuclear power plant and the broad classes
of events that could cause a release of radioactive material. In Section 2, we discussed how
design basis accidents are identified and selected. Based on these concepts, we can describe
accidents by major phenomena, as follows (grouped by primary or direct cause):
1. Reactivity accidents
 Bulk loss of reactivity control
 Loss of reactivity control from distorted flux shapes
2. Decrease of reactor coolant flow
 Loss of Class IV power
 Partial loss of Class IV power
 Single pump trip or single pump seizure
3. Increase of reactor coolant pressure
 Loss of primary pressure and inventory control (increase)
4. Decrease of reactor coolant inventory
 Large heat-transport system LOCA
 Small heat-transport system LOCA
- Single-channel events
- Single steam-generator tube rupture
- Multiple steam-generator tube rupture
 Loss of primary pressure and inventory control (decrease)
5. Increase of secondary-side pressure
 Loss of secondary-side pressure control (increase)
6. Loss of secondary-side heat removal
 Main steam-line break
 Feed-water line break
 Loss of feed-water pumps
 Spurious closure of feed-water valves
 Loss of secondary-side pressure control (decrease)
 Loss of shutdown heat sink
7. Moderator and shield-cooling system failures
 Pipe break
 Loss of forced circulation
 Loss of heat removal
8. Fuel-handling accidents
 Fuelling machine on-reactor

 Fuelling machine off-reactor
 Accidents in the irradiated fuel bay
 Loss of heat removal
 Loss of water
Severe core-damage accidents involving an initiating event and failure of at least two mitigating
systems fall into a separate category because the phenomena of severe core damage are not
strongly coupled to the initiating event. We will cover these later.
Safety analysis is the method by which we can show that the predicted consequences of a
postulated accident meet regulatory and design goals. It involves mathematical modelling of
the major systems in the plant to predict the behaviour of their components in an accident.
Typically, the focus is on:
 Demonstrating that the physical barriers to release of radioactive material are not
further damaged, or have sustained only limited damage, beyond the initiating fail-
ure
 Predicting on-site doses to ensure that credited operator actions are feasible in the
post-accident environment
 Predicting dose to the public.
In general, in traditional safety analysis for CANDU, the mathematical models are best-estimate,
and pessimism is introduced by means of “conservative” assumptions and data. This raises the
issue of margins.
6.2 Margins
A discussion of margins must start by defining the essential physical barriers that need to be
preserved, or their degradation limited, to meet regulatory dose limits with confidence.
Typically, these barriers are the fuel matrix, the fuel sheath, and the heat-transport system
including the pressure tube, the calandria tube, and the calandria shell. To prevent failure of an
essential barrier, physically based quantitative parameter limits are chosen for each failure
mechanism—e.g., Tfuel < 2800°C is a parameter limit (no fuel melting) which will prevent one
mechanism of failure of the pressure-tube barrier through local strain. More conservative
surrogate criteria are chosen if the parameter limit is uncertain. Analysis limits are chosen to
ensure that the parameter limit or surrogate criterion is met. This framework, along with the
use of conservative assumptions and data in the safety analysis, gives the margin to the barrier
failure point, as shown in Figure 44.
Figure 44 Margins
Of necessity, many assumptions about plant operation are made in safety analysis. These are
usually chosen to envelop actual or expected operation—for example, safety analysis will use a
maximum bundle power that is greater than that expected in operation. The plant must oper-
ate consistently with the assumptions in the safety analysis. The operators therefore define a
safe operating envelope (SOE) of parameters under operational control6. Operating the plant
within this envelope ensures consistency with the safety analysis assumptions and incidentally
provides another margin to the results of safety analysis. It is important that the safety analyst
not make assumptions that are too restrictive because this will unnecessarily limit operation.
Other methods of safety analysis are best estimate with analysis of uncertainties (BEAU), which
we will cover later, and best-estimate analysis, which is used for severe accidents.
6
In other countries, a set of Technical Specifications serves a similar purpose.
6.3 Major Computer Analysis Tools Required for DBAs

CANDU safety analysis requires a comprehensive suite of physical models. The mathematical
foundations of these models are presented in simplified form in Section 7. This section lists the
capabilities that these codes should have. A detailed description of individual codes is well
beyond the scope of this chapter; references are given to enable the student to investigate
further any codes of interest.
Reactor physics analysis requires a transient three-dimensional model, especially for larger
CANDU cores. The most demanding application is a large LOCA because the positive void
coefficient leads to relatively fast kinetics and because of the spatial effects associated with flux
tilts and shut-off rod (or liquid absorber) insertion. Three-dimensional effects are also impor-
tant in slow loss of reactivity control starting from distorted flux shapes. Current codes used
[Roy, 2004] include
 WIMS-IST, a two-dimensional lattice-cell code
 RFSP-IST, a reactor code for CANDU full-core 3-D static and dynamic analysis
 MCNP – a “Monte-Carlo” code which tracks large numbers of individual neutron his-
tories to achieve an answer. The accuracy of the answer is limited only by the num-
ber of histories and the quality of the input data from experiments which measure
nuclear cross sections.
The system thermo-hydraulics code is typically a two-fluid, one-dimensional non-equilibrium
network code. The two fluids are water and steam; recent codes also incorporate a third non-
interacting fluid, e.g., hydrogen, as produced in severe accidents. One spatial dimension suf-
fices for CANDU because the system consists largely of linear flow in pipes (feeders, channels,
large pipes above the core) and there are no vessels in which complex three-dimensional
behaviour occurs. However, the flow in the headers can be quite complex, and tools based on
visualization tests are being developed to model it more accurately. The thermo-dynamic non-
equilibrium aspect is important in modelling fuel rewetting and channel refilling after a LOCA
because the flow can be stratified (steam and water flowing separately due to the effect of
gravity) during that time, and each phase can have its own temperature and flow. Finally, a
network capability is clearly a necessity in CANDU, with its multiple parallel paths (e.g., many
channels are connected to one header, and ECC is connected through parallel paths to each
header). The reactor physics calculations must be coupled with the system thermo-hydraulics
code for a large LOCA because the voiding transient determines the power pulse, which in turn
has a second-order effect on the voiding transient. Current system thermo-hydraulic codes
used are CATHENA [Hanna, 1998] and TUF [Liauw, 1997].
Fuel thermo-mechanical models consist of a code for normal operation, which predicts the
initial fuel conditions before the accident (sheath strain, fuel-to-sheath heat-transfer coefficient,
fission gas release, initial fuel and sheath temperatures, etc.) and a transient thermo-
mechanical code for accidents. The latter includes sub-models for fuel failure mechanisms due
to fuel-sheath strain, beryllium braze penetration, sheath embrittlement due to oxidation,
athermal strain, and excessive fuel energy content. Because of the need to predict the dose for
each accident, the models must be able to estimate the percentage of fuel sheaths that fail in
an accident (if any) and the release of fission products to the fuel channel. Codes used include
ELESTRES-IST for initial fuel conditions and ELOCA-IST for transient behaviour [Lewis, 2009].
Under certain circumstances, such as a large LOCA combined with loss of ECC injection, the
pressure tube will overheat and (depending on the internal pressure) sag or strain into contact
with the calandria tube. This requires models of the pressure-tube thermal-mechanical tran-
sient behaviour to predict the extent of deformation and the pressure-tube temperature and
internal pressure when or if it contacts the calandria tube. Such models are now part of the
system thermo-hydraulics code.
The behaviour of a channel subsequent to such contact depends on the heat transfer from the
calandria tube to the moderator. Further deformation will not occur if the calandria-tube outer
surface does not dry out, or at least does not go into widespread film boiling. This in turn
depends on the heat transfer coefficient between the pressure tube and the calandria tube, and
the local moderator sub-cooling. For the first, there are theoretical models and comparisons
with experiment – see, e.g., [Currie, 1984] and [Currie, 1986]. For the second, a two- or three-
dimensional prediction of moderator temperatures (and hence flows) is required. Of most
interest is the steady-state distribution at the time of contact, although transient calculations
are required for in-core breaks. The current industry code is MODTURC-CLAS. See [Xu, 1999]
for a brief description of MODTURC-CLAS and a good summary of thermo-hydraulic codes used
in CANDU.
Following the release of fission products from the fuel, their movement through the heat
transport system to containment and then within containment should be predicted. To date,
CANDU safety analysis has not used models for fission-product transport within the HTS and for
deposition on surfaces such as end fittings and feeder piping, although this is clearly an area
which could be included and has been the subject of intensive R&D over a number of years.
However, the partitioning of fission products between steam and water phases at the break and
within containment has been modelled, as has long-term formation and transport of organic
iodides within and from the water pool. See [Wren, 1999] and [Wren, 2001].
The containment-pressure transient calculation uses the transient energy release from the
break and includes sub-models for dousing, containment air coolers, fission products, hydrogen
transport, and natural and forced circulation. Multi-node, multi-fluid (water, steam, air, hydro-
gen) three-dimensional containment models are used for this analysis. The GOTHIC interna-
tional code is used for CANDU; see, e.g., [Andreani, 2010].
The final step is calculation of dose to the public. The atmospheric dispersion model typically
uses a Gaussian plume model to predict exposure as a function of distance from the station; the
input is the predicted transient release of radionuclides from containment for each accident.
See [Boss, 2006] for an example. The weather assumed is traditionally the worst weather
occurring more than 10% of the time at the site. Exposure-to-dose calculations use standard
ICRP-recommended conversion factors ([ICRP, 2010], [ICRP, 2012]).
Figure 45 shows this whole process as a flowchart.
Figure 45 Safety analysis codes
6.4 Code Validation and R&D

It is essential that the results of safety analysis codes reflect reality, or at least a version of
reality in which the predictive bias is understood. Extensive R&D is done to support safety
analysis codes, in facilities which range from small- to full-scale. The purpose of R&D is not to
simulate an accident, but to provide the fundamental scientific and engineering data against
which the codes can be validated (i.e., their results compared to experimental data). Experi-
ments generate scientific knowledge relevant to the phenomena in a nuclear reactor accident.
This knowledge is used to build engineering models, which in turn are used in digital computer
codes. The codes are then validated against the knowledge (using different data from those
used to develop the models). The validation can include comparison against small-scale and
integral tests, as well as real nuclear power-plant transients where data are available. The
codes are also independently verified to confirm that the mathematical and engineering models
have been represented correctly in the code language logic. Finally the verified, validated codes
are used to prepare the safety and licensing analysis procedures for planned or operating plants
or to assist in the design of advanced plants. This process is shown schematically in Figure 46.
Figure 46 Code validation process

Separate effects must be distinguished from integral tests. The former tend to focus on one
physical mechanism and are generally small-scale. Integral tests combine many phenomena
expected in the reactor and are generally large-scale; however, the interaction among the
phenomena can be complex. Hence, integral tests are used in the later stages of validation of
system models.
It is not possible within the scope of this Chapter to give a full accounting of all the R&D facili-
ties that support CANDU safety technology. A few of the major facilities are discussed briefly.
References are given to enable the student to follow up on any facilities of interest.
6.4.1 Reactor physics

The ZED-2 reactor [AECL, 2011] was described in the problem set in Section 2.7. It is a heavy-
water tank reactor used for fundamental physics experiments on natural uranium and advanced
CANDU fuels and lattices. It is used for measurements of reactivity coefficients such as coolant
void and temperature, and generates basic data for validation of physics codes. At the other
end of the scale, real plant transients such as shutdowns are used to validate some aspects of
CANDU kinetics at full scale.
6.4.2 Thermo-hydraulics
The major integral facility used for validation of system thermo-hydraulic codes is RD-14M,
shown in Figure 47 [Buell, 2003]. It is a full-elevation model of a typical figure-of-eight CANDU
reactor heat-transport system, at full pressure and temperature conditions, and with ten full-
length electrically heated channels. All HTS components are simulated: channels, end-fittings,
feeders, headers, and steam generators. Mass flux, transit times, and pressure and enthalpy
distributions are similar to CANDU. Tests cover all phases of a large or small LOCA scenario,
including blowdown and refill, natural circulation in single-phase and two-phase, and loss of
shutdown cooling.
6.4.3 Fuel
Separate-effects fuel testing is performed out-of-reactor. For integral tests, the NRU reactor has
a vertical in-core loop that can be loaded with CANDU fuel elements and subjected to a LOCA
(Blowdown Test Facility). Fuel behaviour during extended post-dryout operation has also been
studied in the NRU loops.
Figure 47 RD-14M schematic
6.4.4 Moderator thermo-hydraulics

The moderator thermo-hydraulic codes have been validated against a wide variety of tests,
ranging from pseudo-two-dimensional “slices” of a model calandria to in-reactor probes. The
most sophisticated experiments have taken place at Chalk River Laboratories in the Moderator
Test Facility, which is a one-quarter scale CANDU calandria containing 480 heaters to represent
the 480 channels in the Bruce/Darlington design. The facility can measure a grid of local tem-
peratures and local flow velocities in three dimensions.
6.4.5 Fuel channel thermo-mechanical behaviour

One of the advantages of the pressure-tube concept is that channels can be tested at full
diameter. The phenomena exhibited during pressure-tube strain to contact the calandria tube
(for large LOCA, or LOCA + LOECC) have been examined using full-diameter channels, in which
the pressure-tube is internally heated and expands under pressure. The entire pressure tube /
calandria tube assembly is immersed in a water environment which represents the moderator
(Figure 48) [Sanderson, 2003].
Figure 48 Contact boiling tests

By controlling the internal pressure, heating rate, and moderator temperature, a map can be
derived of the regions of heat transfer on the outside of the calandria tube just after pressure-
tube contact (Figure 49), which can be used to validate the channel thermo-mechanical codes.
Figure 49 Heat-transfer regimes on outside of calandria tube
6.4.6 Containment thermo-hydraulics

Three large-scale facilities have been used to validate containment codes, particularly hydrogen
behaviour (combustion and detonation characteristics and limits) [Krause, 2007].
The Large-Scale Containment Facility at Chalk River Laboratories has a room height of 10 metres
with a total volume of ~1,575 m3. It is used for studying hydrogen mixing behavior at high
temperature and high humidity, using helium as a hydrogen simulant. It is also used to study
the behaviour of wet aerosols (Figure 50).
The Large-Scale Vented Combustion Facility at Whiteshell (Figure 51) is a rectangular enclosure
with an internal volume of 120 m3 used for studying hydrogen burning in air/steam mixtures at
various temperatures. The combustion chamber can be subdivided into two or three compart-
ments, with vent openings of various sizes between the compartments and the outside.
The Containment Test Facility at Whiteshell has also been used to investigate combustion
phenomena in single and interconnected vessels, including detonation and the transition to
detonation. It consists of a 6 m3 sphere and a 10 m3 cylinder, which can be connected by ducts
(Figure 52).
Figure 50 Large-scale containment facility
Figure 51 Large-scale vented combustion facility
Figure 52 Containment test facility
6.5 Selection of Initial Conditions

A number of key parameters are chosen in a “conservative” direction for licensing analysis.
These include fundamental core property parameters, initial plant conditions, system perform-
ance measures, and assumptions on the unavailability of portions of mitigating systems. There
is no magic formula for doing this: too much conservatism may require unnecessary design
changes or severely limit operation; too little may not cover uncertainties in the models used or
the station parameters. This dilemma is largely resolved by a BEAU approach, discussed later, in
which all parameters are set at their best-estimate values and the uncertainties are combined
and propagated through the calculation.
Table 11 gives a sample list of parameters and assumptions and how they might be chosen
“conservatively”. Note that what is conservative in one application (e.g., minimizing the num-
ber of containment coolers credited in calculating peak containment pressure) may be non-
conservative in another (calculating high containment-pressure trip effectiveness). Note also
that whatever safety analysis assumptions are made become limits to operation.
Table 11 Some conservative assumptions and parameters

Parameter Conservative Direction Rationale
Reactor thermal power High Minimize time to use

up cooling water
inventory, minimize
margins to critical heat
flux, etc.
Reactor regulating system Normal operation or Choose so as to delay

inactive, whichever is reactor trip
worse; setback is
generally not credited
unless it tends to
“blind” the trip
Radionuclide operating load in the Highest permissible Maximize radionuclide

HTS operating iodine release from station
burden (and associ- and public dose
ated noble gases) plus
any “spiking” at the
time of reactor shut-
down, and end-of-life
tritium concentration
Pressure-tube radial creep Highest expected over Reduce margin to

the time scale over critical heat flux (due
which the safety
to flow by-pass around
analysis is to be validthe fuel bundle in a
crept tube) and
increase value of void
reactivity
Steam generators Clean and fouled cases Reduce reactor trip
effectiveness
Steam-generator tube leak rate Maximum permitted Increase radioactive

during operation, plus material release
assessment of any
consequential effects
due to the accident
HTS flow Low Reduce margins to

critical heat flux
HTS instrumented channel flow High Reduce low-flow trip

effectiveness
Coolant void reactivity coefficient High; Maximize overpower

transient;
Low Delay HTS high-

pressure trip
Fuel loading Equilibrium; Maximize fuel tem-

peratures, radioactive
material releases;
Maximize overpower
Fresh transient
Shutdown system Back-up trip on less Delay shutdown-

effective shutdown system effectiveness
system using the last
of three instrumenta-
tion channels to trip
SDS2 injection nozzles Most effective nozzle Reduce shutdown-

unavailable system reactivity
depth
SDS1 shut-off rods Two most effective Reduce shutdown

rods unavailable system reactivity
“bite” and depth
Maximum channel/bundle power High Earliest time to dryout

and maximize fuel and
sheath temperature
Reactor decay power High Minimize time to use

up cooling water
inventory
Initial flux tilt High Maximize fuel and

sheath temperatures
Moderator initial local maximum Low Minimize margin to

sub-cooling critical heat flux on
calandria tube
Number of operating containment Low; Maximize containment

air coolers and other heat sinks pressure;
High Delay high-pressure

trip and maximize
likelihood of hydrogen
combustion due to
rapid condensation of
steam
Containment leak rate High (typically 2× to Maximize public dose;

10× design leak rate);
Low Maximize containment

pressure
Containment by-pass leakage Pre-existing steam- Maximize public dose

generator tube leak
Weather Least dispersive Maximize public dose

weather occurring
>10% of the time
Operator actions Not credited before 15 Ensure adequate time

minutes after a clear for diagnosis
indication of the
event, for actions that
can be done from the
control room; and not
credited before 30
minutes, for actions
that must be done “in
the field”
6.6 Accident Walk-Through: Large LOCA

As an example, we shall walk through the phenomena of a large HTS pipe break. Then we will
cover more briefly a few more accidents which are representative, but not exhaustive.
6.6.1 Initiating event

A large LOCA in a CANDU is conventionally defined as one where the break area is larger than
twice the cross-sectional area of the largest feeder pipe. Therefore, a large LOCA can be located
only in the large piping above the core. There are three representative locations (Figure 53):
reactor inlet header (RIH), reactor outlet header (ROH), and pump suction line (PSH). Other
possible locations are lines connected to the pressurizer, the shutdown cooling lines, and the
header interconnect lines.
It is conventionally assumed that the pipe break is instantaneous, an assumption which bears
little relationship to reality, but is selected to ensure conservatism in that it maximizes the
predicted coolant voiding rate and hence the coolant void-reactivity insertion and reactor
power pulse. This in turn presents the greatest challenge to shutdown-system effectiveness.
Steam Generators
HTS Pumps
1) Pump Suction
Feeders
2) Inlet Header
Reactor/Fuel Channels
Figure 53 HTS layout
6.6.2 Event sequence (simplified)

 A large HTS pipe break is postulated to occur instantaneously.
 Steam discharges to containment, causing a rise in containment pressure and tempera-
ture.
 HTS pressure drops rapidly, causing a rapid decrease in local saturation temperature.
 Fuel channels downstream of the break experience flashing (rapid boiling). Fuel sheaths
may dry out.
 The reactor power rises quickly due to the positive void coefficient, causing a rapid in-
crease in fuel and sheath temperatures. Note that there are inherent limits to the speed
of voiding, which are set by the subdivision of some CANDUs7 into two separate PHTS
loops, by the fluid inertia in the fuel channels, and by the arrangement of each loop of
the PHTS in a “figure-of-eight” configuration. For further reading on this topic, see
[Popov, 2013].
 Flow in the downstream channels falls due to the change in pressure drop across the
channel (the pressure gradient from the pumps is offset by the pressure gradient due to
the break).
 The reactor is shut down by either shutdown system (~2 sec).
 Containment ventilation paths are isolated.
 The fluid inventory in the channels decreases and flow remains low, so that fuel and fuel
sheaths continue to rise in temperature. The rise is limited by steam cooling.
7
Pickering A and B, CANDU 6, and Darlington
 Fuel overheating increases the gas pressure inside the sheath relative to the coolant
pressure, which is falling, and can force the sheath to strain. Some fuel sheaths may fail.
The temperature excursion depends on the break size, which is traditionally chosen to
“stagnate” the flow in the downstream channels, e.g., 30%-40% RIH breaks8 in Figure 54.
Figure 54 Core flow vs. break size for a group of channels

 Containment dousing is initiated and starts to reduce containment pressure. In multi-
unit sites with a common vacuum building, valves to the vacuum building open, and
pressure starts to go sub-atmospheric.
 ECC injection is signaled at about 20 seconds. Valves on the gas tanks open to drive wa-
ter from the accumulator tanks into the HTS (or high-pressure ECC pumps start). Steam-
generator rapid (“crash”) cooldown is also initiated automatically, and the MSSVs open
to discharge secondary-side steam to atmosphere.
 Fuel temperatures stabilize and fall as the headers, and then the channels, refill.
 After the initial injection phase, ECC switches to medium-pressure and then to low-
pressure ECC, in which the water from the break is recovered from the building sumps,
cooled, and re-injected. This gives a stable end-state which can last for months.
The time scale is shown in Figure 55.
8
The percentage is twice the cross-sectional area of the pipe; hence, a 100% break is twice the cross-sectional
area, allowing for discharge of coolant from both ends of the broken pipe. Note that other breaks may also
stagnate at different times; which one results in the highest fuel and sheath temperatures is determined by
analysis.
Figure 55 Typical LBLOCA timescale

A typical containment-pressure transient for CANDU 6 is shown in Figure 56. The pressure rises
rapidly and is suppressed by dousing sprays. In the long term, the air coolers (and heat removal
by ECC) stabilize the pressure. This is quite different behaviour from a vacuum containment,
which will go sub-atmospheric once the valves to the vacuum building open and will stay sub-
atmospheric for days.
Figure 56 Containment pressure transient for 100% ROH LOCA
6.6.3 Barrier protection

For a LBLOCA, recall that some sheaths may fail, but that the amount of failure must be limited.
The fuel barrier may be degraded, but not to the extent that it jeopardizes the pressure-tube
barrier. The pressure-tube barrier and the containment barrier must not fail. Further failures in
the HTS barrier must be prevented. Moreover, systems which protect these barriers must not
fail. This gives rise to the following requirements (simplified—this is not a complete list) for
barrier protection, which the safety analysis must demonstrate:
 Fluid from the broken pipe will create jet forces, and reaction forces can cause pipe
whip. The effect of these forces on other HTS pipes, on the shutdown systems, on ECC
pipes, and on containment must not lead to further barrier failures.
 Continued fuel heating after reactor trip gives a potential for sheath failure due to exces-
sive sheath strain or sheath embrittlement, and for channel flow area reduction due to
sheath strain. Sheath failures must be limited to meet public dose limits; sheath failure
modes such as embrittlement or severe channel-flow blockage must be prevented to
protect the pressure-tube barrier (i.e., to maintain coolable bundle geometry).
 Fuel melting must be prevented to protect the pressure-tube barrier.
 Pressure tubes may overheat and strain or sag until they contact the calandria tube.
Failure of the pressure tube before contact with the calandria tube must be prevented.
After contact, heat transfer to the moderator must be sufficient to prevent failure of the
channel barrier.
 The containment barrier and its internal structures must not be jeopardized by excessive
internal pressure.
These requirements for barrier protection can be translated into analysis limits, as discussed
earlier. A detailed list of analysis limits is outside the scope of this Chapter.
6.6.4 A note on reactivity coefficients

A large LOCA is the limiting accident in terms of shutdown-system speed. This is due to the
positive void coefficient, and more specifically to the rapidity with which void is produced in a
large LOCA. CANDU has sometimes been criticized for its positive void coefficient; by contrast, a
LOCA in an LWR shuts the reactor down due to loss of the common coolant/moderator. How-
ever, a broader perspective is more useful. The following discussion is taken from [Popov, 2012]
and [Popov, 2013].
As discussed in Chapters 4 and 5, reactivity coefficients measure the amount of change in
reactivity per unit of change in the parameter of interest, while other parameters are kept
unchanged.
We shall develop three themes:
 all reactors have mechanisms to add positive reactivity
 reactivity changes can be slow or fast
 the reactivity changes of most interest to safety are those that are large and fast.
The sign of a reactivity coefficient is less important because the parameter can al-
ways be reversed.
The first theme is readily apparent: addition of positive (and negative) reactivity through control
rods, for example, is essential in power-reactor operation. Inherent reactivity feedback from
fuel temperature, coolant density, moderator density, etc. (see Chapters 4 and 5) can also
introduce positive reactivity, with the details depending on the reactor design and the direction
in which the particular parameter is changing.
As for the second theme: the speed of reactivity feedback in an accident determines the re-
quired countermeasures. Slow reactivity changes are relevant for normal operation in different
reactor states and are introduced in the reactor core by: a) burnup and refuelling, b) planned
control-device movement, and c) xenon changes in the core during operation. In CANDU,
fuelling is done on-line, and therefore the control system does not have to compensate for fuel
burnup. In other words, the reactivity worth of control devices is small (the total worth for all
devices together is 32 mk, and each device is typically worth less than a mk – see Chapters 4
and 5). Reactivity addition from control devices is inherently slow because they are not subject
to pressure-assisted ejection. In PWRs, fuelling is done in batch mode (about every 18 months),
and the excess initial reactivity after fuelling is suppressed by a combination of poison in the
coolant and fuel, and insertion of control rods. As a result, control rods may have high individ-
ual reactivity worth. A typical PWR has a total worth of all control rods of about 110 mk, and
the maximum worth of a single rod (using the Westinghouse Advanced Passive 1000 (AP 1000)
PWR as an example) can be of the order of 7.5 mk under the most pessimistic conditions (hot
zero power in a core just about to be refueled) [Westinghouse, 2011]. This is larger than the
delayed neutron fraction, and therefore rod ejection assisted by coolant pressure can result in
prompt criticality. This means that whereas in normal operation, control rods add reactivity
slowly, in an LWR accident, reactivity can be added very quickly.

Fast reactivity changes are of interest in accidents. These are induced by fast changes in pa-
rameters such as: a) coolant and moderator volume, temperature, or density, b) fuel tempera-
ture, c) location of reactivity devices, or d) reactor core geometry. As discussed in Chapters 4
and 5 and in our “lessons learned” from the SL-1 accident (Section 3.1.1.3), the response to a
large and fast reactivity addition depends on (1) the reactivity added compared to the delayed
neutron fraction , and (2) the prompt neutron lifetime lp. PWRs and CANDUs have similar
delayed neutron fractions: from 0.0054 to 0.0073 mk in a PWR, depending on core burnup, and
0.0053 for an equilibrium-core CANDU. However, the prompt neutron lifetimes are very differ-
ent. For a PWR, lp is typically 20–30 μs; for a CANDU, lp is about 900 μs, or 30–45 times longer
than in a PWR: because heavy water does not absorb neutrons as much as light water, neutron
lifetime in CANDU reactors is much longer.
Figure 57 shows the relationship between reactivity ρ and reactor period for various values of
prompt neutron lifetime (adapted from [Glasstone, 1994]). For ρ << β, the period for CANDU
and PWR reactors for the same reactivity insertion is approximately the same. As ρ approaches
β, the period for PWR reactors decreases sharply, whereas for CANDU, it also decreases, but
fairly smoothly. For ρ > β, the period for PWR reactors is about 40 times shorter (i.e., faster)
than for CANDU. Figure 57 also shows typical values of reactivity changes during certain acci-
dent classes for a PWR and a CANDU reactor, as shown by the coloured areas.
From [Popov, 2012]:
“Consider the fastest reactivity accident in PWR, i.e., rapid ejection of the largest-worth
control rod in a PWR. Using the value of 7.5 mk for “worst” rod worth, the reactor pe-
riod can be estimated of the order of a hundredth of a second following rod ejection. It
is not practically possible for engineered shutdown systems to stop such a fast rate of
rise in power. For PWRs, it is not necessary to have such fast shutdown system because
of a strong negative fuel temperature coefficient. Hence, PWRs can partly mitigate this
event through negative feedback, but engineered shutdown is still required after the ini-
tial transient”.
Consider now the fastest reactivity accident in CANDU: the large LOCA. Since there are
inherent physical limits to the speed and extent of channel voiding, the net result is that
the maximum rate of reactivity addition is approximately 4 mk/sec, and the reactor pe-
riod is about one second or so. At that rate of reactivity insertion, engineered shutdown
systems can be reliably designed to terminate the power rise.”
Current CANDU safety analysis does not predict prompt criticality for a large LOCA, but even if it
did, the period would not change dramatically. Prompt criticality is not the threshold of a
sudden change in behaviour for a CANDU.
Figure 57 Response of PWR and CANDU to reactivity increase

As for the third theme: ideally, all accidents would induce negative reactivity feedback. How-
ever, this is not possible because a large negative coefficient in one event can become a large
positive one for the complementary event. In other words, for every physical change that
produces a negative reactivity effect, the opposite change produces positive reactivity. For
example, insertion of a control rod produces negative reactivity; withdrawal (or ejection) of a
control rod produces positive reactivity. Likewise, for a reactor with a positive coolant void (or
density) coefficient, a decrease in coolant density produces a positive reactivity, whereas void
collapse gives a negative reactivity. Conversely, if the coolant void reactivity is negative, a
decrease in coolant density produces negative reactivity, whereas void collapse gives positive
reactivity.
Table 12 compares the size and sign of reactivity coefficients in PWR and CANDU.
Table 12 Reactivity effects in PWR and CANDU
Reactivity Effect PWR CANDU

Fuel temperatureLarge Negative ~0
increases (-0.023 to -0.029 mk/oC)
Fuel temperatureLarge Positive ~0
decreases (+0.023 to +0.029 mk/oC
Coolant voids Large Negative Large Positive
(-2.5 mk/% void) (+0.15 mk/% void)
Coolant void collapses Positive Negative
(+2.5 mk/% void) (+0.15 mk/% void)
Coolant temperature Negative Small Positive
increases (-0.09 mk/°C at BOC to - (+0.04 mk/°C)
0.54 mk/°C at EOC)
Coolant temperature Positive Small Negative
decreases (+0.09 mk/°C at BOC to (-0.04 mk/°C)
+0.54 mk/°C at EOC)
We can now summarize the impact on
safety of the speed of reactivity feedback
and the size and sign of reactivity coeffi-
cients:
 Only reactivity coefficients driven
by fast phenomena are highly im-
portant to safety; if the phe-
nomenon driving a coefficient
acts slowly, whatever its sign or
size, its impact can easily be com-
pensated for by reliably engi- Figure 58 Importance of reactivity effects
neered reactivity control and
shutdown mechanisms.
 The size of the reactivity coefficient is clearly important to safety because it determines
the magnitude of inherent effects and the performance of the engineered systems that
must control or compensate for them. Generally, a small reactivity coefficient is not im-
portant because the amount of compensation needed is also small, regardless of the
sign or speed. Conversely, a large reactivity coefficient is important, regardless of the
sign.
 The sign of a reactivity coefficient is less important to safety because the physical
mechanism can almost always be reversed, as discussed above.
This concept is illustrated in Figure 58.
The use of inherent and engineered safety defences in PWRs and CANDU is summarized for
some typical accidents in Table 13. Relevant coefficients are shown in brackets.
Table 13 Examples of reactivity response to accidents for PWR and CANDU

Accident PWR CANDU
Large LOCA Power decrease, ECC boration required in
Rapid power increase
medium term to ensure shutdown terminated by engi-
(void reactivity coefficient, fuel temperature
neered shutdown
reactivity coefficient) (void reactivity
coefficient)
Cold-Water Power increase, terminated by a combination Power decrease
Injection of inherent feedback and engineered shut- (coolant temperature
down, e.g., boron addition and void reactivity
(moderator/coolant temperature and fuel coefficients)
temperature reactivity coefficients)
Main Steam-Line Power increase, terminated by a combination Power decrease,
Break of inherent feedback and engineered shut- engineered shutdown
down; possible return to criticality termi- required in medium
nated by boron addition term
(moderator/coolant temperature and fuel (coolant temperature
temperature reactivity coefficients) and void reactivity
coefficients)
Control-Rod Rapid power increase with local power Not physically possi-
Ejection peaking terminated by combination of ble
inherent feedback and engineered shutdown
(delayed neutron fraction, moderator/coolant
temperature, and fuel temperature reactivity
coefficients)
For further information, a detailed and quantitative comparison of power transients in LWRs
and CANDUs has been given by Meneley and Muzumdar in two papers. The first [Meneley,
2009] gives an overview comparison of positive reactivity insertion accidents, whereas the
second [Muzumdar, 2009] looks specifically at large LOCA events in CANDU. These papers are
recommended reading for those wishing to understand further the safety differences and
similarities between the two reactor concepts.
6.7 Accident Walk-Through: Small LOCA

A CANDU has much more small-diameter piping than large-diameter piping because of the large
number of parallel channels, each served by an inlet and outlet feeder. A break in any of these
pipes (or a similar-sized break in the larger pipes) is considered a small LOCA. This includes a
break in a steam-generator tube.

Generally, the event sequence is similar to a large LOCA, but much extended in time. The rate
of rise of void reactivity is slow enough to be compensated for by the control system, and
therefore shutdown systems are triggered on process trips (see Table 8) rather than neutronic
trips. Typical time scales are a few minutes before the ECC set-points are reached, and similarly
the high-pressure and medium-pressure ECC phases are prolonged.
For a steam-generator tube break, the radionuclides contained in the HTS coolant can be
released outside containment. The break must be isolated in the longer term because the
water lost through the steam generator is unrecoverable. Because the discharge is small, the
operator has enough action time to perform HTS cooldown and steam-generator isolation.
Back-flow of (light) water from the secondary side to the primary side, after the latter has been
cooled down and depressurized, causes negative reactivity and is not a safety concern. Recent
CANDUs have manually operated main steam isolation valves which can be used as one of the
means to isolate the affected steam generator in the longer term.

Because a small LOCA may occur during the station lifetime, the safety systems are designed to
prevent fuel-sheath failure both before reactor trip (to limit the period of pre-trip overheating
due to dryout as the channels void) and afterwards, during ECC.
6.8 Accident Walk-Through: Single-Channel Event

The channel nature of the CANDU design means that certain accidents can affect one channel
only, e.g., partial or complete flow blockage due to a foreign object in the HTS, a pressure-tube
failure (which may or may not lead to a calandria-tube failure), a break in an individual feeder
pipe, or assumed failure of the pressure-tube-to-end fitting rolled joint, which can cause an
end-fitting ejection.

The last three initiating events listed above behave like a small LOCA as far as the heat-transport
system is concerned. The safety systems (shutdown, ECC, and containment) are triggered on
small LOCA signals. Flow blockage may not result in a channel failure; for blockages up to about
95% of the channel flow area, the fuel in the channel may fail and could be badly damaged, but
the pressure tube will not fail. For more severe blockages, the fuel will melt, and the channel
will fail.
Each type of single-channel event has certain unique characteristics in addition to those of a
small LOCA. These characteristics are the focus of the safety analysis for these events.
 A break in an inlet feeder of exactly the right size to cause the flow to stagnate tempo-
rarily in a channel can result in channel overheating and failure.

 Any event which leads to channel failure will cause a discharge of the HTS into the mod-
erator. Besides pressurizing the calandria, the discharge can dilute any poison in the
moderator used for reactivity suppression, as discussed in Section 5.2.3.
 The jet and pipe-whip forces from an in-core break can damage a limited number of
shut-off rod guide tubes, as discussed in Section 5.2.7.
 A severe flow blockage or feeder stagnation break can also cause discharge of molten
fuel into the moderator.
 An end-fitting failure can cause discharge of the fuel in the channel into the reactor vault
(i.e., directly into the containment).

A fundamental safety requirement of a channel reactor is that a single-channel failure must not
propagate to other channels; otherwise, a relatively benign and moderate-frequency initiating
event could cascade to become a severe core-damage accident. This conclusion has been
supported for CANDU by years of R&D and code development.
Shutdown depth must remain adequate, accounting for damaged shut-off rod guide tubes.
Moderator pressure increases after an in-core break due to the discharge of high-enthalpy fluid
into the moderator. The pressure is relieved by discharge through the four calandria relief
pipes, which are designed to prevent structural failure of the calandria shell. Recent work
continues to support this conclusion even when significant amounts of molten fuel have been
ejected into the moderator after a severe flow blockage.
6.9 Accident Walk-Through: Loss of Reactivity Control

A malfunction in the reactor regulating system (RRS) is assumed to drain zone controllers, drive
out absorber/adjuster rods, or both. Two types of accidents are considered: continued increase
in reactivity at up to the maximum possible rate and to the maximum degree allowed by the
physical configuration of the devices; and a slow power increase from both normal and dis-
torted flux shapes that terminates just below the overpower trip set-points.

An increase in reactor power causes a flow/power mismatch which has the potential to damage
the fuel. If the ramp is continuous, shutdown-system bulk overpower trips can be designed to
prevent fuel damage.
However, an increase in power from a distorted flux shape, which is either very slow or stops at
an elevated power, could permit fuel to remain in a dryout condition even if the bulk reactor
power is below the average overpower trip set-point. Analysis of such events determines the
trip set-points for the spatially distributed regional overpower (ROP) flux detectors on each
shutdown system.

As with a small LOCA, because loss of reactivity control may occur during the station lifetime,
the shutdown systems are designed to prevent sheath failure.
6.10 Accident Walk-Through: Loss of Forced Circulation

Loss of Class IV electrical power to the HTS pumps causes them to run down and eventually
stop. Partial loss of forced circulation is also possible; particular cases include a partial loss of
Class IV power (to two HTS pumps), a single HTS pump shaft seizure, and a single pump trip.

The flow reduction causes a mismatch between reactor power and coolant flow that can lead to
fuel overheating and HTS pressurization. The power mismatch also causes void formation in the
channels, leading to an increase in reactor power. This initial phase is terminated by a reactor
trip. In the long term, the fuel decay heat is removed by thermo-siphoning to the steam gen-
erators until an alternate heat sink (such as shutdown cooling) can be initiated.

Because this is an event that is expected to occur during the station lifetime, the shutdown
systems are designed to prevent sheath failure and to limit the stress levels in the HTS due to
overpressure. Ideally, the fuel temperature excursion should be limited enough that plant
operation can be resumed without the need for a fuel inspection.
6.11 Accident Walk-Through: Loss of Secondary-Side Heat Removal

We give a particular case as an example: failure of one of the main steam lines inside or outside
the reactor building (or the steam balance header) on the secondary side. This accident is more
limiting than feed-water line failures or loss of feed-water pumps.

A large steam line break causes a rapid depressurization of the affected steam generator and of
the remaining steam generators as well because they are connected at the balance header.
Initially, depressurization of the secondary side causes a corresponding depressurization and
cooling of the primary side because of enhanced heat transfer through the steam-generator
tubes; this causes negative reactivity and power decrease and is not a safety concern. The
reactor will trip early in the accident on signals such as high containment pressure (if applica-
ble9), low steam-generator level, or low feed-water line pressure. As the steam generators lose
9
For Bruce and Darlington, the steam line runs entirely outside the containment boundary.
inventory, heat removal from the heat-transport system degrades, and the operator is required
to initiate an alternate heat sink such as the shutdown cooling system. Note that because the
shutdown cooling system is a high-pressure heat-removal system, the operator does not need
to depressurize the HTS before initiating this alternate heat sink in emergencies.

The safety assessment must demonstrate that three key requirements are met:
 The operator has enough time to initiate an alternate heat sink and maintain fuel-sheath
integrity.
 For main steam-line breaks outside containment, in the turbine hall, one must assess the
design to show that equipment which is required and assumed to mitigate the event (if
located in the turbine hall) is not damaged by the forces, the steam, or the high tem-
perature caused by the break, nor is there damage to the turbine-hall structure.
 For main steam-line breaks inside containment, the containment pressure rises rapidly,
and an additional safety aspect is the preservation of building integrity, including the in-
tegrity of reactor-building internal structural walls. For new reactor designs, it is a re-
quirement that the peak containment pressure be below design pressure.
Large steam-line breaks are limiting in terms of early containment peak pressure and time
available to introduce an alternate heat sink. Small steam-line breaks test the trip coverage and
(for designs where part of the steam line is within containment) can lead to long-term contain-
ment pressurization after the containment dousing water has been exhausted.
6.12 Accident Walk-Through: Fuel-Handling Accident

A fuelling machine carrying spent fuel may be either on the reactor (attached to a channel), or
off the reactor (in transit to the spent-fuel port, or attached to the spent-fuel port and discharg-
ing fuel). The spent fuel must be kept cooled, and hoses are attached to the fuelling machine
through which high-pressure D2O cooling water is pumped. Should one or more hoses fail, the
integrity of the contained fuel is threatened. The consequences to the reactor of a failure of a
fuelling machine on-reactor (e.g., spurious detachment from a channel) are broadly similar to
those of a single-channel event, except that more fuel bundles may fail.

Assume that the fuelling machine is off-reactor. Following loss of coolant through severance of
one or more hoses, the D2O inventory in the fuelling machine will boil off, the bundles will be
uncovered, and fuel may fail, releasing fission products to containment.

A fuelling-machine failure when off-reactor cannot be mitigated by reactor shutdown or ECC;
the only relevant safety system is containment, so that the safety analysis focuses on fission
product release from the fuel bundles and containment effectiveness.
6.13 Accident Walk-Through: Loss of Moderator Inventory or Heat Removal

Moderator system failures include moderator pipe break, loss of forced circulation, and loss of
heat removal.

The moderator does not contain fission products, but does contain activation products (tritium).
Failures in the moderator system can therefore release tritium to containment and cause flux
distortions if the moderator level falls, leading potentially to excess power in some reactor fuel
channels before the reactor is tripped.

Fuel-sheath failures must be prevented (by tripping the reactor in case of excessive flux distor-
tion). The containment capability is analyzed to show that it can limit dose to the public from
tritium.
6.14 Severe Accidents

It is useful on CANDU to distinguish two categories of severe accidents:
1. Severe fuel-damage accidents in which the core geometry is preserved (fuel may be
damaged, but remains inside intact pressure tubes). Decay heat is removed from the
channels to the moderator. Example: loss of coolant plus loss of ECC with moderator
cooling available. A variant on this is “limited core damage”, in which one or a few fuel
channels fail, but the others remain intact. Either configuration is unique to a channel
reactor where the heat can be removed by the moderator; there is no analogy to this
“half-way” point in LWRs, in which a severe accident implies loss of core geometry.
2. Severe core-damage accidents, in which the fuel channels fail and collapse to the bot-
tom of the calandria, i.e., the moderator heat sink is ineffective. Example: loss of cool-
ant plus loss of ECC plus loss of moderator heat removal.
Analyzes of accidents in the first category use similar tools to those for design basis accidents
and will not be discussed further.
CANDU is characterized by large volumes of water around the core, consisting of the moderator
and the reactor vault or shield tank (Figure 59). Even with no active heat removal from either
system, it takes about 20 hours for the water to boil off and for the debris to end up on the vault
floor [Snell, 1988].
Figure 59 Sources of water near the fuel

Blahnik [Blahnik, 1991], using the MAAP-CANDU code, has characterized the degradation of a
CANDU core with no cooling and gradual boiling-off of the moderator. The uncovered channels
heat up and slump under their own weight until they are held up by the underlying channels.
Eventually, as successive layers of channels pile up, the supporting channels (still submerged)
collapse, and the whole core slumps to the bottom of the calandria vessel (Figure 60, Figure 61).
See [IAEA, 2008] and [Blahnik, 2012] for further reading.
Details of the channel failure mode are as follows: the axial creep rate of the pressure-tube
material (Zr-2.5% Nb) increases rapidly with temperature, and excessive sagging of the channel
is expected to occur above ~1200°C. In Blahnik’s model, a sagging channel comes into contact
with the next lower row of channels. The lower row of channels may or may not be cooled
adequately by the moderator, depending on whether it is submerged in the moderator.
Figure 60 Channel collapse in severe core-damage accident
Figure 61 Core collapse

As the moderator level continues to decrease, the lower row of channels is uncovered and sags
under its own weight as well as that of the supported channels [Simpson, 1996]. This process
continues as more channels are uncovered. As sagging increases, channel segments separate
near the bundle junctions by sag-induced local strain. A suspended debris bed is thus formed
which moves downward with the falling moderator level. Because a submerged channel can
support only a certain number of channels, the ends of those channels are expected to fail by
shear. This process will increase the loading on the channels below, leading to progressive
failure of the channels and resulting ultimately in the collapse of the reactor core into the water
pool in the bottom of the calandria vessel.
To address the plant state once the debris has collapsed to the bottom of the calandria, Rogers
et al. [Rogers, 1995] developed an empirically based mechanistic model (Figure 62) of the
collapse process, which showed (assuming that the reactor vault is kept full of water) that the
end-state consists of a bed of dry, solid, coarse debris irrespective of the initiating event and the
core-collapse process. Heating of the debris bed is relatively slow because of the low power
density of the mixed debris and the spatial dispersion provided by the calandria shell, with
melting possibly beginning in the interior of the bed about two hours after the start of bed
Figure 62 Mechanistic model of channel collapse

heating. The upper and lower surfaces of the debris remain well below the melting point
(Figure 63), and heat fluxes to the shield-tank or reactor-vault water are well below the critical
heat flux of 200W/cm2 under existing conditions (Figure 64). The calandria vessel is protected
by a solid crust of material on the inside and by water on the outside, and therefore it can
prevent the debris from escaping. Should the shield tank or reactor vault water not be cooled
or replenished, it will boil off, and the calandria vessel will eventually fail by melt-through, but
this will not occur in less than about a day.
Figure 63 Debris heating transient
,
,
Figure 64 Heat flux on calandria wall

In short, compared to a core melt in an LWR, a severe core-damage accident in CANDU is
inherently:
 Slow, because of all the passive heat sinks around the core
 Incoherent, because the channels fail individually over time, forming a coarse debris,
rather than rapid “candling”
 In a favourable geometry, either retained in the calandria shell or, should the shell
fail, in the shield tank / reactor vault. Only if these also fail will debris be in contact
with the containment boundary
 Subcritical
 Amenable to mitigation by adding water to either the moderator or the reactor vault
/ shield tank.
Some validation of CANDU severe-accident phenomena has been carried out at both large and
small scales. Results obtained from the RASPLAV facility in Moscow have been used to support
heat transfer through vessel walls and retention of debris in a vessel: see [Ader, 1998] and
[Rogers, 2002]. See [Mathew, 2004] for a summary.
6.15 Problems
1. Consider loss of cooling of the spent-fuel bay due to loss of electrical power to the
cooling system. What operating parameters would have the most influence on the outcome,
and how would you select conservative values? What tools would you need to analyze the
accident (i.e., what capabilities would they have to have)?
2. Estimate the evolution (using hand calculations if necessary) of the following severe
accident in CANDU: small loss of coolant plus loss of ECC (assume that crash cooldown is
available because it is on a redundant signal) plus loss of moderator cooling. Write down the
expected event sequence (based on the list below) and estimate the approximate time of:
 reactor trip
 start of fuel overheating
 failure of first channel
 core collapse
 shield-tank failure
 containment behaviour.
Only an approximate answer is sought (to do this accurately could take weeks). If you cannot
obtain the physical data in some cases, especially for the last item, use symbols to show how
you would do the calculation. One approach is to use heat balance.
3. Consider a CANDU that has undergone a severe core-damage accident. Assume that the
core has collapsed into a debris bed at the bottom of the calandria and is being cooled by
boiling of the shield-tank water. Calculate the depth of the debris bed in the calandria (assum-
ing zero porosity) and the average heat flux through the calandria wall. Do the calculation at 12
hours after the event. You will need to look up typical CANDU geometry [AECL, 2005] and
calculate the core decay heat. Do not strive for high accuracy.
7 Safety Analysis – Mathematical Models

In this Section, we describe the basic science underlying safety analysis and present in simplified
form the types of models used. We shall not derive each model from first principles. The
reader is referred to numerous textbooks and reports for details (see Section 11). We borrow
heavily from lecture notes by D. Meneley, J. T. Rogers, and W. Garland.
We focus on the large LOCA because it covers most of the disciplines used.
7.1 Reactor Physics

The fundamental equations in reactor physics have been presented in Chapters 3 to 5. The
main aspects of physics used in safety analysis are reactor kinetics and diffusion theory. For the
kinetics of a (point) reactor:
dN f (t ) k (    ) M
 N f (t )   i Ci , (32)
dt lp i 1
where:
Nf(t) is the number of neutrons as a function of time t
k is the multiplication factor for an infinite reactor
ρ is the reactivity
 is the total delayed neutron fraction
lp is the prompt neutron lifetime
M is the number of delayed neutron groups (chosen by the analyst, typically 6)
i is the decay constant of the ith delayed neutron group
Ci is the number of neutrons of the ith delayed neutron group,
and for each delayed neutron group:
dCi  i N f (t )
  i Ci , (33)
dt l p (1   )
where i is the delayed neutron fraction of each group i.

However, a CANDU cannot be considered a point reactor in accidents. Strong spatial effects
result from flux tilts (especially for a large LOCA, if only half the core is voided), insertion of
shut-off rods from the top of the reactor, or other events. Therefore, in practice, one must use
three-dimensional diffusion equations in conjunction with point kinetics. The diffusion model
tracks the neutrons as a flow through a medium subject to scattering, absorption, and leakage.
The continuity equation, for uniform systems with a single neutron velocity v, becomes the
neutron diffusion equation:
1 
D 2   a  s  , (34)
v t
where:
D is the diffusion coefficient
a is the absorption cross-section

 is the neutron velocity
(r,t) is the neutron flux.
s(r,t) is the source distribution function, i.e., the number of neutrons emitted per unit
volume per unit time by sources at point r and time t.
This is very similar to the diffusion of heat through a medium, with the addition of a source
term (due to fission).
Transient neutron diffusion plus kinetics is the basis of many of the physics codes used in
accident analysis. These equations are solved using spatial finite-difference methods with the
reactor core broken up into many nodes, in each of which the diffusion equation is applied at
each time step.
For a LOCA, for example, time-dependent cross sections representing coolant voiding transient,
fuel temperature transient, reactor control-system action, and shut-off rod movement are the
driving functions for the transient in the flux. The average power transient for the reactor is
extracted from the calculation, along with peaking factors for the hot channel, hot bundle, and
hot element in the loop being analyzed. The system reactivity as a function of importance-
weighted void fraction is also calculated. These data are used as input to a system thermo-
hydraulics code. This code calculates the distribution of coolant void as a function of time from
the pipe break for each pass of the coolant loop. The importance-weighted average void
fraction calculated by the thermo-hydraulics code is combined with the previously calculated
reactivity function, and the cycle is iterated as necessary. The thermo-hydraulics code and the
physics code are combined into one calculation so that manual iteration is not necessary.
7.2 Decay Power

We indicated in Section 3.2 that shutting down a reactor was not sufficient to remove all safety
concerns: the decay heat must be removed. This heat comes from the radioactive decay of
fission fragments and obviously cannot be controlled. In principle, it can be calculated from the
power history: the composition of fission products can be predicted at any time, and their half-
lives and decay chains are known, so that:
Pd (t )   ni (t )Ei , (35)
i
where:
Pd(t) is the power produced by all decaying fission products at time t
ni(t) is the number of atoms decaying per unit time of fission product i at time t
Ei is the average energy produced by the decay of each atom of fission product i.
This is more complex than it seems because ni(t) will depend on the irradiation history, each
fission product may have more than one decay chain (with different energies), and many fission
products are generated. In practice, such a fundamental calculation is done as a reference,
assuming a rapid shutdown after equilibrium operation. In most cases, the evolution of the
accident (because it is short) does not materially change the decay power, so that the results
from the fundamental calculation can be fitted using a series of exponentials, and the curve fit
is simply added to the fission power predicted by the physics codes. In exceptional cases, such
as an accident occurring during power manoeuvring, the result will not be very accurate, but as
with any analysis, an upper bound can be assumed (e.g., assuming the decay power appropriate
to prolonged steady-state operation at the maximum operating power level before the acci-
dent). Such a trade-off between simplified assumptions which give a conservative bound to an
answer, and more realistic assumptions which require more sophisticated analysis tools, is very
common in safety analysis. Because safety analysis resources, like all other resources, are finite,
part of the art of safety analysis is in judging when to use bounding assumptions and when to
use realistic complex models. The difficulty is, of course, that bounding models may lead to
unnecessary restrictions on operation, as discussed in Section 6.2.
7.3 Fuel
The key safety parameters related to fuel are:
 fuel-sheath integrity
 fission product inventory in (and release from) the fuel, and
 fuel temperature.
Prediction of fuel temperature in CANDU is important because:
 it drives sheath temperature and hence sheath integrity;
 high fuel temperatures drive pressure-tube deformation rates;
 limited fuel melting can lead directly to fuel-sheath failure;
 extensive melting can lead to pressure-tube failure;
 release of fission products from the fuel increases with fuel temperature;
 reactivity feedback from fuel temperature does occur, although it is small in CANDU.
7.3.1 Fission products in the fuel

During normal operating conditions, all fission products are formed within the fuel grains; they
are trapped there (with the exception of a few which recoil directly into inter-grain and gas
gaps) until they are released from the grains by diffusion. Those volatile fission products which
are released form a gas mixture inside the fuel sheath. It is therefore useful to categorize fission
products into three groups: (a) bound inventory, (b) grain boundary inventory, and (c) gap
inventory (recall Section 1.7).
The gap inventory includes the fission product gases in the pellet dishes, in the pellet/sheath
gap, and in the sheath end cap. If the fuel sheath fails, the gap inventory escapes quickly, the
grain boundary inventory much more slowly, and the bound inventory even more slowly. At
relatively low fuel temperatures, these diffusion processes are very slow, so that almost all
isotopes remain in the grains. (An exception to this occurs when uranium dioxide is exposed to
air at moderate temperatures; in this case, oxidation to higher states takes place, and the grain
structure is destroyed. Much of the fission products are then released. Such a circumstance
could occur after failure of an end-fitting and ejection of the channel contents into the reactor
vault). As temperature and fuel burnup increase, the amount of fission gas in the gap increases
to a maximum of about 10% for typical CANDU operating conditions [Lewis, 1990].
At any given burnup, a larger fraction of the fission product gases is released near the centre of
the pellet, where the temperature is highest. All volatile fission products tend to migrate down
the temperature gradient toward the outside of the pellet. Their diffusion is assisted by the fact
that the pellet cracks under the influence of the temperature gradient; this cracking increases
with fuel burnup and pellet centre temperature.
In summary:
(a) at low burnup or temperature, nearly all fission products are trapped in fuel grains;
(b) fission products trapped at grain boundaries increase with temperature and burnup;
(c) the gas-gap inventory increases steadily with fuel temperature and burnup.
7.3.2 Fuel temperatures

Consider three-dimensional steady-state heat conduction in a medium. Assuming that energy is
being produced in the medium at a volumetric rate H, energy conservation can be stated as:
(Rate of change of internal energy) = (rate of energy release into medium) - (rate of en-
ergy loss from conduction),
or at any point,
T
c  H  k  2T , (36)
t
where T is temperature, k thermal conductivity, ρ the density, and c the specific heat. Note the
similarity to the neutron diffusion Equation (34).
Consider one-dimensional steady-state heat conduction in a cylindrical fuel pin. In cylindrical
co-ordinates,
d 2T 1 dT H
2
  , (37)
dr r dr kF
where kF is the fuel thermal conductivity. Integrating,
Hr 2
T (r )  T (0)  , (38)
4kF
giving the typical parabolic radial distribution of temperatures within an oxide fuel element, as
shown in Figure 65.
Figure 65 Temperature distribution across a fuel pin

We can apply the same method to the temperature drop ΔTS across the sheath (for which the
internally generated heat Hs=0):
Ha 2 ln[(a  b) / a]
TS  TSi  TSo  , (39)
2k S
where the indices i and o refer to the inner and outer surfaces of the sheath respectively. It
would be tempting to conclude that the temperature drop from the fuel centreline to the outer
part of the sheath is simply ΔTF +ΔTS. Not so—the gap between fuel and sheath provides a
further thermal resistance, so we can write:
q  hg (TF  TSi )
, (40)
where hg is the gap heat-transfer coefficient, which is a very complex function of surface rough-
ness, contact pressure, and temperature and is determined by experiment.
The relationship between sheath and coolant temperature can be again derived from:
q  h(TSo  TC ) , (41)
where q is the heat flux per unit area to the coolant, TC is the coolant temperature, and h is the
convective heat-transfer coefficient. At steady state, all the heat produced in the fuel is trans-
ferred to the coolant, so that for a length of element ℓ,
H a 2
q , (42)
2 (a  b)
so that
Ha 2
TSo  TC  . (43)
2h(a  b)
Typical values in CANDU fuel are (see [Page, 1972] for a summary):
kF = 0.004 kW/m.°C
kS = 0.017 kW/m.°C
hg = 7–60 kW/m2.°C
a = 6.07 mm
b = 0.42 mm.
Heat capacity, c, does not enter the steady-state equations, but is important in transients. The
values for uranium dioxide and Zircaloy-4 are respectively 0.5 and 0.4 J/g-°C. The corresponding
heats of fusion are 27 and 42 J/g. The relatively large value of c and the high melting point
(2840°C) for UO2 represent important safety characteristics: typically, one can almost double the
stored energy relative to the normal operating point, inside UO2 before it melts. Metal fuel, on
the other hand, has little heat capacity and melts at a lower temperature, but has much higher
thermal conductivity, typically ten times more than UO2.
7.3.3 Internal gas pressure

CANDU fuel has a “collapsible” fuel sheath, which creeps down plastically onto the pellet during
irradiation due to the excess external coolant pressure. The small enclosed gas space in the
element results in high sensitivity of gas pressure to fuel-sheath geometry. A small amount of
fill gas is added to this space on assembly to achieve the proper sheath stress distribution
during operation. As burnup increases, gas pressure causes the sheath once again to lift off the
fuel; the gap heat-transfer coefficient decreases because the sheath creeps away from the fuel
pellet. This decrease leads to higher peak fuel temperature, greater fission gas release from the
fuel, and finally higher gas pressure. A new equilibrium point is reached. In an accident, gas
pressure is the driving force for sheath strain. Gas pressure can be modelled using the ideal gas
law. The volume depends on the transient behaviour of the gas gap, which results from com-
plex models of sheath and fuel behaviour and is beyond the scope of this Chapter.
7.3.4 Sheath strain – large LOCA example

Consider a large LOCA as an example. One of the objectives in LOCA analysis is to predict the
number of fuel sheaths which fail; this determines the amount of radioactive material released
to the coolant and is input into the calculations of release into containment and public dose.
Because the fuel heats up quickly, the gas pressure inside the sheath increases relative to the
coolant pressure and can force the sheath to strain. A uniform strain of at least 5% will not lead
to sheath failure; however, greater strains can lead to local instability, ballooning, and failure.
The strain-rate equations are complex functions of material composition, material state, tem-
perature, irradiation, and transverse stress and are determined from experiments. Typically,
their form is established as follows.
Define the transverse stress  across the sheath as:
Pr
 , (44)
w
where:
P is the pressure differential across the tube,
r is the tube radius,
w is the tube thickness.
The strain rate is then expressed in terms similar to the following:
d
   A n e k /T  B me /T , (45)
dt
where  is strain, A, B, k, l, m, n are determined by experiment, and T is temperature.
7.4 Heat-Transport System

The behaviour of the heat-transport system is predicted by solving the equations of mass,
energy and momentum conservation for non-equilibrium transient two-phase flow in a network
in one dimension. “Two-phase” means that we consider steam and water. “Non-equilibrium”
means that the steam and water phases, even in the same location, can have different pres-
sures, temperatures, and flow rates. “Transient” means that the desired behaviour is a function
of time. Current CANDU thermo-hydraulic codes are one-dimensional, although sometimes
phenomenological two- or three- dimensional models are used for components such as headers
and channels. “Network” means that we model parallel paths (e.g., fuel channels, ECC) and
several components connected together at the same point (e.g., at the headers).
Codes that model the heat-transport system cover:
 equations of state for the various phases
 component models for steam generators, fuel channels, fuel, headers, secondary
side, valves, pumps, etc.
 correlations for pressure drop, heat transfer (including critical heat flux (CHF)), flow
regimes
 constitutive relationships: these are equations describing transfers of mass, momen-
tum and energy between the phases. The sets of constitutive relationships depend
on the flow regime.
 efficient numerical solution schemes
 plant controllers.
Most thermo-hydraulic simulation codes for reactors break the circuit up into nodes containing
mass, and therefore energy, and links joining the nodes, which are characterized by flow, length,
roughness, diameter, etc. Mass and energy conservation equations are written for the nodes,
and the momentum conservation equation is written for the links. We shall show simplified
examples here (one-dimensional flow in a level pipe).
Figure 66 shows two nodes, i and j, connected by a link k.
Figure 66 Node/link structure

The mass conservation equation for node i is:
dM i
 Wk , (46)
dt k
where the Wk are all the mass flows into and out of node i along links k.
The conservation of momentum equation is applied to link k and is (ignoring gravity):
dWk Ak   f k Lk  Wk 2 
  i
( P  P )    k k  2
, (47)
 2 g c  Ak 
j
dt Lk   Dk
where for the link:
W is the mass flow in link k (in Figure 66, between nodes i and j)
A is the flow area
P is the pressure in the nodes connected to the link
L is the length
D is the hydraulic diameter
f is the friction factor
(fL/D + k) is a pressure-loss coefficient which accommodates changes in flow area, en-
trance effects, etc.
The first term is simply Newton’s law applied to a fluid to which a pressure difference is applied;
the second term is the loss due to friction plus other effects that cause pressure loss. Terms for
pumps and gravity can be added.
Conservation of energy is applied to each node i as follows:
dU i
 Win ein  Wout eout  Qi , (48)
dt
where
Ui is the internal energy of node i
Qi is the heat generated in node i.
The summations are the rates of energies coming into and out of node i due to mass transfer.
These three equations are applied to each phase. There are four unknowns, e.g., mass, mo-
mentum, temperature, and pressure. The fourth equation needed is the equation of state for
each fluid, typically in the form

  f ( P,T ) .
In safety analysis codes, the equation of state is used in the form of massive detailed tables of
water and steam properties or is fitted by correlations.
7.5 Fuel Channels

When we look at the behaviour of fuel channels in accidents, we emphasize three basic disci-
plines: heat transfer, stress-strain behaviour, and hydrogen chemistry.
7.5.1 Heat transfer

Heat can be transferred from the fluid to the pressure tube by conduction and convection and
from the fuel to the pressure tube by conduction (if the fuel sags into contact with the pressure
tube) (Figure 67).
Figure 67 Heat transfer to pressure tube

At high temperatures, the fuel can transfer heat to the pressure tube by radiation, as character-
ized by the Stefan-Boltzmann law,
E   T f4  TPT4  , (49)
where
ε is emissivity
T is temperature in degrees Kelvin
σ is the Stefan-Boltzmann constant (1.36 × 10-4 kilo-calories per metre2-second-K)
f refers to fuel and PT to pressure tube.
A similar equation would apply to radiation from the pressure tube and from the steam itself,
although these quantities tend to be small.
In practice, radiation becomes important at sheath temperatures of approximately 800°C or
more and is approximately equal to the decay power in the fuel around 1200°C–1400°C. In
practice, the hard part is working out the geometry: computer codes break the complex fuel
bundle and pressure tube into smaller pieces, calculate the “view factor” (how much of the
pressure tube or neighbouring fuel element each piece can “see”) for each piece, and calculate
the radiation heat transfer piece by piece.
Heat can also be transferred from the calandria tube to the moderator, which is significant after
pressure-tube contact (see below). The heat-transfer characteristics follow a pool boiling curve.
Limited patches of film boiling can be tolerated for short periods after contact; lengthy or
extensive dryout will lead to calandria-tube (and pressure-tube) failure.
7.5.2 Strain
The pressure tube can be heated by conductive, convective, and radiative heat transfer from the
coolant and the hot fuel. If the pressure tube heats up beyond 800°C or so, it will start to
deform plastically. If the channel pressure is high (> 6MPa), the strain highly localized, or both,
the pressure tube may burst; if not, it will either strain radially or sag under gravity to contact
the calandria tube and transfer heat to the moderator water. The equations for pressure-tube
strain are, not surprisingly, similar to those for sheath strain.
7.5.3 Hydrogen
At high temperatures, Zircaloy oxidizes in steam to produce heat and hydrogen:
Zr  2 H 2O  ZrO2  2 H 2  heat . (50)
This is a quadruple threat:
 hydrogen collects in containment (significant amounts of hydrogen can be produced
quickly only by a LOCA with ECC failure or in a severe core-damage accident) and
can, under the right circumstances, become flammable or detonate; hydrogen can
also be produced slowly over the long term by radiolysis (radiolytic decomposition)
of ECC water as it circulates through the core;
 the heat generated by the chemical reaction increases fuel and pressure-tube tem-
peratures;
 the presence of a non-condensable gas in large quantities can impede ECC water if
the operator tries to recover from an impairment of ECC by injecting water late10;
and
 the formation of zirconium dioxide may embrittle the sheaths so that they may
fragment and block cooling flow if ECC is eventually restored.
Correlations of the reaction rate of steam and Zircaloy therefore form an essential part of fuel-
channel codes.
The reaction rate becomes autocatalytic around 1400°C–1500°C, that is, the heat it generates
10
This was the fear behind the “gas bubble” within the reactor vessel during the Three Mile Island accident. The
concern was not burning within the vessel because no oxygen was present. The hydrogen which escaped to
containment, which did contain oxygen, formed a flammable mixture of 9% hydrogen and did indeed burn without
serious consequences [Jaffe, 1979].
keeps the chemical reaction going with no further heat input. In light-water reactors, this is a
major concern: the fuel elements are close together, so there is no place to which the heat can
radiate. Regulatory practice in the United States sets a strict limit on sheath temperature in a
design basis accident—namely, 1200oC—which was chosen to allow little possibility of an
autocatalytic reaction. In CANDU, the presence of the colder pressure tube less than a few
centimetres from each fuel element moderates the reaction. Embrittlement is still a concern,
but regulatory practice permits calculation of actual oxidation rates and thicknesses, rather than
setting a criterion based on temperature alone; the particular criterion used in Canada is: “the
oxygen concentration in the cladding must remain below 0.7 wt.% over at least half of the
cladding thickness” (see [Grandjean, 2008] for a broad review).
7.6 Moderator
The modelling requirement for the moderator is to predict the transient local water tempera-
ture at each point. The objective is to show that local moderator sub-cooling at any location
where the pressure tube contacts the calandria tube in an accident is sufficient to prevent
prolonged film boiling on the outside of the calandria tube. The physical problem is therefore
solution of three-dimensional fluid flow with heat addition in a porous medium; the medium is
not continuous because of the presence of fuel channels, reactivity devices, etc. The mass,
momentum, and energy equations described above are therefore generalized to three dimen-
sions. Experimental validation is, of course, a must because of the complex geometry.
7.7 Containment
As far as fundamental equations are concerned, containment behaviour is governed by the
same physical phenomena as the heat-transport system, with a few complications:
 the containment volume is compartmentalized, and the flow within the larger com-
partments is three-dimensional;
 a number of different fluids coexist: air (the normal contents of containment), steam
and water from a pipe break, and hydrogen if the sheaths are heavily oxidized;
 heat is added by steam and hot water and also by radioactive decay of any fission
products carried into the containment atmosphere by the discharging fluid; heat is
removed by water sprays, condensation on containment and equipment surfaces,
containment air coolers, and indirectly by ECC in recovery mode;
 pressure can also be influenced by use of the vacuum building (in multi-unit plants),
leakage from containment through cracks, and deliberate venting through filters.
Containment codes usually have sub-models for each of these phenomena.
Many containment codes also track the movement of fission products along with other fluids.
Fission products can exist as:
 noble gases, which interact very little with water or surfaces;
 tritium oxide (specifically the mixed oxide DTO) from the coolant or moderator.
 iodine, cesium, strontium, etc. which interact strongly with water (dissolve and ion-
ize) and tend to plate out on surfaces;
 actinides such as plutonium. These are released from the fuel in quantity only if the
core is massively destroyed.
Generally, iodine-131 is the significant radioisotope of concern because of its short half-life (8.1
days), high-energy gamma emission, and ability to get into the food chain as described in
Section 7.8.3.
As long as the pH of water inside containment is high, iodine will stay dissolved in the water.
High pH can be engineered through storage of chemicals such as tri-sodium phosphate in areas
likely to flood in an accident. However, a fraction of the iodine will react with organic material
in containment and form methyl iodide, which is volatile, not very soluble, and hard to capture
on filters. In an accident, any iodine-131 which does leak from containment is therefore likely to
be in this chemical form.
7.8 Fission Products, Atmospheric Dispersion, and Dose

7.8.1 Fission-product source term
The fission-product inventory in an operating reactor can be estimated as follows. Suppose that
the reactor has been operating at a power of P MW(th). If the recoverable energy per fission is
assumed to be 200 MeV, the total number of fissions occurring per second is
106 joule fissions MeV
Fission rate = P(MW)    . (51)
MW  sec 200MeV 1.60 1013 joule
If the cumulative yield of the ith fission product (the yield of the fission product itself plus the
yields of all its short-lived precursors) is γi atoms per fission, then the rate of production of this
nuclide is
Rate of production = 3.13  1016 P i atoms/sec . (52)
The activity at time t of that fission product while in the core of the reactor is
 i  3.13  1016 P i (1  e  i t ) disintegrations/sec , (53)
or in Curies,
1Ci
 i  3.13 1016 P i (1  e i t )Bq  . (54)
3.7 1010 Bq
If the activity is saturated, that is, if λi >>1, Equation (54) reduces to
 i  8.46  105 P Ci . (55)
Table 14 gives the inventories of the most important noble gases and iodine fission products
computed for a typical 1000 MWe (PWR) plant at the end of a fuel cycle. This example is
illustrative: the time-average inventory would be different in a 1000 MWe CANDU partly be-
cause of on-power fuelling and lower burnup.
7.8.2 Atmospheric dispersion

Consider a reactor containment after an accident in which the concentration of a particular
nuclide is C Bq/m3. Assume a leak rate of V m3/ s into the atmosphere at a height of h metres,
as shown in Figure 68. The release rate Q is given by:
Q( Bq / Sec)  C ( Bq / m3 ) V (m3 / sec) . (56)
This release is dispersed into the surrounding area by the release plume. For given weather
conditions with wind velocity u and other data, the concentration of radioactive material at
some distance and direction from the source can be calculated from the Gaussian dispersion
model [CSA, 2008]:
1/2
 2  Q f  h2 /2 z 2
   e , (57)
    zu  x
where (Figure 68):
χ is the sector-averaged long-term concentration in Bq/m3 at a distance of x metres from
the source and will be uniform through the sector
Q is the release rate in Bq/s from a source of height h metres
σz is the vertical diffusion coefficient in metres
θ is the angle subtended by the sector in radians
f is the fraction of the time that the wind blows into the sector
u is the mean wind velocity in m/s.
Table 14 Typical core inventory of volatile fission products
Nuclide* Half-life† Fission yield‡ Curies (×10-8§)

85m
Kr 4.4 h 0.0133 0.24
85
Kr 10.76 y 0.00285 0.0056
87
Kr 76 m 0.0237 0.47
88
Kr 2.79 h 0.0364 0.68
133
Xe 5.27 d 0.0677 1.7
135
Xe 9.2 h 0.0672 0.34
131
I 8.04 d 0.0277 0.85
132
I 2.28 h 0.0413 1.2
133
I 20.8 h 0.0676 1.7
134
I 52.3 m 0.0718 1.9
135
I 6.7 h 0.0639 1.5
* Superscript m refers to a nuclide in an isomeric state (see Section 2.8).
†
Figure 68 Atmospheric dispersion
7.8.3 Dose
Release of radioactive material from containment leads to an external dose to humans (often
called “cloud” dose) which depends on the ambient radiation level and an internal dose which
depends on inhaled species (and, in the longer term, ingestion of contaminated foods). External
dose is a function of exposure time to an ambient radiation level. Internal dose is a function of
radiation uptake and residence time in the body.
For example, the dose from I131 can be delivered through the following route:
 released from containment
 deposits on grass
 eaten by cows
 excreted in milk
 drunk by people.
Alternatively, I131 can deposit on plants which are eaten directly by people. The external (direct)
dose from atmospheric release of I131 can also be significant.
Detailed tables have been prepared for each radioisotope which make it possible to convert
from external concentration, inhalation, and ingestion to dose. See [ICRP, 2010] for the most
recent tables. However, these are given by type of radiation rather than by radionuclide; for an
older report that does give the conversion factors by radionuclide, see [EPA, 1993].
Sample Problem
Consider a major release of radioactive Ar41 causing an air concentration χ of 108 Bq/m3. A
person is immersed in this cloud for an hour. What is his external dose?
Answer
From [EPA, 1993], Table III.1, we see that the effective dose coefficient hD for air immersion in
Ar41 is 6.5×10-14 Sv/Bq-s-m3. Hence, the dose D for an immersion of t seconds is:
D  hD  t  6.5  10 14 Sv/Bq-s-m 3  108 Bq/m 3  3600sec  0.23Sv . (58)
Note that this assumes a continuous release because we have not accounted for decay; the half-
life of Ar41 is 1.83 hours.
7.9 Problems
1. Calculate the dose to a person due to the release of 1000 Ci/hour of xenon-133 from a
CANDU. Assume that the release point is 20 m high and the receptor is 1 km distant. Assume
further that the person stays in that location for 15 minutes. Consult CSA-N288.1 for any
models you need. You should be able to find dispersion factors in the (public) annual environ-
mental monitoring reports issued, e.g., by OPG.
2. How many grams does 1000 Ci of iodine-131 represent? (Hint: remember or look up the
half-life).
3. Calculate the average volumetric heat generation of the fuel in a 600MW CANDU at
100% power. At what percentage power (assuming the same heat removal from the fuel as in
normal operation) would the centre of the average pin melt?
4. Do the following:
(a) Calculate the amount of hydrogen produced by oxidation of 25% of the Zircaloy in the
sheaths in a CANDU (this is not atypical of a severe accident such as a LOCA + LOECC);
(b) Calculate the amount of energy released (you will need to look up the heat of reaction)
(c) Assume that this energy is released starting from 30 minutes after the accident and ending
two hours afterwards. Compare the energy to the decay heat produced in the same time.
(d) Now assume that the hydrogen is transported into containment and burns. Calculate the
energy produced by the burn.
8 Safety of Operation
In this section, we shall touch briefly on some high-level aspects of safety in operation. No
matter how well a plant is designed, in the end, its safety strongly depends on how, and by
whom, it is operated. We shall then touch briefly on future trends in safety, notably use of
passive systems.
8.1 Safety Culture

The International Atomic Energy Agency (IAEA) was set up in 1957 as the world's “Atoms for
Peace” organization within the United Nations family. It had a dual mandate: if nations would
eschew the path of nuclear weapons development, those countries which already had nuclear
weapons would assist them in developing a civilian nuclear power programme. The objectives
were thus: safeguards, aimed at preventing the proliferation of nuclear weapons, and promo-
tion (including safety), aimed at assisting non-nuclear-weapons states.
The promotional side of the IAEA has been increasingly focussed on safety, especially since the
Chernobyl accident. Early on, the IAEA developed a series of Safety Guides, which enunciated
good safety practices in all areas of the nuclear fuel cycle, from design to waste management.
These Guides were prepared in a collaborative and consensual manner by IAEA members.
Because of this, they tended to contain useful advice, but they were not specific enough to
affect design and operation in a fundamental way. They were generally adopted by both
purchaser and vendor nations. Because safety remains the responsibility of each country, the
Guides have no legal force internationally, but tend to be incorporated informally or adopted
formally as part of each country’s safety regulations.
Chernobyl caused a fundamental rethinking of the effectiveness of the guidance that the IAEA
was offering. The first action taken was by an international group of independent experts, who
provided advice to the Director-General of the IAEA—the International Safety Advisory Group,
or INSAG. They produced a number of key documents which gained widespread international
acceptance and have strongly influenced the development of safety since then.
INSAG-1 [INSAG, 1986] initiated safety culture as a meme; however, this report was withdrawn
because of deficiencies in the Soviet account of Chernobyl, on which it relied. [INSAG, 1992]
entitled Basic Safety Principles for Nuclear Power Plants, tried to set down in one place what its
title implied: what were the underlying and fundamental safety principles of nuclear power
plants. Five levels of safety principles were defined, in a hierarchy going from the general and
all-encompassing to specific technical practices. The document was written in the present
tense, as if all reactors followed the safety principles—a clear message that if they did not, they
should be modifying at least their operating practices. Details are beyond the scope of this
Section.
INSAG recognized implicitly that no plant design is so safe that it cannot be damaged by incom-
petent operation; this is why plant safety is fundamentally the responsibility of the plant opera-
tor, not the designer and not the regulator. It is even more important that an unforgiving design
requires cautious operation, which in case of uncertainty always opts for the prudent course of
action. In this framework, Chernobyl was an unforgiving design run by an organization deficient
in safety culture.
The difficulty with safety culture was that it was hard to “get hold of”—like personal character,
one could sense when it was deficient, but it was hard to measure. INSAG therefore hastened
to elaborate on the term in a subsequent report [INSAG, 1991] entitled Safety Culture. The
term was redefined as:
“Safety culture is that assembly of characteristics and attitudes in organizations and in-
dividuals which establishes that, as an overriding priority, nuclear plant safety issues re-
ceive the attention warranted by their significance”.
The three concepts in this definition are that safety culture is attitudinal as well as structural;
that it relates both to organizations and individuals; and that it matches all safety requirements
with appropriate perceptions and action. Note that it gives safety a priority to the extent
warranted—it is not an absolute.
In simpler terms, the best design and the most carefully written procedures will not help if staff
does not place safety first in their thoughts and actions. See [IAEA, 2002] and [INSAG, 1999] for
more detail.
Other organizations have used variant definitions. The U.S. Nuclear Regulatory Commission
[USNRC, 2011a] defines it as “the core values and behaviors resulting from a collective com-
mitment by leaders and individuals to emphasize safety over competing goals to ensure protec-
tion of people and the environment.”
Informal but revealing definitions include “Safety culture is what you do when the boss isn’t
looking”, and “Safety culture is the way we do things around here”.
To assist in recognizing a lack of safety culture, the IAEA has defined the stages of organizational
decline, as shown in Table 15.
Table 15 Stages of organizational decline

Stage Name of stage Characteristic of stage
1 Over-confidence Good past performance leading to self-satisfaction
2 Complacency Occurrence of minor events that are subjected to minimum
self-assessment, and delay in improvement programmes
3 Denial Number of minor events increases, with possibly a more
significant event. These are treated as isolated events.
Findings from audits are considered invalid. Root-cause
analysis not used.
4 Danger Several potentially serious events occur, but management
and employees reject criticism from audits or regulators by
considering their views biased. The oversight function is
afraid to confront management.
5 Collapse Regulator intervenes to implement special evaluations.
Management is overwhelmed and may need to be re-
placed. Major and very costly improvements need to be
implemented.
Audits by organizations such as INPO (Institute of Nuclear Power Plant Operations, an associa-
tion of operating organizations) have distilled—through auditing utilities which were subse-
quently forced into extended plant outages due to management deficiencies—a list of symp-
toms of a poor safety culture [INPO, 2010]. It is the rare organization which does not recognize
at least part of itself in the list.
Finally, IAEA Safety Guides are being rewritten to make them more detailed, so that they state
the increased safety expectations of the 21st Century. “Requirements” on both design [IAEA,
2012] and operations [IAEA, 2011a] are becoming the de facto minimum standard for nuclear
power plants. Indeed, the IAEA design “requirements” report is the basis of the current Cana-
dian top-level design requirements for new builds [CNSC, 2008].
8.2 International Nuclear and Radiological Event Scale

No discussion of safety would be complete without a summary of the International Nuclear and
Radiological Event Scale (INES). This is a ranking of accidents so that the safety significance of
events in nuclear power plants is reported in a consistent way all over the world [INES, 2008].
The seven levels in the scale are shown in Table 16. The table is self-explanatory, but should be
read carefully.
Table 16 INES event scale

Description People and the environ- Radiological barriers and Defence in depth
and INES ment controls at facilities
Level
Major Acci- Major release of radioac-
dent Level 7 tive material with wide-
spread health and envi-
ronmental effects
requiring implementation
of planned and extended
countermeasures.
Serious Significant release of
Accident Level radioactive material likely
6 to require implementa-
tion of planned counter-
measures.
Accident with Limited release of radio- Severe reactor core
wider conse- active material likely to damage.
quences require implementation Release of large quanti-
Level 5 of some planned coun- ties of radioactive mate-
termeasures. rial within an installation
Several deaths from with high probability of
radiation. significant public expo-
sure. This could arise
from a major criticality
accident or fire.
Accident with Minor release of radioac- Fuel melt or damage to
local conse- tive material unlikely to fuel resulting in more
quences Level result in implementation than 0.1% release of core
4 of planned countermea- inventory.
sures other than local Release of significant
food controls. At least quantities of radioactive
one death from radiation. material within an
installation with high
probability of significant
public exposure.
Description People and the environ- Radiological barriers and Defence in depth
and INES ment controls at facilities
Level
Serious Exposure in excess of ten Exposure rates of more Near accident at a
Incident Level times the statutory than 1 SV/h in an operat- nuclear power plant with
3 annual limit for workers. ing area. no safety provisions
Non-lethal deterministic Severe contamination in remaining.
health effect (e.g., burns) an area not expected by Lost or stolen highly
from radiation. design, with a low prob- radioactive sealed
ability of significant public source.
exposure. Misdelivered highly
radioactive sealed source
without adequate
radiation procedures to
handle.
Incident Exposure of a member of Radiation levels in an Significant failures in
Level 2 the public in excess of 10 operating area of more safety provisions, but
mSv. than 50 mSv/h. with no actual conse-
Exposure of a worker in Significant contamination quences.
excess of the statutory within the facility into an Found highly radioactive
annual limits. area not expected by sealed orphan source,
design. device or transport
package with safety
provisions intact.
Inadequate packaging of
a highly radioactive
sealed source.
Anomaly Overexposure of a
Level 1 member of the public
above statutory limits.
Minor problems with
safety components with
significant defence in
depth remaining.
Low-activity lost or stolen
radioactive source,
device or transport
package.
No safety significance (Below scale/Level 0)
8.3 Safety Aspects of Future Designs

In the final part of this Section, we shall be discussing some of the safety characteristics of
modern advanced designs. These designs fall into two categories: evolutionary and advanced.
1. Evolutionary designs, as their name suggests, build on operating experience in current

plants and add those improvements which are warranted by experience.
2. Advanced designs use passive concepts in their operation and particularly with respect
to safety. The original ideas behind the use of passive safety was to:
 simplify the design and make it cheaper to build, operate, and maintain
 increase the real safety of the plant through systems which were less complex and
more reliable because they used “natural” forces
 increase the perceived safety of the plant for the same reasons.
In its broadest sense, passive safety emphasizes the use of “natural” forces (gravity, self-
correcting neutronic feedback) and de-emphasizes systems which require large amounts of
electricity (pumps), rapid automatic response, complex logic, or high energy. A clear under-
standing of the meaning of terms such as “passive” and “inherent” is essential; we shall follow
[IAEA, 1991].
8.3.1 Definitions
Inherent safety refers to the achievement of safety through the elimination or exclusion of
inherent hazards through the fundamental conceptual design choices made for the nuclear
plant. This is impossible for practical reactor sizes because it requires elimination of systems
(because they would not be needed) to remove or compensate for decay heat, excess reactivity,
and high energy releases. It is possible to have inherent safety for low-energy pool reactors
such as the 20kW(th) SLOWPOKE-2 research reactor [AECL, 1976]—the total potential reactivity
addition can all be compensated for without fuel damage by the inherent negative feedback
from fuel and coolant temperatures, and the power at which this compensation is achieved can
be absorbed indefinitely by the pool and the surroundings. For larger reactors, elimination of
one or more of these hazards does, however, give a reactor an inherently safe characteristic.
Note that the hazard must be eliminated deterministically, not probabilistically; for example, a
plant is inherently safe against fires if it has no combustible material.
An ideal passive component does not need an external input to operate; a passive system is
composed of passive components and structures. Ideally a passive system has no reliance on
external mechanical or electrical power, signals, or forces. It does rely on “natural” laws,
properties of materials, and internally stored energy. Therefore, heat removal from a reactor by
thermo-siphoning to an elevated tank of water is passive, at least until the water runs out. In
practice, most passive designs do allow active signals because there is usually a need to switch
from the active heat-removal systems used in full-power operation to passive decay-heat
removal systems after an accident. CANDU shutdown systems are passive in this respect: once
they receive a signal, they actuate by gravity or stored energy.
An active component or system is one which is not passive.
Fail-safe means that a given failure leads to a safe condition; the component or system is then
fail-safe with respect to that condition. The fail-safe characteristic is specific to the failure
mode: for example, Shutdown System 2 in CANDU is fail-safe with respect to loss of electrical
power to the valves, but not to loss of gas pressure.
Grace period is the period of time during which a safety function is ensured after an accident
without the necessity for human intervention. Colloquially, it used to be called “walk-away
safe” for whatever grace period was involved—this term has unfortunate connotations (opera-
tors are not expected to walk away from an accident) and is best not used. A grace period can
be achieved through active or passive means; usually, the first line of defence is assumed to
function properly in determining the grace period. Therefore, the grace period for loss of feed
water in first-generation CANDUs is about 30 minutes because after this period, the operator
must manually valve in an alternate heat sink. For new CANDU designs, the grace period is
extended to days through use of automatic steam-generator depressurization followed by
automatic connection of the elevated reserve water tank to the steam generators. For modern
reactor designs, evolutionary or passive, a grace period of three days for most single failures is
expected.
8.3.2 Categories of Passive Safety

Very few systems are totally passively safe. To recognize the range of possibilities, the IAEA
[IAEA, 1991] defined four categories of passivity, as summarized in Table 17.
Table 17 Categories of passive safety characteristics
Characteristic Category A Category B Category C Category D
Signal Inputs of Intelli- No No No Yes
gence
External power sources No No No No
or forces
Moving mechanical No No Yes Either
parts
Moving working fluid No Yes Yes Either
Example Barriers such Heat removal Rupture disk Shutdown
as fuel clad, by natural or spring- Systems 1 and
containment; circulation to loaded valve 2 in CANDU
core cooling heat exchang- for overpres-
relying only on ers in water sure protec-
radiation or pools, from tion; accumu-
conduction to the core or lator isolated
outer struc- containment by check valve
tural parts
For many passive designs, even those for which the “execution” is passive, the actuation may
involve an electrical signal. Part of the justification for this is that such signals are highly reliable
and can use backup power from batteries if main power fails. Note that even some valves can
be actuated on battery power. However, details of how this is implemented are being revisited
after Fukushima.
8.3.3 Passive safety desiderata

A passive design strives to ensure that the three major safety functions (other than monitoring)
can be carried out in a passive or pseudo-passive manner. Recall that these functions are to
shut down the reactor, to remove decay heat, and to contain any fission products. We describe
in general terms how each of these might be accomplished.
8.3.3.1 Shut down the reactor

CANDU shutdown systems are passive in the sense that once they are actuated by a signal, the
devices themselves are inserted into the core by gravity (shut-off rods) or stored energy (spring
assist to the shut-off rods and gas-driven poison injection). This places them into IAEA Category
D above. More passive approaches could be developed based on change in material properties
with temperature, e.g., a shutdown system consisting of tubes inserted into the reactor, with a
low-melting-point neutron absorber within them, initially located above the core, and which
would melt and insert into the core on increasing temperature. This system needs no external
“intelligence”, but does have a moving fluid, placing it in Category B. Even more basic, fuel with
a strong negative temperature-feedback coefficient is certainly a passive form of reactivity
compensation. If it truly shut down the reactor, it would be in Category A. However, negative
feedback does not shut down the reactor after, say, an inadvertent insertion of positive reactiv-
ity (control rod withdrawal); it simply allows the power to rise and then equilibrate at a level
where the negative reactivity due to the higher fuel temperature offsets the reactivity addition
of the control rod. One still needs to be sure that the power can be removed somehow (by
passive means) and that the fuel is not damaged. A strong negative coolant-temperature
feedback works the same way, but raises the added concern that fast insertion of cold water
could cause a rapid power increase before the negative fuel or coolant temperature had time to
compensate for it. The SLOWPOKE 2 reactor dealt with this problem by physically limiting the
amount of reactivity that it was possible to add; an alternative is to limit physically the rate at
which it can be added, so that negative feedback has time to take effect. Removal of moderator
in LWRs after a LOCA is an example of passive shutdown, but the ECC water that replaces it
must be borated to prevent recriticality and therefore is an active means of shutdown.
8.3.3.2 Remove decay heat

In passive designs, removal of decay heat from the fuel is normally done by thermo-siphoning to
an elevated heat sink, usually a heat exchanger in a large supply of water high up in the build-
ing. Alternatively, the entire core and its surroundings can be flooded by pouring water by
gravity from an elevated supply; the core heat is then turned to steam, which flows to and is
removed passively from containment. In some, but not all, passive designs decay heat is
removed at low pressure; therefore, some means of depressurizing the heat-transport system is
required first. This is done using a Category “D” device. CANDU offers inherent characteristics
for passive heat removal in severe accidents, as discussed in Section 6.14. Existing and future
CANDUs therefore incorporate the features desired in next-generation reactors: spreading of
core debris so it can be easily cooled (the calandria shell and/or the reactor vault/shield tank)
and passive cooling of the damaged core (moderator and/or reactor vault water).
8.3.3.3 Contain radioactive material

The containment structure is already passive, category A. However, for a number of other
containment functions, a passive approach can be taken.
8.3.3.3.1 Ventilation isolation

The building can simply be sealed during operation, or the isolation system can be Category D
(e.g., a spring-loaded valve which fails closed on loss of signal).
8.3.3.3.2 Decay heat removal

This is an extension of core thermo-siphoning. The idea is to move the heat into an elevated
tank of water. To achieve this, heat exchangers can be placed in the building, with the tube side
connected to the tank and the shell side exposed to containment atmosphere. This requires
two natural convection loops: one in the containment atmosphere, transporting heat from the
core (e.g., steaming from a LOCA) to the heat exchangers, and one transporting heated water
from the heat exchangers to the elevated tank. One variation on this idea is used in the
AP1000: the metal shell of the building becomes the heat exchanger, and the elevated tank
trickles water down the outside [Westinghouse, 2011b].
8.3.3.3.3 Hydrogen removal

Here, passive autocatalytic recombiners can be used. These simply offer a catalyst-coated
surface to the containment atmosphere and as the air/steam/hydrogen mixture flows through,
the hydrogen and oxygen are catalytically recombined. An alternative is to inert the contain-
ment atmosphere, inherently removing the possibility of hydrogen combustion since there will
be no oxygen.
8.3.4 Summary
This Section has indicated the direction that safety may take in the future. Passive safety is
attractive because of its simplicity, public appeal, and aura of high reliability. Evolutionary
designs have also incorporated safety enhancements while both remaining economical and
posing less of an “innovation” risk to owners and operators. Once adequate (or even more than
adequate) safety is achieved, factors such as economics and proven performance may become
the determinants of the choice of technology, particularly as electricity markets become more
deregulated.
8.4 Problems
1. Place the following accidents on the International Nuclear and Radiological Event Scale,
with brief reasons for your choice. You may need to research some of these if they are
not covered in the text:
1. NRX accident, 1952
2. SL-1 accident
3. Pressure-tube failure in Pickering A (G-16)
4. Fire in Narora plant in India
5. Chernobyl accident (power runaway)
6. Three Mile Island accident (core melt)
7. Fukushima tsunami-induced accident
8. Erosion/corrosion of Davis-Besse vessel head
2. If you work for a design organization or a regulator or a nuclear power plant, evaluate
its safety culture in terms of the safety culture framework defined by one or more of
INSAG, IAEA, INPO, etc. Give reasons and evidence, not just opinion.
3. Look up information on any two modern designs (e.g., Enhanced CANDU 6, Westing-
house AP1000, EPR). Compare and contrast them as follows: for each of the systems
performing the fundamental safety functions (shut down, cool, contain radioactive ma-
terial), categorize them as active or as passive categories A, B, C, or D (give reasons).
9 Review
This Chapter has summarized the basic concepts of safety for nuclear power plants. Starting
with the hazards inherent to nuclear power, it has described tools to identify possible accidents,
in particular the top-down and bottom-up approaches. Real experience has been a powerful
driver in the approach to nuclear safety, and the Chapter has described seminal historical
accidents and lessons learned—and as Fukushima has shown, learning lessons is an ongoing
process. We presented risk assessment, first in terms of safety goals for acceptable safety
performance, and then developing the probabilistic tools used to show that the goals have been
achieved. We then described (for CANDUs) mitigating systems to shut down the plant, remove
decay heat, and contain fission products in response to the needs identified by both determinis-
tic and probabilistic analyzes. We looked at accident phenomenology, using a large LOCA as a
model, and extended this discussion to CANDU characteristics in severe core-damage accidents.
Then we described (at a high level) the mathematical models underlying the safety-analysis
codes used to predict plant behaviour in accidents. Finally we looked—all too briefly—at the
safety role of the plant operators and indicated the options for the safety characteristics of
future designs.
10 References
[Ader, 1998] C. Ader, G. Heusener, and V. G. Snell, “Strategies for the Prevention and Mitigation
of Severe Accidents”, IAEA International Symposium on Evolutionary Water-Cooled Reac-
tors: Strategic Issues, Technologies, and Economic Viability, IAEA Paper No. IAEA-SM-
353-16, Korea, 1998.
[AECB, 1980] Atomic Energy Control Board, Requirements for the Safety Analysis of CANDU
Nuclear Power Plants, Consultative Document C-6, Proposed Regulatory Guide, June
1980.
[AECL, 1976] R. E. Kay, J. W. Hilborn, and N. B. Poulsen, The Self-Limiting Power Excursion
Behaviour of the Slowpoke Reactor, Atomic Energy of Canada Limited Report AECL-4770,
January 1976.
[AECL, 2005] Atomic Energy of Canada Limited, CANDU 6 Technical Summary, May 2005 (Primer
on CANDU 6 design).
[AECL, 2011], ZED-2 User Facility Proposal Package, AECL report ZED2-123110-REPT-001,
Revision 0, May 2011.
[Andreani, 2010] M. Andreani and D. Paladino, “Simulation of Gas Mixing and Transport in a
Multi-Compartment Geometry using the GOTHIC Containment Code and Relatively
Coarse Meshes”, Nuclear Engineering and Design, Vol. 240, No. 6, pp. 1506–1527 (2010).
[Blahnik, 1991] C. Blahnik, et al., “Modular Accident Analysis Program for CANDU Reactors”,
Proc. 12th Annual Canadian Nuclear Society Conference, Saskatoon, Saskatchewan, Can-
ada, June 9-12, 1991, pp. 235-242.
[Blahnik, 2012] C. Blahnik, M. J. Brown, T. Nitheanandan, and S. M. Petoukhov, “Perspective on
Analyses of Core Damage in CANDU Reactors”, 24th Nuclear Simulation Symposium, Ot-
tawa, Ontario, Canada, October 14-16, 2012.
[Boss, 2006] K. Aydogdu and C. R. Boss, “Radiation Physics and Shielding Codes and Analyses
Applied to Design-Assist and Safety Analyses of CANDU and ACR Reactors”, PHYSOR-
2006, ANS Topical Meeting on Reactor Physics, Vancouver, BC, Canada, September 10-
14, 2006.
[Brooks, 1980] G. L. Brooks and E. Siddall, “An Analysis of the Three Mile Island Accident”,
presented at the CNS First Annual Conference, Montreal, Canada, June 1980.
[Buell, 2003] J. Buell, J. D. Dormuth, P. Ingham, and R. Swartz, RD-14M Facility Description, 153-
112020-UM-001, 2003; also USNRC ML031690499.
[Cameron, 1996] I. Cameron, Nuclear Physics and Reactor Theory 1.1—Module 9, Source Neu-
tron Effects, Course given at Chulalongkorn University, Thailand, 1996. See CANTEACH:
https://canteach.candu.org.
[Chan, 1987] P. Chan, et al., “The Chernobyl Accident: Multidimensional Simulations to Identify
the Role of Design and Operating Features of the RBMK-1000”, Conference on Probabilis-
tic Safety Assessment and Risk Management, Zurich, Switzerland; September 1987. Also
Atomic Energy of Canada report AECL-9246.
[Charak, 1995] I. Charak and P. H. Kier, CANDU Reactors, Their Regulation in Canada, and the
Identification of Relevant NRC Safety Issues, Argonne National Laboratory report
NUREG/CR-63 15 and ANL-9515, April 1995 (prepared for U.S. Nuclear Regulatory Com-
mission).
[Chen, 2007] W. L. Chen, et al., “Effects of Cobalt-60 Exposure on Health of Taiwan Residents
Suggest New Approach Needed in Radiation Protection”, Dose Response, Vol. 5, No. 1,
pp. 63–75 (2007).
[CNSC, 2008] Canadian Nuclear Safety Commission, Design of New Nuclear Power Plants,
Regulatory Document RD–337, November 2008.
[CNSC, 2008a] Canadian Nuclear Safety Commission, Safety Analysis for Nuclear Power Plants,
Regulatory Document RD–310, February 2008
[CNSC, 2011] Canadian Nuclear Safety Commission, CNSC Staff Integrated Safety Assessment of
Canadian Nuclear Power Plants for 2010, INFO-0823, September 2011.

[CSA, 2008] Canadian Standards Association, Guidelines for Calculating Derived Release Limits
for Radioactive Material in Airborne and Liquid Effluents for Normal Operation of Nu-
clear Facilities, CSA N288.1-08, 2011.
[Cross, 1980] W. G. Cross, “The Chalk River Accident in 1952”, Historical Perspective on Reactor
Accidents, Seattle, Washington, July 1980.
[Currie, 1984] T. Currie and J. T. Rogers, “Experimental and Numerical Studies of Heat Transfer
Through Contact Areas of High Ellipticity”, ASME National Heat Transfer Conference, Pa-
per 84-HT-85, 1984.
[Currie, 1985] T. C. Currie, J. T. Rogers, and J. C. Atkinson, A Study of the Technical and Economic
Feasibility of Using a SLOWPOKE-3 Nuclear Reactor for Building Heating and Cooling at
Carleton University, Study for AECL; November, 1985.
[Currie, 1986] T. C. Currie and J. T. Rogers, “Heat Transfer between Rough Surfaces in Contact
over a Highly Elliptical Contour Area: Comparison of Experimental and Numerical Re-
sults”, Proceedings of Eighth International Heat Transfer Conference, San Francisco, Au-
gust 1986.
[Cuttler, 2009] J. M. Cuttler and M. Pollycove, “Nuclear Energy and Health: The Benefits of Low-
Dose Radiation Hormesis”, Dose Response, Vol. 7, No. 1, pp. 52–89, 2009.
[EPA, 1993] K. F. Eckerman and J. C. Ryman, Federal Guidance Report No. 12 - External Exposure
to Radionuclides in Air, Water, and Soil, Environmental Protection Agency report EPA-
402-R-93-081, September 1993.
[Glasstone, 1994] S. Glasstone and A. Sesonske, Nuclear Reactor Engineering, 4th Edition,
Chapman & Hall, 1994.
[Gonzales, 2013] A. J. Gonzalez, et al., “Radiological Protection Issues Arising During and After
the Fukushima Nuclear Reactor Accident”, J. Radiol. Prot., Vol. 33, pp. 497–571, 2013.
[Grandjean, 2008] C. Grandjean and G. Hache, A State-of-the-Art Review of Past Programmes
Devoted to Fuel Behaviour Under Loss-of-Coolant Conditions, Part 3: Cladding Oxidation,
Resistance to Quench and Post-Quench Loads, Institut de Radioprotection et de Sûreté
Nucléaire (IRSN) report DPAM/SEMCA 2008-093, 2008.
[Hanna, 1998] B. N. Hanna, “CATHENA: A Thermal-Hydraulic Code for CANDU Analysis”, Nuclear
Engineering and Design, Vol. 180, 113–131, 1998.
[Howieson, 1987] J. Q. Howieson and V. G. Snell, Chernobyl: A Canadian Technical Perspective,
Atomic Energy of Canada Limited publication AECL-9334, January 1987. Also Nuclear
Journal of Canada, Vol. 1, No. 3, September 1987.
[HSE, 2006] Health and Safety Executive (United Kingdom), Safety Assessment Principles for
Nuclear Facilities, Bootle, 2006 Edition, Revision 1.
[Hurst, 1953] D. G. Hurst, The Accident to the NRX Reactor, Part II, Atomic Energy of Canada
Limited, Report AECL-233, October 1953.
[Hurst, 1972] D. G. Hurst and F. C. Boyd, “Reactor Licensing and Safety Requirements”, Paper 72-
CNA-102, 12th Annual Conference of the Canadian Nuclear Association, Ottawa, Ontario,
June 1972.
[Huterer, 1984] J. Huterer, E. C. Ha, D. G. Brown, and P. C. Cheng, “Darlington GS Vacuum Build-
ing: Containment Shell”, Nuclear Engineering and Design, Vol. 85, 131-140, 1985.
[IAEA, 1991] International Atomic Energy Agency, Safety-Related Terms for Advanced Nuclear
Plants, IAEA-TECDOC-626, September 1991.
[IAEA, 2000] International Atomic Energy Agency, Safety of Nuclear Power Plants: Design, IAEA
report NS-R-1, 2000.
[IAEA, 2002] International Atomic Energy Agency, Safety Culture in Nuclear Installations, IAEA-
TECDOC-1329, December 2002.
[IAEA, 2008] International Atomic Energy Agency, Analysis of Severe Accidents in Pressurized
Heavy-Water Reactors, IAEA-TECDOC-1594, June 2008.
[IAEA, 2010] International Atomic Energy Agency, Development and Application of Level 1
Probabilistic Safety Assessment for Nuclear Power Plants, IAEA-SSG-3, April 2010.
[IAEA, 2011] International Atomic Energy Agency, International Fact Finding Expert Mission of
the Fukushima Dai-Ichi NPP Accident Following the Great East Japan Earthquake and
Tsunami, IAEA Report, June 16, 2011.
[IAEA, 2011a] International Atomic Energy Agency, Safety of Nuclear Power Plants: Commission-
ing and Operation: Specific Safety Requirements, IAEA report SSR-2/2, July 2011.
[IAEA, 2012] International Atomic Energy Agency, Safety of Nuclear Power Plants: Design:
Specific Safety Requirements. IAEA report SSR-2/1, January 2012.
[ICRP, 1990] International Commission on Radiological Protection, 1990 Recommendations of
the ICRP, report ICRP-60, table B.
[ICRP, 2010] International Commission on Radiological Protection, Conversion Coefficients for
Radiological Protection Quantities for External Radiation Exposures, ICRP Publication
116, Ann. ICRP 40(2–5), 2010.
[ICRP, 2012] International Commission on Radiological Protection, Compendium of Dose Coeffi-
cients based on ICRP Publication 60, ICRP Publication 119, Ann. ICRP 41 (Suppl.).
[INES, 2008] International Atomic Energy Agency and OECD/Nuclear Energy Agency, INES - The
International Nuclear and Radiological Event Scale - User’s Manual, 2008 Edition, IAEA,
Vienna, 2009.
[Inhaber, 1978] H. Inhaber, Risk of Energy Production, Atomic Energy Control Board report AECB
1119, March 1978.
[INPO, 2010] Institute of Nuclear Power Operations, Field Guidance: Organizational Effective-
ness Evaluation and Assistance, INPO report, Revision 2, March 2010.
[INSAG, 1986] International Nuclear Safety Advisory Group, Summary Report on the Post-
Accident Review Meeting on the Chernobyl Accident, International Atomic Energy Agency
Report 75-INSAG-1, Vienna, 1986.
[INSAG, 1991] International Nuclear Safety Advisory Group, Safety Culture, International Atomic
Energy Agency Report 75-INSAG-4, 1991.
[INSAG, 1992] International Nuclear Safety Advisory Group, The Chernobyl Accident: Updating
of INSAG-1 – INSAG 7, International Atomic Energy Agency Report INSAG-7, 1992.
[INSAG, 1996] International Nuclear Safety Advisory Group, Defence in Depth in Nuclear Safety,
International Atomic Energy Agency Report INSAG-10, 1996.
[INSAG, 1999] International Nuclear Safety Advisory Group, Basic Safety Principles for Nuclear
Power Plants, International Atomic Energy Agency Report INSAG-12, 1999.
[Ionescu, 2009] S. Ionescu, O. Uta, M. Pârvan, and D. Ohâi, “Pressurized Heavy Water Reactor
Fuel Behaviour in Power Ramp Conditions”, Journal of Nuclear Materials, Vol. 385, 387–
391, 2009.
[Jaffe, 1979] L. Jaffe, Staff Reports to the President's Commission on the Accident at Three Mile
Island, Reports of the Technical Assessment Task Force, Volume II, 1979.
[Japan, 2011] Government of Japan, Nuclear Emergency Response Headquarters, Report of the
Japanese Government to the IAEA Ministerial Conference on Nuclear Safety: The Acci-
dent at TEPCO’s Fukushima Nuclear Power Stations, June 2011.
[JNTI, 2011] Japan Nuclear Technology Institute, Examination Committee on Accident at Fuku-
shima Daiichi Nuclear Power Station, Examination of Accident at Tokyo Electric Power
Co., Inc.’s Fukushima Daiichi Nuclear Power Station and Proposal for Countermeasures,
October 2011 (there are many more similar reports from IAEA and others).
[Kaplan, 1979] S. Kaplan and B. J. Garrick, “On the Use of Bayesian Reasoning in Safety and
Reliability – Three Examples”, Nuclear Technology, Vol. 44, 231, 1979.
[Kemeny, 1979] J. G. Kemeny, Report of the President's Commission on the Accident at TMI,
1979.
[Krause, 2007] M. Krause, “Hydrogen Program at AECL”, 2nd European Review Meeting on Severe
Accident Research (ERMSAR-2007), Forschungszentrum Karlsruhe GmbH (FZK), Germany,
June 12-14, 2007.
[LANL, 2000] Los Alamos National Laboratory, A Review of Criticality Accidents, report LA-13638,
2000. (See page 74.)
[Larson, 1961] E. A. G. Larson, A General Description of the NRX Reactor, Atomic Energy of
Canada Report AECL-1377, 1961.
[Lewis, 1953] W. B. Lewis, The Accident to the NRX Reactor on December 12, 1952, Atomic
Energy of Canada Limited, Report AECL-232, July 1953.
[Lewis, 1990] B. J. Lewis, C. E. L. Hunt, and F. C. Iglesias, “Source Term of Iodine and Noble Gas
Fission Products in the Fuel-to-Sheath Gap of Intact Operating Nuclear Fuel Elements”,
Journal of Nuclear Materials, Vol. 172, 197-205, 1990.

[Lewis, 1999] B. J. Lewis, P. Tume, L. G. I. Bennett, M. Pierre, A. R. Green, T., Cousins, B. E.
Hoffarth, T. A. Jones, and J. R. Brisson, “Cosmic Radiation Exposure on Canadian-Based
Commercial Airline Routes”, Radiat. Prot. Dosim., Vol. 86, No. 1, pp. 7–24, 1999.
[Lewis, 2009] B. J. Lewis, F. C. Iglesias, R. S. Dickson, and A. Williams, “Overview of High-
Temperature Fuel Behaviour with Relevance to CANDU Fuel”, Journal of Nuclear Materi-
als, Vol. 394, pp. 67–86 (2009).
[Liauw, 1997] W. K. Liauw, W. S. Liu, R. Y. Chan, “TUF Validation against Reactor Trip Data at
Darlington NGS”, Canadian Nuclear Society Conference: Powering Canada’s Future, To-
ronto (Canada), June 8-11, 1997.
[Liverman, 1979] D. M. Liverman and J. P. Wilson, “The Mississauga Train Derailment and
Evacuation, 10–16 November 1979”, Canadian Geographer, Vol. 25, No. 4, pp. 365–375,
1981.
[Mathew, 2004] M. Mathew, T. Nitheanandan, and S. J. Bushby, Severe Core Damage Accident
Progression within a CANDU Calandria Vessel. AECL Chalk River Laboratories. Reactor
Safety Division, Forschungsbericht (2004).
[McCormick, 1981] N. J. McCormick, Reliability and Risk Analysis, Academic Press, 1981, ISBN 0-
12-482360-2.
[Meneley, 2003] D. A. Meneley, Nuclear Safety and Reliability, Lecture notes for a course at the
University of New Brunswick, available at the CANTEACH repository,
https://canteach.candu.org/Pages/Welcome.aspx, 2003.
[Meneley, 2009] D. A. Meneley and A. Muzumdar, “Power Reactor Safety Comparison – A
Limited Review”, Proceedings, CNS 30th Annual Conference, Calgary, AB, May 31–June 3,
2009.
[Morison, 1987] W. G. Morison, et al., “Containment Systems Capability”, Nuclear Journal of
Canada, Vol. 1, No. 1, pp. 53-68, 1987 [from CANTEACH].
[Muzumdar, 2009] A. J. Muzumdar and D. A. Meneley, “Large LOCA Margins in CANDU Reactors:
An Overview of the COG Report”, Proceedings, CNS 30th Annual Conference, Calgary, AB,
May 31–June 3, 2009.
[NISA, 2011] Nuclear and Industrial Safety Agency (NISA) and Japan Nuclear Energy Safety
Organization (JNES), The 2011 Pacific Earthquake off the Pacific Coast of Tohoku and the
Seismic Damage to the NPPs, April 4, 2011.
[OMSC, 1981] Ontario Ministry of the Solicitor General, The Mississauga Evacuation – Final
Report, November 1981.
[Ott, 1981] K. O. Ott and J. F. Marchaterre, “Statistical Evaluation of Design-Error Related Nu-
clear Reactor Accidents”, Nuclear Technology, Vol. 52, No. 2, 179-188, 1981.
[Page, 1972] R. D. Page, Nuclear Power Symposium - Lecture No. 5: Canadian Power Reactor
Fuel, Atomic Energy of Canada Limited, 1972 [from CANTEACH].
[Popov, 2012] N. K. Popov and V. G. Snell, “Safety and Licensing Aspects of Power Reactor
Reactivity Coefficients”, Proceedings, 20th International Conference on Nuclear Engineer-
ing (ICONE20), Anaheim, California, USA, July 30–August 3, 2012.
[Popov, 2013] N. Popov, V. Snell, R. Tran, and G. LeRoy, “Thermo-Hydraulic Aspects of Reactivity
Coefficients in CANDU”, 15th International Topical Meeting on Nuclear Reactor Thermal-
Hydraulics (NURETH-15), Pisa, Italy, May 12–17, 2013.
[Rizkalla, 1984] S. H. Rizkalla, S. H. Simmonds, and J. G. MacGregor, “Pre-Stressed Concrete
Containment Model”, Journal of Structural Engineering, Vol. 110, No. 4, pp. 730–743,
1984.
[Rizkalla, 1986] S. Rizkalla, “Limit States Behaviour of Pre-Stressed Concrete Containment
Structures”, Proceedings, Second International Conference on Concrete Technology for
Developing Countries, Tripoli, Libya (Splaj), October 27–30, 1986;. Vol. 2, El-Fateh Univer-
sity, Civil Engineering Department, 1986.
[Rogers, 1987] J. T. Rogers, “Insights from Chernobyl on Severe Accident Assessment of CANDU
Reactors”, Nuclear Journal of Canada, Vol. 1, No. 2, pp. 107–118, 1987.
[Rogers, 1995] J. T. Rogers, D. A. Meneley, C. Blahnik, V. G. Snell, and S. Nijhawan, “Coolability of
Severely Degraded CANDU Cores”, ICHMT International Seminar on Heat and Mass
Transfer in Severe Reactor Accidents, Cesme, Turkey, May 21–26, 1995. Also Atomic En-
ergy of Canada Ltd. Report AECL-11110.
[Rogers, 2002] J. T. Rogers and M. L. Lamari, “Application of the CANDU Molten Core Model
‘DEBRIS.MLT’ to RASPLAV Test Results”, 22nd Nuclear Simulation Symposium, Canadian
Nuclear Society, Ottawa, November, 2002
[Rogers, 2004] J. T. Rogers, Options For Coal-Fired Power Plants in Ontario, monograph at
Canadian Nuclear Society, http://www.cns-snc.ca/cns_media, September 27, 2004.
[Rogers, 2004a] J. T. Rogers, “CANDU Power Reactor Behavior Following Reactor Trip”, Presenta-
tion at Panel Discussion on Islanding, Cogeneration Conference, Ottawa, October 26,
2004.
[Rogovin, 1980] M. Rogovin and G. T. Frampton, Three Mile Island—A Report to the Commis-
sioners and to the Public, Nuclear Regulatory Commission Special Inquiry Group, January
1980.
[Roy, 2004] R. Roy, et al., “Reactor Core Simulations in Canada”, PHYSOR 2004 -The Physics of
Fuel Cycles and Advanced Nuclear Systems: Global Developments, Chicago, Illinois, April
25-29, 2004, American Nuclear Society, Lagrange Park, IL (2004)
[Rzentkowski, 2013] G. Rzentkowski, Y. Akl, and S. Yalaoui, “Application of Probabilistic Safety
Goals to Regulation of Nuclear Power Plants in Canada”, 34th Annual Conference of the
Canadian Nuclear Society, Toronto, June 9–12, 2013.
[Sanderson, 2003] B. Sanderson, Fuel Channel Behaviour, presented to U.S. Nuclear Regulatory
Commission, May 6–7, 2003, USNRC accession ML031340404.
[Santamaura, 1998] P. A. Santamaura, J. G. Tielemans, T. H. Nguyen, H. S. Shapiro, R. E. B.

Henderson, and B. A. duQuesnay, “Severe Core Damage Frequency and Insights from
CANDU 6 Level 1 Probabilistic Safety Assessment”, Pacific Basin Nuclear Conference
(PBNC98), Banff, Alberta, Canada, May 1998.
[Semonov, 1983] B. A. Semonov, “Nuclear Power in the Soviet Union”, IAEA Bulletin, Vol. 25, No.
2, June 1983.
[Siddall, 1981] E. Siddall, Risk, Fear, and Public Safety, Atomic Energy of Canada Limited Report
AECL-7404, April 1981.
[Simpson, 1996] L. A. Simpson, P. M. Mathew, A. P. Muzumdar, D. B. Sanderson, and V. G. Snell,
“Severe Accident Phenomena and Research for CANDU Reactors”, Proc. 10th. Pacific Ba-
sin Nuclear Conference, October 20–21, 1996, Kobe, Japan.
[Slovic, 1987] P. Slovic, “Perception of Risk”, Science, Vol. 236, No. 4799, pp. 280-285, April 1987.
[Smithsonian, 2004] Smithsonian National Museum of American History, Three Mile Island: The
Inside Story, 2004.
[Snell, 1978] V. G. Snell, “Evolution of CANDU Safety Philosophy”, Proceedings, Canadian Nu-
clear Society Symposium on CANDU Reactor Safety Design, November 1978.
[Snell, 1988] V. G. Snell, S. Alikhan, G. Frescura, J. Q. Howieson, F. King, J. T. Rogers, and H.
Tamm, “CANDU Safety under Severe Accidents: An Overview”, IAEA/OECD International
Symposium on Severe Accidents in Nuclear Power Plants, Sorrento, Italy, March 1988.
Also Atomic Energy of Canada Ltd. Report AECL-9802.
[Snell, 1989] V. G. Snell, J. W. Hilborn, G. F. Lynch, and S. McAuley, “Safety Aspects of District
Heating Reactors”, invited paper presented to the IAEA Workshop on the Safety of Nu-
clear Installations of the Next Generation and Beyond, Chicago, August 1989.
[Stacy, 2000] S. M. Stacy, Proving the Principle, Idaho Operations Office of the Department of
Energy, Idaho Falls, Idaho, report DOE/ID-10799, 2000. This is a history of Idaho Nuclear
Laboratories. See Chapters 15 and 16 for SL-1.
[StatsCan, 2009] Leading Causes of Death, Total Population, by Age Group and Sex, Statistics
Canada, 2009.
[Talbot, 2003] E. O. Talbot, A. O. Youk, K. P. McHugh-Pemu, and J. V. Zborowski, “Long-Term
Follow-Up of the Residents of the Three Mile Island Accident Area: 1979–1998”, Envi-
ronmental Health Perspectives, Vol. 111, No. 3, March 2003.
[TEPCO, 2011] Tokyo Electric Power Company, Evaluation Status of Exposed Dose of Employees
Engaged in Emergency Work in Fukushima Daiichi Nuclear Power Station, Exhibit, July
13, 2011.
[TEPCO, 2011a] Tokyo Electric Power Company, Effects of the Earthquake and Tsunami on the
Fukushima Daiichi and Daini Nuclear Power Stations, May 24, 2011.
[Thompson, 1965] T. J. Thompson and J. G. Beckerley, The Technology of Nuclear Reactor Safety,
MIT Press, 1965.
[Tregoning, 2008] R. Tregoning, L. Abramson, and P. Scott, Estimating Loss-of-Coolant Accident

(LOCA) Frequencies through the Elicitation Process, United States Nuclear Regulatory
Commission report NUREG-1829, April 2008.
[UNSCEAR, 2000] United Nations Scientific Committee on the Effects of Atomic Radiation,
ANNEX J, Exposures and Effects of the Chernobyl Accident, UNSCEAR 2000 Report to the
General Assembly, 2000.
Hereditary Effects of Radiation, UNSCEAR 2001 Report to the General Assembly, with
Scientific Annex, 2001.
ANNEX D, Health Effects due to Radiation from the Chernobyl Accident, UNSCEAR 2008
Report to the General Assembly, 2008.
[USAEC, 1962] U.S. Atomic Energy Commission, Idaho Operations Office, Nuclear Incident at the
SL-1 Reactor, report IDO-19302, January 1962.
[USAEC, 1962a] U.S. Atomic Energy Commission, Final Report of SL-1 Recovery Operation, report
IDO-19311, July 1962.
[USAEC, 1962b] U.S. Atomic Energy Commission, Additional Analysis of the SL-1 Excursion,
report IDO-19313, November 1962.
[USDOE, 1986] U.S. Department of Energy, Report of the U.S. Department of Energy’s Team
Analyses of the Chernobyl-4 Atomic Energy Station Accident Sequence, DOE/NE-0076,
November 1986.
[USNRC, 1981] United States Nuclear Regulatory Commission, Fault Tree Handbook, Report
NUREG-0492, January 1981.
[USNRC, 2001] United States Nuclear Regulatory Commission, Systematic Radiological Assess-
ment for Source and By-Product Materials, report NUREG-1717, June 2001.
[USNRC, 2002] A. Howell, et al., United States Nuclear Regulatory Commission, Degradation of
The Davis-Besse Nuclear Power Station Reactor Pressure Vessel Head: Lessons Learned
Report, USNRC report, September 30, 2002.
[USNRC, 2011] United States Nuclear Regulatory Commission, Recommendations for Enhancing
Reactor Safety in the 21st Century: The Near-Term Task Force Review of Insights from the
Fukushima Dai-Ichi Accident, USNRC report, July 12, 2011.
[USNRC, 2011a] United States Nuclear Regulatory Commission, Final Safety Culture Policy
Statement, USNRC report NRC–2010–0282; see Federal Register / Vol. 76, No. 114 /
Tuesday, June 14, 2011 / Notices.
[USNRC, 2013] United States Nuclear Regulatory Commission, Title 10, Code of Federal Regula-
tions, Part 50, Appendix A, Criterion 54.
[USSR, 1986] USSR State Committee on the Utilization of Atomic Energy, The Accident at the
Chernobyl Nuclear Power Plant and its Consequences, Information compiled for the IAEA
Experts’ Meeting, Vienna, 25-29 August 1986; Parts I and II, August 1986. Also see In-
formation on the Accident at the Chernobyl Nuclear Power Station and its Consequences,
prepared for IAEA, translation of Atomnaya Energiya, Vol. 61, No. 5, pp. 301-320, No-
vember 1986.
[Venart, 2004] J. E. S. Venart, “Flixborough: the Explosion and its Aftermath”, Trans IChemE, Part
B, March 2004; Process Safety and Environmental Protection, Vol. 82, Issue B2, pp. 105–
127, 2004.
[Westinghouse, 2011] Westinghouse, AP1000 Design Control Document, Tier 2, Revision 19,
Section 15.4.8.2.1.7, USNRC.
[Westinghouse, 2011a] Westinghouse, AP1000 Design Control Document, Tier 2, Revision 19,
Section 15.6.5.2, USNRC.
[Westinghouse, 2011b] Westinghouse, AP1000 Design Control Document, Tier 2, Revision 19,
Section 1, USNRC.
[Wren, 1999] J. C. Wren, J. M. Ball, and G. A. Glowa, “The Chemistry of Iodine in Containment”,
Nuclear Technology, Vol. 129, No. 3, pp. 297-325, 2000.
[Wren, 2001] J. C. Wren and J. M. Ball, “LIRIC 3.2—An Updated Model for Iodine Behaviour in
the Presence of Organic Impurities”, Radiation Physics and Chemistry, Vol. 60, 577–596,
2001.
[Xu, 1999] J. Xu, V. S. Krishnan, P. J. Ingham, B. N. Hanna, and J. R. Buell, “Thermal-Hydraulic
Design Methods and Computer Codes”, China Journal of Nuclear Power Engineering, Vol.
20, No. 6, 1999.
[Yaremy, 1979] E. M. Yaremy, “A Review of the Incident at the Three Mile Island Nuclear Power
Plant”, presented to the Canadian Electrical Association Thermal and Nuclear Power Sec-
tion Meeting, Calgary, Alberta, October 1979.
11 Further Reading
[Duderstadt, 1976] J. J. Duderstadt and L. J. Hamilton, Nuclear Reactor Analysis, New York, NY:
Wiley, 1976. ISBN: 9780471223634.
[IAEA, 2001] International Atomic Energy Agency, Thermo-Hydraulic Relationships for Advanced
Water-Cooled Reactors, IAEA-TECDOC-1203, April 2001.
[OECD-NEA CSNI, 1989] Thermo-Hydraulics of Emergency Core Cooling in Light Water Reactors,
a State-of-the-Art Report (SOAR) by a Group of Experts of the NEA Committee on the
Safety of Nuclear Installations, OECD-NEA report Rep. No. 161, Paris.
12 Glossary
AECB Atomic Energy Control Board
AECL Atomic Energy of Canada Limited
AOO Anticipated Operational Occurrence
BDBA Beyond Design Basis Accident
BEAU Best Estimate with Analysis of Uncertainties
BECS Boiler Emergency Cooling System
CHF Critical Heat Flux
CNSC Canadian Nuclear Safety Commission
DBA Design Basis Accident
DBE Design Basis Earthquake
DSA Deterministic Safety Analysis
ECC Emergency Core Cooling
ECI Emergency Coolant Injection (a subsystem of ECC)
EPR European Pressurized Reactor
EWS Emergency Water System
FMEA Failure Modes and Effects Analysis
HPECC High-Pressure ECC
HTS Heat-Transport System (see also RCS)
IAEA International Atomic Energy Agency
INES International Nuclear Event Scale
INPO Institute of Nuclear Power Operations
INSAG International Nuclear Safety Advisory Group
LLOCA Large Loss-of-Coolant Accident
LOCA Loss-of-Coolant Accident
LPECC Low-Pressure ECC
LWR Light-Water Reactor
MCR Main Control Room
MPECC Medium-Pressure ECC
MSSV Main Steam Safety Valve
MWe Megawatt electric

OPG Ontario Power Generation
PRV Pressure-Relief Valve
PSA Probabilistic Safety Analysis
PWR Pressurized-Water Reactor
R&D Research and Development
RBMK Реактор Большой Мощности Канальный (High Power Channel-Type Reactor)
RCS Reactor Coolant System (see also HTS)
ROP Regional Overpower (Protection System)
S/C Suppression Chamber
SCA Secondary Control Area
SCDF Severe Core-damage Frequency
SDS Shutdown System
SG Steam Generator
SOE Safe Operating Envelope
SOR Shut-Off Rod
Sv Sievert, a unit of radiation dose
TMI Three Mile Island
USNRC United States Nuclear Regulatory Commission
13 Appendix 1 – Basic Rules of Boolean Algebra

This is a quick refresher in the basics of Boolean logic.
Boolean algebra is a set of laws formulated by the British mathematician George Boole. It deals
with statements (here represented by A, B, C, etc.) which can be either true or false and are
often denoted by the numbers 1 and 0 respectively. So, for example, A=0 means that the
statement A is false.
For example, let A represent the statement, “the sun is the centre of the universe”. Then A=0.
13.1 Operators
Several operators can act on these statements:
AND is an operator which gives the answer C=1 only if both of its inputs are 1, and 0 otherwise.
It should be represented as:
A.B or AB or A∙B or A  B or A.AND.B, or graphically as in Figure 69. C
It can be thought of as the intersection of sets A and B.
Examples:
AND
If A is true and B is true, then A  B should be true.
If A=1 and B=0, then A.B=0.
A B
OR is an operator which gives the answer C=1 only if either or both of its inputs Figure 69 AND gate
are 1, and 0 otherwise. It is represented as:
A+B or A∪B or A.OR.B, or graphically as in Figure 70.
C
It can be thought of as the union of sets A and B.
Examples:
OR
If A is true and B is false, then A∪B is true.
If A=1 and B=0, then A+B=1.
A B
NOT is an operator which inverts the input. Figure 70 OR gate
It is represented as .NOT.
The output of .NOT.A is denoted as A’ or Ā.
It is represented graphically as in Figure 71.
It can be thought of as “the opposite of” or “the complement of” set A.
Examples: Figure 71 NOT gate
If A is true, then .NOT.A is false.
There are also four lesser-used operators (.NAND., .NOR., .XOR., and .XNOR.)
13.2 Basic Principles
A=0 or A=1
(i.e., something can only be true or false, not both)
If A=0, A  A=0
(i.e., if A is false, and because A  A=A, then A  A is false. Sometimes this is written as 0.0=0)
If A=1, A∪A=1
(i.e., if A is true, and because A∪A=A, then A∪A is true. Sometimes this is written as 1+1=1.
Remember, you are not doing arithmetical addition!)
If A=0, then A∪A=0

(This is obvious if you think of set theory as we did above. Sometimes it is written 0+0=0).
If A=1, then A  A=1

Again, obvious from set theory; also written 1.1=1
If A=1 and B=0, then A  B = B  A = 0

(i.e., if A is true and B is false, then the intersection of A and B can never be true, and vice versa.
Sometimes this is written as 1.0 = 0.1 = 0)
If A=1 and B=0, then A∪B = B∪A = 1

(i.e., if either one of A or B is true, then the union of A and B (A.OR.B) is always true. This can be
written 1+0 = 0+1 = 1)
13.3 Theorems
These sound abstract, but are obvious once you draw Venn diagrams with set-theory symbols.
You can substitute the mathematical symbols + and ∙ for ∪ and  if this is more intuitive for
you.
Commutative Law
A∪B = B∪A
AB = BA
Associative Law
(A∪B)∪C = A∪(B∪C)
(A  B)  C = A  (B  C)
Distributive Law
A  (B∪C) = (A  B) ∪ (A  C)
A∪(B  C) = (A∪B)  (A∪C)
Identity Law
A A = A
A∪A = A
Completeness
(A  B)  (A  B)  A
(A  B)  (A  B)  A
Redundancy
A∪(A  B) = A
A  (A∪B) = A
Mathematical
1+A=1
1∙A=A
0+A=A
0∙A=0
A+Ā=1
A∙Ā=0
A + ĀB = A + B
A(Ā + B) = A∙B
DeMorgan’s Theorem
(A  B)  A B
(A  B)  A  B
These will be useful when you work out fault trees mathematically.
13.4 Combining Probabilities

If event A occurs x times out of n repeated experiments, then:
P ( A)  Probability of event A
x
 lim n
n (59)
Obviously:
0  P( A)  1
P( A)  P( A)  1 (60)
In other words, an event must either occur or not occur; there is no third possibility.
13.4.1 Probability of both events occurring

A1A2 means that both events occur, and therefore P(A1A2) is the probability that both events
occur. The product rule for probabilities states that:
P(A1A2) = P (A1 | A2) P (A2) = P (A2 | A1) P (A1). (61)
For example, if A1 is the probability that part 1 fails and A2 is the probability that part 2 fails,
then
P(A1A2) = probability that both part 1 fails and part 2 fails
= probability that part 2 fails  (probability that part 1 fails, given that part 2 fails)
= probability that part 1 fails  (probability that part 2 fails, given that part 1 fails).
Because probabilities are numbers, the  does mean simple multiplication in these equations.
It is conventionally omitted; i.e., P(A1)P(A2) is understood to mean P(A1)  P(A2).
The conditional probability P (A1 | A2) means the probability of event A1 given that event A2 has
occurred.
Figure 72 shows this graphically; yellow represents all events; green those events with outcome
A1; blue with outcome A2; red, with outcome both A1 and A2.
Figure 72 Probability of both of two events

If the events are independent,
P(A2 | A1) = P(A2). (62)
In general:
P( A1 A2 ...AN )  P( A1 ) P( A2 A1 )...P( AN A1...AN 1 ) . (63)
If events are independent:

P( A1 A2 ...AN )  P( A1 ) P( A2 )...P( AN ) . (64)
13.4.2 Probability of either event occurring

The union of two events, A1 and A2, is denoted as:
A1  A2 or A1 + A2 or A1 OR A2.
This means the cases where either event occurs, including cases where both events occur. Note
that:
P(A1 + A2) = P (A1) + P (A2) - P (A1A2). (65)
Figure 73 Probability of either of two events

as shown in Figure 73. Note that the + sign on the left of this equation is not the same as the +
sign on the right: the former refers to the union of two sets, the latter to arithmetic addition.
The reason for subtracting P(A1A2) is that what you want is the total area encompassed by the
combination of the blue and green ovals in the diagram. If you simply add P(A1) and P(A2), you
count the intersection where both events occur (in red) twice. Therefore, you have to subtract
one of them away.
In general:
N N 1 N
P( A1  A2  ...  AN )   P( AN )    P( A A n m )  ...  (1) N 1 P( A1 A2 ...AN ) . (66)
n 1 n 1 m  n 1
If events are independent:

N
1 P( A1  A2  ... AN )  [1  P( AN )] . (67)
n 1
13.4.3 Rare independent events

Assuming (as is often but not always the case for nuclear failure probabilities) that the events
are both independent and rare, or P(AN)<<1, then:
N
P( A1  A2  ...  AN )   P( AN ) . (68)
n 1
14 Appendix 2 – Common-Cause Failures – An Example

Consider two shut-off rods, each of which has a probability of failure of 0.001 per demand.
What is the probability that they both fail when required?
If they are independent, then P(A1A2) = P(A1)P(A2) = (0.001)2 = 10-6 per demand.
Suppose that a common-cause (CC) failure occurs 10% of the time. That is:
P(A1) = P(A2) = 0.0009 (random) + 0.0001 (CC),
and therefore the probability of one rod failing given that the other has failed is 90% random
and 10% common cause (with probability 1):
P(A1|A2) = 0.9 * 0.001 + 0.1 * 1 = 0.1009
or
P(A1A2) = P (A1 | A2) P (A2) = 0.1009 * 0.001 = 0.0001009 ~ 10-4.
Therefore, a 10% common-cause probability has increased the combined probability of failure
by a factor of 100.
15 Appendix 3 – Why a Reactor Cannot Explode Like an Atomic Bomb

This Appendix provides further technical detail on the difference between a nuclear weapon
and a reactor power excursion that is not shut down. It assumes that you have read the reactor
physics Chapter. Some material is taken from [Meneley, 2003]. All material is in the public
domain.
Assume a pure plutonium core surrounded by a shaped charge of high explosive. The core is
initially subcritical and is rendered supercritical by compression from the explosive. The neu-
tron kinetics follows the familiar equation (delayed neutrons are irrelevant here):
dN (t )   
 N (t ) , (69)
dt 
where N(t) is the neutron density,  is the reactivity,  is the delayed neutron fraction, and  is
the prompt neutron generation time.
Solving,
 
( )t
N (t )  N (0)e 
. (70)
For a weapon, a typical value of  -  is ~1.7 (so that  is pretty well irrelevant). Because critical-
ity is achieved by fast, not thermal fission,  is about 10-8 seconds or one “shake”. The e-folding
time (when the population of neutrons increases by a factor of 2.7) is therefore 1.7 shakes.
After 33 shakes, the number of neutrons is:
8
N (33 108 )  1 e1.710 3310  2.3 1024 .
8
(71)
Using very crude assumptions, such as not adding in the contribution of previous generations of
neutrons and not counting energy losses and core expansion, we can estimate the order of
magnitude of the energy produced. If each fission event generates 180MeV of usable energy,
and if for each incident neutron absorbed in Pu239, two fast neutrons are produced, the number
of neutrons in Equation (71) is accompanied by an energy release at 56ns of:
1kT
E  2.3 1024 neutrons  0.5 fissions / neutron 180Mev / fission   8kT , (72)
2.6 1025 MeV
where kT is kilotons of TNT equivalent. This is 34 × 106MJ.
The reaction is self-limiting because the thermal energy causes the core to expand rapidly and
stops the chain reaction.
By contrast, in a thermal reactor such as CANDU, the largest  -  (from a large LOCA) is about
0.011;  is about 0.001 sec. The timescale is measured in tens of seconds rather than nanosec-
onds, and the reaction is terminated by damage to the core lattice after about ten seconds, with
an energy input of ~50 full-power seconds. The total energy produced for an 1800MWth
reactor is therefore 9×104 MJ, or more than two orders of magnitude smaller than the value for
a bomb.
16 Acknowledgments
We gratefully thank the following reviewers for their hard work and excellent comments during the
development of this Chapter. Their feedback has much improved it. Of course the responsibility for any
errors or omissions is entirely the author’s.
Glenn Archinoff
Glen McGee
Dan Meneley
Terry Rogers
We also thank Diana Bouchard for expertly editing and assembling the final copy.
1
CHAPTER 14
Nuclear Plant Materials and Corrosion
prepared by
Dr. Derek H. Lister and Dr. William G. Cook
Summary:
The choice of materials of construction of a nuclear reactor, while important in terms of plant
capital cost, is crucial to the safe and economic operation of the unit throughout its design
lifetime; it also affects decisions about plant life extension. Most of the failures of nuclear
systems involve the degradation of materials as they interact with their environments, indicating
that chemistry control within systems should be formulated as materials are selected. All of the
three major process systems of a CANDU reactor – the primary coolant, the moderator and the
secondary coolant – have a variety of materials of construction, so the control of system chemis-
try in each is a compromise based on the characteristics of the interactions between the system
materials and the environment, including the effects of irradiation on both the material and the
coolant or moderator. Ancillary systems, such as those providing condenser cooling water or
recirculating cooling water, have similar constraints on material choice and chemistry control.
Components such as concrete structures, cabling and insulation, although not necessarily
associated with process systems, may also be critical to safety or plant operation and must have
their materials chosen with care. This chapter describes the materials of construction of the
main systems and components of a CANDU reactor and shows how they interact with their
environments.
Table of Contents
1 Introduction ............................................................................................................................ 3
2 Materials for Nuclear Applications ......................................................................................... 4
2.1 Metallurgy and Irradiation Effects............................................................................................... 4
2.2 Materials of Construction............................................................................................................ 7
3 CANDU Process Systems and their Materials ....................................................................... 21
3.1 Major Process Systems.............................................................................................................. 21
3.2 Ancillary Process Systems.......................................................................................................... 23
3.3 Supporting Structures and Components ................................................................................... 25
4 Corrosion and Material Degradation in Nuclear Reactor Systems ....................................... 27
4.1 General Corrosion...................................................................................................................... 27
4.2 Flow-Accelerated Corrosion (FAC)............................................................................................. 35
4.3 Galvanic Corrosion..................................................................................................................... 39
4.4 Stress-Corrosion Cracking & IGA ............................................................................................... 42
4.5 Crevice Corrosion & Pitting ....................................................................................................... 45
4.6 Fretting Wear & Flow-Induced Vibration .................................................................................. 48
4.7 Hydrogen Effects ....................................................................................................................... 49
©UNENE, all rights reserved. For educational use only, no assumed liability. Nuclear Plant Materials and Corrosion – September 2014
4.8 Microbiologically-Induced Corrosion (MIC)............................................................................... 49

5 References............................................................................................................................. 51
List of Figures
Figure 1. Unit cell configurations of some common metallic crystal structures (after Callister) ... 5
Figure 2. Grain structure for low-alloy (UNS G10800), austenitic (316 SS) steels and Zircaloy 4
(ASM, 2004). ..................................................................................................................... 5
Figure 3. Effect of neutron irradiation on the stress/strain properties of materials with cubic
crystal structures (after Was, 2006).................................................................................. 6
Figure 4. Partial iron-carbon phase diagram (after Callister)........................................................ 10
Figure 5. Schematic Cross-Section through Magnetite Film on Carbon Steel [Lister, 2003] ........ 31
Figure 6. Scanning Electron Micrograph of Magnetite Film formed on Carbon Steel during an
1100-h Exposure in High-Temperature Water [Lister, 2003] .......................................... 31
Figure 7. Schematic diagram of Zircaloy corrosion in high-temperature water, showing three
regions (dotted line indicates early assumed pre-transition cubic and post-transition
linear kinetics) [Hillner et al, 2000]................................................................................. 33
Figure 8. Scalloped surface of outlet feeder from Point Lepreau CANDU.................................... 36
Figure 9. (a) Effect of temperature and flow on FAC at pH25ºC 9.0 with ammonia; (b) Effect of
temperature and pH on the solubility of magnetite. ..................................................... 38
Figure 10. The overall mechanism of flow-accelerated corrosion................................................ 38
Figure 11. Examples of transgrannular (TGSCC) (a) and intergrannular (IGSCC) (b) stress
corrosion cracking in stainless steel and brass respectively (after Fontana 1986)......... 42
Figure 12. Typical sensitization mode and IGA of Type 304 stainless steel (after Fontana 1986) 44
Figure 13. Initial (A) and final (B) stages in the development of crevice corrosion of a metal M
(Fontana, 1986)............................................................................................................... 46
Figure 14. Self-perpetuating corrosion pit (after Fontana, 1986)................................................. 47
Figure 15. An undercut pit and surface wastage from overlapping pits....................................... 47
List of Tables
Table 1. Compositions of commonly used zirconium alloys in CANDUs......................................... 8
Table 2. Standard compositions of commonly used ferritic, martensitic and austenitic steels in
CANDUs........................................................................................................................... 14
Table 3. Compositions of commonly used nickel alloys in nuclear power plants......................... 15
Table 4. Severity of radiation-induced damage of common polymers......................................... 20
Table 5. The galvanic series for important metals ........................................................................ 41
Table 6. Biocides used in industrial water systems....................................................................... 51
Nuclear Materials and Corrosion 3
1 Introduction
The performance of power-producing systems has always been limited by the properties of
engineering materials and their interactions with the environment. Since the early days of
industrial steam generation in the seventeenth and eighteenth centuries, for example, the ever-
increasing need for power for new factories created a steady demand for larger boilers and
increasingly severe steam conditions. Materials selection was based on limited knowledge of
corrosion, especially stress corrosion cracking as the consequence of chloride and oxygen
contamination of the water used in the boilers and catastrophic failures of equipment occurred
frequently.
As late as the end of the nineteenth century, hundreds of steam plant explosions accompanied
by large numbers of casualties were being recorded every year in Europe and North America.
The causes were generally linked to the failure of riveted joints or poorly worked steel plate in
fire-tube boilers. The innovation of the water-tube boiler and the understanding of localized
corrosion of steels in high-temperature water were major factors that led to much safer equip-
ment - even as operating conditions continued to become more severe. The major conse-
quence of these failures was the development of codes and standards for pressure retaining
components, the ASME Pressure Vessel Code in North America for example, which dictate the
acceptable design and construction practices for materials produced to particular specifications.
Today, the risk of catastrophic failure of a power-producing system is low. This can be attributed
to the strict adherence to the standards that are imposed on component designers and manu-
facturers as well as on plant operators. The setting of these standards clearly involves a tho-
rough knowledge of the properties of the materials of construction and an understanding of
their behaviour in the local environment. That environment may itself be adjusted for overall
optimum performance by specifying optimum chemistry control strategies. The prime example
of such chemistry control is the specification of an alkalinity level in the feedwater systems of
steam-raising plants, including nuclear secondary coolants, which is necessary to minimize
corrosion of piping and components and to keep systems clean of dissolved and particulate
corrosion products and impurities. In a nuclear reactor core, materials must be able to with-
stand not only the operational conditions (pressure, temperature and water chemistry) but
must show minimal degradation from the effects of radiation (high gamma and neutron flux).
Radiation effects on materials may include loss of ductility, shape change from radiation en-
hanced creep and growth and enhanced corrosion resulting from hydrogen ingress (deuterium
ingress for the case of heavy water systems).
While the current performance record of power plants is generally good, technology is not
standing still. The push for bigger returns on capital investment and the accompanying trends
towards higher plant efficiencies and longer component lifetimes lead to even more severe
operating conditions in power systems. Inevitably, the demands on the materials of construc-
tion escalate. The predominant materials of construction in steam-raising equipment and
nuclear systems are the metals and alloys. Their interaction with the operating environment
very much dictates the chemistry control to be practised by plant operators.
2 Materials for Nuclear Applications
As described above, materials of construction for nuclear applications must be strong, ductile
and capable of withstanding the harsh environment to which they are subjected. Furthermore,
for materials used in the core of a nuclear reactor, it is important that they have specific proper-
ties such as low neutron absorption and high resistance to radiation-induced creep, hardening
and the associated loss of ductility so that reactors can operate for the decades expected by
plant owners. Thus, nuclear materials must be selected or specified based upon their strength
and interactions with the environment (including the effects of radiation), all of which are
dependent upon the metallurgy of the material.
2.1 Metallurgy and Irradiation Effects

2.1.1 Crystal Structure and Grain Boundaries
All metals used in power plant construction are crystalline in nature, meaning that they have a
defined and consistent crystal structure. Three of the basic and most observed crystal struc-
tures are shown in
Figure 1, namely: face-centred cubic (FCC), body-centred cubic (BCC) and hexagonal close-
packed (HCP). Upon cooling from a melt during the production of a metal or alloy, individual
crystals nucleate and coalesce as the material solidifies. As the individual crystals grow together
they eventually meet and combine as an assemblage of crystal grains, which will likely not have
the same orientation of the crystal lattice. This leads to the development of grain boundaries in
the material, where the atoms along the boundary between the two grains have elongated
metallic bonds leading to slightly elevated boundary energies than those within the bulk of the
individual crystal. The grain boundaries are thus slightly more active than the individual grains
and can act as diffusion short-circuits or locations for the development of precipitates such as
carbides. The materials used in nuclear power plants are thus termed polycrystalline due to the
grain/grain-boundary structure.
(a) FCC structure (b) BCC structure (c) HCP structure

Figure 1. Unit cell configurations of some common metallic crystal structures (after Callister)
The grains of a metal or alloy may be engineered through appropriate heat treatments (hot and
cold working and annealing for example) to be of a specific size and/or orientation to meet a
specified purpose. Small-grained metals (1 – 25 m) will have more grain boundaries and may
be more affected by phenomena such as high-temperature creep or radiation-induced creep.
Examples of the metallurgy and microstructure of ferritic steel, austenitic stainless steel and
zirconium are shown in Figure 2. Note that the grain structure, size and orientation may play
significant roles in the specific materials properties such as strength, conductivity and creep
resistance.
Figure 2. Grain structure for low-alloy (UNS G10800), austenitic (316 SS) steels and Zircaloy 4
(ASM, 2004).
2.1.2 Irradiation effects on materials

The effects of irradiation on materials may be manifest in several ways, the most dramatic being
the dislodging or dislocation of a random atom in the crystal lattice to a new location. During
neutron bombardment, the energy dissipated by the neutron upon collision with a metal atom
can create new defects in the crystal structure, typically “interstitial” sites where the dislocated
atoms come to rest between the regular bonding sites in the crystal structure and “vacancy”
sites left where the atoms were originally stationed. The pair (interstitial/vacancy) is termed a
Frenkel pair and is a key phenomenon associated with radiation damage in polycrystalline
materials. The damage induced in the material is additive and substantial, considering that in a
CANDU reactor the overall neutron flux may be 2x1017 neutrons/m2s and the flux in enriched
reactors (light water reactors, breeders etc.) may be several times higher. Radiation damage is
typically measured in “displacements per atom” (dpa) and the dose can be as high as 10 – 30
dpa after 30 years of operation.
The fact that neutron bombardment creates interstitials and vacancies within the material gives
rise to changes in the physical dimensions and the mechanical properties of the material when
placed under stress. The creation of vacancies within the material leads to swelling, elongation
and growth whereas production of interstitials can cause the material to shrink. The energy
dissipated by radiation may induce phase changes within the material or promote segregation of
alloying constituents, both phenomena will affect the material’s strength and ductility. Overall,
the mechanical properties and dimensions of a material under irradiation will change and the
effect ultimately leads to challenges for the integrity of the components that make up a nuclear
reactor. Each mechanical effect is described briefly below, more details can be found in Funda-
mental of Radiation Materials Science (Was, 2007).
Irradiation Hardening
When a polycrystalline material is placed under stress it will stretch elastically with increasing
stress/strain up to the yield point of the material. Additional strain beyond the yield strength
leads to plastic deformation whereby the original shape and dimensions of the material can no
longer be attained with the relaxation of the strain. Plastic deformation, at the atomic level, is
induced when the atoms in the crystal lattice slip, which typically occurs on slip planes in the
matrix. When a material is irradiated by particles (neutrons) producing Frenkel pairs, the effect
of the interstitital/vacancy sites is to provide barriers to the normal slip planes within the
material. This effectively increases the yield strength, which will increase continuously in
proportion to the increasing radiation dose (dpa). Figure 3 shows the effect of neutron irradia-
tion on the yield stress/strain for materials with the FCC and BCC crystal structures [Was, 2006].
Figure 3. Effect of neutron irradiation on the stress/strain properties of materials with cubic
crystal structures (after Was, 2006).
Irradiation Creep and Growth

All metals will exhibit creep over time when exposed at high temperatures and under opera-
tional stress. Thermal creep is a diffusion-based migration of atoms and vacancy sites within the
metal’s lattice, which typically requires a minimum operating temperature before its effects are
observed – generally the operating temperature must be greater than 0.3Tm, where Tm is the
melting temperature of the metal. Since diffusion is the primary mechanism at play for thermal
creep it typically increases exponentially with increasing temperature following an Arrhenius
law. As described above, particle bombardment or irradiation of a metal creates interstitials and
vacancies within the metal’s lattice, effectively mimicking the effects of thermal creep of the
metal only occurring at much lower temperatures than those required for significant thermal
creep alone. Since the diffusion rates of vacancies and interstitials are unaffected by irradiation
(diffusion is a temperature dependent phenomenon), irradiation creep is more complex in its
mechanism and involves the preferential absorption of interstitials under the action of strain in
the material and phenomena such as the “climb and glide” of the dislocations over barriers to
the diffusion process. Irradiation creep is preferential in directions perpendicular to the applied
stress i.e. pipes will tend to elongate due to irradiation creep.
For materials that do not have a cubic crystal structure (zirconium being the most relevant with
an HCP structure below 862oC), grain distribution and orientation (described as the material’s
texture) play significant roles in irradiation creep and growth. The distinction made between
creep and growth is that the former is the elongation of a material under applied stress while
the latter occurs even when no external stress is applied. The material’s texture gives rise to
anisotropy in the preferred crystallographic planes by which the material will stretch under an
induced strain as well as grow in a preferred direction during creep. Thus, the creep characteris-
tics of non-cubic crystal materials (zirconium and zirconium alloys) are highly dependent upon
the applied stress as well as the operational temperature.
Irradiation Embrittlement
As a metal gains strength upon particle irradiation, it will also tend to lose its ductility, becoming
hard and brittle with increasing damage and radiation dose. The material’s yield strength
increases through build-up of dislocations along barriers to atom migration, but other effects
such as the precipitation of secondary and tertiary phases within the material can change the
mechanical properties dramatically. Materials that are typically ductile and not susceptible to
brittle fracture in an un-irradiated environment may cleave and fracture excessively upon
irradiation due to the increasing hardness and loss of ductility.
2.2 Materials of Construction

2.2.1 Zirconium
Zirconium is the eleventh most abundant element in the earth’s crust making it more prevalent
than the common transition metals copper, lead, nickel and zinc. It occurs naturally as the
minerals zircon (zirconium silicate – ZrSiO4) and zirconia (zirconium oxide – ZrO2) and always in
conjunction with 0.5 – 2.5% hafnium (Hf), a metal that has very similar chemical and physical
properties making it difficult to separate the two elements. Zirconium or, more specifically, the
alloys fabricated from it are the most important of the nuclear reactor materials. They are
resistant to corrosion in many process environments, nuclear heat transport systems in particu-
lar, and they have excellent nuclear properties making them the predominant materials for
construction of in-core components (fuel sheathing, pressure tubes and calandria tubes).
Zirconium itself has a neutron absorption cross-section of 0.18 barns, making it nearly transpa-
rent to the thermal neutrons in a water-cooled and -moderated reactor. Zirconium alloys used
in nuclear reactors must be highly processed to keep the hafnium concentration below 100 ppm
since its neutron absorption cross-section is 102 barns, nearly 600 times that of zirconium.
Higher concentrations of hafnium would increase the parasitic absorption of neutrons in the
reactor core and reduce the fuel burn-up. Pure zirconium has an HCP structure that transforms
to BCC at 862C before melting at 1850C. It is a reactive metal that combines easily with
oxygen, hydrogen, nitrogen, carbon and silicon and in air retains its metallic lustre because of a
very thin but protective zirconia (ZrO2) oxide layer. A protective but thicker zirconia layer is
retained during exposure to high-temperature water, making zirconium alloys ideal for service in
water-reactor cores. The pure metal is alloyed with small amounts of elements such as tin,
chromium, iron, nickel and niobium to improve mechanical strength, corrosion resistance and to
reduce hydrogen pickup. In CANDUs, the alloy Zircaloy-2 is used for the calandria tubes and was
the initial choice as the pressure tube material. Zircaloy-2 contains 1.2 – 1.7 weight percent Sn,
0.07 – 0.2 weight percent Fe, 0.05 – 0.15 weight percent Cr and 0.03 – 0.08 weight percent Ni. It
is also the material used for BWR fuel cladding. For fuel sheathing in CANDUs and fuel cladding
in PWRs, Zircaloy-4 was developed and used because it was observed to have a lower overall
corrosion rate and a reduced tendency to pick up hydrogen. It has a similar composition to that
of Zircaloy-2 except for the nickel content, which is reduced to a maximum of 0.007 weight
percent, and the iron content, which is increased to 0.18 – 0.24 weight percent. The composi-
tions of zirconium alloys used in CANDU reactors are shown in Table 1. Comparatively recent
developments in fuel cladding technology for PWRs are the Westinghouse alloy ZIRLO, which
contains 0.7 – 1.0 weight percent Sn and 1.0 weight percent Nb, and the Areva variant M5,
which has no Sn and 0.8 – 1.2 weight percent Nb. A Zr-1%Nb alloy was developed and used for
decades as the fuel cladding material in the Russian VVERs and the Westinghouse and Areva
alloys are modifications from the original Russian composition. They are reputedly more
corrosion-resistant and have lower hydrogen pick up than Zircaloy-4 under high fuel burn-up.
Similarly, for BWR cladding, Westinghouse has developed the alloy ZrSn, which is Zr with 0.25
weight percent Sn. The pressure tubes in CANDUs are made of Zr-2.5%Nb, an alloy of Zr and Nb
with the latter within the range 2.4 – 2.8 weight percent. This too is a modification of Russian
alloys developed for the RBMK reactors.
Table 1. Compositions of commonly used zirconium alloys in CANDUs.

Common UNS # chemical composition (wt%)
designation Sn Nb Fe Cr Ni O Hf
(ppm) (ppm)
Zr 702 R60702 Fe + Cr < 0.2 < 0.16% 4.5%
Zircaloy 2 R60802 1.2-1.7 0.07- 0.05- 0.03- 1200- 100
0.2 0.15 0.08 1400
Zircaloy 4 R60804 1.2-1.7 0.18- 0.07- 0.03- 1200- 100
0.24 0.13 0.08 1400
Zr-2.5%Nb R60901 2.4-2.8 1200- 100
1400
2.2.2 Steels
Carbon Steels
The most widely used materials in power-producing systems are the steels, which are basically
alloys of iron and carbon but which may also contain other metals as minor or major alloying
elements. The alloying elements impart desired properties such as strength, hardness or
corrosion resistance to the steel.
Pure iron at ambient temperatures has a body-centred cubic (BCC) crystal structure, in which
the atoms making up the unit cell of the crystal lattice are conveniently pictured as occupying
the corners of a cube with an atom at the cube centre. Figure 1 shows illustrations of unit cell
configurations of typical metals. The BCC phase is called α-iron or ferrite, which in pure iron is
stable below 912°C. Above this temperature the -iron or austenite phase with a face-centred
cubic (FCC) crystal structure is stable up to 1394°C. The FCC unit cell has an atom at each corner
of a cube and one at the centre of each face as shown in Figure 1. Above 1394°C another BCC
phase, δ-iron, is stable up to the melting point at 1538°C. The relatively small atoms of carbon
that can exist in solid solution in these structures are in random interstitial positions between
the iron atoms. The carbon concentration determines the stability limits of the phases.
As the portion of the iron-carbon phase diagram that is pertinent to steels indicates, carbon has
a low solubility in ferrite. Figure 4 illustrates the iron-carbon phase diagram, in which the
regions of stability of the different steel phases are shown in terms of their composition and
temperature. The maximum in ferrite is 0.02 weight percent and occurs at a temperature of
727°C. At temperatures much above and below this, the solubility declines to almost zero. The
common carbon steels, however, have carbon contents ranging up to about 2.0 weight percent,
and they are quite stable at temperatures below 727°C. To accommodate the carbon, these
steels are composed of more phases than just the ferrite. As the phase diagram indicates, for
carbon contents up to 6.7 weight percent, which is the amount of carbon in the compound Fe3C
(iron carbide or cementite), two distinct phases exist together in equilibrium - ferrite and
cementite. The proportions depend upon the overall amount of carbon in the mixture but are
not affected much by temperature below the 727°C region.
Figure 4. Partial iron-carbon phase diagram (after Callister)
In the aggregate of metal grains and intercalating grain boundaries that make up a carbon steel,
some grains may be ferrite or cementite and some will be a two-phase ferrite-cementite com-
posite. The composite grains are formed from alternating microscopic layers or lamellae of
ferrite and cementite in a structure called pearlite. The size of the grains, the proportions of the
different phases, and the details of the pearlite structure govern the mechanical properties of
the steel. In general, higher carbon contents produce steels that can be made stronger and
harder by heat treatment and mechanical working to control grain structure and properties. At
carbon contents above the 2 weight percent value, however, the material enters the range of
the cast irons and starts to become brittle and lose properties such as ductility and toughness
that are important for many power plant applications. For most boiler applications and piping in
general, steels with carbon contents of 0.20 weight percent or less are employed. Equipment
such as pumps, valves, turbine components etc. requiring stronger or harder material commonly
use steels with carbon contents up to 0.50 weight percent or more.
The phase diagram in Figure 4 indicates that the high-temperature -phase of iron, austenite,
can contain a relatively large amount of carbon in solid solution - up to 2.1 weight percent at
1148°C. This is because the FCC structure of austenite, although having a higher atomic packing
fraction than the BCC structure of ferrite (0.74 versus 0.68), has more interstitial sites to ac-
commodate the small carbon atoms (by contrast, the more open structure of BCC leads to
higher rates of solid-state diffusion). Note that austenite is stable down to the eutectoid tem-
perature 727°C, below which it transforms into the lamellar pearlite. A heat treatment may
then consist of heating a carbon steel to a temperature that puts it in the austenite region,
where the carbon is completely miscible. Cooling at a controlled rate to below 727°C will then
promote the transformation to pearlite, pearlite-ferrite or pearlite-cementite, depending upon

whether the total amount of carbon is respectively equal to, less than or greater than the
eutectoid composition of 0.77 weight percent.
If, during the heat treatment, the austenitized steel at high temperature is cooled suddenly or
quenched (by plunging into cold water, oil, etc.) rather than cooled in a slow controlled fashion,
a different phase called martensite is produced. This has a body-centred tetragonal (BCT) crystal
lattice, the unit cell of which can be viewed as a BCC structure stretched in the direction of one
of the cube axes as shown in Figure 1. Carbon atoms can then be accommodated in interstitial
positions between iron atoms in the direction of elongation or in the centres of the basal planes
and can thereby stay in solid solution. Martensite is a very hard, strong but quite brittle materi-
al that has a metastable structure (and therefore does not appear on the phase diagram, which
shows equilibrium states only). It nonetheless survives indefinitely at ambient temperature.
The formation of martensite can impart very useful properties to steel. Thus, by selective
austenitizing, quenching and then tempering for a given period at temperatures near the
eutectoid point (727°C), the amount of martensite in a steel can be controlled and a hard, strong
material that is not too brittle can be obtained. Other iron-carbon phases called bainite and
spheroidite are also formed during selective heat treatments to produce steels with specific
characteristics of strength, ductility and toughness.
Alloying elements
Metals other than iron may be added to steel to impart desired mechanical properties or
corrosion resistance - particularly useful for service at high temperature. The atoms of alloying
metals are generally in substitutional solid solution, meaning that they take the place of iron
atoms in the crystal lattice. If the metal addition is less than 5 weight percent what is commonly
called a low-alloy steel results, while additions greater than 5 percent produce the intermediate-
and high-alloy steels. Note, however, that the plain carbon steels often contain small amounts
of elements such as manganese, phosphorus or sulphur - generally less than 1 weight percent -
as residuals from the steelmaking process or they may be added for specific reasons such as
improved machinability. Manganese in particular can impart desirable mechanical properties to
steel and in fact may be added deliberately to form low-alloy manganese steel with a marked
resistance to wear. Table 2 shows the typical compositions of various steels and stainless steels
commonly used in CANDUs.
Plain carbon steel and manganese steel are used extensively for general construction steelwork
in power systems and in non-nuclear boiler construction for tubing and components that
experience temperatures below about 500°C. They are also employed in the secondary or
steam-raising systems of nuclear power plants and constitute a major proportion of the piping in
the primary coolant systems of heavy water reactors such as the CANDU reactor, which operate
below about 320°C, and for steam generator tubes and piping in carbon-dioxide-cooled Magnox
reactors, which operate below about 360°C. As the service temperature increases the amount
of alloying is necessarily increased to impart resistance to creep and scaling. For example, boiler
superheater tubes operating at temperature up to 580°C are made of alloys such as 23 weight
percent Cr, 1 percent Mo or even 9 percent Cr, 1 percent Mo. For still higher temperatures, the
use of these low- or intermediate-alloy ferritic steels would require excessive wall thicknesses to
withstand high pressures and to offset scaling, and their cost advantage would be lost. The
highly alloyed materials - in particular, the austenitic stainless steels, “super alloys” and Ni-Cr-Fe
alloys such as the Inconels and Incoloys - are therefore employed.
Stainless steels
The addition of small amounts of selective elements such as copper or chromium to steel can
impart substantial resistance to corrosion, so that some of the low-alloy steels have superior
performance to plain carbon steel in atmospheric environments. True passivity in single-phase
water systems, as achieved with the stainless steels, requires at least 12 weight percent of
chromium as an alloying element. At least 20 weight percent chromium is required for true
passivity in high-temperature gaseous oxidation. [Birks, 2006]
The stainless iron-chromium alloys are basically ferritic with a BCC crystal structure. They
typically have chromium contents of 12-25 weight percent with about 1 percent manganese and
silicon and up to 0.2 percent carbon, to which other elements such as the strong carbide-
formers molybdenum, vanadium and niobium are added to enhance properties such as creep
resistance. Like the carbon steels they can be heat-treated and conditioned, depending upon
their content of carbon and other elements, to provide desired properties of strength, hardness
and toughness. There are also grades that are commonly used in the martensitic condition,
which typically contain 12-20 weight percent chromium, up to 1.0 percent carbon and often 1-2
percent nickel or up to 1.0 percent molybdenum.
The stainless steels used for corrosion resistance and for highest temperature service in power
systems are the austenitic grades. For example, the cladding for the fuel sheaths in the carbon-
dioxide-cooled AGR nuclear reactor, operating at temperatures up to about 650°C, is a 20 weight
percent chromium, 25 percent nickel-niobium stainless steel. The Russian design of pressurized
water reactor, the VVER, also uses austenitic stainless steel for the steam generator tubes
operating below about 320°C. Austenitic steels have the FCC structure, which is stabilized at low
temperature by the presence of nickel at concentrations between about 6 and 20 weight
percent. Resistance to corrosion is again provided by the chromium, which is added at a con-
centration between 16 and 30 weight percent. For most power plant applications, stainless
steels based on the common 18-8 or 18-12 (18 weight percent chromium and 8 percent or 12
percent nickel) grades are employed. In the designation of the American Society for the Testing
of Materials (ASTM) they are types 304 and 320, respectively, of the general 300-series austenit-
ic grades. Manganese is usually present at 2 weight percent and carbon at up to 0.2 percent,
though special low-carbon grades with less than 0.03 percent carbon are available when a
particular resistance to the deterioration of heat-affected regions such as welds by intergranular
attack in corrosive media is required (see intergranular attack Section 4.4). The addition of small
amounts of titanium or niobium to form the stabilized grades can also counteract such weld
deterioration, which is caused by the removal of the protective chromium from solid solution
within the grain boundaries by solid-state precipitation as chromium carbide over a specific
temperature range – a process called sensitization. For resistance to pitting corrosion and to
improve creep resistance, molybdenum is added in the range of 2.0 to 3.0 weight percent to
form type 316 stainless steel. It should be noted that the austenitic stainless steels cannot be
heat-treated like the ferritic steels to improve strength or hardness; however, cold working can
change the microstructure and form small amounts of martensite, for example, that change the
properties and may even induce ferromagnetism. Annealing at high temperature will relieve the
effects of cold-working.
2.2.3 Nickel Alloys

For very high temperature applications (greater than 600°C or so) in power systems where
strength, creep resistance and oxidation resistance are required, such as furnace or turbine
fittings, or for environments where resistance to aqueous corrosion - particularly localised
corrosion - is a major consideration, such as nuclear reactor coolant systems, nickel alloys may
be employed.
Elemental nickel has the FCC crystal structure, which it tends to promote in its alloys - austenitic
stainless steel being the prime example described above. Oxidation resistance of the nickel
alloys is provided by chromium, which is present in the common alloys at concentrations
between about 14 and 25 weight percent. For highly aggressive environments, such as strong
acid solutions, the so-called “super alloys” with molybdenum additions up to 20 weight percent
in alloys of the Hastelloy type, as produced by the Haynes International company, are effective
and even higher levels (at about 30 percent with little or no chromium) produce the alloy
Hastelloy B2 with superior resistance to concentrated hydrochloric and other reducing acids at
all temperatures up to the boiling point. These materials may however have limited application
in high-temperature environments in power systems.
Table 2. Standard compositions of commonly used ferritic, martensitic and austenitic steels in
CANDUs.

designation C-max Fe Cr Ni Mn Mo Si P Cu
Low-alloy steels
A106 gr.B 0.3 Bal. < 0.4 < 0.4 0.29- < 0.15 > 0.1 <0.035 <0.4
1.06
A106 gr.C 0.35 Bal. < 0.4 < 0.4 0.29- < 0.15 > 0.1 <0.035 <0.4
1.06
Martensitic stainless steels
403 S40300 0.15 Bal. 11.5- < 0.6 < 1.0 < 0.5 < 0.04
13
410 S41000 0.15 Bal. 12.5 < 1.0 < 0.04
Austenitic stainless steels
304 S30400 0.08 Bal. 18-20 8-10.5 <2 < 1.0 <
0.045
304L S30403 0.03 Bal. 18-20 8-12 <2 < 1.0 <
0.045
316 S31600 0.08 Bal. 16-18 10-14 < 2.0 2-3 < 1.0 <
0.045
316L S31603 0.03 Bal. 16-18 10-14 < 2.0 2-3 < 1.0 <
0.045
A class of nickel alloys that should be mentioned here, even though they have not achieved wide
use in thermal power systems but are extensively employed for gas turbine applications, is that
of the super alloys, briefly mentioned above. They have outstanding strength, creep resistance
and oxidation resistance at high temperature. Chromium concentrations around 20 weight
percent provide their oxidation resistance, which may be enhanced by molybdenum at concen-
trations up to 10 percent in some cases. Grain-boundary properties, which affect creep, are
improved by small additions of metals such as niobium, zirconium or hafnium while strength is
imparted by additions of aluminium and titanium. The last two metals, usually present at levels
of a fraction of a percent, strengthen the FCC crystal structure by precipitating with nickel as an
intermetallic compound called gamma prime (’). This precipitation hardening or age hardening
process is carried out at temperatures approaching 900°C on material that has been quenched
from solution treatment in a high-temperature region where all the constituents are in solid
solution. A typical alloy of this type is Inconel X750, as produced by the International Nickel
company, which contains 70 weight percent nickel, up to 17 percent chromium, up to 9 percent
iron, about 1 percent niobium, 2.25 - 2.75 percent titanium and 0.4 - 1.0 percent aluminium.
Alloy X750 in wire form has been used to form toroidal spacers to separate the hot pressure
tubes from the surrounding cool calandria tubes. The alloy loses practically all its ductility under
neutron irradiation but maintains sufficient strength to carry the necessary loadings. Composi-
tions of some nickel alloys commonly used in nuclear power plants are shown in Table 3.
The most widespread application of nickel alloys in the nuclear industry is for the tubing of
steam generators in water-cooled nuclear reactors of the CANDU or pressurized water reactor
(PWR) type. In order to ensure resistance to stress corrosion cracking and general corrosion,
Alloy-400, a nickel-copper alloy based on Monel-400 (63-70 weight percent nickel, 28-34
percent copper, with 1-2 percent iron, as produced by International Nickel), was employed in
early CANDU reactors. The more rigorous operating conditions in later CANDUs prompted a
move to alloys with better resistance to general corrosion in a somewhat oxidizing water
coolant, so Alloy-600 (72 weight percent nickel, 14-17 percent chromium and 6-10 percent iron)
or Alloy-800 (30-35 percent nickel, 19-23 percent chromium with the balance of iron, which is
still classed as a nickel alloy, even though the major constituent is iron) were employed. The last
two materials, respectively based on Inconel and Incoloy as produced by International Nickel,
are used widely in the PWRs, though Alloy-600 is being replaced with Alloy-690 (61 weight
percent nickel, 30 percent chromium and 9 percent iron) as a material more resistant to crack-
ing. It should be noted that the nickel-chromium alloys used for steam generator tubing im-
prove their resistance to cracking when they are thermally treated to relieve the stresses
imparted during tube forming.
Table 3. Compositions of commonly used nickel alloys in nuclear power plants.

designation C- Fe Cr Ni Mn Mo Si P Cu Ti
max
600 N06600 0.15 6-10 14-17 Bal. < 1.0  < 0.5  < 0.5 
690 N06690 0.05 7-11 27-31 Bal. < 0.5 8-10 <0.5 <0.015 < 0.05 
X750 N07750 0.08 5-9 14-17 Bal. < 1.0 2.8- < 0.5  < 0.5 2.25-
3.3 2.75
800 NG N08800 < Bal. 19-23 30-35 < 1.5  < 1.0  < 0.75 0.15-
0.02 0.6
2.2.4 Copper Alloys
Copper has an important engineering role in electrical equipment and conductors, stemming
from its high electrical conductivity (second only to that of silver, which is about 10 percent
higher). In addition, its high thermal conductivity, good mechanical properties and corrosion
resistance in many environments make it an important material for heat exchangers in the
power industry and elsewhere.
The pure metal melts at 1082C and at all temperatures below this it has an FCC crystal struc-
ture as shown in Figure 1. It is readily worked and machined and has good resistance to atmos-
pheric corrosion and corrosion in clean waters. For heat exchanger applications, however, it is
alloyed most commonly with zinc - to form the brasses - or with nickel - to form the cupronick-
els. The alloys are harder and stronger than elemental copper and offer good resistance to
aqueous corrosion - particularly in marine environments.
Zinc itself has an HCP crystal structure and it dissolves in copper to a maximum of about 37
weight percent at 500C without changing the structure from FCC. The resulting solid solutions
are in the α phase, which is the basis of the α-brasses. Zinc concentrations above the 37 weight
percent limit precipitate a second phase, the BCC β phase, which becomes the sole equilibrium
phase between about 46 and 50 percent zinc. Above this limit another phase, the phase,
precipitates but forms brittle alloys that are of little commercial utility.
The most widely used brass in the power industry is based on the single-phase 70-30 alloy (70
weight percent copper, 30 percent zinc), since this has the optimum combination of mechanical
properties and corrosion resistance and is not subject to the problems of reduced ductility and
difficulty of working experienced by the two-phase α-β or the single-phase β brasses. While the
α-brasses are easily formed into tubing etc., they tend to lose ductility with increasing cold work
and must be annealed at high temperature to avoid stress corrosion cracking (known as season
cracking) in service. Brasses are also susceptible in aqueous environments to a type of corrosion
called dezincification, in which the alloy is transformed in its original shape to porous copper
having little rigidity or strength. This phenomenon is stifled by as little as 0.04 weight percent
arsenic, which is a standard additive to α-brasses. A common material for heat exchangers and
condensers is Admiralty brass, having 71 weight percent copper, 28 percent zinc, 0.02 to 0.06
percent arsenic and small amounts of lead and iron. Antimony and phosphorus are similar to
arsenic in their corrosion inhibition of brass.
Nickel has a similar FCC crystal structure to that of copper, and the two elements are completely
soluble in each other in all proportions. The resulting α-phase solid solutions therefore have
melting points between that of pure copper (1082C) and that of pure nickel (1452C). The
cupronickels (so-called when the copper is the major constituent) are tough, readily worked and
have a high corrosion resistance. Alloys based on 70 weight percent copper and 30 percent
nickel, the 70-30 cupronickels, find widespread use in heat exchangers for sea water service,
when additions of iron and manganese to about 2 percent provide extra resistance to erosion-
corrosion. The 90-10 cupronickels have better thermal conductivity than the 70-30 alloy and are
more resistant to aqueous corrosion at higher temperatures, so they have been commonly used
in feedwater heaters in steam-raising systems.
2.2.5 Titanium Alloys
Pure titanium has an HCP crystal structure (the α phase) at temperatures up to 884C, where it
transforms to the BCC β phase, which is stable up to the melting point of 1677C. Alloying
elements stabilize one phase or the other. Aluminum, carbon, oxygen and nitrogen, for exam-
ple, stabilize the α phase and raise the α-β transition temperature, whereas copper, chromium,
iron, molybdenum and vanadium lower the transition temperature and may even stabilize the β
phase at room temperature.
The principal titanium alloys are therefore of three types: alpha or near-alpha, alpha-beta and
beta. Their high strength-to-weight ratio (especially of the alpha or near-alpha and alpha-beta
alloys) has put them in high demand in the aerospace industry, while in the power industry the
exceptional resistance to corrosion in a wide range of corrosive waters - particularly sea water -
has made titanium an important material for condenser tubes and plate heat exchangers.
Titanium is actually a very reactive metal and must be melted and machined with care. Its
resistance to corrosion is afforded by the thin, protective layer of oxide (TiO2) that spontaneous-
ly forms in air to produce an effective barrier against the environment, thereby putting titanium
in the class of passive metals. The rather specialized metal refining and component manufactur-
ing process for titanium tubing at one time made the material an expensive proposition for
condensers. The growing demand for titanium products, however, has had the effect of lower-
ing the price relative to competitive materials - especially the copper alloys and steels. Also, the
development of seam-welded tubing suitable for condenser service and the fact that no thick-
ness allowance need be made for corrosion on the tube wall makes titanium an attractive
alternative.
The titanium used mostly in condenser applications is the commercially pure material, which is
therefore an alpha-phase material. It exists in five grades distinguished by their impurity
content. Thus, the maximum amounts of carbon, nitrogen, hydrogen, iron and oxygen are
specified for each grade at levels of the order of 0.01 to 0.1 weight percent, and each grade has
particular mechanical properties. Titanium condenser tubing is relatively immune to corrosion
in salt or brackish water and in all types of polluted water. Marine organisms stick to its smooth
surface with some difficulty and surfaces that do foul or silt are not susceptible to pitting or
crevice corrosion. On the condensing steam side of the tubes used for power plant condensers,
there is no effect on titanium from common gases such as ammonia, carbon dioxide or oxygen.
In fact, the main concern with the use of titanium for condenser tubes is that less noble compo-
nents of the condenser may suffer galvanically-accelerated corrosion and must be protected
either cathodically or by coating or both. It must be borne in mind that in aqueous service the
metal absorbs hydrogen readily, so that its exposure as the cathodic component of a metallic
couple can promote degradation as the titanium-hydride is formed and the material becomes
embrittled.
2.2.6 Concrete
Concrete is an inorganic composite consisting of a coarse aggregate of gravel, crushed rock or
slag and a fine aggregate of sand, all held together with a cement. Angular particles or rock
fragments are preferred to rounded ones for the aggregate because they tend to interlock and
produce a more bonded and therefore stronger structure, although they have more surface
defects that can initiate cracks or voids. The larger the aggregate the better in order to minimize
the defects, bearing in mind that the size of the particles or fragments should be no more than
~20% of the structure’s thickness. The aggregate material can be tailored to the requirements of
the structure. Lightweight concrete incorporating steel-making slag as aggregate, for instance, is
a better thermal insulator than one incorporating normal silica-based rock, and heavy concrete
incorporating dense mineral aggregate such as ilmenite, barytes or magnetite or even metal
shot is used for radiation shielding. Ideally, the sand is a silica-based mineral with particle size
0.1 – 1.0 mm. It partially fills the spaces between the coarse aggregate and reduces the porosity
of the final concrete, in turn reducing the tendency to disintegrate during alternate freeze-thaw
cycles. The cement is a mixture of fine mineral particles that sets and hardens after being made
into a paste with water. The most common type of cement is Portland cement, named after the
cliffs at Portland in England, which the hardened material resembles. Portland cement is a type
of hydraulic cement, i.e., one that hardens upon reaction with water. It is a mixture of calcium
silicate (conveniently designated as 60-65% CaO and 20-25% SiO2), iron oxide and alumina (7-
12%). Adding water hydrates the mixture to produce a solid gel, releasing heat as it does so. The
main chemical reactions can be generalized to:
2CaO × SiO2 + xH2O ®Ca2 SiO4 .xH2O+heat (1)

3CaO× Al2O3 +6H2O ®Ca3 Al2 (OH)12 +heat (2)
3CaO+ SiO2 +( x +1)H2O ®Ca2 SiO4 .xH2O+Ca(OH)2 +heat (3)
The hardened cement binds the aggregate particles together, so there must be enough cement
added to coat all of the particles completely if a firm composite is to be obtained. A typical
cement proportion is 15 vol% of total solids. Both the time to harden and the strength of the
final concrete depend on the composition of the initial mixture; a rapid-setting concrete gener-
ally is weaker than a slow-setting one. Complete curing is attained after 28 days or so, when
concrete reaches its maximum compressive strength.
Like other ceramic-based materials, concrete is strong in compression but weak in tension, so
for most construction purposes it has to be strengthened. Reinforced concrete has steel rods
(“rebars”), wires or mesh introduced into the structure to take the tensile loads, while the
concrete supports the parts under compression. Such a combination is effective for components
undergoing bending forces. In prestressed concrete the reinforcing steel is pulled in tension
between an anchor and a jack while the concrete is poured and cured. After the concrete has
set, the external tension on the steel is removed and as it relaxes the forces are transferred to
the surrounding restraining concrete, placing it in compression. The component can now resist
higher tensile or bending forces. Post-stressed concrete also has induced compressive stresses,
but they are imposed by running steel cables or rods through tubes set in the structure and
pulling and maintaining them in tension after the concrete has hardened. Reinforced concrete
has many applications in nuclear reactor construction – building foundations, primary contain-
ment and shielding, supports for large components such as steam generators, etc. Concrete
containment buildings are pre- or post-stressed, and for the LWRs are lined with steel plate.
Existing CANDU containment buildings are lined with epoxy.
Concrete structures undergo various modes of degradation. Rebar in civil installations, typically
bridges and road overpasses, corrodes in contact with aqueous solutions such as road salt and
in time can cause serious deterioration of the structure. Pre- and post-stressing steel can also
rust and lose strength and over time the compressive stresses may relax. The concrete itself may
be attacked; chlorides, nitrates and sulphates react with the cement and cycles of freezing and
thawing can cause cracking. An insidious mode of degradation for concretes with aggregate
containing silica in a non-crystalline or glassy form is the co-called alkali-aggregate reaction
(AAR). This is an internal process whereby alkaline constituents of the cement react over time
with the silica to form a hydrated calcium silicate. This compound has a larger volume than its
precursors and the resultant swelling causes the structure gradually to spall and crack. The
hydro-electric dam at Mactaquac on the St. John river in New Brunswick has a particularly
pernicious form of this degradation. The concrete swelling puts such severe compressive forces
on the structure, which is firmly anchored between rock abutments, that the stresses have to be
relieved periodically by sawing vertically through the dam from top to bottom – a process that
can continue only for about another decade. The plinth supporting the turbine at the Gentilly-2
CANDU is cracking because of this form of degradation. Concrete used as radiation shielding
around a nuclear reactor relies on its water content to maintain not only its strength but also its
effectiveness for stopping neutrons. Over time, the absorption of radiation and the resulting
heating can lead to dehydration, involving water both bound chemically with hydrous minerals
and free as excess molecules. To minimize these losses, reactor shielding structures are conti-
nuously cooled, often with cooling water pipes embedded in the structure.
2.2.7 Polymers
Polymers for industrial or domestic consumption are generally called plastics. Because of their
ease of manufacture and formation, their physical and chemical properties and their low cost
they have a huge number of applications. The transparency of several types makes them useful
as substitutes for glass whilst the elasticity of others puts them in the class of elastomers. They
can be produced as homogeneous material or as both the matrix and the fibre reinforcement
for composites. In reactor systems they are used for electrical insulation, coatings, seals, etc.
and occur in a myriad of common articles such as textiles, containers, tools, etc.
Most polymers, and the ones discussed here, are composed of “giant” molecules consisting of
chains of simpler units bonded together. They are usually organic carbon-based compounds,
although some are inorganic, based on silicone Si-O units. Molecular weights can range from
104 to 106 g/mol or higher. A linear polymer has its molecules as single strands in a tangled
arrangement, rather like a portion of spaghetti. In a branched polymer, each “spaghetti” strand
has one or more shorter strands branching off. Linear or branched polymers can be cross-linked,
whereby atoms in individual strands are strongly bonded covalently to atoms in other strands,
either directly or via linking atoms.
Thermoplastics are composed of linear or branched molecular strands that are weakly con-
nected to each other by van der Waals bonds and that can be partially separated or untangled
by applying tensile stress. Thermoplastics therefore are ductile and deform plastically. When
heated they at first soften and then melt, making them easily moulded into useful articles and
readily recycled. Polyethylene is a common example of a thermoplastic. By contrast, thermoset-
ting plastics have their linear or branched molecular strands strongly bonded covalently to each
other in a rigid, three-dimensional network. They are therefore stronger than thermoplastics but
more brittle. On heating, they do not melt but decompose, making them difficult to recycle.
Polyurethanes are an example of a thermosetting plastic. Elastomers are thermoplastic or lightly
cross-linked polymers that have their molecular strands in a folded or coiled configuration.
Under stress the molecules deform reversibly, rather like a spring, so that the original form is
regained when the stress is relaxed. Natural rubber is the prime example of an elastomer which
can be stretched elastically to >200%. The soft, thermoplastic latex from the rubber tree is
made into useful rubber by “vulcanising”, involving reacting it with sulphur to cross-link the
highly folded molecular strands. Depending on the degree of vulcanisation, a tougher and more
elastic material is formed, one that does not soften and become sticky when heated to mod-
erate temperatures. Low (1 – 3%) sulphur contents form soft commercial rubber, while high (23-
35%) contents form hard rubber or ebonite.
Polymers are severely degraded by radiation; the familiar embrittlement of polyethylene by

prolonged exposure to sunlight is a mundane example. In nuclear reactors, the deterioration of
electrical insulation in radiation fields, affecting cable runs and instrument connections, is a
burgeoning problem, necessitating frequent inspections and often replacements as plastic
components have deteriorated, sometimes to the extent of cracking and disintegrating in-situ.
The reactor vault area is especially vulnerable, and when reactor control cabling is affected the
deterioration becomes a safety issue. Ion-exchange resins (see Section 2.3 in Chapter 15) in
clean-up systems for active circuits such as the moderator (in CANDUs) and the primary coolant
are based on polystyrene and are subject to radiation damage as they remove radionuclides
from the water. This has been a factor in the decision not to regenerate active resins. Both
natural and synthetic rubbers have their tensile strengths reduced but are affected by radiation
in different ways. For example, natural rubber is somewhat more resistant to radiation than
butyl rubber but is hardened while butyl rubber is softened. Consequently, diaphragms and O-
rings made of either material in control valves are replaced frequently to avoid problems arising
from radiation-induced deterioration.
The damage mechanism is the radiation-induced rupture of the chemical bonds holding the
polymer together, and the subsequent re-forming of the bonds to create an altered structure.
Depending on the material, the degree of cross-linking can be increased and polymerization
increased or reduced, oxidation in air can be promoted and deterioration with gas evolution
may occur. Serious changes in properties such as conductivity, strength, hardness, ductility, etc.
result. Table 4 indicates the severity of damage of several common polymers caused by radia-
tion.
Table 4. Severity of radiation-induced damage of common polymers
polyethylene
polypropylene
nitrile rubber
natural rubber
polystyrene
plexiglass
mylar
cellulose
PTFE
PVC
neoprene
nylon
10-3 10-2 10-1 100 101 102
Absorbed Dose MGy
minor moderate severe
It is noteworthy that polytetrafluoroethylene (PTFE – or the DuPont trade name Teflon®), a most
useful product in many domestic and industrial applications because of its non-stick and low
friction properties, is the most severely affected by radiation in the list in Table 4 (in any case, its
use in reactor systems is proscribed because it tends to release fluorine species as it degrades,
and fluorine is particularly damaging to zirconium and its alloys).
3 CANDU Process Systems and their Materials

3.1 Major Process Systems
Diagrams of the various systems are presented in other chapters. The materials of construction
of those systems are listed here. Their operating experience has been detailed by Tapping
(2008).
3.1.1 Primary Heat Transport System

The primary heat transport system consists mainly of the in-core fuel channels connected to the
steam generators by a system of feeder pipes and headers. It has a purification system for the
heavy water (isotopic purity ≥95.5wt%) and is connected to a pressurizer.
The fuel channel components in contact with the heavy water primary coolant comprise the
pressure tubes of Zr-2.5%Nb, the end-fittings of type 403 stainless steel, the liner tubes of type
410 stainless steel, the shield plugs of ductile iron and the seal plugs of Cr-Fe stainless steel. The
fuel bundles are of elements sheathed with Zircaloy-4 that are separated by spacers of Zircaloy-
4 that are resistance-welded to the end-plates, which are also made of Zircaloy-4. Bundles sit
inside the pressure tubes on bearing pads of Zircaloy-4 which, like the spacers, are attached to
the elements with a braze of Zr-Be.
The feeders and headers are made of carbon steel. Originally, when a low cobalt content was
specified to minimize radiation-field build-up around the PHT circuit, a low-chromium material
(A106 grade B with ~0.02wt%Cr) was selected. As described in Section 4.2 of this chapter and in
Section 3.2 of Chapter 15, this gave rise to flow-accelerated corrosion (FAC) of the feeders at the
reactor outlet. Replacement outlet feeders are specified as A106C with a chromium level of
about 0.3wt%Cr, which, based upon extensive laboratory testing, should be enough to extend
the life of the outlet feeders by about 50% and significantly reduce system fouling from trans-
ported corrosion products. A degradation phenomenon in some outlet feeders before replace-
ment was a type of cracking, which occurred only at the Point Lepreau CANDU-6. Incidences of
feeder cracking at other CANDUs have been extremely low – confined to one repaired field weld
– making the Point Lepreau experience unique and rather mysterious. By 2005, nine feeders had
been removed because of deep cracks from the inside surfaces, in some cases they were
through-wall (Slade and Gendron, 2005), and a total of 12 feeders had been removed and
replaced up to the time of the refurbishment shut-down that began in 2008. Cracks were all in
the first and/or second tight-radius bends downstream of the fuel channel outlet, where
residual stresses were high and where FAC was prevalent. Detailed examination revealed
extensive shallow cracks on the outside surfaces also, and these outside-surface cracks were
found to be widespread in feeder bends across the reactor face. The cracking mechanism was
not conclusively identified before Point Lepreau was shut down for the refurbishment, which
included replacement of the feeders with the A106C material mentioned above and with stress-
relieved bends. However, it was postulated that the inside-surface cracks possibly originated
from stress-corrosion cracking caused by slightly oxidizing conditions at the reactor outlet due to
insufficient suppression of radiolysis and exacerbated at reactor start-up, especially if air ingress
had occurred. The outside-surface cracking was possibly low-temperature creep cracking,
facilitated by atomic deuterium diffusing through the metal lattice due to FAC (Slade and
Gendron, 2005).
The purification is a feed-and-bleed system taken off the connection between the coolant loops
or, at Bruce A (which has only one primary loop) directly off the reactor headers. It has heat
exchangers with tubes of nickel alloy (e.g., Alloy-600 or Alloy-800) and vessels such as the ion-
exchange columns of 300-series stainless steel. The pressurizer, also taken off the connection
between the coolant loops, is a carbon-steel vessel with immersion heaters clad with nickel
alloy.
The steam generators are tubed with nickel alloy – Alloy-400 at Pickering, Alloy-600 at Bruce and
Alloy-800 at the CANDU 6s and Darlington and the refurbished units at Bruce (Units 1 & 2). The
replacement units at Bruce, like all the other CANDU steam generators except the original
designs at Bruce A, are of the “light-bulb” design (see Figure 17 in Chapter 8) with the steam
separator and dryer separate for each unit above the tube bundle (each original Bruce A design
has four steam generators connected to an integral steam drum at each end of the reactor). The
tube sheets in the primary heads are clad with the tube material, but the heads themselves are
carbon steel. The divider plates in the steam generator heads have been replaced with stainless
steel to eliminate FAC. All the steam generators except those at Bruce have integral preheaters,
but at Bruce the preheaters are separate heat exchangers of carbon steel with tubes of Alloy-
600 and tube sheets in contact with the primary coolant clad with the same material (the Bruce
preheaters remove heat from the primary coolant serving the inner core fuel channels, which
run at a higher power than the outer core ones).
3.1.2 Secondary Heat Transport System

The secondary coolant system consists of the shell side of the steam generators, fed with light-
water coolant from the condenser and feed water train and supplying steam to the turbine and
thence back to the condenser. CANDUs have been following a policy of eliminating copper alloys
from the circuit to avoid corrosion problems from carry-over to the steam generators. The only
components remaining with copper alloys are brass-tubed condensers at Gentilly 2 (currently
shut down and due for decommissioning), as listed below, although it is of note that Pickering
changed its condensers from brass to stainless steel because of release of copper to Lake
Ontario.
As mentioned above, steam generator tubing is Alloy-400 at Pickering, Alloy-600 in the original
units at Bruce – Alloy-800 in replacement units and at Darlington and the CANDU 6s. The tube
supports at Pickering A are lattice bars of carbon steel and at Pickering B are broached plates of
carbon steel; at Bruce the original supports are broached plates of carbon steel but, like Darling-
ton and the CANDU 6s (except Embalse, Point Lepreau and Wolsong 1), the replacement units
have lattice bars of type 410 stainless steel. Embalse has broached plates of carbon steel, Point
Lepreau has broached plates of type 410 stainless steel and Wolsong 1 has lattice bars of Alloy
600. Steam generator shells, shrouds, etc. are of low-alloy/carbon steel.
The feedwater and steam cycle piping is generally carbon steel, although some regions of the
system susceptible to flow-accelerated corrosion, such as piping carrying two-phase steam-
water mixtures or highly-turbulent feedwater, may have alloy steel or even stainless steel.
All CANDUs except Embalse have a deaerator in the feed water train. These, the high-pressure
feedwater heaters and the moisture-separator/reheater in the steam circuit are made mostly of
carbon steel, although stainless steel may be employed for some reheater tubes. The low-
pressure feedwater heaters are made of stainless steel.
The condensers at Point Lepreau and the Qinshan CANDUs are tubed with titanium since they
draw from sea water while the Embalse plant continues to use Admiralty brass (with stainless
steel for the outer tubes of the bundles). The remaining CANDUs use Type 304 stainless steel as
their condenser tubes, Bruce units 1 & 2 being the most recently converted during their refur-
bishment project that was completed in 2012.
3.1.3 Moderator
The moderator system is the piping and components containing and processing the heavy water
moderator (isotopic purity ≥99.75wt%). It includes the cover gas system containing the helium
cover gas and maintaining its content of deuterium and oxygen (formed by heavy-water radioly-
sis in the reactor) at safe levels of <2wt% via a catalytic recombiner.
All reactors have a calandria vessel made of type 304L stainless steel and calandria tubes made
of Zircaloy-2. Most of the rest of the circuits (including the heavy water expansion tank) are of
stainless steel, except the heat exchangers, which are nickel alloy – primarily Alloy 800.
3.2 Ancillary Process Systems

3.2.1 Condenser Cooling Water
CANDU condensers are cooled with whatever natural water source is available – sea water at
coastal sites or fresh/brackish water at riverside or lakeside sites. Water is drawn into the system
through a steel screen arrangement to remove coarse debris, pumped through a circuit of
concrete ductwork and steel piping to the condenser, then discharged back to the source. No
CANDU to date has employed cooling towers to enable recirculation of condenser cooling water.
The condenser tubing materials have already been listed above. Like the condenser shells, the
inlet and outlet “water boxes” are of carbon steel and they may be cathodically protected from
galvanic corrosion caused by contact with tube sheets of more noble metals.
3.2.2 Recirculating Cooling Water (RCW)

This system of purified water kept usually at high pH (typically adjusted using the AVT strategy of
the steam cycle) and low oxygen with hydrazine additions cools components such as the mod-
erator heat exchangers. It comprises its own heat exchangers of nickel alloy – normally Alloy 800
– and carbon steel piping fed directly from the local water source.
3.2.3 Shield Tank

The shield tank encloses the calandria circumference, providing shielding and cooling for the
reactor. In the early reactors it is a water-cooled carbon steel vessel but in the CANDU 6 it is the
concrete vault filled with water. The end shields are water-cooled tanks filled with ⅜”- ½”
diameter carbon-steel balls that provide shut-down biological shielding for the reactor faces and
support for the fuel channel assemblies. They are made of type 304L stainless steel and have
lattice tubes (to accommodate the fuel channel end-fittings) of the same material. Like the main
shield tank, they are filled with ⅜”- ½” diameter carbon-steel balls. Note that the Pickering
reactors have cooled steel slabs as end shields rather than water-cooled tanks.
The purified water, maintained at high pH with lithium hydroxide, is circulated through heat
exchangers tubed with nickel alloy or even copper alloy, where it is cooled by plant service water
(RCW).
3.2.4 Spent Fuel Bay

The spent fuel bay is the swimming-pool arrangement for accepting and storing spent fuel under
water directly from the reactor. About ten-year’s accumulation (at ~80% capacity factor of the
plant output) of fuel bundles can be stored with about 4m of water above. The water cools the
bundles and provides biological shielding. There is an intermediate bay for handling and tempo-
rary storage. The construction is reinforced concrete lined with epoxy-coated fiberglass and with
a stainless steel floor.
3.2.5 Emergency Cooling

The emergency core cooling system (ECC) is a light-water system designed to supply cooling
water to the reactor core via the primary coolant system emergency injection. It is triggered by a
loss in pressure following a LOCA (loss-of-coolant accident) and for later reactors with individual
containment buildings, is supplied initially from the ECC water tanks which are located outside
the containment. For high-pressure operation as occurs with a small leak, the water is injected
into the primary coolant headers by automatically pressurizing the ECC tanks with nitrogen (the
loop is automatically isolated from the other one). As the system pressure drops, the ECC pumps
start and stored water is then injected into the headers from the dousing tank, which is below
the roof slab. During these phases, the heat sink is the steam generators. Finally, when the
dousing tank is empty and the system pressure has dropped even further, collected water is
pumped into the system from the sump via the ECC heat exchanger, which is cooled with service
water. This phase is designed to continue indefinitely. The multi-unit stations with a vacuum
building have the ECC storage as the dousing tank below the roof slab of the vacuum building.
The dousing tank below the roof slab is made of reinforced concrete and the piping system of
carbon steel and stainless steel. In the later reactors the shell-and-tube ECC heat exchangers to
reject heat to the outside are alloy-tubed but early designs used titanium plate-type heat
exchangers.
3.3 Supporting Structures and Components
3.3.1 Containment
In the four-unit CANDUs at Darlington and Bruce A and B, the reactor of each unit is contained
within a reinforced concrete vault, which extends to shielded rooms above containing the steam
generators, heat transport pumps and reactivity drive mechanisms. Main steam piping exits the
containment for the turbine building. Openings in the floor of the vault allow entry of the
fuelling machine, which serves all four units via a duct running below grade for the length of the
station and connecting to a fuel handling facility. The overall building for each unit has a flat,
steel roof.
The four-unit CANDUs at Pickering A and B, like the CANDU 6 reactors, each have the reactor,
steam generators, etc., in a separate containment building of pre-stressed, post-tensioned
concrete with a domed roof. The building is lined with epoxy, but future designs could incorpo-
rate a steel liner. The main steam pipes exit the containment directly for the turbine building.
3.3.2 Vacuum Building (Multi-unit Plants)

Bruce A and B and Darlington each have the four reactor containments separately connected to
a normally sealed building that is kept at about 7kPa absolute pressure. At Pickering, the eight
units at the A and B sites are connected to one such vacuum building. In the event of an over-
pressure in a reactor containment, isolation valves connecting that reactor with a concrete duct
running to the vacuum building open and the atmosphere of the containment is vented. A spray
from the ECC storage in the vacuum building roof condenses any vapour.
The vacuum building is constructed of reinforced concrete. It is supported inside with an ar-
rangement of columns and beams that also hold up the ECC storage tank and the vacuum
chambers.
3.3.3 Fuelling Machine

Each CANDU reactor is served by two fuelling machines, with one machine connecting to each
end of a fuel channel that is to be refueled. A machine has a rotating magazine to hold fuel and
channel seals and plugs, a snout assembly to lock onto the channel and a ram assembly to
remove, store and replace channel seals, plugs and fuel bundles. Figures 10 – 12 in Chapter 8
show the arrangement of the fuelling machine.
The machines simultaneously clamp onto the end-fittings of a channel, unlock, remove and
store the seal plugs, remove and store the shield plugs then push new fuel bundles into the
pressure tube from one end and remove and store spent fuel bundles from the other end,
replace the shield plugs and finally replace and lock the seal plugs. During the operation the
machine is in contact with the primary coolant at pressure and sustains a flow of heavy water
from the PHT supply tanks. The seal plug sealing faces must be maintained in good condition to
prevent expensive heavy water leaks.
Fuelling machines are large pieces of equipment comprising many components that remotely
perform a complex series of tasks. They are made of many engineering materials. Of note are
the mechanical bearings that support the rotating magazine, which have traditionally been
made of the extremely hard-wearing cobalt alloy, Stellite®. This has been responsible for a
significant component of out-core radiation fields, since even very small releases of cobalt to the
primary coolant during a refuelling operation may give rise to the production of 60Co in the
reactor core, and this distributes around the circuit and contaminates out-core components,
particularly at the reactor faces and in the steam generator heads. Recent research is identifying
alternative materials for the bearings such as silicon nitride (see for example Qiu, L., 2013).
Additionally, activation products of antimony (primarily Sb-124) originate from bearings in the
heat transport pumps and as impurities in steel and may contribute significantly to the overall
activity levels throughout the primary heat transport circuit.
3.3.4 Fuel
The fuel materials in contact with the primary coolant, viz. the element sheaths, end plates,
bearing pads, spacers and the braze attaching the appendages to the sheath have already been
described. The fuel pellets themselves within a sheath are of high-density, sintered UO2 of
natural isotopic content (0.7% 235U) and the inside surface of the sheath is coated with a thin
layer of graphite to reduce sheath-pellet interactions (the CANLUB system). The standard bundle
is of 37 similar elements; however, more-efficient designs such as the CANFLEX® bundle with 43
elements of different diameters have been proposed; thus, for future applications, the Ad-
vanced CANDU Reactor employs low-enriched uranium (LEU) fuel of UO2 with 3-4% 235U in a
CANFLEX®-type bundle with a central element containing a neutron absorber. Note that the
flexibility of the CANDU system makes it suitable for a wide variety of fuel materials, and trials
with thorium, recovered uranium (from reprocessing of LWR-type fuel) and plutonium mixed-
oxide (with uranium) have been carried out successfully.
3.3.5 Reactor Control Mechanisms

Liquid zone controllers are vertical thimble tubes of type 304 stainless steel attached vertically
to the top of the calandria with zone control tubes of Zircaloy extending downwards into the
moderator between the calandria tubes. They control reactivity during normal operation. Six of
them per reactor contain among them fourteen compartments of circulating light water as the
absorber, and the amount of absorber can be controlled by adjusting an overpressure of helium
in each compartment and the inlet flow rate.
There are mechanical adjuster rods (21 in the CANDU 6s), also inserted vertically into the
moderator, to act as neutron absorbers. They are tubes of stainless steel or Zircaloy containing
“shim rods” of steel/cadmium or cobalt that are normally fully inserted to flatten the flux profile
but can be withdrawn mechanically via motor-driven winches (sheaves) and cables as override
for xenon poisoning following reactor shutdown or a power reduction.
Shut-off rods and mechanical control absorbers are of similar design, penetrating the calandria
vertically between the calandria tubes in similar fashion to the adjuster rods. Each employs a
cable, winch and clutch to hoist, hold or release an absorber tube containing cadmium as a
neutron absorber. They are normally parked in thimbles above the calandria but can be driven
vertically downwards into the core to reduce reactivity through perforated guide tubes of
Zircaloy that are anchored at the bottom of the calandria. Shutdown System #1 comprises 28
such assemblies arranged throughout the reactor; they release their rods by de-energising the
clutch and drive them into the core by spring-assisted gravity to achieve rapid power decrease.
Four other assemblies can be used for reactor control to supplement the liquid zone controllers.
They are normally parked out of the core but are driven in at variable speeds as required to
reduce reactor power.
4 Corrosion and Material Degradation in Nuclear Reactor Systems
As described above and elsewhere in this text, a nuclear reactor is a complex system of different
connected materials that must behave optimally in unison to ensure the safe and efficient
operation of the plant. Even under the strictest chemistry control practices and careful plant
operation, the corrosion and subsequent degradation of the plant components are the inevita-
ble consequences of thermodynamics; the metals that make up the pipes, valves, fittings and
vessels all tend to revert back to their thermodynamically stable state – usually an oxide,
haematite (Fe2O3) for example, the basis of rust on iron and steel. In some cases, the degrada-
tion of system components is exacerbated by vibration, fatigue or fretting or by the interaction
with high-velocity coolant creating flow-accelerated corrosion (FAC); in all cases, component
lifetimes are shortened. Thus, knowledge of the basic forms of corrosion is a prerequisite to
understanding the materials selection and chemistry requirements of CANDUs.
4.1 General Corrosion

Metals that are in every-day use for engineering applications are fundamentally unstable. They
tend to react with their environment to assume a more stable thermodynamic state, which in
the coolant systems of nuclear reactors is normally an oxide. General corrosion is the term for
attack that is spread more or less uniformly over the entire component surface; it is driven by an
electrochemical reaction in the case of ionic environments such as molten salts or aqueous
coolants, or by a chemical reaction in the case of non-ionic coolants such as gases like carbon
dioxide. (Note, however, that any oxidation is strictly speaking electrochemical, since it is a
chemical reaction involving the transfer of electrons.) The metal thins gradually and may fail if
left without being inspected for any length of time, although the rates of such corrosion are low
and are generally predictable from tabulated data. Components are designed with adequate
thickness (incorporating a “corrosion allowance”) that will last for the desired lifetime; failures
are rare.
4.1.1 General corrosion in gas-cooled systems

As an example of a direct chemical oxidation, consider the alloy steel Fe-9%Cr as the boiler-tube
material in Advanced Gas-Cooled Reactors (AGRs). In the CO2 coolant at 580oC this material
corrodes uniformly, forming a duplex film of the oxide Fe-Cr spinel (based on iron chromite,
FeCr2O4) overlaid with the inverse spinel magnetite (Fe3O4). Spinels have a crystal structure in
which the unit cell contains 32 oxide ions in a face-centred cubic arrangement and 24 metal ions
dispersed in the interstices to neutralise the charge. The interstices comprise 64 sites with
tetrahedral symmetry and 32 sites with octahedral symmetry. In so-called normal spinels, the
eight divalent metal cations (Fe2+ in the example of FeCr2O4) are in tetrahedral sites and the 16
trivalent cations (Cr3+ in FeCr2O4) are in octahedral sites. In inverse spinels, the divalent cations
(Fe2+ in Fe3O4) are distributed along with half the trivalent cations (Fe3+ in Fe3O4) in octahedral
sites, while the remaining trivalent cations are in tetrahedral sites. The relative stabilities of the
normal and inverse spinel lattices influence the extent to which magnetite can incorporate Cr3+
ions and chromite can incorporate Fe3+ ions.
The mechanism for the formation of the duplex oxide film on the AGR steel has been postulated
to be the direct reaction of metal with CO2 that diffuses through pores and microchannels in the
growing oxide [Kofstad, 1985]. For example:
Fe  2Cr  2CO2  FeCr2O4  2C (4)
The carbon is distributed through the oxide and may carburise the metal. The development of
the oxide porosity is governed by stresses created at the metal-oxide interface (M-O) by the
difference between the density of the oxide and that of the metal (quantified as the “Pilling-
Bedworth” ratio, which is defined as the ratio of the volume oxide formed to the equivalent
volume of metal consumed; it is greater than unity for the common transition metals). We note,
however, an alternative mechanism [Taylor et al., 1980], whereby corrosion is controlled by the
protective nature of the inner layer; thus, the magnetite forms at the oxide-coolant interface (O-
C) via direct chemical reaction with CO2 of Fe that has migrated through the oxide:
3Fe + 2CO2 ® Fe3O4 + 2C (5)
although the details will involve reactions with diffused ionic species:
Fe 2+ + 2Fe 3+ + 2CO2 + 8e - ® Fe3O4 + 2C (6)
The kinetics of general corrosion governed by the increasing protection afforded by a conti-
nuously-thickening oxide film are often described by a parabolic expression, derived from the
principle that diffusion rate, and therefore corrosion rate and oxide film growth rate (dm/dt), is
inversely proportional to film thickness (m):
dm k
= (7)
dt m
which integrates to:
m2 -mo2 = kt (8)
where mo is the initial film thickness. Modifications of (8) such as (9) are occasionally quoted:
m = mo +kt 0.5 (9)
CANDU reactors have two main gas systems, although several ancillary systems also use inert
cover gases to help prevent corrosion and maintain safety envelopes. The two main gas systems
include the moderator cover gas system (already described) and the annulus gas system (AGS).
The AGS circulates CO2 through the annulus between the calandria tube and its companion
pressure tube. The system is continually monitored for moisture by dew point measurements as
a warning of a leak from the PHT system or from the calandria end shield joint. Deuterium
produced from the corrosion of the pressure tube on the PHT-side, diffuses into the AGS
forming heavy water vapour when combined with oxygen present in the CO2 so the system
needs to be regularly purged to reduce the moisture content. The oxygen added to the CO2 in
the AGS also performs an important role of oxidizing the outside of the pressure tube. By
oxidizing the outside pressure tube surface it helps to maintain a barrier to the ingress of
deuterium into the pressure tube and promotes oxide formation on the X750 garter-spring
spacers.
4.1.2 General corrosion in water-cooled systems

Examples of electrochemical reactions are provided by the general corrosion of metals in water.
Thus, carbon steel in contact with high-temperature water develops protective magnetite films
according to the overall scheme:
3Fe + 4H2O ® Fe3O4 + 4H2 (10)
The carbon steel feeders in CANDU primary coolant corrode accordingly, except that the coolant
is D2O rather than H2O and deuterium rather than hydrogen is evolved. Mechanistically, (10)
can be regarded as two half-reactions, the anodic or oxidation process:
3Fe  Fe 2   2Fe 3   8e  (11)

and the cathodic or reduction process:
4H2O+ 8e - ® 4O 2- + 4H2 (12)
The half-reactions occur simultaneously and at the same rate. According to the classic Wagner-
Traud theory [Wagner and Traud, 1938] it is postulated that the anodic sites on the metal,
where metal ions dissolve (reaction (11)), are separate from the cathodic sites, where electrons
are discharged, oxide ions form and hydrogen is released (reaction (12)). The electrons are
transported between the sites and the metal ions and oxide ions interact and precipitate as
magnetite. Clearly, for the oxidation to be uniform, the sites must be mobile across the surface.
In reality, it is understood that in high-temperature water the oxidation at the M-O produces
ferrous ions according to:
3Fe ® 3Fe 2+ +6e - (13)
and that the hydrogen is produced in the accompanying reduction:
6H2O+6e - ® 6OH - +6H (14)
where H is in the form of hydrogen atoms that migrate through the metal. It is important to
note that if dissolved oxygen is introduced into neutral or alkaline water, reaction (14) is re-
placed by the more favourable reaction (15):
1.5O2 + 3H2O+6e - ® 6OH - (15)
and hydrogen production ceases.
The ferrous ions and hydroxide ions from (13) and (14) or (15) interact to form soluble ferrous
hydroxide:
3Fe 2+ +6OH - ® 3Fe(OH)2 (16)
but this oxidises and precipitates as magnetite according to the Schikorr reaction [Schikorr,
1928]:
3Fe(OH)2 ® Fe3O4 + 2H2O+H2 (17)
Again, if oxygen is present, the oxidation of ferrous hydroxide proceeds via the more favourable:
3Fe(OH)2 +0.5O2 ® Fe3O4 + 3H2O (18)
There is only room at the M-O (metal-oxide interface) for about half of the ferrous hydroxide to
precipitate there as inner-layer magnetite; the other half, along with any hydrogen molecules
generated at the M-O by (17), diffuse through the developing magnetite film to the O-C (oxide-
coolant interface). There, if the coolant is reducing and saturated in dissolved iron as in the
carbon steel inlet feeders of CANDUs, it precipitates as outer-layer magnetite according to (17);
the molecular hydrogen is then dispersed in the bulk of the flowing coolant. In the presence of
dissolved oxygen, the outer layer precipitates according to (18) and the magnetite film develops
completely without the evolution of hydrogen. In time, it may itself oxidise to the Fe(III) oxide
haematite:
2Fe3O4 +0.5O2 ® 3Fe2O3 (19)
and if the conditions are oxidising enough haematite may be formed directly without the
intermediate formation of magnetite.
The duplex films of magnetite generally consist of a fine-grained inner layer of particles ten or so
nanometres in size, overlaid with crystallites of sizes up to about ten micrometres. The outer
layer crystallites are generally facetted octahedra, obviously grown from solution (see Figure 5
and 6).
Figure 5. Schematic Cross-Section through Magnetite Film on Carbon Steel [Lister, 2003]
Figure 6. Scanning Electron Micrograph of Magnetite Film formed on Carbon Steel during an
1100-h Exposure in High-Temperature Water [Lister, 2003]
The general corrosion of the higher alloyed materials in widespread use in water reactor sys-
tems, such as the stainless steels and nickel alloys, also leads to the development of duplex
oxide films. The oxide layers are generally thin and compact, conferring a high degree of corro-
sion resistance. The precise compositions of the oxides depend on the coolant chemistry,
particularly the oxidising propensity. In general, it is the inner layers that “passivate” the alloy;
these are based on the normal spinel FeCr2O4, which has a very stable lattice structure and a low
solubility in the chemically reducing coolants of the pressurised water reactors (PWRs) and
CANDUs [Lister, 1993]. Variants of the iron chromite structure contain elements arising from
other alloy constituents such as nickel, viz. NixFe1-xCr2O4, where x depends on the composition
of the alloy; small amounts of other metals such as cobalt may also be incorporated and in the
primary coolant this gives rise to radiation fields from 58Co and 60Co. The outer-oxide layers are
variants of the inverse spinel magnetite and are designated as NixFe3-xO4; again, x depends on
the composition of the underlying alloy but also, since the layer is precipitated from solution, it
depends upon the metals dissolved in the coolant and originating in the rest of the circuit [Cook
et al., 2000]. In PWRs, the composition of the outer layers on the alloys approximates
Ni0.6Fe2.4O4, a variant of nickel ferrite (or trevorite, NiFe2O4). This also approximates the compo-
sition of the particulate matter (called “crud”) that circulates in suspension in the primary
coolant at concentrations of the order of a ppb (part per billion) or less and that forms deposits
on in-core fuel assemblies. In CANDU primary coolants the large surface area of carbon steel
feeders makes magnetite the predominant constituent of crud.
The boiling water reactors (BWRs) traditionally operate under conditions of normal water
chemistry, NWC, which exposes their alloys to high-temperature water that is slightly oxidising.
This favours the formation of soluble Cr(VI) rather than the Cr(III), which is the basis of the
chromite oxides that are widespread in the PWRs. Consequently, the stainless steels that
constitute much of the BWR circuits develop oxides with less FeCr2O4; they also contain the
Fe(III) oxide Fe2O3, haematite. In the move to less-oxidising coolant chemistry in the BWRs,
accomplished by adding hydrogen to the feedwater to create hydrogen water chemistry (HWC),
the oxides on stainless steel approach those found in the PWRs. It is interesting to note that
detailed in-situ studies of corrosion rates on type SA 316L stainless steel indicate that NWC
conditions are somewhat less corrosive than HWC conditions [Ishida and Lister, 2012]. The
corrosion for the first 100 – 150 h under NWC follows logarithmic kinetics whereas under HWC it
follows a joint parabolic-logarithmic law.
4.1.3 Corrosion and degradation of zirconium

In the water-cooled reactors, the major in-core materials in contact with the coolant are alloys
of zirconium. The CANDUs have fuel bundles of the dioxide of natural uranium (i.e., 99.3% 238U,
0.7% 235U as UO2) sheathed in Zircaloy-4 and the bundles are contained in pressure-tubes of Zr-
2.5%Nb. Typical fuel burnups in CANDUs are about 7 MWd/kg and on-power refuelling is
necessary to maintain the core critical. The PWRs traditionally use Zircaloy-4 to clad their fuel
assemblies containing enriched UO2; however, with the higher enrichment of 3-4% 235U in
advanced fuel, which can achieve burnups up to 52 MWd/kg, alloys containing Nb that are more
corrosion- and hydriding-resistant, such as Zirlo© , tend to be employed. The BWR fuel assem-
blies can reach somewhat lower burnups of up to 42 MWd/kg and have their fuel clad with
Zircaloy-2. Both types of LWR (light water reactor – PWR or BWR) have batch refuelling.
The corrosion of zirconium alloys has been postulated as occurring in three stages, as shown in
Figure 5. The stages have been characterized as [Hillner et al., 2000]:
 Pre-transition, when a thin, black, tightly-adherent oxide thickens according to sub-

parabolic (often assumed cubic) kinetics.
 Transitory, apparently corresponding to successive “cubic” periods of decreasing dura-
tion, when the thickening film cracks and becomes porous and new oxide grows at the
M-O.
 Post-transition,
transition, when kinetics become rapid and linear and white or grey oxide is
formed.
As indicated in Figure 7, the transitory period was not recognized in early models, although the
omission has little effect on the general description. The transition occurs at film thicknesses of
2–5 µm.
Figure 7. Schematic diagram of Zircaloy corrosion in high-temperature

temperature water, showing three
regions (dotted line indicates early assumed pre
pre-transition cubic and post--transition linear
kinetics) [Hillner et al, 2000]
Much of the early research leading to our understanding of the cor corrosion
rosion and hydriding of
zirconium and its alloys was conducted by Cox [Cox, 1976]. The basic electrochemical reaction is
the formation of a compact oxide film based on zirconia (ZrO2) via:
Zr + 2H2O ® ZrO2 +H2 (20)
The pre-transition
transition film is adherent and compact zirconia of tetragonal crystal structure that
grows via the solid-state
state diffusion through the oxide of ions/vacancies and d electrons/holes.
Electron diffusion is rate-determining
determining and drives the cathodic reaction by reducing water to
form hydrogen. The hydrogen genera
generated (in overall terms by (20))
)) enters the metal in atomic
form and at high enough concentrations precipitates within the lattice as flakes of hydride of
nominal composition between ZrH and ZrH2, making the material brittle. The post-transition
post
oxide has monoclinic crystal structure and is much less protective than the pre
pre--transition oxide,
hence the accelerated corrosion.
orrosion. Since these processes occur with no oxide dissolution or
release of cations to the coolant, the weight
weight-gain
gain of laboratory specimens is an accurate gauge
of corrosion.
The oxide thicknesses on Zircaloy

Zircaloy-4 cladding on high-burnup
burnup PWR fuel may reach 100 µm or
more but on the advanced alloys are generally less than half that amount. Concentrations of
LiOH of more than ~4 ppm (parts per million) may exacerbate corrosion rate. BWR exposures of
Zircaloy-2-clad
clad fuel induce nodular corrosion, whereby white spots of thick oxide appear on the
underlying black oxide and grow with increasing exposure. The severity of nodular corrosion
depends upon the metallurgy of the alloy and the aggressiveness of the environment, but rarely
causes problems (although spalled oxide has occasionally caused abrasion damage in BWR
control-rod drive mechanisms). So-called “shadow corrosion”, where galvanic effects thicken
ZrO2 films in proximity to components made of non-zirconium alloys such as stainless steel
control-blade handles, is an additional mechanism in the low-conductivity and low-hydrogen
environment of the BWR. BWR fuel is also subject to increased corrosion when the surface
temperature is increased by the deposition of copper-containing crud that impairs the nuclea-
tion of steam bubbles. Such “CILC” (crud-induced localized corrosion) has led to fuel failures.
These mechanisms of corrosion of zirconium alloys are reviewed in detail in the report of
Adamson et al. [2007].
Because of its relatively low burn-up, CANDU fuel sheathing develops a very thin, protective
oxide and failures from coolant-induced general corrosion are almost non-existent. The Zr-
2.5%Nb pressure tubes in CANDUs and other PHWRs also experience low corrosion rates and
their general corrosion over the 20-30-year lifetime in-reactor is acceptable. The oxidation,
however, does force hydrogen (actually, deuterium from the heavy-water coolant) into the
metal lattice, and this is much aggravated by deuterium that migrates into the pressure-tube
ends across the rolled-joint between it and the stainless-steel end fittings, which also experience
general corrosion. Hydriding (deuteriding) can then be a problem if concentrations exceed the
“terminal solid solubility” (TSS), leading to precipitation of hydride “platelets”, brittleness and
possible cracking if excessive stresses are experienced [Simpson et al., 1974]. In fact as men-
tioned earlier, in the past, pressure tubes have failed because of delayed hydride cracking (DHC),
so they must be continually monitored for uptake of hydrogen/deuterium during service [Per-
ryman, 1978].
The Zr-Nb alloy of CANDU pressure tubes is subject to several other forms of degradation
(growth and creep), and these will be life-limiting for the components in one form or another
and are the primary reason a mid-life refurbishment (after approximately 30 years of operation)
is required on CANDU reactors. As described in Section 2.1 above, neutron irradiation causes an
increase in strength and a loss of ductility (or loss of toughness) in Zr-2.5%Nb alloys (as it does in
all metals) but this does not prevent the pressure tubes from performing their role as a pressure
boundary. The neutron irradiation causes additional defects (dislocations) to be produced in the
metal lattice, which move through the metal’s crystal structure either under the action of stress
(known as radiation induced creep) or by the action of the neutron flux alone (radiation induced
growth). The migration of the dislocations results in extension or contraction of the tube in its
three major directions but because of the tube’s texture or predominant grain orientation,
growth is concentrated in the longitudinal direction. Providing sliding bearings on each end
fitting has been demonstrated to accommodate the radiation induced growth and elongation. In
early CANDU designs, channel extension was only sufficient to allow for thermal and pressure-
based expansion. Design changes have substantially increased the channel extension capability.
Fixing one end of the fuel channel now accommodates full-life channel elongation allowing for
extension to occur at the opposite, free end of the channel. Typically, at the midway point
through the reactor core’s operating lifetime, the fuel channel is released at the initially fixed
end and then pushed to its maximum inboard position and fixed at the formerly free end,
allowing for the full scope of the design extension to be used. Fuel channel elongation is great-
est in the high power channels in the centre of the reactor resulting in a dish-shaped profile of
fuel channel extension. Differential elongation could cause feeder pipes at reactor ends to
contact and chafe each other, which must be avoided. Because of the material condition of the
Zr-2 calandria tubes, they do not elongate to the extent of applying significant loads on the end
shields.
The pressure tubes also sag with increased exposure time due to the elongation described
above and from diametral creep. The diametral creep meant that fuel channels in the middle of
the core, where the combination of neutron flux, pressure and temperature are highest, would
suffer from significant portions of the coolant flow by-passing the fuel bundles, reducing the
heat transfer efficiency and leading to reductions in the margins to dry-out. The remedy was
then to derate – reduce the reactor power. Pressure-tube/calandria-tube sag also meant that
the pressure tubes in some reactors approached the calandria tubes closely, and occasionally
may have contacted one of the horizontal reactivity mechanism control tubes spanning the
calandria. The pressure tube and its matching calandria tube are separated by wire wound
toroidal spacers or garter springs that accommodate relative movement between the two. The
pressure tube will sag more between the spacers than will the calandria tube due to the higher
temperature and pressure operation and the weight of the fuel bundles. Should the spacers
move out of position from vibration during operation, there is a risk that the hot pressure tube
touchs the cool calandria tube inducing a thermal gradient or “cold spot” on the wall of the
pressure tube. The will cause any hydrogen/deuterium in the vicinity to diffuse down the
temperature gradient and form a hydride or hydrogen blister at the surface. Above a certain
size, the blister will crack to accommodate the volume change and may initiate a delayed
hydride crack that could penetrate the tube wall. As these degradation problems worsened with
time of operation, pressure-tube rehabilitation programs were undertaken. REFAB (Reposition-
ing End Fitting And Bearing) involved shifting pressure tubes within the core so that they were at
the start of their travel on the bearings to accommodate elongation, as described above. SLAR
(Spacer Location and Repositioning) involved using electromagnetic induction to find the
spacers separating the pressure tubes and calandria tubes and then shift them to their design
positions to counteract sag and avoid subsequent contact. Severely affected fuel channels could
have their tubes replaced, and in the extreme case LSFCR (Large Scale Fuel Channel Replace-
ment) could be undertaken, in which a full core is replaced. This was done during the early
operation at Pickering, when the pressure tubes suffered delayed hydride cracking (DHC)
because of faulty installation; it is an integral part of reactor refurbishment to double the
operating life of the CANDU plant.
4.2 Flow-Accelerated Corrosion (FAC)

Flow-accelerated corrosion (FAC), sometimes called flow-assisted corrosion, was formerly
referred to as erosion-corrosion, but the last term is now usually confined to attack in a corro-
sive medium aggravated by mechanical forces such as those arising from abrasion by solid
particles. FAC is a form of general corrosion, but it is classified on its own here because of its
particularly aggressive nature. Attacking the carbon steel piping of feedwater systems, it is one
of the most widespread causes of shutdown in all kinds of steam-raising plant and on occasions
has led to pipe rupture with accompanying injuries and even deaths of plant personnel. Fatal
accidents due to FAC in condensate systems have occurred in at least two nuclear plants – the
Surry PWR in the US in 1986 and the Mihama 3 PWR in Japan in 2004 – and at fossil-fired plants
such as the Pleasant Prairie coal-fired station in the US in 1995. Unexpected and sometimes
excessive pipe-wall thinning from FAC has also been an endemic problem in the outlet feeders in
the primary coolant systems of the CANDU reactors that were in operation before the two
CANDU 6 models in Qinshan, China, that went into service in 2002 and 2003. Some feeders have
had to be replaced as they reached their minimum wall thickness long before their design
lifetime, Although these replaced feeders represent a small fraction of the total number in
service, the necessity of regular inspections to ensure suitable margins still remains for the pipe
wall thickness; feeder FAC represents a considerable cost to plant operators through planning,
conducting and verifying the field inspections in the restricted, high radiation dose area.
The basic mechanism of FAC in power plants is the dissolution and “wearing” of the normally-
protective film of magnetite that develops on the corroding steel in high-temperature water. As
its title suggests, the phenomenon is exacerbated by fluid flow – specifically, turbulence. As it
proceeds, the attack “sculpts” the metal surface into a characteristic pattern of shallow hollows,
called scallops, the size of which is governed by the turbulence (see Figure 8). The magnetite
dissolves at its interface with the coolant (i.e., at the O-C, by the reverse of Equation 17) and is
replenished by corrosion at the metal-oxide interface (M-O) by reactions described above for
general corrosion (in overall terms by Equation 7). Also as described above, not all the corroded
iron can be accommodated as oxide at the M-O – about half diffuses through the oxide to the O-
C. In an FAC environment the coolant is depleted of dissolved iron, no outer layer of magnetite
can precipitate and the diffused iron is transported to the bulk coolant along with that produced
by the magnetite dissolution. The transport to the bulk is effected by mass transfer, which is
governed by the convective processes in the coolant – i.e., by the turbulence. Since the diffusion
of iron through the oxide film governs the corrosion rate, the thickness of the film attains a
steady-state which then controls the FAC at a constant rate.
Figure 8. Scalloped surface of outlet feeder from Point Lepreau CANDU
This simple view of FAC, in which there are two processes in series – the dissolution of the
magnetite film and the transport of iron to the bulk coolant – has been described by Equation
21 [Berge et al., 1980]:
R=
(
kmkd C s -Cb ) (21)
0.5km +kd
where: R is the FAC rate (gFe/cm2.s) , km is the mass transfer coefficient (cm/s), kd is the magne-
tite dissolution rate constant (cm/s), Cs is the magnetite solubility (gFe/cm3) and Cb is the
concentration of iron in the bulk coolant (gFe/cm3). High alkalinity reduces FAC in feedwater
systems by reducing magnetite dissolution (by reducing Cs and kd) – although at the higher
temperatures in primary coolant systems increasing alkalinity tends to increase magnetite
dissolution and is expected to increase FAC accordingly. The effects of flow and pH in the
mechanism of FAC are shown in Figure 9.
We note that if dissolution controls, kd<< km and Equation 21 reverts to:
(
R = 2kd C s -C b ) (22)
Conversely, if mass transfer controls, km << kd and we have:
(
R = km C s -C b ) (23)
However, mass transfer cannot be the sole mechanism responsible for FAC, since dependences
on flow rate much greater than those conventionally described by empirical mass-transfer
correlations are frequently encountered; moreover, FAC often occurs in circumstances where
the dissolution of the magnetite is much slower than mass transfer and should control the rate,
leading to the untenable situation of the process’s being described by Equation 22, which has no
flow dependence at all. These observations have led to the postulate of a mechanism for FAC
wherein rapidly-flowing coolant that is unsaturated in dissolved iron dissolves the magnetite (as
in the conventional description) but the dissolution loosens the nano-crystals of oxide at the O-C
which are then stripped from the surface (eroded or spalled) by the force of the fluid shear
stress at the pipe wall [Lister and Uchida, 2010]. A schematic view of the overall mechanism,
which also involves some of the reactions of magnetite and hydrogen along with the diffusion of
dissolved iron through the (now single-layered) oxide film as described earlier for general
corrosion, is presented in Figure 10.
(a) after Bignold et al, 1982 (b) after Sturla, 1973

Figure 9. (a) Effect of temperature and flow on FAC at pH25ºC 9.0 with ammonia; (b) Effect of
temperature and pH on the solubility of magnetite.
Figure 10. The overall mechanism of flow-accelerated corrosion.
The mathematical model describing this mechanism was first applied to a CANDU primary
coolant system, for which measured values of kd [Balakrishnan, 1977] were so much less than
km that the conventional explanation of FAC via Equation 21 was clearly inadequate. The FAC in
several outlet feeders at the Point Lepreau CANDU was modelled, leading to predictions of wall
thinning rate around the first couple of bends downstream of the end-fitting that corresponded
very well to plant measurements. At the same time, the oxide spalling terms led to realistic
estimates of crud levels (concentrations of suspended oxide particles) in the primary coolant.
After replacement of the dissolution terms in the model with precipitation terms and insertion
of appropriate coolant flow rates and temperatures, the corrosion and oxide build-up on inlet
feeders were also predicted very well, [Lister and Lang, 2002].
Traces of some transition metals (copper, molybdenum and especially chromium) in the carbon
steel have long been known to reduce FAC rates; concentrations of Cr as low as a fraction of a
percent have been demonstrated to lower FAC rates drastically [Bouchacourt, 1987]. Unfortu-
nately, the existence of a lower limit of Cr content of 0.03 – 0.04%, below which FAC rates are
not affected, has been propounded and is apparently incorporated into some examples of FAC
software. Recent work, however, has demonstrated a reduction of ~60% in FAC rate between
steels containing 0.001% Cr and those containing 0.019% Cr [Lister and Uchida, 2010].
The FAC of CANDU outlet feeders has now been mitigated in replacement piping and in new
reactors by specifying a carbon steel with a higher chromium content [Tapping, 2008]. The
original material had been specified with a low cobalt content to minimise the build-up of out-
core radiation fields from 60Co during reactor operation, and this resulted in a low chromium
content since no Cr concentration was specified; for example, the Point Lepreau feeders were
made of A106 Grade-B steel with 0.019% Cr. Replacement material has been A106 Grade-C steel
with a Cr content of ~0.33% and this has led to a reduction in FAC rate of around 50%, which
should be enough to assure the integrity of the piping for the lifetime of the plant.
The lower temperatures in secondary coolants and feedwater systems lead to higher magnetite
solubilities and correspondingly higher FAC rates than in primary coolants (both solubility and
FAC rate peak at ~130-150oC, irrespective of pH or flow – see Figure 9). Thus, even with a steel
containing 0.3% Cr and carrying feedwater dosed with amines, the 56 cm-diameter piping at the
Mihama-3 PWR at 140oC corroded at an average rate of 0.3-0.4 mm/year over the 20 years or so
of operation before rupturing when the wall thickness reached about 2 mm. The mechanisms of
FAC in feedwater systems of Mihama and elsewhere, including the effects of steel composition,
have been successfully modelled (Phromwong et al., 2011).
4.3 Galvanic Corrosion

The aqueous corrosion of metals involves the transfer of electrons – it is an electrochemical
process. Thus, for the corrosion of iron or steel in pure water, Equation 13 represents the anodic
“half-cell” process forming iron ions in solution and electrons that are conducted to adjacent
sites and react with water molecules (via Equation 14 – the cathodic “half-cell” process) to form
hydroxide ions and hydrogen. The anodic reaction must be balanced by the cathodic reaction.
The ions in solution interact and form iron hydroxide that will precipitate as the basis for rust or,
at high temperature, will form magnetite.
A metal in contact with a solution establishes an electrical potential with respect to the solution.
Such a potential cannot be measured directly, since a conductor introduced into the solution to
serve as a measuring electrode to indicate the potential difference with the metal will establish
its own potential. The best we can do is then to measure the metal’s potential with respect to an
electrode that is designated a standard. Such a standard is the “standard hydrogen electrode”
(SHE), which is platinum metal in a solution of hydrogen ions at unit activity (i.e., an acid at pH
0) in contact with hydrogen gas at unit activity (i.e., at 1 atm pressure or partial pressure). The
SHE is given the potential zero, and the potentials of metals in equilibrium with solutions of
their own ions at unit activity are listed against the SHE potential (see Table 5). This list is the
galvanic series for industrially-important metals, and those with high potentials are termed
more “noble” than those “active” ones with low potentials. An active metal can displace a more
noble metal from solution – so in principle, a metal with a negative potential can displace
hydrogen from solution as it dissolves or corrodes; a familiar example is the coating of an iron
nail with copper as it is dipped into copper sulphate solution. Note, however, that some active
metals such as aluminum and chromium are corrosion-resistant in many aqueous and atmos-
pheric environments. This is because they form a very protective oxide that resists corrosion and
makes their behavior more noble (they “passivate”).
Two dissimilar metals in contact in an aqueous environment can therefore act as an electro-
chemical cell, in which the more noble metal acts as the cathode and the more active metal acts
as the anode; i.e., the more active metal dissolves – it corrodes. The greater the difference in
the potentials, the greater the reaction rate. This is the principle of batteries such as the Daniel
cell, in which electrodes of copper and zinc immersed in sulphate solution generate a potential
difference of ~1.1 V as the Zn anode oxidizes and dissolves while Cu species are reduced and
precipitated as metal at the Cu cathode. These processes generate an electric current in an
external circuit. Galvanic coupling of metals is to be avoided in power plants, otherwise the
more active metal will tend to corrode adjacent to the contact point. The severity of the corro-
sion depends upon the conductivity of the solution and the integrity of the joint between the
metals. Oxide films may act as insulators, for example, and reduce the galvanic effect. On the
other hand, galvanic processes can be used to protect metals from general corrosion. Galvanis-
ing steel by coating with zinc, for example, promotes a galvanic process, since the more active
Table 5. The galvanic series for important metals

Metal-metal-ion Potential (25oC)
Equilibrium Volts vs SHE
Au3+ + 3e -  Au +1.498
Pt 2+ +2e -  Pt +1.200
Pd 2+ +2e -  Pd - +0.987
Ag+ +e -  Ag More Noble +0.799
Hg22+ +2e -  2Hg +0.797
Cu2+ +2e -  Cu +0.377
2H + +2e -  H2 0.000
Pb2+ +2e -  Pb -0.126
Sn2+ +2e -  Sn -0.136
Ni 2+ +2e -  Ni -0.250
Co2+ +2e -  Co More Active -0.277
Cd 2+ +2e -  Cd ¯ -0.403
Fe 2+ +2e -  Fe -0.440
Cr 3+ +3e -  Cr -0.744
Zn2+ +2e -  Zn -0.763
Al 3+ +3e -  Al -1.662
Mg2+ +2e -  Mg -2.363
Na+ +e -  Na -2.714
K + +e -  K -2.925
Li + +e -  Li -3.040
zinc becomes the anode when the couple is immersed in a solution and lowers the potential of
any areas of exposed steel, protecting them cathodically. Similarly, components prone to
corrode in aqueous environments may be protected cathodically by connecting them to an
electrical source or by attaching a sacrificial anode of more active metal to lower their potential.
It is important to note that in most galvanic couples it is the cathodic component that controls
the reaction rate and therefore the corrosion rate. Consequently, if dissimilar metals must be in
contact, the area of the cathode should be small relative to that of the anode. It follows that to
protect an active metal in contact with a more noble one by coating, painting, etc., the cathode
(more noble component) should be the one coated. Horror stories abound of components such
as carbon steel heat exchanger shells corroding rapidly adjacent to tube sheets of more noble
metals such as titanium because the steel was coated as a protective measure rather than the
titanium.
In summary, measures to minimize galvanic corrosion include: selecting metals as close togeth-
er as possible in the galvanic series; avoiding small-anode/large-cathode combinations (e.g.,
choosing fasteners of more noble material)

material); insulating dissimilar metals (e.g., sleeving bolts in
flanged joints, as well as using insulating washers); applying coatings carefully, and keeping them
in good condition (especially those on anodes); avoiding threaded joints where possible; design-
ing for the anodic members (making them thicker, easily replaceable, etc.); and, installing a third
metal that is anodic to both in the couple.
4.4 Stress-Corrosion
Corrosion Cracking & IGA
Stress-corrosion
corrosion cracking (SCC) occurs in many metals in many environments; even nominally
ductile materials can be affected. The attack may be trans
trans-granular or inter-granular,
granular, depending
upon whether the crack path traverses the metal grains or follows the grain boundaries (see
Figure 11).
(a) (b)
Figure 11.. Examples of transgrannular (TGSCC) (a) and intergrannular (IGSCC) (b) stress
corrosion cracking in stainless steel and brass respectively (after Fontana 1986)
There are three factors required for inducing SCC: the environment must be sufficiently aggres-
aggre
sive, which involves temperature since high temperatures generally increase a material’s suscep-
susce
tibility; the composition and microstructure must be suscept
susceptible,
ible, which includes effects such as
sensitisation as discussed earlier; and, the metal must be under tensile stress, which may be
residual stress from manufacturing and/or operational stress from contained pressure. Note,
however, that threshold levels of these factors do not necessarily exist if there is a long enough
exposure time of the material (Andresen et al., 2000).
The inside surface of Zircaloy sheathing of nuclear fuel elements can crack during operation
under the influence of mechanical strain and of fission products, particularly iodine, released
during irradiation. The CANLUB process for coating the insides of CANDU fuel sheaths with a
thin layer of graphitic material was designed to counteract SCC by acting as a lubricant and to
some extent as a getter for iodine. It has made CANDU fuel tolerant to a range of conditions
imposed by power ramping and is applicable to both UO2 and ThO2 fuels (Hastings et al., 2009).
The alloys most susceptible to cracking in nuclear coolants have been the austenitic steels such
as the 300-series stainless steels used for LWR piping and component cladding and the austenit-
ic nickel alloys such as Alloy 600 (Inconel-600) used for steam generator tubing and pressure
vessel penetrations. Thus, in the first years of operation of BWRs, the coolant in the reactor
recirculating system was treated with “normal water chemistry” – NWC, which was nominally
pure water that became oxidizing with ~200 ppb dissolved oxygen because of radiolysis in the
core. The recirculation piping (made of type ASA 304 stainless steel) was then found to be prone
to IGSCC. Extensive pipe replacements were necessary using, where possible, stress-relieved
material containing low amounts of carbon to minimize the presence of sensitized material in
service, especially in components where cold work and/or welding were used in fabrication.
Lately, to make the coolant less oxidizing and therefore less aggressive to the austenitic steels
around the core circuit, the chemistry has been changed to “hydrogen water chemistry” – HWC,
which involves injecting hydrogen into the feedwater to counteract the radiolysis in the core. A
level of ~2.0 ppm has been used to reduce the electrochemical corrosion potential (ECP) of the
steel from the region of +100 to +300 mV with respect to the standard hydrogen electrode
(SHE), attained under NWC conditions, to -230 mV (SHE) or lower – the measured threshold for
SCC of stainless steel under BWR conditions. To achieve such low ECPs on all susceptible com-
ponents in the reactor vessel without raising the hydrogen additions to unacceptable levels,
General Electric’s NobleChemTM treatment has been applied, involving the addition of noble
metal salts to the coolant to create a catalytic environment on surfaces (Hettiarachchi, 2005).
Note that in CANDU reactors, the main use for the austenitic stainless steels is in the moderator
system (the calandria vessel and the piping) and, at the low temperature of the circulating heavy
water (<60oC), there have been no incidences of SCC since it was recognized early in the design
of the CANDU reactors that the low carbon versions (L-grade) of the austenitic stainless steels
were to be specified to avoid known problems associated with sensitization, especially during
welding.
Early steam generators in the PWRs and the CANDUs at Bruce A and the demonstration CANDU
NPD were tubed with Alloy 600, which was known to be somewhat resistant to TGSCC (trans-
granular stress-corrosion cracking) from chlorides, although research had indicated that under
some circumstances at high temperature it could crack in pure water (Coriou et al., 1960).
Unfortunately, most of those tubes had generally received only a cursory heat-treatment during
manufacture; they had been “mill-annealed” (MA) and as a result were supplied in a sensitized
state that in many instances led to IGSCC during operation. Early failures occurred on the
primary side in further-stressed locations such as the tight-radius U-bends and the rolled area
near the tubesheet. Also, IGSCC occurred on the secondary side because of contact with high
concentrations of alkaline impurities in crevices with tube support plates and under “sludge
piles” on the tube sheet. Full replacement of most of the PWR steam generators tubed with MA
Alloy 600 became necessary, so that now only a few remain, the rest having been replaced with
those tubed with thermally-treated
treated (TT – to relieve stresses) Alloy 690 – a nickel alloy containing
more chromium and less nickel. In CANDUs and several European PWRs, Alloy 800 has been
preferred for SG tubing. The Alloy 600 SGs in Units 1 & 2 at Bruce A have been replaced with
those tubed with Alloy 800 following the refurbishment projects
projects.. This contains more chromium
and iron than Alloy 600 and, having less than 50% nickel, is not strictly speaking a nickel alloy,
although it is usually classified as such for comparison with the others. It has been relatively
resistant to SCC.
Stainless steels and austenitic alloys containing moderate to high concentrations of carbon (i.e.,
>0.03%)
03%) are vulnerable to corrosion of their grain boundaries if they have undergone a heating
process such as welding that raises the temperature to the region of 500500-800oC. Heating to this
range “sensitises” the material by preferentially precipitating the chromium in the grain bounda-
bound
ries as chromium carbide (Cr23C6), thereby leaving the metal surface above the grain boundaries
depleted in chromium that would otherwise form the passivating oxide based on Cr2O3. As
shown in Figure 12, the grain boundaries are then subject to corrosion via intergranular attack
(IGA).
In particularly aggressive environments, IGA can be severe enough to loosen metal grains which
then fall out, leaving a roughened and even more vulnerable surface exposed. In welded materi-
mater
al, the attack is often called weld decay.
Figure 12.. Typical sensitization mode and IGA of Type 304 stainless steel (after Fontana 1986)
Note that some situations may tolerate sensitised material; thus, sen
sensitised
sitised stainless steel may
be used in non-aggressive
aggressive environments such as those found in architectural applications. For
exposures to potentially aggressive conditions, however, sensitised material should not be used.
A low-carbon
carbon grade of material (e.g., type ASA 304L stainless steel, with L indicating <0.03% C) is
normally specified and since the carbon content is much reduced there is little available to
precipitate as the carbide.
Austenitic material in general is usually supplied in the solution-quenched (also called solution-
annealed or quench-annealed) condition, indicating that it was heated to above the carbide-
precipitation temperature to dissolve the carbides into the metal matrix and then quenched to
cool rapidly through the precipitation range of temperature so that carbide has no time to
precipitate. Note that subsequent welding can re-sensitise the material, in which case the
component must be solution-quenched again. There are also “stabilised” alloys such as type ASA
347 or type ASA 327 stainless steels which, respectively, contain small amounts of the strong
carbide formers niobium or titanium. These have been heat-treated to precipitate the carbon as
the niobium or titanium carbide at a temperature above the range where the chromium carbide
would precipitate, leaving no carbon to combine with the chromium.
4.5 Crevice Corrosion & Pitting

Crevice corrosion and pitting are similar in that each proceeds via an electrochemical mechan-
ism involving the formation of differential concentration cells arising from the occlusion of
chemical species within restricted areas. Beneath gaskets in flanged connections, under bolt
heads and washers, within riveted and lapped joints and under deposits of sludge or silt are
common areas for attack, which can occur in a wide range of environments, including cooling
waters, and which is greatly exacerbated by the presence of chloride ions. As with other types
of localized corrosion, metals that rely on a passivating oxide for corrosion protection, such as
titanium, stainless steel, etc., are especially vulnerable, although active metals such as carbon
steel are also affected. In fact, all metals are said to undergo crevice corrosion given the right
conditions. The mechanism proceeds as ingress of a corroding solution into a crevice such as a
lapped joint in steel, for example, initially causes metal ions (in this example Fe2+) to be released
into the restricted space by the straightforward corrosion reaction. If oxygen is present in the
solution it is rapidly depleted in the enclosed space and a differential aeration cell between the
interior and exterior is established, promoting internal anodic metal dissolution and cathodic
reduction of oxygen on surfaces external to the crevice. Even without oxygen present, the
accumulated Fe2+ ions hydrolyse to Fe(OH)2 and create internal acidic conditions while Cl- ions
are drawn into the space to maintain charge neutrality. Any oxide films within the space are
degraded, the anodic processes are accelerated and attack is perpetuated (Figure 13 is a genera-
lized illustration of the development of crevice corrosion in an aqueous environment contami-
nated with NaCl).
A B
Figure 13. Initial (A) and final (B) stages in the development of crevice corrosion of a metal M
(Fontana, 1986)
In the 1980s, the basic mechanism of the denting of the steam generators of Westinghouse
design, which had drilled-hole
hole support plates of carbon steel,, was the corrosion of the steel
forming the continuous crevicess around the tubes. The heat transfer and boiling within the
crevices drew in impurities such as chlorides and when these concentrated, especially in the
presence of oxygen, aggressive acid conditions attacked the steel. The resulting growth of
corrosion-product magnetite in the crevices exerted enough force to compress (“dent”) the
tubes and deform the plate itself. Several steam generators have had to be replaced because of
denting. The prime
rime example in CANDUs is the crevice corrosion of the inner surfaces of pressure
p
tubes beneath fuel bearing pads
pads, exacerbated by localized boiling which concentrates solutes
such as LiOH and leads to aggressive conditions.
The similarity of the mechanism of pitting with that of creviccrevice

e corrosion can be seen in the
illustration in Figure 14. Note that the precipitation of corrosion products (e.g., hydroxides in the
illustrations) can occur at the mouth of the crevice and around the rim of the pit and in the
latter example can grow into tubes or tubercules
tubercules. Pits usually occur on upward-facing
upward surfaces,
less frequently on vertical surfaces and rarely on downward
downward-facing
facing surfaces, indicating an effect
of gravity on the occluded high-density
density solution
solution.
Figure 14. Self-perpetuating corrosion pit (after Fontana, 1986)
Pitting in particular can be extremely damaging. The pit mouths are often quite small and may
be accompanied by extensive undercutting, culminating in perforation in severe cases, and pits
may run into each other to create a rough surface degradation called general wastage (Figure
15). Beneath deposits, pits can go undetected and escape cursory inspections.
Figure 15. An undercut pit and surface wastage from overlapping pits
The pitting of heat exchanger tubes under silt deposits in an environment of static aerated
cooling water polluted with chlorides is a pervasive problem in most industries, including
nuclear power generation. Another common example in reactor systems is the pitting of the
secondary side of steam generator tubes beneath sludge piles on the tubesheet. This was a
particular problem with the German PWRs that had steam generators tubed with Alloy-800 and
operated with chemistry controlled with additions of sodium phosphate; the concentration of
aggressive species beneath the sludge led to pitting, in some cases extending to wastage. The
steam generators at the CANDU 6 at Point Lepreau, also tubed with Alloy 800, plugged many
tubes because of pitting during the period of dosing with phosphate. The change from phos-
phate to all-volatile treatment seems to have arrested the degradation.
4.6 Fretting Wear & Flow-Induced Vibration

The rubbing of one material against another can damage surfaces; in a corrosive environment
the damage is exacerbated. The vibration of foreign objects trapped within components such as
heat-exchanger tube bundles (known examples are condensers and steam generators) can cause
rapid deterioration of the tubes and loose objects within steam generator channel heads have
been known to damage tube sheets. In CANDUs and PWRs the severe conditions of flow of
steam-water mixtures in the secondary side of steam generators can create excessive vibrations
that cause fretting between the tubes and the tube supports. The U-bend sections of the tube
bundles, where the steam quality and coolant velocities are high, are particularly susceptible.
Experiments to characterise the fretting behavior of material combinations at various steam
generator conditions (temperature, coolant chemistry, etc.) typically record the rate of surface
damage, in terms of amount of material removed per unit time, as a function of the work rate,
which is the energy dissipated by sliding contact between two surfaces per unit time. The ratio
of the two, wear rate V to work rate W, defines the wear coefficient, K:
V
K= (24)
W
K is usually quoted in units of m2/N, or Pa-1.
AECL at Chalk River performed an extensive series of fretting-wear tests on steam generator
tube samples in contact with tube support samples. The factor having the greatest effect on
tube damage was temperature; K at a steam generator operating temperature of 265°C was 20
times higher than at 25°C. Other factors, namely chemistry (ammonia, hydrazine, phosphate or
boric acid), type of tube support (flat bar or broached hole), tube material (Alloy 800 or Alloy
690) or support material (type 410 or type 321 stainless steel) had relatively little effect. These
results led to the suggestion that an average value of K of 20×10-15 Pa-1 and a conservative value
of 40×10-15 Pa-1 may be used over a range of work rates for typical steam generator geometries
and operating conditions (Guérout and Fisher, 1999).
Fretting wear also affects in-core components. In PWRs, it has been estimated that over 70% of
fuel leaks are caused by fretting between the Zircaloy fuel rods and support grids – the so-called
GRTF, grid-to-rod fretting (EPRI, 2008). The failures usually occur on final-cycle fuel assemblies,
which are located at the core periphery. Similar failures occur in the WWER reactors, which have
stainless steel grids.
The localized accelerated corrosion of CANDU pressure tubes because of the crevice effect
beneath fuel bundle bearing pads has already been mentioned. When there is relative motion
between the bearing pad and the pressure tube, resulting from flow-induced vibration for
example, fretting can exacerbate the damage. This has rarely been seen except in reactors
having 13 fuel bundles, when significant damage is usually at the end of the pressure tube,
where the cross-component of the coolant flow may cause more vigorous vibration. On the
other hand, debris caught beneath fuel bearing pads can lead to significant fretting wear
anywhere along the pressure tube and objects caught between elements within the fuel bundle
can lead to fuel failures.
4.7 Hydrogen Effects

The most common form of hydrogen damage in reactor systems is the degradation of hydride-
forming materials, in particular the zirconium alloys. In reactors in general these make up the
fuel cladding, but in CANDUs they also make up the pressure tubes and the calandria tubes. As
described earlier, delayed hydride cracking (DHC) and hydride blistering have been responsible
for major replacements of CANDU pressure tubes in the past. Titanium also can be degraded by
hydrogen, so it should be borne in mind that the contact of titanium components (condenser
tubes, tube-sheet cladding, etc.) with carbon steel components (condenser shell, tube supports,
etc.), which may promote galvanic corrosion of the steel, will drive cathodically-generated
hydrogen into the metal. Similarly, hydriding of titanium condenser tubing in contact with a
aluminum-bronze tube sheet occurred at the CANDU 6 at Point Lepreau. The condenser was
completely re-tubed during the refurbishment outage and the Al-bronze tube sheet replaced
with titanium. It follows that cathodic-protection measures such as sacrificial anodes, which
lead to the production of hydrogen on cathodic components, can also promote hydriding of
susceptible metals.
4.8 Microbiologically-Induced Corrosion (MIC)

Although MIC has been known for over 70 years, it has only been recognized as a serious
corrosion mechanism in nuclear and fossil plants for 30 or so years (EPRI, 1988). Early references
to the subject were general descriptions of practical situations (Uhlig, 1948) and mechanistic
interpretations of observations were attempted from the 1960s onwards (Videla and Herrera,
2005).
There are many types of microbe that influence corrosion and they occur in many environments,
especially natural waters. Consuming a variety of nutrients, they can grow under a wide range of
temperatures, pressures and alkalinities/acidities. Their spores are even more resistant to
adverse conditions and can be transported around systems and lie dormant for years before
being revived to germinate, attach to surfaces and create localized colonies that exude corrosive
waste matter (Licina and Cubicciotti, 1989). The biofilms created by the colonies consist of the
so-called exopolymers, largely consisting of polysaccharides that develop as slimes containing
corrosion products and debris carried from elsewhere in the system. They impede the transport
of corrosion inhibitors to the metal surface and to some extent they shield the organisms from
attack by treatment chemicals such as biocides. They can be tenacious and difficult to remove
by fluid forces alone and in some cases – often carbon steel piping systems – can be thick
enough to cause severe pressure drops.
The microbial species may exist in anaerobic (e.g., deaerated) conditions or in aerobic (e.g.,
aerated) conditions, and in some situations the one type creates local environments that sustain
the other type, giving cycles of corrosion activity. Stagnant or low-flow conditions are conducive
to MIC, so components buried in damp or wet soil and water-containing piping such as fire
dousing systems are particularly vulnerable. Locally stagnant environments, such as under dirt
or deposits and in crevices, also promote MIC. Reactor design features with redundant, static
water systems along with long shut-down periods make nuclear plants particularly susceptible.
Many materials are attacked, including the common metals in reactor systems such as stainless
steels, nickel alloys, carbon steel, copper, and copper alloys. Typical features of attack in carbon
steel are corrosion products deposited as tubercles and in stainless steel and nickel alloys are
small pits, often stained with rust, that can be deep enough to be through-wall and cause small
leaks especially at welds.
Sulphate-reducing bacteria (SRB) are responsible for many failures in water systems. The species
desulfovibrio and desulfomaculum are among the most widespread and economically important
organisms causing corrosion. They are anaerobic because their metabolism involves the con-
sumption of the sulphate ion in deaerated waters, especially under deposits, although they can
survive limited exposures to aerated conditions. They operate in steel systems by promoting the
overall mechanism:
( )
4Fe+ SO42- + 4H2O ® 3Fe OH +FeS + 2OH -
2
(25)
in which the microbes are instrumental in reducing the SO42- to S2-. The early postulate that the
microbes act by depolarizing the cathodic areas and catalyzing the recombination of adsorbed
hydrogen atoms to molecular H2 has been largely superseded by mechanisms involving the
formation of H2S and the interaction with iron sulphide corrosion products (Videla and Herrera,
2005).
Aerobic species act by creating differential aeration and concentration cells through their
biofilms. In seawater-cooled systems, for example, species such as Pseudomonas proliferate and
can cause extensive damage to carbon steel and stainless steels. Detailed examinations of type
304 stainless steel surfaces pitted by Pseudomonas in seawater have shown more localized
depletion of chromium in the passive films than similar control surfaces exposed in the absence
of the bacteria (Yuan and Pehkonen, 2007).
Counteracting MIC is difficult. System cleanliness is quoted as a desired remedy but it should be
noted that deionised water systems are susceptible. Moreover, removing slimes and biofilms by
flushing is seldom effective, even with pulsating flows, and mechanical scouring with sponge
balls etc. can damage protective films. Biocides are therefore a preferred recourse in many
situations. Videla and Herrera (2005) list common biocides in industrial applications:
Table 6. Biocides used in industrial water systems.

Biocide Properties Usual concentration
(mg/L)
Chlorine Effective against bacteria and algae; oxidizing; pH 0.1–0.2 (continuous)
dependent
Chlorine Effective against bacteria, less so against fungi and 0.1–1.0
dioxide algae; oxidizing; pH-independent
Bromine Effective against bacteria and algae; oxidizing; 0.05–0.1
wide pH range
Ozone Effective against bacteria and biofilms; oxidizing; 0.2–0.5
pH-dependent
Isothiazolones Effective against bacteria, algae and biofilms; non- 0.9–10
oxidizing; pH-independent
a
QUATs Effective against bacteria and algae; non-oxidizing; 8–35
surface activity
Glutaraldehyde Effective against bacteria, algae, fungi and bio- 10–70
films; non-oxidizing; wide pH range
b
THPS Effective against bacteria, algae and fungi; low
environmental toxicity; specific action against SRB
a
Quaternary ammonium compounds.
b
Tetrakis-hydroximethyl phosphonium
5 References
R. Adamson, F. Garzarolli, B. Cox, A. Strasser and P. Rudling. Corrosion mechanisms in zirconium
alloys. Zirat 12 Special Topic Report. A.N.T. International (2007).
P.L. Andresen, T.M. Angeliu, R.M. Horn, W.R. Catlin and L.M. Young. Effect of deformation on
SCC of unsensitized stainless steel. Proc. CORROSION 2000 Conf., Orlando, Fl., USA.
NACE Intern. (2000 March 26-31).
ASM International, ASM Handbook: Volume 9 – Metallography and Microstructures, GF. Vander
Voort, Editor, ASM, Materials Park, Ohio, 2004.
P.V. Balakrishnan. A radiochemical technique for the study of dissolution of corrosion products
in high-temperature water. Can. J. Chem. Eng., 55, 357 (1977).
P. Berge, J. Ducreux and P. Saint-Paul. Effects of chemistry on corrosion-erosion of steels

in water and wet steam. Proc. 2nd Intern. Conf. onWater Chemistry of Nuclear Reactor Systems,
Bournemouth, UK. BNES. (1980 October 14-17).
G.J. Bignold, K. Garbett and I.S. Woolsey. Mechanistic Aspects of the Temperature Dependence
of Erosion-Corrosion. Proc. Specialists Meeting on Erosion-Corrosion in High Tempera-
ture Water and Wet Steam. Les Renardières, France. EDF. (1982).
N. Birks, G.H. Meier, F.S. Pettit, Introduction to the high-temperature oxidation of metals, 2nd
Edition, Cambridge University Press, 2006.
M. Bouchacourt. Identification of key variables: EdF studies. Proc. EPRI Workshop on Erosion-
Corrosion of Carbon Steel Piping. Washington DC, USA. (1987 April).
W.D. Callister, D.G. Rethwisch, Materials Science and Engineering: An Introduction, 8th Edition,
Wiley, 2009.
W.G. Cook, D.H. Lister and J.M. McInerney. The effects of high liquid velocity and coolant
chemistry on material transport in PWR coolants. Proc. 8th Intern. Conf. onWater Chemi-
stry of Nuclear Reactor Systems, Bournemouth, UK, BNES, October 22-26, (2000).
H. Coriou, L. Grall, Y. Legall and S. Vettier. Stress corrosion cracking of Inconel in high tempera-
ture water. 3e Colloque Metall., Saclay, 1959. North Holland Publ., Amsterdam. p. 161
(1960).
B. Cox. Mechanisms of zirconium alloy corrosion in nuclear reactors. J. Corr. Sci. Eng.,Vol. 6,
Paper 14 (2003).
EPRI. Sourcebook for microbiologically influenced corrosion in nuclear power plants. Electric
Power Research Institute. Technical Report NP-5580 (1988).
EPRI. Fuel reliability guidelines: PWR grid-to-rod fretting. Electric Power Research Institute,
Product ID 1015452. (28th July 2008).
M.G. Fontana. Corrosion Engineering. McGraw-Hill. (1986).
A. Garlick, R. Sumerling and G.L. Shires. Crud-induced overheating defects in water reactor fuel
pins. J. Br. Nucl. Energy Soc., 16(1), 77 (1977).
F.M. Guérout and N.J. Fisher. Steam generator fretting-wear damage: a summary of recent
findings. Trans. ASME, 121, 304-310 (1999 August).
I. Hastings, K. Bradley, J. Hopwood, P. Boczar, S. Kuran and S. Livingstone. We can use thorium.
Nucl. Eng. Intern. (2009 August 28).
S. Hettiarachchi. BWR SCC mitigation experiences with hydrogen water chemistry. Proc. 12th
Intern. Conf. on Environmental Degradation of Matls. in Nuclear Power Systems. Salt
Lake City, Utah, USA. TMS. (2005 August 14-18).
E. Hillner, D.G. Franklin and J.D. Smee. Long-term corrosion of Zircaloy before and after irradia-
tion. J. Nucl. Matls., 278(2-3), 334-45 (2000).
K. Ishida and D.H. Lister. In-situ measurement of corrosion of Type 316 Stainless Steel in 280ºC
pure water via the electrical resistance of a thin wire. J. Nucl. Sci. Tech. 49 (11), (2012
November).
P. Kofstad. On the formation of porosity and microchannels in growing scales. Oxidation of

Metals, 24(5-6), 265-276 (1985).
G.J. Licina and D. Cubicciotti. Microbial-induced corrosion in nuclear power plant materials. J.
Nucl. Matls., 41(12). 23-27 (1989).
D.H. Lister. Activity transport and corrosion processes in PWRs. Nucl. Energy, 32(2), 103- 114
(1993).
D.H. Lister. Some aspects of corrosion in cooling water systems and their effects on corrosion
product transport. Proc. EUROCORR 2003, Budapest, Hungary, (2003 September 28-
October 2).
D.H. Lister and L.C Lang. A mechanistic model for predicting flow-assisted and general corrosion
of carbon steel in reactor primary coolants. Proc. Chimie 2002; Intern. Conf. Water
Chem. Nucl. Reactor Systems. Avignon, France. SFEN. (2002).
D.H. Lister and S. Uchida. Reflections on FAC mechanisms. Proc. Intern. Conf. on Flow-
Accelerated Corrosion (FAC) in Fossil and Combined Cycle / HRSG Plants, Arlington, VA,
USA. (2010 June 29-July 1).
E.C.W. Perryman. Pickering pressure tube cracking experience. Nucl. Energy, 17, 95–105 (1978).
P. Phromwong, D.H. Lister and S. Uchida. Modelling material effects in flow accelerated corro-
sion. Proc. Intern. Conf. on Environmental Degradation of Matls. in Nuclear Power Sys-
tems, Colorado Springs, CO. TMS. (2011 August 7-11).
E.C. Potter and G.M.W. Mann. Oxidation of mild steel in high-temperature aqueous systems.
Proc. 1st Intern. Congress Metallic Corros. London. p. 417 (1961).
Qiu, L. Dissolution of silicon nitride in high-temperature alkaline solutions. Proc. Intn’l. Conf.
Props. Water and Steam. IAPWS. Greenwich, London. Sept. 1st-5th (2013).
P.R. Roberge. Handbook of corrosion engineering. McGraw Hill (2000).
G. Schikorr. On the iron-water system. Z. Elektrochem. 35, 62 (1928).
C.J. Simpson and C.E. Ells. Delayed hydrogen embrittlement of Zr-2.5wt.%Nb. J. Nucl. Mater., 52,
289–295 (1974).
J.P. Slade and T.S. Gendron. Flow accelerated corrosion and cracking of carbon steel piping in
primary water – operating experience at the Point Lepreau Generating Station. Proc. In-
tern. Conf. on Environmental Degradation of Matls. in Nuclear Power Systems, Salt Lake
City, UT. TMS. (2005).
P. Sturla. Oxidation and deposition phenomena in forced circulating boilers and feedwater
treatment. Proc. Fifth National Feedwater Conf. Prague, Czech. (1973).
R.L. Tapping. Materials performance in CANDU reactors: The first 30 years and the prognosis for
life extension and new designs. J. Nucl. Materials. 383, 1-8 (2008).
M.R Taylor, J.M Calvert, D.G. Lees and D.B. Meadowcroft. The mechanism of corrosion of Fe-
9%Cr alloys in carbon dioxide. Oxidation of Metals, 14(6), 499-516 (1980).
H.H Uhlig. Corrosion handbook. John Wiley, New York. 466-481 (1948).
H.A. Videla and L.K. Herrera. Microbiologically-influenced corrosion: Looking to the future.
Intern. Microbiol., 8. 169-180 (2005).
C. Wagner and W. Traud. On the interpretation of corrosion processes through superposition of

electrochemical partial processes and on the potential of mixed electrodes. Z. Elektro-
chem. Ang. Physik. Chemie, 44, 398 (1938).
G.S. Was, Fundamentals of Radiation Materials Science: Metals and Alloys, Springer, 2007.
S.J. Yuan and S.O. Pehkonen. Microbiologically influenced corrosion of 304 stainless steel by
aerobic Pseudomonas NCIMB 2021 bacteria: AFM and XPS study. Colloids and Surfaces
B: Biointerfaces, 59(1), 87-99 (2007).
6 Acknowledgements
Kit Colman
Mark Daymond
Ed Price
Bob Tapping
1
CHAPTER 15
Chemistry in CANDU Process Systems
prepared by
Dr. William G. Cook and Dr. Derek H. Lister

Summary:
The efficient and safe operation of a CANDU reactor is highly dependent upon the selection and
proper implementation of chemistry control practices for the major and ancillary process systems
such as the primary and secondary coolants and the moderator. The materials of construction of
the various systems are selected in consideration of the neutron economy while keeping the
proper chemical environment in mind in order to keep corrosion and degradation low and to
ensure desired plant operating lifetimes. This chapter begins with an overview of the basic
chemistry principles required to manage chemistry in CANDU reactors and then provides a
detailed description of the chemistry control practices and the reasons behind their use in the
major and ancillary process systems. The chapter concludes by examining the current practices
of component and reactor lay-up for maintenance shut-downs and refurbishments and a descrip-
tion of heavy water purification and upgrading.
Table of Contents
1 Introduction ............................................................................................................................... 4
1.1 Overview ......................................................................................................................... 4
2 Chemistry Principles Applied to Reactor Coolants .................................................................... 4
2.1 pH, pD and Apparent pH (pHa)........................................................................................ 5
2.2 Solution Conductivity...................................................................................................... 7
2.2.1 Specific Conductivity................................................................................................... 7
2.2.2 Cationic Conductivity or Conductivity After Cationic Exchange (CACE) ................... 10
2.3 Purification and Ion Exchange Resin ............................................................................. 11
2.3.1 Structure of Ion Exchange Resin ............................................................................... 11
2.3.2 Ion exchange capacity............................................................................................... 11
2.4 Water Radiolysis............................................................................................................ 12
2.5 Solubility of Gases in Water – Henry’s Law................................................................... 15
3 Primary Heat Transport System ............................................................................................... 17
3.1 Chemistry Control in the PHTS...................................................................................... 17
3.2 Corrosion Issues in the PHTS......................................................................................... 20
3.2.1 Feeder Pipe FAC and Cracking................................................................................... 20
3.2.2 Delayed Hydride Cracking of Zirconium Alloys ......................................................... 21
3.3 Activity Transport.......................................................................................................... 21
4 Secondary Heat Transport System........................................................................................... 23
4.1 Chemistry Control in the Secondary System................................................................. 23
©UNENE, all rights reserved. For educational use only, no assumed liability. Chemistry in CANDU Process Systems – September 2014
4.2 Corrosion Issues in the Secondary System ................................................................... 27

4.2.1 Boiler crevices ........................................................................................................... 27
4.2.2 Flow-Accelerated Corrosion in the Feedwater and Steam Extraction Piping........... 28
5 Moderator System ................................................................................................................... 31
5.1 Chemistry Control in the Moderator System ................................................................ 31
5.1.1 Reactor shims............................................................................................................ 32
Gadolinium Nitrate (Gd(NO3)3) ......................................................................................... 32
Boric Acid .......................................................................................................................... 33
5.1.2 Guaranteed Shutdown State (GSS) ........................................................................... 33
Gadolinium Oxalate........................................................................................................... 33
5.1.3 Shutdown System 2 (SDS2) ....................................................................................... 34
5.2 Moderator Cover Gas.................................................................................................... 34
6 Auxiliary Systems ..................................................................................................................... 36
6.1 Calandria Vault and End Shield Cooling System............................................................ 36
6.2 Liquid Zone Control....................................................................................................... 37
6.3 Annulus Gas .................................................................................................................. 38
6.4 Emergency Core Cooling Systems ................................................................................. 38
6.5 Service Water ................................................................................................................ 39
7 Lay-up Practices ....................................................................................................................... 40
7.1 Dry Lay-up ..................................................................................................................... 41
7.2 Wet Lay-up .................................................................................................................... 41
8 Heavy Water Systems .............................................................................................................. 42
8.1 Upgrading...................................................................................................................... 42
8.2 Clean-up ........................................................................................................................ 44
Tritium Removal ................................................................................................................ 44
9 Summary of Relationship to Other Chapters........................................................................... 46
10 References ............................................................................................................................... 46
11 Acknowledgements ................................................................................................................. 48
List of Figures
Figure 1. Temperature dependence of pKw / pKD and the neutral pH / pD for the dissociation of
light and heavy water. ............................................................................................................. 7
Figure 2. Simplified flow diagram of a PHT purification loop (courtesy of AECL)......................... 19
Figure 3. Magnetite solubility calculated from thermodynamic data optimized to the Tremaine &
Leblanc data [Tremaine, 1980]. ............................................................................................ 20
Figure 4. Mass balance on steam generator impurity inventory. ................................................. 26
Figure 5. Alloy 800 Recommended Operating Envelope (Tapping, 2012). .................................. 28
Figure 6. Effect of steam voidage on FAC rate of carbon steel at 200°C under three chemistries
(neutral, pH25°C with ammonia and with ETA) at a flow rate of 0.56 L/min (the effect of
reduced flow rate on FAC with ammonia also indicated). [Lertsurasakda et al., 2013] ....... 30
Figure 7. Normal moderator cover gas system volume and relief ducts (CANTEACH)................. 35
Figure 8. Schematic diagram of the Calandria vault and reactor assembly.................................. 36
Figure 9. Locations requiring attention during maintenance outages and lay-up of a fossil plant.
[after Mathews, 2013] .......................................................................................................... 40
Figure 10. Oxide/steel volume ratios. ........................................................................................... 41
Chemistry in CANDU Process Systems 3
Figure 11. Schematic representation of a GS dual temperature absorption column (Benedict et

al., 1980). .............................................................................................................................. 43
Figure 12. Schematic diagram of the Darlington tritium removal facility. (Busigin & Sood) ........ 45
List of Tables
Table 1. Conductivity at infinite dilution in H2O [CRC, 2014].......................................................... 9
Table 2. Henry’s constants for select gases at 25oC (calculated from IAPWS , 2004)................... 16
Table 3. Target chemistry parameters in the PHTS. ...................................................................... 18
Table 4. Common radioactive isotopes and their source in the PHTS.......................................... 22
Table 5. Target chemistry parameters in the secondary system (all-ferrous materials)............... 24
Table 6. Comparison of delta pH for two chemical dosing strategies in a typical steam cycle. ... 29
Table 7. Target chemistry parameters in the moderator system.................................................. 32
Table 8. Typical Chemistry Parameters of the Calandria Vault and End Shield Cooling System... 37
1 Introduction
1.1 Overview
This chapter explains the current state-of-the-art of CANDU system chemistry. It begins with an
overview of the basic chemistry principles, as relevant to the major process systems of a CANDU
reactor, and goes on to describe the modes of operation of the primary heat transport circuit,
the secondary heat transport circuit, the moderator system and plant auxiliary systems with
respect to system chemistry. It draws on knowledge of overall system configuration and mate-
rials selection as described in chapters 2 and 11 of this text and, through examples of plant
chemistry specifications, creates a detailed knowledge of the reasoning behind the combined
selection of materials and chemistry. The necessity of minimizing corrosion and degradation of
the auxiliary systems supporting the operation of a CANDU reactor are also highlighted along
with a description of current chemistry dosing practices for these systems. Since every reactor
undergoes frequent maintenance outages and a mid-life refurbishment, the lay-up practices
during these outages are important factors affecting the overall lifetime of the plant; current
practices are described. Finally, the isotopic degradation and tritiation of the heavy water
coolant and moderator necessitate the use of clean-up and upgrading, issues that are dealt with
in the final sections of this chapter.

 The relation between system chemistry conditions and material selection
 The current chemical dosing practices for the various cooling systems
 The primary reasoning behind specific chemistry and materials selection
 The current best practice for system layup during maintenance outages
2 Chemistry Principles Applied to Reactor Coolants
The chemistry of CANDU reactor system coolants is generally kept quite simple with the intent
of maintaining highly pure water with low concentrations of chemical additives to maintain low
corrosion rates on the materials in the systems. Before the specific chemistry control strategies
of the various nuclear and non-nuclear coolant systems are described, it is beneficial to define
several basic chemistry concepts that are particularly relevant to chemistry control in the plant.
These include a reminder of the measure of acidity/alkalinity (pH) and how it is typically used in
heavy water systems as “apparent pH or pHa”, the definition and calculation of the conductivity
of solutions containing dissolved ionic compounds and a discussion of water radiolysis or the
breakdown of water when exposed in a radiation field. Each of these topics will be dealt with in
turn to begin the chapter and will be used throughout the remainder of the chapter in describ-
ing the chemistry of the various reactor systems.
2.1 pH, pD and Apparent pH (pHa)

The pH in a heavy water system is more correctly the pD, where D is the deuterium isotope.
Before the implications of using heavy water in the CANDU systems are discussed, it is beneficial
to review some basic chemistry principles. Recall that the definition of pH is:
pH = -log10 aH+ ( ) (1)
where aH+ is the activity of the hydrogen ion in the aqueous solution, which is dependent upon
the concentration of the hydrogen ion (mH+ - mol/kg), the activity of the standard reference state
(mo = 1 mol/kg) and the ionic activity coefficient (+/-) as shown in equation 2.
mH+
aH+ = g± (2)
mo
Under dilute conditions, the activity of ions in solution can be approximated by their concentra-
tion since the mean ionic activity coefficient (+/-) is dependent upon the ionic strength of
solution and is nearly unity (1) for concentrations below ~ 10-3 mol/kg.
For neutral, light water (H2O), there is an equilibrium established between the water molecules
and the dissociation products H+ (or H3O+) and OH- as shown in equation 3.
H2O  H+ + OH- (3)
By definition, the equilibrium constant for this chemical dissociation reaction is given by the
ratio of the activities (concentrations) of the product species to that of the reactant (H2O);
Equation 4 defines Kw, the dissociation constant or auto-ionisation constant for water, which has
a value of 10-14 at 25 oC [IAPWS, 2007, Bandura & Lvov, 2006]. The activity of water (aw) is
typically regarded as unity in dilute solutions, thus neutral water, defined as the point where the
activity of H+ equals that of OH-, will give a pH of 7 as calculated through equation 4.
aH+ aOH-
Kw = = 10-14 @ 25oC (4)
aH O
2
Increasing the pH of a solution is as simple as adding hydroxyl anions from a base such as lithium
hydroxide (LiOH), which will shift the equilibrium between H+ and OH-. Lithium hydroxide is a
strong base that is essentially completely dissociated in water up to reactor operating tempera-
tures and thus can be used at low concentrations to achieve considerable changes in pH.
Exercise 1 demonstrates the calculation.
The dissociation of water (and other ionic compounds) is dependent upon temperature. The
Bandura & Lvov correlation, as recommended by the International Association for the Properties
of Water and Steam (IAPWS) is commonly used to relate this temperature dependence, giving
the result shown in Figure 1. In the temperature range 0 – 310 oC the pKw (defined as -log10Kw)
decreases significantly before starting to increase again around 250oC. This implies that the
neutral point of water is lowered as temperature is increased, as illustrated by simple calculation

– the pH of neutral water at 250oC is ~ 5.57 and at 300oC is about 5.64.
In the heavy water systems of the CANDU reactor, the proper equations describing the dissocia-
tion of heavy water would be:
D2O  D+ + OD- (5)

and the pD is:
pD = -log10 aD+ ( ) (6)
where aD+ is the activity of the deuterium ion in aqueous solution; this is analogous to the
definition of pH in equation 1. As in light water, the activity coefficient under sufficiently dilute
conditions is nearly unity so we can consider the activities of D+ and OD- to be equal to their
molal (mol/kg) concentrations. Unfortunately, the direct measurement of pD is usually not
possible since pH electrodes are typically constructed with light-water-based fill solutions and
are calibrated using light water buffers. Thus, for heavy water systems, it is standard practice to
quote a pHa or “apparent” pH, which identifies the observed or measured pH value of the heavy
water using traditional pH electrodes and buffer solutions. A general relation [Mesmer &
Herting, 1978] between the dissociation of heavy water and that of light water leads to the
simple correlation shown in equations 7 and 8:
pHa = pD - 0.41 (7)
or, when considering an alkaline solution such as CANDU systems adjusted with LiOH:
pHa = pH+ 0.456 (8)
This equation is the direct result of the difference between the dissociation of heavy water and
that of light water when considering that pOD ≈ pOH at room temperature. Also shown in
Figure 1 is the pKD expressing the dissociation equilibrium constant for heavy water [Mesmer &
Herting, 1978]. It is apparent that heavy water is less dissociated than light water under equiva-
lent conditions, leading to higher equivalent values of the neutral point of heavy water at the
various temperatures. This difference results in the approximate 0.41 unit shift between pH of
light water and the pD of heavy water.
Figure 1. Temperature
mperature dependence of pKw / pKD and the neutral pH / pD for the dissociation of
light and heavy water.
2.2 Solution Conductivity

The fact that water dissociates into H+ and OH- ions (equation 3) gives rise to a finite conductivity
value, which means that water in its pure state can act as an electrical conductor, albeit a very
poor one. We recall that the passage of current or electricity is the movement of charge meas- mea
ured in Amperes (A), which is the movement of one Coulomb of charge per second (i.e. 1 A = 1
C/s). Typically the charge carriers are envisaged as electrons moving down a conducting cable
but the migration of ions in solution also carries charge from one point to another down a
potential gradient. The conductivity (or invers
inversely,
ely, the resistivity) of a solution is dependent upon
the concentration of the charge
charge-carrying
carrying ions in solution and their overall mobility. Thus, a
measurement of the conductivity of an aqueous solution will give an indication about the
quantity of the ions
ns that it contains and, in the case of nominally pure CANDU process system
water with small concentrationss of additives to increase or lower the pH, this measurement
should be directly related to the concentration of the specific cations and anions in the solution.
2.2.1 Specific Conductivity
The molar conductivity (S.m2/mol) of an aqueous solution is given by:
k
Lm = (9)
c
where  is the specific conductance of the electrolyte (S/m or 1/.m)

m) and c is the stoichiometric
molar concentration of the solution. Measuring the conductivity (or resistivity) of an aqueous
solution is typically accomplished by placing two plate electrodes in the solution having fixed
separation between them and with known cross-sectional area and then measuring the absolute
resistance between them.. The specific resistivity (()) follows from the calculation of the cell
©UNENE, all rights reserved.

d. For educational use only, no assumed liability. Chemistry in CANDU Process Systems – September 2014
resistance shown in equation 10 and specific conductivity () is related through equation 11.
r´ 
R= (10)
A
1
k= (11)
r
The molar conductivity of the solution, and hence the solution specific conductivity, can be
broken down into the sum of the individual contributions of the ions, each one of which will be
dependent upon its concentration and overall mobility. Essentially, the specific conductivity of
the solution is the sum of the specific conductivities of the individual ions:
k = å ki (12)
i
We define the molar conductivity of the ion in solution (i) in an analogous way as the overall
molar conductivity of the solution, thus:
ki
l m,i = (13)
ci
In an ideal solution, the molar conductivity would vary linearly with increasing concentration
(and specific conductivity); however, in reality the solution molar conductivity has a non-linear
dependence on solution concentration, particularly in more concentrated solutions. This stems
from the fact that ion-to-ion and ion-to-solvent interactions play larger roles in concentrated
solutions. Thus, the molar conductivity for an individual ion in solution may only be truly
measured if there are no other interfering ions present, i.e. in an infinitely dilute solution.
Inserting equation 13 into 12 and subsequently into 9 yields:
åc l
1
Lm = (14)
c i i i
If we note for a M+X- electrolyte, where + and - are the valences of the cation and anion
respectively, that the concentration of each ion will depend upon the total concentration of the
salt in solution, at infinite dilution this results in the simple relation of equation 15:
L ¥m = n+ lM¥+ + n- l ¥X- (15)
The ionic molar conductivities at infinite dilution are tabulated in various sources and allow for
the calculation of the solution specific conductivity directly [CRC, 2014]. It should be noted that
the above equations apply for solutions at infinite dilution but may be used directly with small
error up to total solution concentrations of approximately 10-3 mol/L. Table 1 shows the ionic
molar conductivity of cations and anions relevant to the operation of CANDU reactors. Note
that the ionic molar conductivity of each ion is per equivalent since the total migration of an ion
subjected to an electric field will be dependent upon its charge (valence). Note also the compa-
ratively large conductivities of H+, D+, OH- and OD-, which is due to the Grotthus “hopping”
mechanism for these ions in aqueous solution where the individual ions themselves do not
physically migrate but they are exchanged through the interconnecting hydrogen-bonded
structure of liquid water. These ions tend to dominate the conductivity of the relatively pure
CANDU process solutions.
Table 1. Conductivity at infinite dilution in H2O [CRC, 2014].

Cations (S cm2 mol-1 eq-1) Anions (S cm2 mol-1 eq-1)
H+ 349.7 OH- 198
D+ 249.9 OD- 119
+ -
Li 38.7 Cl 76.31
+ 2-
Na 50.1 ½ SO4 80.0
3+ -
1/3 Gd 67.3 HCO3 44.5
+ 2-
NH4 73.5 ½ CO3 69.3
½ Fe2+ 54 NO2- 71.8
+ -
N2H5 59 NO3 71.48
Example 15.1- calculating solution conductivity from pH and values at infinite dilution.
Lithium hydroxide is added to the light water in the calandria vault and end shield cooling
system to control the pH and minimize corrosion of the carbon steel components. The pH
specification is 9.5. Calculate the required lithium ion concentration in the water (in ppm) to
attain the pH specification and estimate the solution conductivity assuming no other impurities
are present.
Solution
Since the pH specification is given (9.5), the concentration of H+ and OH- are readily evaluated
from the definition of pH and Kw:
pH = -log10 ([H+ ])
thus: [H+ ] = 10-pH = 10-9.5 = 3.16 ´10-10 mol/kg
recalling the dissociation of water (Kw, eq. 4), the OH- concentration is:
- Kw 10-14
[OH ] = + = -10
= 3.16 ´10-5
[H ] 3.16 ´10 mol/kg
In this case, lithium hydroxide is a strong base and completely dissociates into lithium and
hydroxyl ions:
LiOH  Li+ + OH-
thus, the concentration of lithium must be equal to the hydroxyl concentration in the water, i.e.:
[Li+]=[OH-] = 3.16 x 10-5 mol/kg
Converting to ppm (mg/kg):
mol Li 6.94 g 1000 mg

[Li+ ] = 3.16 ´10-5 ´ ´ = 0.219 mg / kg
kg H2O mol g
Ans.
To calculate the expected conductivity, combining equations 12 and 13 results in:
k = å c il i
i
Conductivity values for Li+ and OH- can be attained from Table 1, and since [Li+]=[OH-]:
(
k = cLiOH lLi¥+ + l ¥OH- )
Since the units of conductivity are in cm2, the concentration is converted to mol/cm3 for consis-
tency:
mol 1000 kg m3 L 1mL

c LiOH = 3.16 ´10-5 3
= 3.16 ´10-8 mol / cm3
kg m 1000 L 1000 mL 1cm3
Calculating the conductivity and converting to standard units of mS/m:
mol æ Scm2 ö
k = 3.16 ´10 -8 ç 38.7 +198 ÷ = 7.48 ´10-6 S / cm
cm3 è moleq ø
thus, the conductivity of the cooling water is expected to be 0.748 mS/m. Ans.
2.2.2 Cationic Conductivity or Conductivity After Cationic Exchange (CACE)
While the conductivity of a solution provides an indication of the concentration of impurity ions
that are present, it is not species- or ion-specific; thus, there is no way of determing from a
conductivity measurement what are the individual dissolved ionic species present in the solu-
tion. One way of providing a better indication of the concentration of anionic impurities in an
aqueous solution is to use the cationic conductivity (or conductivity after cationic exchange -
CACE) [IAPWS, 2012], which gives a direct measure of the anionic impurities in the system that
may be aggressive to the corrosion of plant materials. Cationic conductivity is simply the
conductivity of an aqueous solution after it has been passed through a strong acid ion-exchange
column to remove the cations (such as Na+, K+, Ca2+ etc.), which replaces them with the proton
(H+). The solution that elutes from the cation exchange column will contain the same concentra-
tion of anions (Cl-, SO42- etc) as the original sample and can provide a rapid and direct indication
of the rate of increase of these impurities. This gives a very useful online measurement of
impurity ingress, specifically in the case of a condenser leak.
2.3 Purification and Ion Exchange Resin

Filter media and ion exchange resins are used extensively in the nuclear and power generating
industry for removing particulate and ionic impurities from a solution. All power plants will have
a water treatment plant that includes ion-exchange as a final “polishing” method to ensure
extremely pure water is available for use in the various process systems and each process system
will contain filters of various configurations to collect suspended solids. Ion exchange media
come in many forms and can be tailored for removal of specific impurity cations or anions as
appropriate. The main classes of ion exchange media are designated as strong acid, weak acid,
strong base and weak base and each may be used in a water treatment plant or a purification
stream in the various process systems of a nuclear power plant.
2.3.1 Structure of Ion Exchange Resin

An ion exchange resin consists of a co-polymer, typically polystyrene and divinyl-benzene that
have functionalized “exchange” groups attached to each benzene ring in the polymer matrix. A
strong acid resin, suitable for removing alkali earth elements and transition metals from a water
stream, is typically sulfonated producing an SO3-/H+ functional group on each benzene ring (note
that a perfect one-to-one aromatic ring to exchange group ratio is not achievable). The proton is
the ion-exchange cation that will be exchanged with other cations from the solution. Strong
base resins typically consist of a quaternary ammonium exchange site attached to two or three
methyl groups whereby the functional exchange site is a chloride (Cl-) or hydroxyl anion (OH-)
that acts to exchange with the impurity anions in the process stream to be purified. For nuclear
plant process systems, where water purity is paramount, the resins are typically supplied and
used in the H+ and OH- forms, although different suppliers can provide functionalized exchange
groups such a lithiated (Li+) strong cation resin for use in particular systems such as the primary
heat transport system. It is important to note that before use of an ion exchange resin in the D2O
of the primary coolant or the moderator system, resins must be deuterated to the D+ and OD-
form by plant operators to ensure that the isotopic purity of the D2O is not significantly affected
by H2O exchange with the resin (see discussion in Section 8.2).
Once fabricated, cation and anion exchange resins appear as small beads (typically ~0.5 – 1.5
mm in diameter) and may be slurried in and out of their process vessels for easy replacement
and regeneration. Traditional resins were specified as “gel-type” as they were effectively a
wrapped-up sphere of polymer chains with significant porosity and tended to break apart if
continuously reused or placed in high flow rate process streams. Newer, “macroporous”, resins
are much stronger and have shown good resilience in harsh environments.
2.3.2 Ion exchange capacity

The purification of a process water stream depends upon the type and quantity (or volume) of
ion exchange resin used as well as its capacity to remove various ions from the water. Ion
exchange capacity is defined as the number of ionic “equivalents” that can be exchanged per
litre of resin used. For example, a strong acid resin in the protonated form (H+) can have an
exchange capacity of around 2 eq/L; thus, for every litre of the resin used in a purification vessel
two equivalents (or moles) of H+ cations are available to exchange with cationic impurities in the
process stream. If sodium (Na+) is used as an example, one litre of this strong cation resin may
remove 2 mols of Na+ cations, or effectively 46 grams of sodium.
The exchange capacity is also dependent upon the particular cation being exchanged with higher
valence cations showing the greater preference for adsorption on the exchange site. An example
of cationic preference is:
Fe3+ > Ca2+ > Ni2+ > Cu2+ > Fe2+ > K+ > NH4+ > Na+ > H+
For a strong base ion exchange resin, an example of anionic preference follows:
SO42- > CO32- > HSO4- > NO3- > Cl- > HCO3- > CHOO- > CH3COO- > F- > OH-
Thus, a ferrous or cuprous cation would tend to displace a sodium cation from the ion exchange
bed or a sulfate anion would displace a chloride or hydroxyl anion. The capacity of an ion-
exchange bed can be described in a manner similar to a chemical equilibrium equation, as
shown in equation 17 using the exchange of the sodium cation with the proton as an example.
Na+ +RH  H+ +RNa (16)
The equilibrium constant, in this case called the selectivity coefficient, is given by:
[RNa][H+ ]
K Na = (17)
H
[RH][Na + ]
where [RNa] and [RH] designate the concentration (mol/L) of the specific cation that is adsorbed
on the resin and [Na+] and [H+] are the incoming sodium ion and proton concentrations in the
solution to be purified. It should be noted that a fresh ion exchange column will have a fixed
number of exchange sites based upon the resin’s exchange capacity and the amount of resin
present; the balance between the [RNa] and [RH] must always be equal to the initial total
exchange capacity. Once most or all of the proton exchange sites available have been used or
taken up by another cation, the resin will start to “throw” the cations of lowest selectivity back
into the process water instead of removing them. This is obviously not a desirable situation as
the intent of the purification system is to remove the impurity cations from the process water
stream; at this point, the ion exchange column or vessel is said to be “spent”.
2.4 Water Radiolysis

The chemistry of CANDU process systems is specified to protect the reactor core and steam
generator materials from localized corrosion, minimize deposition of corrosion products on the
fuel and limit the corrosion of system components. The pH of irradiated CANDU process sys-
tems is typically adjusted by using lithium hydroxide to produce mildly alkaline conditions and
some irradiated systems are kept under reducing conditions by excluding oxygen to prevent
elevated electrochemical corrosion potential (ECP) and excessive corrosion. For in-core systems,
the water is continuously bombarded by an intense radiation field of high-energy gamma and
neutrons, which breaks the chemical bonds and produces highly reactive radical species. This
process is known as water radiolysis. The radiolysis of pure water results in the production of
hydrogen (H2) (or deuterium D2, for heavy water), oxygen (O2) and hydrogen peroxide (H2O2) (or
deuterium peroxide D2O2 in heavy water) [Spinks & Woods, 1990]. The radiolytic production of
these oxidizing species has a direct impact on the corrosion of system components; they elevate
the ECP, increasing the possibility of stress-corrosion cracking of the alloy steam generator tubes
and, under low-temperature shut-down conditions when the radioactive decay of in-core
components and fuel still produces intense gamma radiation, they promote the pitting of carbon
steel. Radiolytically generated hydrogen peroxide will also degrade ion-exchange resins.
The radiolytic production of hydrogen, oxygen and hydrogen peroxide can be managed by
adding hydrogen to the system. The elementary chemical reactions and kinetics are complicated
but the net result is that the added hydrogen molecules react with the radical species and
mitigate the radiolytic production of oxygen and hydrogen peroxide. The basic mechanisms for
these processes are explained below with the overall result being suppression of the net radi-
olytic production of oxidizing species. Further details of the radiation kinetics associated with
water radiolysis can be found in the AECL report by David Bartels [Bartels, 2009] or in the book
chapter by George Buxton [Buxton, 1987].
Upon absorbing the energy dissipated by a particle or photon, water breaks down into several
radical species as primary products of the irradiation process; on a timescale of approximately
microseconds, the overall result is:
H2O + radiation ® e-aq ,H+ , ×H, × OH, ×HO2 ,H2 ,H2O2 (18)
Reactions of these primary products can re-form water, and produce further hydrogen and
hydrogen peroxide through equations 19 - 21:
×H+ ×OH ® H2O (19)

×H+ ×H ® H2 (20)
×OH + ×OH ® H2O2 (21)
Oxygen is produced as a secondary product through reactions of the radical species with the
hydrogen peroxide molecule:
×OH+ H2O2 ® ×HO 2 +H2O (22)

×OH + ×HO 2 ® O2 + H2O (23)
×HO2 + ×HO2 ® O2 + H2O2 (24)
As mentioned above, providing sufficient added hydrogen to the water can mitigate the net
production of oxygen and hydrogen peroxide. Hydrogen reacts with the hydroxide radical
forming a hydrogen atom that readily combines with hydrogen peroxide and/or oxygen, effec-
tively removing them from the solution and mitigating their overall production as shown in
equations 25 - 27. The overall effect is a chain reaction where additional hydroxide radicals are
produced (eq. 27) continuing the reaction chain, provided sufficient hydrogen is present. From a
kinetic standpoint, the concentration of hydrogen in the water needs to be sufficient to make its
rate of reaction with the hydroxyl radical (eq. 25) faster than the rate of reaction of the hydroxyl
radical with hydrogen peroxide (eq. 22). This results in the reduction of the concentration of the
hydrogen peroxide molecule and effectively suppresses the conversion of hydrogen peroxide to
oxygen, removing both species from the solution. When sufficient hydrogen is added to an
irradiated aqueous system, the very simplified chemical relation shown in equation 28 is the net
result of the radiolysis process.
×OH+H2 ® ×H+H2O (25)

×H+ O2 ® ×HO 2 (26)
×H+H2O2 ® ×OH +H2O (27)
H2O +H2 (excess) + radiation ® H2O +H2 (excess) (28)
The fact that added hydrogen is required in the water to promote the desired recombination
reactions implies that there must be a minimum hydrogen concentration by which the overall
rate of the reaction in equation 25 is just sufficient to overcome the reaction of the hydroxyl
radical with hydrogen peroxide (equation 22) and hence suppress the production of oxygen.
This minimum concentration is called the “critical hydrogen concentration” (CHC) and has been
measured in flow loops (both fuelled reactor systems and in gamma cells) to be around 0.5
mL/kg (~2x10-5 mol/L). [Elliot & Stuart, 2008]. Operating with the hydrogen concentration
above the CHC with no oxidants initially present leads to a system in net radiolytic suppression
while operation with an initial hydrogen concentration below the CHC (or with the presence of
significant quantities of oxidants or impurities in the water) leads to net radiolysis occurring
within the system.
Impurities dissolved within water will influence the water radiolysis processes. For example,
impurities such as nitrate and nitrite anions will interfere with the water recombination reac-
tions since they tend to react with and consume the hydrogen atoms and hydroxyl radicals
[Yakabuskie, 2010], effectively lowering their concentrations and the overall rate of the oxygen
and hydrogen peroxide consumption reactions (eq. 26 & 27). This will result in an increase in
the quantity of oxygen and hydrogen produced by water radiolysis in the presence of nitrate
and/or nitrite ions. Chloride ions interfere with the hydrogen recombination reactions since
they readily exchange their outer valence electron with the hydroxyl radicals interfering with the
primary recombination reaction (eq. 19). The chloride ion is then regenerated through electron
exchange with water molecules and can continuously impede the desired recombination
processes. Organic impurities from compressor leaks, oil ingress from pumps or other mechani-
cal systems will break down rapidly in a radiation field. High molecular weight organics undergo
polymerization reactions, extending their size and weight leading to plugging of filters and other
flow paths. Lower molecular weight organics decompose producing hydrogen and carbon
dioxide that can rapidly lead to hydrogen excursions and reduction in system pH.
The chemistry specifications for a water-filled, irradiated CANDU system must therefore be
designed to prevent the build-up of the water radiolysis products in the system, or the effects of
water radiolysis must be continuously managed to maintain system chemistry specifications.
Some systems, such as the primary heat transport system and the end shield cooling system, are
operated under net radiolytic suppression by ensuring the hydrogen concentration in the water
is above the CHC. Other systems, such as the moderator, liquid zone control and the spent fuel
bay, are operated with net water radiolysis occurring. For these systems, the hydrogen and
oxygen produced are managed by ensuring impurity concentrations are kept extremely low and
by providing catalytic recombiner units in their cover gas flow paths to ensure flammable
concentrations are not exceeded. Further details of the operation of each of these systems are
found later in this chapter.
2.5 Solubility of Gases in Water – Henry’s Law

Gases are typically sparingly soluble in water due to the vapour-liquid equilibrium established
between a gas mixture covering the liquid sample. From thermodynamics, the chemical poten-
tial of the chemical species must be equal at the interface between the gas and the liquid
meaning that, even though the gas solubility may be low in the liquid phase, it is finite and is an
important parameter to consider in nuclear plant process streams. Many CANDU process and
auxiliary systems have a head tank and/or a cover gas space that operates at a particular
pressure and the solubility of the gas in the liquid phase and its mass transfer and migration to
or from the cover gas space needs to be readily monitored. Using the example of hydrogen
production by the water radiolysis processes described above, the continuous irradiation of
CANDU coolant streams will result in the build up of hydrogen in the water that will equilibrate
with its surroundings resulting in hydrogen release to the cover gas space. This needs to be
continuously monitored to ensure that the hydrogen concentration is maintained below the
flammability limits.
The equilibrium established between a gas dissolved in water and its partial pressure in the gas
space above the liquid is given as:
ygas p gas
kH = » (29)
x gas c gas
where ygas is the mole fraction of the given gas in the vapour space and xgas is the mole fraction
in the liquid phase. Since the gases of interest in CANDU process systems are sparingly soluble
the mole fractions are typically equated to partial pressure (pgas) and concentration (cgas) for the
gas phase and liquid phase respectively giving units of pressure (atm or MPa) per molarity
(mol/L). Rearrangement of equation 29 shows that, under low to intermediate partial pressures,
there is a linear dependence between the partial pressure of the gas above the liquid and that
dissolved in it. Thus, by knowing the equilibrium constant, in this case called Henry’s Law
constant, the concentration of gas dissolved in water may easily be evaluated as demonstrated
in Example 15.2. Henry’s law constants at 25oC are shown in Table 2 for some gases commonly
encountered in CANDU process systems. IAPWS has released detailed correlations for determin-
ing the Henry’s Law constant for various gases as a function of temperature [IAPWS, 2004].
Table 2. Henry’s constants for select gases at 25oC (calculated from IAPWS , 2004).
Gas kH in H2O (in D2O)
(atm/M)
O2 775.7
H2 1263
N2 1522
He 2541 (2343)
CO2 29.40
D2 (921.0)
Example 15.2 - Henry’s law calculation – D2 in heavy water
The heavy water storage tank for the primary heat transport system contains D2O at roughly
room temperature and is purged with a helium cover gas to maintain the hydrogen concentra-
tion below 4% (by volume). The specification for dissolved deuterium in the PHTS is 3 – 10
mL/kg, calculate the equilibrium concentration that would be attained in the cover gas if helium
is not frequently added and purged and the dissolved hydrogen is maintained at the upper limit
of the specification.
Solution
If 10 mL/kg of dissolved deuterium is maintained in the PHTS, Henry’s law can be used to
estimate the equilibrium concentration in the storage tank cover gas. First the dissolved deute-
rium concentration must be converted to consistent units with Henry’s law (i.e. Molar or mol/L)
using the conversion factor that 1 mol gas = 22.4 L gas at standard temperature and pressure
(STP). Thus:
LD 1mol D 1000 kgD O m3

c gas = 0.010 2
´ 2
´ 2
´ = 4.46 ´10-4 mol / L (M)
kg D O
2
22.4 L D2 m3D O 1000 L
2
( )(
pgas = k H c gas = 921atm / M 4.46 ´10 -4 M = 0.411 atm )
For a cover gas operating at approximately atmospheric pressure this would amount to a cover
gas concentration of 41.1%, well above the flammability limit of about 8% in helium!
3 Primary Heat Transport System
As explained in Chapter 10, the purpose of the Primary Heat Transport System (PHTS) in a
CANDU reactor is to remove the heat generated from the fissioning of the reactor fuel and
transport it to the steam generator for production of steam in the Secondary System. The
materials of construction in the PHTS are numerous due to the different functions of the various
components; the particular materials and their properties are described in detail in Chapter 14.
The materials with the largest surfaces areas in contact with the heavy water coolant of the
PHTS are: zirconium-based alloys in the core of the reactor (Zircaloy-4 fuel cladding, Zr-2.5Nb
pressure tubes), nickel alloys for the steam generator tubing (Alloy 400 [Monel] at Pickering,
Alloy 600 [Inconel] at Bruce, Alloy 800 [Incoloy] at Darlington and the CANDU-6s), type 410
stainless steel for the fuel-channel end fittings and carbon steel for the feeder pipes joining the
fuel channels through the headers to the carbon-steel channel heads in the steam generators.
Thus, as with any complex system, the chemistry control practices are a compromise among the
optimum chemistries for each of the major materials used.
The Zircaloy-4-clad fuel bundles, Zr-2.5Nb pressure tubes and nickel-alloy steam generator tubes
exhibit low general corrosion rates over a wide range of pH and temperature [Cox, 2003]. A
major role of chemical control is therefore to protect against localized corrosion, in particular
the stress-corrosion cracking of the steam generator tubes and the hydriding and cracking of the
pressure tubes. As described earlier in the section on radiolysis, adding hydrogen maintains
reducing conditions and this minimizes the possibility of nickel-alloy cracking, but to minimize
the potential for hydriding of Zr-2.5Nb the hydrogen is kept within strict limits (3-10 mL/kg).
A further role of the chemical control in the PHTS is the protection of the large surface area of
carbon steel. As described in Section 6.1 in Chapter 14, the general corrosion of iron produces
mixed and, at low temperature, often hydrated oxides (Fe2O3, Fe3O4, FeOOH etc.) that become
the thermodynamically stable phases at a pH (at room temperature) greater than 9 or so. It is
well known that carbon steel corrosion is minimized in mildly alkaline, deaerated water. Howev-
er, at low temperature with static water that may occur during a shut-down, oxidising conditions
can induce severe pitting of carbon steel and should be avoided. Note that during reactor
operation the inlet feeders in CANDUs are exposed to coolant saturated in dissolved iron
because of the heat transferred producing steam in the steam generators and the consequent
drop in iron solubility, so they undergo general corrosion and develop thick magnetite films as
the result of magnetite precipitation from the oversaturated coolant. On the other hand, the
outlet feeders see coolant undersaturated in iron because of the heating in the core and as a
consequence undergo flow-accelerated corrosion (FAC), which leads to very thin magnetite films
and loads the system with iron that largely deposits as magnetite in the steam generators.
Oxidising conditions tend to mitigate FAC, but the greater need to avoid localized corrosion of
the alloy components in the core and the steam generators dictates the use of reducing condi-
tions.
3.1 Chemistry Control in the PHTS

As explained above, one of the primary objectives of chemistry control in the PHTS of a CANDU
reactor is to minimize the corrosion of the carbon steel, zirconium and alloy surfaces, which
involves operating under alkaline conditions and mitigating the radiolytic production of oxidizing
species. Other objectives include minimizing deposition of corrosion products on the fuel
(fouling) and minimizing and controlling the concentration of activiated corrosion products and
fission products (both gaseous and dissolved) in the system. These objectives are accomplished
through dosing and control of the primary coolant’s pHa and deuterium concentration through
regular additions of lithium hydroxide and hydrogen gas. Also, the molecular hydrogen gas
exchanges rapidly in the reactor core with deuterium in the D2O molecules. Although the
additions downgrade the heavy water slightly, this is compensated by periodic isotopic upgrad-
ing of the system’s D2O (see Section 8.1). Each individual plant maintains its own chemistry
practices and operational guidelines but, in general, guidelines for these values are shown in
Table 3.
Table 3. Target chemistry parameters in the PHTS.

Parameter Typical Specification Range
pHa: 10.2 – 10.4
[Li+]: 0.35 – 0.55 mg/kg (ppm)
[D2]: 3 – 10 mL/kg
conductivity: 0.86 – 1.4 mS/m (dependent upon LiOH concentration)
Dissolved O2 < 0.01 mg/kg
[Cl-], [SO42-] etc < 0.05 mg/kg
Isotopic > 98.65 % D2O
Fission products ALARA (< 106 Bq/kg D2O; monitoring I-131 indicative of fuel failure)
During steady-state operation, variations in any of the above parameters are typically small in
the absence of system transients or upset conditions. Oxygen concentrations are typically non-
detectable. The alkalinity is controlled through periodic additions of LiOH through the sampling
system return (for elevating pHa) or by providing a periodic bleed flow through the purification
system using an ion-exchange column containing strong acid cation resins (for lowering pHa).
Lithiated mixed-bed ion-exchange columns (where the strong-acid IX resin D+ sites have been
saturated with Li+; both acid and base resins are deuterated) are run as normal purification for
the PHTS to collect cationic and anionic impurities. A simplified flow diagram for a PHT purifica-
tion circuit is shown in Figure 2. The purification system keeps particulate concentration low and
helps to maintain the anioic impurity concentrations below the specification to minimize the risk
of SCC of the stainless steel and alloy components in the system and to minimize the risk of
strain-induced cracking of carbon steel components [Turner & Guzonas, 2010]. It also serves to
keep the radioacitivity of the PHT coolant low as the filters and ion exchage column may capture
activiated corrosion products or ionic fission products released from failed fuel bundles.
Figure 2. Simplified flow diagram of a PHT purification loop (courtesy of AECL).
Since pHa and conductivity are directly related to [Li+], purity of the PHTS coolant is readily
observed by comparing the measured values to the theoretical values as calculated using the
equations developed earlier in the chapter. If the system is controlled well and has minimal
impurities, then the conductivity and pHa measured should match closely with the calculated
values appropriate to the measured lithium-ion concentration. Deviations in the measured-to-
theoretical value indicate ingress of ionic impurities or problems with system sampling tech-
niques.
The above discussion seems to imply that sampling the high-temperature PHTS system is routine
and simple. While it is routine, collecting and analyzing samples from a high-temperature, high-
pressure system is far from simple. Take, for example, the collection of a sample in an open
polypropylene jar. The sampled water must first be cooled from the operational temperature
and throttled to low pressure. Each of these operations changes the state of the chemical
species present in the sample. Once the valve at the end of the sample line is opened, the
ideally oxygen-free coolant sample is exposed to the atmosphere and will readily absorb nitro-
gen, oxygen and carbon dioxide from the air. Thus, this sample is immediately not appropriate
for obtaining a dissolved oxygen measurement. Additionally, the absorbed CO2 dissociates in
the sample solution to form carbonate (CO32-) and bicarbonate (HCO3-) anions, which disturb the
equilibrium chemistry established between lithium hydroxide and water. The CO2 absorption
effectively produces carbonic acid and lowers the pH (or pHa) of the sample, rendering it com-
promised for system monitoring. Thus, differences in pHa, conductivity and [Li+] from PHTS
samples may imply improper sampling techniques, not problems with system chemistry and
care must be taken to identify the root cause. Several standard methods and recommended
practices are available to properly specify and conduct sampling campaigns including the
Technical Guidance Document on corrosion product sampling recently released by IAPWS
[IAPWS, 2014].
3.2 Corrosion Issues in the PHTS

3.2.1 Feeder Pipe FAC and Cracking
Recent trends for PHTS chemistry practice have focused upon “narrow
“narrow-band”
band” pHa control, within
the 10.2- 10.4 range, in order to help mitigate excessive FAC on the outlet feeder pipes, which is
demonstrated to increase with increasing pHa (Lister, Lister, 2002, Slade & Gendron, 2005). This
observation results from the fact that the magnitude of FAC is partpartially
ially controlled by the degree
of corrosion-product under-saturation
saturation in the coolant, via processes described by equations 21-
23 presented in Chapter 14. Consider the closed
closed-loop
loop PHT system and take the PHT main pump
as a starting point. The temperature at this point is typically around 260oC and is considered to
be the “cold-leg”
leg” of the PHT circuit. In these conditions, the PHT coolant is typically saturated
sat or
over-saturated in corrosion-products
products (mainly dissolved iron), which leads to the precipitation of
magnetite on the surfaces up to the inlet of the reactor core and promotes low corrosion rates
of the carbon steel inlet feeder pipes (the precipitation
tion actually begins within the steam genera-
gener
tor and along the whole inlet section of feeders and headers to the core). Shown in Figure 3 is
the calculated solubility of magnetite at a pH of 10 10.2 and 9.0 (in light water) between the
o o
temperatures of 250 C and 310 C, the approximate temperature of the coolant exiting the core.
It is seen that for the elevated pH, the solubility increases over this temperature range, which is
considered to be beneficial since this will typically prevent d deposition
eposition of corrosion products in
the reactor core. The coolant ultimately leaves the core in an under under-saturated
saturated condition since
most of the in-core surfaces contacting the coolant are zirconium alloys and this leads to a large
driving force for magnetitee (and metal) dissolution in the end fittings and carbon steel outlet
feeder pipes. Compounding the problem is the intricate series of bends that are required for the
feeder pipes to connect appropriately to the end end-fittings – many outlet feeder pipes can have
two tight-radius
radius bends. This acts to increase the fluid turbulence and exacerbates the FAC
phenomenon due to the action of fluid shear and mass transfer.
Figure 3. Magnetite solubility calculated from thermodynamic data optimized to the Tremaine
& Leblanc data [Tremaine, 1980].
The Point Lepreau CANDU-6 6 was the first reactor to exhibit FAC of the outlet feeder pipes. The
feeders were on average corroding at rates greatly in excess of what was assumed and predicted
in providing a suitable corrosion allowance and many sections of outlet feeders needed to be
replaced well before the end of their design lifetime. Unlike other CANDU stations that also
observed FAC in the outlet feeder pipes, the Point Lepreau station also had demonstrated
cracking, which necessitated many removals and replacements prior to the refurbishment
outage that began in 2008. The stress-corrosion cracking (SCC) or environmentally assisted
cracking (EAC) observed at Lepreau was very unexpected as carbon steel is not typically suscept-
ible to significant cracking. Theories supporting mechanisms such as hydrogen embrittlement
due to the high corrosion rates and hydrogen production associated with the outlet feeder FAC
were proposed along with a detailed examination of the construction and operational history
[Slade & Genderon, 2005]. Evidence pointed to high residual stresses from the bending, welding
and fabrication process as a primary culprit to the cracking mechanism and the fact that none of
the bends were stress relieved following fabrication. No direct evidence of a connection with
hydrogen embrittlement related to FAC was found, however it is still suspected to have played a
factor. The material for replacement sections of pipe, and that used for new construction and
refurbishment projects, is now specified to contain a higher chromium content (~0.3 wt%) and
has slightly more carbon (~0.35% vs 0.3%) which, in combination act to significantly reduce the
FAC issue and increases the yield strength respectively, thereby creating more margin to protect
against cracking. All new bends are also fully stressed relieved before being placed into service.
3.2.2 Delayed Hydride Cracking of Zirconium Alloys

Of the corrosion issues of most concern in the PHT system, perhaps the most critical is delayed
hydride cracking (DHC) of the Zr-2.5Nb pressure tubes as they represent the physical barrier and
pressure boundary of the reactor core. As described in Chapter 10, the pressure tubes in a
CANDU reactor are sealed on each end through a rolled-joint connection to the 403SS end-
fittings. Deuterium produced through the limited corrosion of the Zr-2.5Nb pressure tubes has
been shown to accumulate within the alloy; its accumulation is readily measured through scrape
samples obtained during maintenance outages. Over time, should the solution solubility limit of
deuterium in the alloy be exceeded, zirconium-hydride platelets may form within the metal’s
lattice and between grain boundaries. The hydride is a brittle, ceramic-like material that is
prone to cracking under stress as was demonstrated on several of the original Zircaloy 2 pressure
tubes in the Pickering A station that were replaced as a result of DHC in the period between
1974 – 1976 [Urbanic, 1987]. The most prone locations are at the rolled joint areas on each end
of the pressure tube due to the increased deuterium production and migration rate at the
galvanic couple between the pressure tube and the stainless steel end-fitting – deuterium
produced through corrosion of the stainless steel is free to migrate into the pressure tube due to
the intimate contact resulting from the rolled-joint seal. Residual stress due to the rolling
process also plays a large role in these locations.
3.3 Activity Transport

The PHTS coolant in a water-cooled reactor such as CANDU may become activated itself through
the absorption of neutrons as it passes through the reactor core. This is particularly relevant in
CANDU reactors where heavy water is used as the reactor coolant and moderator (in the
separate moderator system described in Section 5) since neutron absorption by the deuterium
atoms in heavy water will directly produce tritium (T or H3), which is radioactive and has a 12.3
year half life. Tritium production and decay in the PHTS heavy water means that, even in the
absence of particulate or ionic impurities, the coolant will be radioactive and the entire PHTS
will have significant concentrations of tritium, typically of the order of 104 Bq/kg of heavy water.
This is several orders of magnitude lower than the steady state tritium concentrations in the
moderator water since only about 3% of the PHTS coolant is in the reactor core at any given time
whereas >95% of the moderator heavy water is continuously being exposed to the high neutron
flux in the reactor core. Thus, all reactor systems that employ heavy water as their coolant will
contain significant tritium activity in all sections of the circuit.
In addition to the production of tritium, the PHTS heavy water coolant will contain trace concen-
trations of dissolved ions and particulates that are the result of the corrosion and wear of the
system materials. These impurities, which include iron, nickel, chromium, cobalt and antimony
can deposit in the core and may become activated by neutron absorption. In addition, zirco-
nium alloy wear products released by the movement of fuel inside the fuel channel during
refuelling, and fission products and actinides released from the (infrequent) failure of the fuel
cladding, can be transported out of the core by the coolant. These species are easily deposited,
adsorbed and incorporated into the oxide layers that form on the surfaces of out-of-core
components leading to elevated levels of activity on components removed from the direct
radiation field of the reactor core, the reactor inlet/outlet headers and feeders and inner
surfaces of the steam generator tubes, for example.
Typical activation and fission products observed in the PHT and their half-lives are shown in
Table 4. Note the relatively long half-life of Co-60, which is a major contributor to radiation
fields in out-of-core components. The fission products are generally kept at very low concentra-
tions in the PHT coolant since they are indicators of failed fuel elements, which are removed as
quickly as possible once identified. The radioiodines are of particular concern since they are
readily absorbed in the thyroid of humans and must be contained in the system and disposed of
appropriately.
Table 4. Common radioactive isotopes and their source in the PHTS.
Activation Typical source Half Life

product
Cr-51 Alloys 27.7 days
Fe-59 Steels and alloys 44.6 days
Sb-124 Impurity in steels and alloys – bearings and wear surfaces 60.2 days
Co-58 Nickel alloys 70.8 days
Mn-54 Steels and alloys 312.5 days
Co-60 Impurity in steels and alloys – hard-facing materials for wear 1924 days
resistance (e.g. Stellites)
I-131 Fission product 8.04 days
Xe-133 Fission product 5.24 days
Xe-135 Fission product 9.1 hours
Kr-85 Fission product 10.73 years
Kr-88 Fission product 2.84 hours
4 Secondary Heat Transport System
The Secondary Heat Transport System (commonly called the secondary system or steam cycle)
produces the steam necessary to drive the turbines and electrical generator. The configuration
of the secondary system is described in detail in Chapter 8 but in summary it contains the
condenser, a series of low-pressure feedwater heaters, a deaerator, a series of high temperature
feedwater heaters, the boilers or steam generators, the steam supply piping and control sys-
tems, and the high- and low-pressure turbines with a moisture separator and steam reheater in
between them. While the secondary system at each plant is unique in terms of its exact configu-
ration and the materials used for the various components there are typically two classifications
of system; all-ferrous and copper-containing. The operating chemistry for the secondary system
is dependent upon the type of materials from which the plant is constructed. For example, as
described in the corrosion section in Chapter 14, it is desirable for iron-based materials to
operate with alkaline chemistry to promote the formation of passive oxide films and minimize
corrosion. This is often done with the addition of ammonia or other volatile amines and target
pH values (measured at room temperature) can be up to 10. In a copper-based system or a
system that contains some copper components, copper corrosion is known to be accelerated
considerably by ammonia, especially when oxygen is present, and maximum pH’s must be kept
to about 9.2 – 9.4 and the use of ammonia minimized or excluded all together. Reducing condi-
tions are maintained, mainly to protect the steam-generator alloys from cracking, by the addi-
tion of hydrazine.
4.1 Chemistry Control in the Secondary System

In order to protect the entire secondary system from corrosion, a volatile pH-controlling agent is
employed, i.e., one that enters the steam phase in the boilers and is carried through the entire
system such that it protects the main steam lines, steam extraction lines, moisture separator and
reheater and the condenser. Previously, the chemical buffer sodium phosphate was in wide-
spread use for dosing feedwater but was effectively non-volatile and concentrated in the steam
generator, where its reactions within crevices and under deposits could lead locally to either
highly alkaline or acid conditions if not properly adjusted and monitored. Extended pitting or
wastage of nickel-alloy steam generator tubes, especially under the “sludge pile” on the tube
sheet, was a not-infrequent occurrence. Nowadays, it is common practice to use “all-volatile”
treatment (AVT), with hydrazine for oxygen control and a volatile base such as ammonia to
distribute the alkalizing agent around the system. Ammonia, however, is so volatile that it
concentrates in the steam phase and can leave the water-touched areas unprotected, so less-
volatile amines such as morpholine (the cyclic compound O(C2H2)2NH, ethanolamine
(HO(CH2)2NH2) or cyclohexylamine ((CH2)5CNH2) may be used instead of, or in combination with,
ammonia. A drawback of these higher-molecular-weight organic compounds is that they break
down at high temperature to simpler substances such as acetic acid and eventually carbon
dioxide, which may be corrosive; their concentrations therefore have to be limited in the steam
generators through strict chemistry control.
Chemistry parameters that are targets for the secondary system chemistry control and are
typically measured at the high pressure feed water heater outlet are shown in Table 5.
Table 5. Target chemistry parameters in the secondary system (all-ferrous materials).

pH 9.5 < pH < 10 (for all-ferrous systems)
Dissolved O2 < 0.01 mg/kg (ppm)
Hydrazine 0.020 – 0.030 mg/kg (ppm)
Na+ < 0.05 g/kg
Cl- ; SO42- < 0.05 g/kg
The distribution of amine retained in the water phase or stripped to the steam phase is de-
scribed by the distribution coefficient, which is dependent upon both temperature and concen-
tration. The true distribution coefficient is defined as the ratio of the vapour-phase mole
fraction of volatile species (yB) to its liquid phase mole fraction (xB):
yB
DB = (30)
xB
Note that, in the true distribution coefficient, the measure is per fraction of the volatile species.
For ammonia this means the neutral species, NH3. However, due to the dissociation of ammonia
and other amines in water, the true distribution coefficient is not typically what is measured; the
measurement is the “apparent” distribution coefficient. The apparent distribution coefficient
accounts for the total concentration (or fraction) of ammonia-based species in the liquid phase
including the dissociation products as shown in equation 31. For example, the equilibrium
established between ammonia and the ammonium cation in water is given by equation 32 and
its equilibrium constant in equation 33. If ammonia is stripped to the vapour phase, its liquid
phase concentration will be reduced and will subsequently affect the concentration of the
ammonium cation retained in solution. As with other equilibrium chemical equations that have
been discussed in this chapter, the equilibrium constant for amine dissociation is temperature-
dependent, as are the distribution coefficients, thus the apparent distribution coefficient is a
complicated function of the true distribution coefficient and equilibrium chemistry.
[NH3 ]Total = [NH3 ](aq) + [NH+4 ] (31)
NH3 +H2O  NH+4 + OH- (32)
aNH+ aOH- [NH+4 ][OH- ]

K NH = 4
» (33)
3
aNH aH O [NH3 ]
3 2
While protection of the entire secondary system is the goal of the chemistry dosing practices,
protection of the materials in the steam generators or boilers is paramount since the boiler
tubes represent the physical barrier between nuclear and non-nuclear sections of the plant. A
boiler contains an array of 1000’s of tubes, tube support plates, tube sheet, and associated
steam driers. Since dissolved species, other than the ammonia or the other amines dosed to
control pH, represent impurities in the feedwater it is important to minimize (or eliminate if
possible) their concentrations entering the boiler. Some impurities of particular importance in
boiler feed water include iron, copper, sodium, chloride and sulphate. The iron (and copper if
the system includes some copper-bearing materials) enters the boiler as corrosion products
released from the predominately carbon steel components in the feed water train and, since
they are not volatile will accumulate in the boiler as steam is produced. Ultimately, boiler sludge
(mainly iron oxides or hydrated oxides) will accumulate on the tube sheet and tube support
plates and must be periodically removed through a physical or chemical cleaning procedure.
This can become difficult if significant copper is retained in the deposits.
Sodium, chloride, sulphate and other inorganic impurities can enter the feedwater system
through leaks in the condenser tubes and must be kept at very low concentrations since all non-
volatile species will tend to concentrate in the boiler, particularly in crevice regions in the tube
sheet or around the tube support plates. The concentration of these species is kept low in the
boiler through a continuous blow-down that removes a small fraction of the circulating water in
the boiler. Example 15.3 demonstrates the calculation on the effectiveness of blow-down to
maintain low levels of contaminants in the boiler. A simple calculation of the feedwater input
flow rate divided by the blow-down rate can give a rough indication of the concentration factor
(CF) that can be achieved in a steam generator crevice. For example, an individual boiler in a
typical CANDU-6 plant will have an incoming flow rate of around 250 kg/s. With a blow-down of
approximately 0.5% of the incoming flow (amounting to 1.25 kg/s) this shows a concentration
factor of 200, meaning that any impurities entering the boiler from the feedwater will be
concentrated 200 times in the bulk water of the steam generator. Regular sampling of the blow-
down stream for sodium, chloride, sulphate and other impurities provides a ready indication of
poor feedwater chemistry.
Impurities in the boiler cause hide-out, whereby they are absorbed and retained in the boiler
crevices or in the deposited sludge piles, which can lead to aggressive chemistry conditions for
the boiler tubes. Alloy 800, the choice material for recent CANDU reactor boiler tubes, has
shown excellent corrosion resistance and little stress corrosion cracking over a wide range of
chemistry conditions. However, if the environment adjacent to the tubes becomes sufficiently
acidic or alkaline due to the ingress and hide-out of ionic impurities, chlorides, sulphate or
sodium for example, the recommended operating envelope for the boiler tubes may be com-
promised. This situation has occurred at various nuclear plants, both CANDUs and PWRs, where
boiler tubes had to be plugged due to through-wall cracking and leakage of primary system
water into the boiler. This is detected by increasing radiation levels in the boiler blow-down or
main steam line.
Example 15.3 - Blow-down calculation.
A sketch of a steam generator with feedwater inlet flow rate (F), steam flow rate (S) and
blowdown flow rate (B) is shown in Figure 4. Assuming steady state operation, feedwater iron
concentration of 5 ppb (mg/kg), feedwater flow rate is 250 kg/s and blowdown rate of 0.5%
(1.25 kg/s), calculate the concentration of iron in the bulk boiler water (CM). If the blowdown
rate was reduced to 0.1%, what effect does that have on bulk boiler water iron concentration?
Figure 4. Mass balance on steam generator impurity inventory.
Solution
A steady state mass balance on the iron inventory using the sketch in Figure 4 results in:
Fe In = Fe Out
F × CF = S × Cs +B × CB
We not that the concentration in the bulk boiler water will be the same as the blow-down
concentration, hence CM = CB. Also, the iron in the boiler water is non-volatile so its concentra-
tion in the steam line will be zero. Applying these conditions and rearranging results in:
F
CM = CB = C
B F
The term F/B is the concentration factor described above and is the ratio of the feedwater flow
rate over the blow-down rate. For the conditions indicated above this results in a concentra-
tion factor (CF or F/B) of 200. Thus:
CM = 200 x CF = 200 x 0.005 mg/kg = 1 mg/kg Ans.
Reducing the blow-down rate to 0.1% effectively increases the concentration factor by five
leading to CM = CB = 5 mg/kg. Ans.
4.2 Corrosion Issues in the Secondary System

4.2.1 Boiler crevices
As described above, the steam generators represent critical locations in a CANDU plant as the
boiler tubes and tube sheets are the physical barriers between the nuclear and non-nuclear
sections of the plant. As a result, the operational chemistry practices in the steam generation
circuit are targeted primarily at protection of the shell side of the boiler while trying to minimize
corrosion issues in the feedwater train and the steam cycle. The recent material of choice for
the steam generator tubing is nuclear grade Alloy 800 (~20%Cr, 35% Ni, balance Fe with ~1%
each Al & Ti), which has shown excellent corrosion resistance and minimal cracking in well over
30 years of operation in CANDU plants and German PWRs. Earlier CANDUs used Inconel 600 or
Monel 400, each showing poorer performance than Alloy 800 in terms of their susceptibility to
underdeposit corrosion, SCC, intergrannular attack and fretting wear [Tapping et. al, 2000]. It is,
however, imperative to keep the feedwater and hence boiler chemistry within specification
limits for pH and oxygen, as these parameters directly affect the electrochemical corrosion
potential (ECP) attained on the boiler tubes and have significant influences on corrosion and
environmentally-assisted cracking (EAC). During operation, crevices in the steam generator
accumulate impurities from the concentrated environment in the bulk boiler water due to the
concentration factor (CF) effect as illustrated in the blow-down example above (crevices as
defined here are not only the physical crevices between the boiler tubes and the tube sheets
and support plates but also the tight-tolerance micro-environments that are created under
boiler sludge deposits and along the tube support plates). As the power level is reduced and the
CF subsequently lowered, the impurities leach out of the crevices and may be consequently
removed via blow-down, a phenomenon known as hide-out return. Failure to keep impurity
concentrations low in the bulk boiler water may lead to chemical imbalances in the boiler
crevices and promote either acidic or alkaline pH’s, both of which are detrimental to the boiler
tubes from a general corrosion perspective. The presence of impurity oxygen in the bulk boiler
water compounds the effect and may lead to cracking, particularly during transient operations
during station run-ups or shut-downs.
AECL has developed a recommended operating environment envelope for Alloy 800 boiler tubes
[Tapping, 2012]. The chart, shown in Figure 5, is the result of decades of reactor operating
experience and ongoing research at the Chalk River Laboratories and elsewhere to evaluate the
cracking and degradation propensity of boiler tubes in out-of-specification chemistry conditions.
While a plant may be within the recommended operating envelope for the majority of its
operational lifetime, there will be periods of out-of-specification chemistry due to ingress of
contaminant ions or elevated concentrations of dissolved oxygen. The effects of these, hopeful-
ly short, periods need to be carefully assessed as they may be the initiating events for pitting
and cracking later on in the life of the plant.
Figure 5. Alloy 800 Recommended Operating Envelope (Tapping, 2012).

2012)
4.2.2 Flow-Accelerated
Accelerated Corrosion in the Feedwater and Steam Extraction Piping
Protection
ection of the boiler is the primary objective of the chemistry dosing practices in the second-
secon
ary system and one of the best ways to accomplish this is to minimize the corrosion of the
materials in the feedwater train and regenerative feedwater heaters. Ex Excessive
cessive corrosion of the
feedwater piping will introduce corrosion products into the boiler water
water,, ultimately dumping the
material as sludge that accumulates on the tube support plates and tube sheet in the steam
generators. Flow accelerated corrosion (FAC), described in detail in Chapter 14, 14 is prevalent in
the feedwater system since the steam that condenses following the low pressure turbines is
ultimately pure water with a relatively low concentration of volatile amine that has been carried
through the steam cycle. This condensate water is effectively iron
iron-free
free or fully undersaturated in
dissolved corrosion products and is an excellent fluid for promoting FAC of carbon steel piping.
The moisture content of wet et steam in the moisture separator and turbin
turbinee steam extraction lines
will be effectively pure water and fully undersaturated in corrosion products as well,well but have
the additional complication of possessing small water droplets that can impinge on piping
surfaces,
urfaces, creating a mechanical erosion mechanism compounding the FAC effect. Critical
locations for FAC are typically downstream of tight radius bends, orifice plates for flow mea- me
surements or any location where turbulence is enhanced and mass transfer from the piping
promoted. The solubility of the primary protective oxide film, magnetite, also plays a role in the
most affected locations since it is seen to go through a maximum at temperatures between 130- 130
o
150 C as shown in Figure 7 in Chapter 14.
Management of FAC in the secondary circuit of power plants is quite advanced and relies upon
computer model predictions (EPRI’s CHECWORKS code for example) to assess critical locations in
the plant and frequent inspections. One of the key factors in protecting the secondary side
piping from FAC is to maintain a suitable alkalinity level at all locations around the steam cycle.
Magnetite solubility is demonstrated to be at a minimum around a pH25°C of 9.5 or so and
maintaining this specification at every location in the steam cycle is highly desirable to minimize
pipe wall thinning by FAC. However, this is difficult to achieve in the field since chemical dosing
typically occurs at a sole location in the feedwater circuit (usually just downstream of the
condensate extraction pump) and the fact that the two-phase sections of the plant can have a
significantly different chemical composition from that of the feedwater due to the distribution
coefficient of the volatile amine used for plant pH control. In principle, the target is to maintain
a delta pH (pH = pHT – pHneutral@T) greater than 1.0 to ensure that sufficient alkalinity is main-
tained at each location and the solubility of magnetite is kept to a minimum.
Table 6 shows the results of a calculation for the effect of different chemical dosing strategies on
the delta pH throughout a common secondary side feedwater and steam circuit. In case 1,
ammonia is used as the pH-controlling chemical and in case 2 morpholine is used, both achiev-
ing a final feedwater pH25C of 9.6. Note that in case 2, 1 mg/kg of ammonia is also assumed to
be present due to the decomposition of the morpholine (and hydrazine) at the higher tempera-
ture locations in the system; such decomposition typically results in a residual ammonium
concentration between 0.5-1.1 mg/kg. As is demonstrated in the table, both chemical dosing
strategies achieve the same target pH25°C and maintain a sufficient delta pH at the outlet of the
HP heater, thus maintaining suitable protection of the feedwater train from FAC. However, due
to the differences in volatility and distribution coefficients, the delta pH achieved with the
morpholine chemistry in case 2 provides much better protection for the steam extraction lines
and moisture separator.
Table 6. Comparison of delta pH for two chemical dosing strategies in a typical steam cycle.
Case 2: 25 mg/kg Morpholine
Case 1: 2.3 mg/kg Ammonia
+ 1 mg/kg Ammonia
(pH25 = 9.6)
(pH25 = 9.6)
Location pHT pHneutral pH pHT pHneutral pH
HP heater outlet
6.71 5.68 1.03 6.87 5.68 1.19
(final feedwater)
Bulk Boiler water 5.92 5.59 0.32 6.34 5.59 0.74
HP turbine extraction to
6.29 5.67 0.62 6.80 5.67 1.14
HPH2
Moisture Separator/Reheater 6.57 5.78 0.80 7.13 5.78 1.35
LP turbine extraction to DA 6.85 5.89 0.96 7.43 5.89 1.54
LP turbine extraction to LPH3 7.11 6.01 1.11 7.73 5.89 1.54
Condensate 8.73 6.83 1.89 9.49 6.83 2.65
Similar effects are realized with other amines of lower volatility, such as ethanolamine (ETA),
which has a relative volatility of about 0.15 at 200°C while that of ammonia is about 5.0. Its
effect on FAC compared with that of ammonia and in neutral chemistry is illustrated
illustrat in Figure 6,
which shows on-line
line measurements of the FAC rate of carbon steel in an experimental loop. loop The
operating conditions were two-phase
phase steam
steam-liquid
liquid flows at 200°C with ammonia and ETA at the
same pH25°C (9.2) and with no additive
additive. The effect of the additives on reducing FAC below that of
neutral chemistry is immediately realized at zero voidage (% steam by volume) and continues up
to 97% voidage. As the voidage increases from zero the FAC rates with the two additives are
roughly the same and increase together
together. At about 80% voidage, the FAC rate with ammonia
increases further as the ammonia partitions to the vapour phase, while tthathat for ETA decreases as
the ETA partitions to the liquid phase (the FAC occurs in the liquid film on the walls of the pipe or
component). The diagram also shows how reducing the flow rate also reduces the FAC rate.
Figure 6. Effect
fect of steam voidage on FAC rate of carbon steel at 200°C under three chemistries
(neutral, pH25°C with ammonia and with ETA) at a flow rate of 0.56 L/min (the effect of reduced
flow rate on FAC with ammonia also indicated). [Lertsurasakda et al., 2013]
5 Moderator System
As described previously, the main purpose of the heavy water moderator is to slow down or
thermalize the high energy neutrons produced through the fission reactions to low, thermal
energies so that they more easily induce further fission reactions. A significant amount of heat
is produced as a result of the high-energy neutrons losing their kinetic energy and this must be
dissipated through the external heat exchanger loops. A secondary purpose of the moderator
system is to serve as an additional reactivity control mechanism for the reactor, which is accom-
plished by dosing the heavy water with low concentrations of neutron-absorbing elements,
boron or gadolinium for example. The third purpose of the moderator system is to act as a
reactor safety system through rapid additions of high concentration of neutron poisons (known
as shut-down system 2 – SDS 2). Thus, the moderator system encompasses a large volume of
heavy water contained within the calandria vessel, the associated side recirculation loops for
heat removal and purification, the moderator cover gas system for removing deuterium gas
produced through the water radiolyisis, and the connections to SDS2. A schematic overview of
the moderator systems is shown in Figures 4 and 5 of Chapter 8.
5.1 Chemistry Control in the Moderator System

The materials of construction in the moderator system comprise mainly stainless steel (calandria
vessel, heat exchangers, etc), Zircaloy (calandria tubes) and nickel-based alloys, so the chemistry
of the moderator is not specified to limit corrosion specifically since the corrosion rates of these
alloys at the low temperature and pressure of the moderator system (~60oC and essentially
atmospheric pressure) is quite low. Thus, the heavy water in the moderator system is highly
purified and the only chemicals added to the system are those used as neutron poisons for
excess reactivity control or to guarantee reactor shut-down. Typically, gadolinium nitrate
(Gd(NO3)3) and/or boric acid (H3BO3) (alternatively boric anhydride – B2O3) are dosed to the
moderator system at low concentrations (a few ppm) and these dissolved ions make up the
primary contributor to the conductivity of the moderator heavy water – so much so that system
conductivity can be translated as a direct measure of reactor negative reactivity. Unlike many of
the other reactor coolant systems, the moderator system is not purposely deareated to maintain
low dissolved oxygen concentrations. Due to the large radiation fields associated with the
reactor core, the heavy water moderator is continuously bombarded with radiation, leading to
high production of deuterium and oxygen gases as the net products of water radiolysis. To avoid
accumulation of these gases, they are constantly removed through the cover gas system and the
deuterium is eliminated by recombination in catalytic re-combiner units, if necessary with
oxygen addition.
Typical operating parameters for the moderator system are shown in Table 7. Unlike the heat
transport systems where the water is kept alkaline to control and minimize the corrosion of
system components, the heavy water in the moderator is kept slightly acidic to ensure that
gadolinium hydroxide (Gd(OH)3) will not precipitate on the surfaces of the moderator compo-
nents and that SDS-2 can function if required.
Table 7. Target chemistry parameters in the moderator system.

D2O isotopic 99.6 – 99.9%
Conductivity 0.005 – 0.1 mS/m
Dissolved D2 < 3 mL/kg
pHa 4.5 – 6.5
total anions < 0.2 mg/kg
TIC / TOC* < 1.0 mg/kg
*Analysis for TIC/TOC has become increasingly important due to concerns for gadolinium oxalate
precipitation during guaranteed shut down conditions (see Section 5.1.2).
5.1.1 Reactor shims
Gadolinium Nitrate (Gd(NO3)3)

Gadolinium has been demonstrated to be an excellent reactor shim and chemical for controlling
excess reactivity during refuelling operations and to ensure reactor shut down conditions. The
isotopes Gd-155 and Gd-157 are 15 and 16% abundant in nature and have huge neutron absorp-
tion cross-sections of 61000 and 255000 barns respectively, which gives them the distinction of
being the naturally occurring isotopes with the largest known neutron absorption cross sections.
Gadolinium nitrate hexahydrate (Gd(NO3)3.6H2O) is easily dissolved in the heavy water modera-
tor. Upon dissolution, the pH of the heavy water will become mildly acidic (pHa between ~5 to
6) due to the interaction of the nitrate salt dissociation (equation 34). It is important to maintain
these slightly acidic conditions within the heavy water moderator because the Gd3+ cations will
readily hydrolyse at neutral to alkaline pH and precipitate as gadolinium hydroxide (equation
35). This is undesirable since the gadolinium must be homogeneously dispersed in the modera-
tor system to achieve its primary goal of reactivity control.
(
Gd NO3 ) 3
 Gd3+ + 3NO-3 (34)
Gd3+ + 3H2O  Gd OH + 3H+ ( ) 3
(35)
Gadolinum concentrations within the moderator system can range from zero to ~ 1 mg/kg under
typical reactor operating conditions, leading to water conductivities of ~ 0.005 mS/m to ~ 0.10
mS/m. Care must be taken during normal reactor operation to limit the conductivity, and hence,
nitrate concentration in the heavy water to below ~ 0.05 mS/m (corresponding to ~ 1 mg/kg
gadolinium nitrate) since the nitrate anions may form reactive intermediate species with the
water radiolysis products and promote the formation of deuterium gas (D2) [Yakabuskie et al.,
2010], which can lead to excessive cover gas concentrations and increased workload for the re-
combiner units (the D2 concentration in the cover gas must be kept below 4%). Some operating
reactors have used gadolinium sulfate (Gd2(SO4)3.6H2O) as their preferred salt, which eliminates
the issues of increased deuterium production from interaction of the radiolysis products with
the nitrate ion (sulfate ions do not form reactive intermediates with short-lived radiolysis
products). However, the sulfate salt is sparingly soluble compared to the nitrate salt thus, if
excess reactivity control is required, such as during start-ups with significant amounts of fresh
fuel in the core, the preferred reactivity control shim is boric acid. Sulphur species may also
promote localized corrosion of alloy components.
Boric Acid
Boron 10 is 20% abundant in nature and has a neutron absorption cross-section of 3838 barns.
Boric anhydride or boric acid is soluble in water and can be an efficient alternative to gadolinium
as a reactivity control mechanism. Its dissolution, however, is very slow and so boric acid is only
employed in the operations of the CANDU plant when necessary. This is typically when concerns
of excess deuterium production are encountered through the use of higher concentrations of
gadolinium nitrate such as during startups with significant volumes of fresh fuel.
5.1.2 Guaranteed Shutdown State (GSS)

During maintenance shut-downs, the reactor is placed in a guaranteed shut-down state (GSS)
whereby a soluble neutron absorbing salt (gadolinium nitrate) is injected into the moderator at
high concentrations in addition to having the reactor at zero power with the shut-down rods.
This provides redundancy and ensures that the reactor will be maintained in a sub-critical state
until the soluble poison is removed via gadolinium and nitrate/nitrite removal in the mixed bed
ion exchange columns of the moderator purification circuit, a simplified schematic of which is
shown in Figure 3 of Chapter 8.
To ensure a guaranteed shut down state, gadolinium concentrations of greater than 15 mg/kg
are typically required and this is the lower limit for gadolinium in the moderator during GSS.
Typical concentrations are targeted at greater than 20 mg/kg and assurances must be made to
the Canadian Nuclear Safety Commission that GSS is achieved and maintained. This assurance is
typically accomplished by manually sampling the moderator and analyzing for gadolinium
concentration at least twice per day.
Gadolinium Oxalate
In 2008 during a planned station outage at a unit of the Pickering B plant, while the unit was
placed in GSS it was observed that the Gd concentration in the moderator was decreasing at an
alarming rate. The over-poisoned condition for the GSS requires assurances that the Gd concen-
tration in the moderator is maintained above 15 mg/kg at all times and the measured loss rate
of Gd was approximately 2 mg/kg per day. The utility, with the assistance of experts from AECL
Chalk River traced the unexpected Gd depletion to precipitation as gadolinium oxalate
(Gd2(C2O4)3), a salt known to be very insoluble in water. Oxalate is normally not present in the
moderator heavy water systems and, upon investigation, its formation was attributed to a
known leak of CO2 from the annulus gas system through the rolled joints at one of the end
fittings, a condition that had been assessed in 2005 and monitored routinely through measure-
ment of total inorganic carbon (TIC). Elevated CO2 concentrations in the moderator heavy water
produce carbonate anions, which are readily converted to oxalate through combination with
primary water radiolysis radicals such as shown in equations 36 & 37.
-
CO2(aq) + e(aq)  ×CO2 (36)
×CO2 + ×CO2  C2O2-

4
(37)
Under typical reactor operating conditions, with little gadolinium present in solution, the oxalate
anions would readily decompose back to CO2. However, under GSS with appreciable concentra-
tion of gadolinium in solution and a high enough concentration of TIC (from the ongoing CO2
leak in the rolled joint), the oxalate anion could readily combine with the gadolinium present as
quickly as it was produced, leading to precipitation of the gadolinium oxalate salt throughout the
moderator system, as shown in equation 38.
2Gd3+ + 3C 2O2-
4
® Gd2 C 2O4 ( ) 3(s)
(38)
The Gd depletion is of concern for maintaining the reactor in GSS but the fate of the oxalate
precipitate throughout the calandria vessel and on the calandria tubes is of greater concern as it
could prevent the reactor from becoming critical upon startup and must be removed. Several
research programs were initiated to formulate a chemical cleaning strategy to remove the
precipitated gadolinium oxalate; however, none were ultimately needed as the precipitate in-
core was found to be readily oxidized and converted to the soluble nitrate salt through action of
the shut-down gamma fields and ultra violet radiation (Cerenkov radiation) in the drained (air-
filled) and humid calandria vessel. The remaining precipitate from out-core surfaces was easily
removed through filtration upon moderator refill and reactor start up [Evans, 2010].
Oxalate production in the moderator system has also been demonstrated to occur by radiolytic
processes involving total organic carbon (TOC), primarily from oil ingress. Two operating CANDU
units have recently experienced issues with radiolytic decomposition of hydrocarbon lubricants
within the moderator system, which will produce decomposition gases such as hydrogen and
carbon dioxide as well as create particulate material through polymerization reactions that is
highly efficient at plugging filters [Ma, 2010]. These issues with elevated TIC and TOC in the
moderator system have led to more strict guidelines and chemistry control practices since the
implications of oxalate formation have been realized.
5.1.3 Shutdown System 2 (SDS2)

Gadolinium nitrate is used in the moderator for rapid reactor shut down in the event of a station
upset or trip. Shutdown System 2 (SDS2) is described in detail in Chapter 8 and comprises the
poison addition tanks which contain approximately 1000 litres each of concentrated gadolinium
nitrate solution (~8000 mg/kg). The poison tanks are connected to a high pressure helium gas to
rapidly inject the gadolinium solution into the moderator heavy water; the systemis able achieve
a reactor shut down in a matter of seconds.
5.2 Moderator Cover Gas

As described above, significant deuterium and oxygen production occurs in the moderator due
to the large volume of heavy water that is continuously exposed to high radiation fields asso-
ciated with the reactor core. The net water radiolysis produces the gases that are initially
dissolved in the coolant but diffuse into the moderator cover gas space at the top of the calan-
dria vessel. If the concentration of deuterium gas were not controlled, flammable concentra-
tions would be reached in the cover gas in a matter of hours under normal operating conditions.
Flammable concentrations of hydrogen (deuterium) in the helium cover gas are ~ 8%, so upper
control limits are set at 4% D2.
The head space in the calandria vessel is separated from the bulk of the moderator heavy water
through a series of calandria relief ducts, as depicted in Figure 7. It is extremely important that
the moderator level be maintained within the relief ducts as these provide the conduit for the
deuterium and oxygen gases to diffuse into the head-space and to the helium cover gas. Since
the surface area for moderator heavy water to cover gas exchange is limited when water level is
maintained within the relief ducts, bulk diffusion rates can be minimized and kept within
controllable margins. If the moderator level were to fall below the relief ducts a large surface
area would then present itself for the D2 and O2 to readily migrate into the gas phase, leading to
a rapid increase in cover gas concentrations and potential for exceeding flammability limits.
Figure 7. Normal moderator cover gas system volume and relief ducts (CANTEACH)
The helium cover gas is maintained at slightly above atmospheric pressure (~ 110 kPa absolute)
and is circulated by several compressors, operating in parallel, through the head-space and on to
the catalytic re-combiner units (RCUs). The hydrogen (deuterium) RCUs are usually AECL-
patented components that contain a supported platinum or palladium catalyst. The re-
combination reaction that ensures deuterium concentrations are kept low in the cover gas
follows equation 39. As can be seen, the reaction requires a 1/2x stoichiometric concentration
of oxygen in order to fully recombine the deuterium gas to heavy water; however, it is common
practice to dose the helium cover gas with oxygen in excess to ensure sufficient recombination
occurs. By ensuring the cover gas oxygen concentration is between 1-2%, sufficient excess is
continually maintained and deuterium concentrations are kept to very low values and are often
undetectable.
D 2 + 12 O2 ¾catalyst
¾¾® D 2O (39)
6 Auxiliary Systems
6.1 Calandria Vault and End Shield Cooling System
The Calandria Vault and End-Shield Cooling system is a light-water system that acts as a biologi-
cal shield from the gamma radiation and neutrons present in the core of the reactor. In this
system, radiation energy is dissipated as heat that is removed from the water in external heat
exchangers. Schematic diagrams of the system are shown in Figure 8 and the system is de-
scribed in more detail in Chapter 8. The system comprises two distinct circuits; the calandria
vault that surrounds the reactor calandria vessel, which is enclosed in concrete; and the end
shields on either face of the reactor through which the fuel channels protrude. The end shields
are packed with carbon steel balls that stop most of the neutrons and gamma emanating from
the reactor core (note, in some CANDUs the end shield is a solid carbon steel plate with integral
cooling ports). The system operates at nearly atmospheric pressure and contains a circulated
nitrogen cover gas. Any gas that is produced from water radiolysis will diffuse into the cover gas,
which needs to be purged periodically to avoid the build up of flammable concentrations of
hydrogen and oxygen.
Figure 8. Schematic diagram of the Calandria vault and reactor assembly.
Typical operating specifications for the calandria vault and end shield cooling system are shown
in Table 8. Alkalinity is controlled by addition of lithium hydroxide to ensure low corrosion rates
of the system materials and the specifications for dissolved hydrogen and dissolved oxygen in
the system are set to ensure flammability limits are not exceeded. If the nitrogen cover gas for
the system reaches the control limit for hydrogen (4%), the system is purged with fresh nitrogen
to ensure safe operation is maintained.
Table 8. Typical Chemistry Parameters of the Calandria Vault and End Shield Cooling System.
pH 9.0 – 10.0
[Li+] 0.07 – 0.7 mg/kg
conductivity 0.24 – 2.4 mS/m
Anionic impurities < 1.0 mg/kg
H2 (vol. % in cover gas) < 4%
O2 (vol. % in cover gas) < 2%
Under normal circumstances, the end shield cooling system operates with a small excess of
dissolved hydrogen produced through the corrosion of system materials. This results in the net
radiolytic production of hydrogen and oxygen being suppressed. Some CANDU reactors have
had chronic issues with build up of hydrogen in the cover gas, which requires a frequent purging
to ensure that flammable concentrations are not reached. The increased hydrogen production
has been linked to the presence of dissolved oxygen in the water due to the addition of higher–
than-normal quantities of air-saturated make-up water. Dissolved oxygen is consumed through
recombination with the dissolved hydrogen, but when there is insufficient dissolved hydrogen
net water radiolysis will recommence to produce hydrogen and more oxygen, which exacerbates
the process. Once net radiolysis is occurring on a long-term basis, simply addressing the aerated
make-up water ingress issue has been shown not to reduce the hydrogen production.
The CANDU industry has investigated the possibility of mitigating excessive hydrogen production
in the calandria vault and end shield cooling system and has demonstrated that adding an
oxygen scavenger, such as hydrazine, to the system will mitigate the hydrogen production and
return the system to net radiolytic suppression [Stuart, 2012]. The dissolved oxygen reacts with
hydrazine to form nitrogen and water, which is a process that is fast at low temperatures as it is
mediated by the radiation field. Note that simply adding dissolved hydrogen to the water would
also promote a radiolysis recombination reaction, however, since the Calandria Vault and End
Shield Cooling system operates at atmospheric pressure, this method is not feasible and hydra-
zine additions are becoming common practice in operating reactors.
6.2 Liquid Zone Control

The liquid zone control system, as described in detail in Chapter 8, is designed to make fine
adjustments to the neutron flux profile in the reactor. This is accomplished by varying the water
level in each of the liquid-zone control tubes by adjusting the inlet water flow rate to the
individual zones. The materials of construction for the liquid zone tubes are primarily Zircaloy 4,
with auxiliary components being 300-series stainless steel; thus, the corrosion of the system
components is minimal under a wide range of chemistry conditions. Since the purpose of the
light water in the liquid zone system is to absorb neutrons to shape the reactor flux, the water is
kept nominally pure with a very low conductivity. The cover gas space is occupied with helium
and no direct method is used to limit the water radiolysis reactions that produce hydrogen and
oxygen. Hydrogen control in the cover gas space is minimized by purging with helium as re-
quired and by recombiners. Water purity in the liquid zone system is paramount to minimize the
hydrogen produced by radiolysis since impurities such as chlorides or nitrate/nitrites will
interfere with the hydrogen recombination reactions and catalyze oxygen and peroxide produc-
tion. Thus, nitrogen in the cover gas (indicative of air ingress) is highly undesirable, since it will
produce nitrate anions in the presence of a radiation field and subsequently produce nitric acid
that can lead to low pH conditions and corrosion of the system materials as well as promote high
production rates of hydrogen.
The typical operating specifications for the liquid zone control system are focused on water
purity as measured through liquid conductivity. For high purity make-up water, it is often
possible to achieve conductivities in the range of 0.006 – 0.008 mS/m, and most utilities will try
to keep the system operating within these limits or at least below 0.01 mS/m. By doing so,
water radiolysis is minimized and the cover gas purge can be done as frequently as necessary to
maintain the cover gas hydrogen concentration below 4%.
6.3 Annulus Gas

The annulus gas system provides a thermal barrier between the Zr-2.5Nb pressure tubes
operating at temperatures between 300-310oC and the low temperature (~60oC) calandria tubes
that separate the moderator from the primary heat transport system. The annulus space in each
fuel channel provides a gap of approximately 5 mm and is continuously purged with carbon
dioxide gas, thus purity of the CO2 is a direct indication of the overall health of the system.
Typical impurities in the CO2 purge gas include deuterium (D2), which diffuses through the
pressure tubes from to the small amount of corrosion occurring on the PHT side of the tube, and
water vapour (D2O). The corrosion of the exterior surface of the pressure tubes and the interior
surface of the calandria tubes is not typically an issue when the CO2 is pure. The presence of
heavy water vapour in the gas can indicate leaks from either the PHT or moderator systems.
Thus, typical chemistry parameters that are monitored in the annulus gas include the volume
percentages of deuterium and oxygen and the dew point (oC). Gradual increases in the dew
point are expected in the system since, ultimately, a small amount of heavy water will migrate
through the rolled-joint seals at each end of the fuel channel. However, accelerating increases in
dew point measurement provide an on-line indication of developing leaks in the system.
Oxygen may also be injected into the CO2 annulus gas to a concentration of approximately 2-4%.
This is beneficial as it promotes protective ZrO2 films on the pressure tubes and calandria tubes
and helps to ensure that the garter springs or calandria tube to pressure tube spacers (Alloy
X750 – a nickel superalloy) are maintained in a suitably oxidized state, also promoting passive
oxide film formation. If no oxygen is present, the garter springs can become extremely brittle as
the reduced oxides or elemental nickel present under these conditions can promote cracking in
the material. This is exacerbated by radiation embrittlement due to the high neutron fluxes to
which the garter springs are exposed. Oxygen injection to the annulus gas also promotes the
recombination with deuterium to form D2O, helping to ensure the deuterium concentrations are
kept below the flammability limits.
6.4 Emergency Core Cooling Systems

The emergency core cooling systems must be ready on a moment’s notice in the event of a loss
of coolant accident (LOCA) from the PHT, thus it is imperative that the system components do
not corrode excessively and accumulate particulate corrosion products that can accumulate as
sludge in the normally stagnant system. Short-term decay heat removal is supplied by the high-
pressure and medium-pressure ECC systems that have their own distinct water supplies. For
longer-term core cooling, the emergency cooling water is supplied via a tank inside the vacuum
building in the multi-unit CANDUs or from the dousing tank in the roof of the containment
building of single-unit stations.
In the event of a LOCA, when a reactor PHTS is vented to the building, water from the dousing
tank is sprayed into the building to condense the steam and reduce the pressure. The supply
tank is connected to the sump below the reactor building that collects water draining from the
system for re-injection via the emergency injection pumps. The system chemistry is maintained
moderately alkaline and is deoxygenated with hydrazine, which will decompose over time to
ammonia producing the alkalinity required to maintain low corrosion rates of the ECC piping and
injection nozzles. Recent regulations have focused on ECC chemistry, particularly the effect of
debris or corrosion products clogging the strainers of the injection pumps. Of primary concern
are high-pH conditions, where the corrosion of aluminum components in contact with the ECC
water (the sump strainers for example) has been the focus of considerable research effort
[Edwards et al, 2010]. At too high a pH, aluminum components may corrode excessively and
form precipitates that could block the strainers. For this reason, an upper limit on dousing
system pH is typically set around 9.5.
6.5 Service Water

A nuclear power plant contains many subsystems that act as coolants for the primary process
systems. These include auxiliary heat exchangers for the emergency core cooling systems,
recirculated cooling water for the turbine/generator lubrication oil, and general service recircu-
lated cooling water (RCW). These systems are chemically-dosed to maintain low corrosion of
the piping and heat exchanger tubing by providing an alkaline environment with minimal
dissolved oxygen. The RCW and TARCW (turbine & auxiliaries recirculated cooling water)
systems are continuously circulated during plant operation, making their chemistry control quite
simple; chemicals may be added as required to meet the pH and dissolved oxygen specifications.
These auxiliary systems will typically operate with the same chemicals as are used in the feedwa-
ter circuit, but care must be taken to limit the operating pH if ammonia is used since the aux-
iliary systems will typically contain some copper-bearing components. In general, pH is con-
trolled by ammonia and/or morpholine additions to maintain a value of around 9.2. Dissolved
oxygen is monitored and managed through hydrazine additions at concentrations between 50 –
100 ppb, ensuring low corrosion of the carbon steel piping. Particulate corrosion products are
also monitored to ensure system corrosion is low; higher particulate concentrations can be an
indication of out-of-specification chemistry conditions that may be remedied by chemical
addition or through a feed-and-bleed procedure to lower overall chemical and impurity concen-
trations.
7 Lay-up Practices
Inevitably, every power plant need
needs to be shut down for regular maintenance outages. In the
case of a CANDU, a mid-life
life refurbishment outage (after ~ 25-30
30 years operation) is required to
remove and replace the Zr2.5Nb pressure tubes that tend to stretch and sag due to the opera-oper
tional temperatures and high atomic displacements caused by the continuous neutron bom- bo
bardment under load.. A regular maintenance outage may last from 2 – 8 weeks depending
upon the scope of work required and refurbishment outages are planned for at least two years.
During any of these plant outages, the systems are typically depressurized and frequently
drained to facilitate the required maintenance work and subsequent inspections. These main- mai
tenance activities can allow air ingress into the closed-loop
loop systems that are intentionally
intenti kept
de-aerated
aerated to minimize corrosion during operation.. The duration and extent of the oxidizing
conditions that are produced when the systems are shut down can play large roles in their
behaviour during startup and subsequent steady power operation. Thus, it is now considered
extremely important to provide a suitable lay
lay-up
up condition to the plant during regular mainten-
ance outages and is paramount for protecting nonnon-refurbished
refurbished components during an extended
shut down during a mid-lifelife refurbishment
refurbishment. Degradation mechanisms and problem locations in
the feedwater and steam circuit of fossil plants are described by Mathews and are shown
schematically in Figure 9 [Mathews, 2013]
Figure 9. Locations requiring attention during maintenance outages and lay-up

lay of a fossil
plant. [after Mathews, 2013]
7.1 Dry Lay-up

It is common practice in the power generating industry to provide, at a minimum, a low humidi-
ty environment for systems that are placed in “dry lay lay-up”. This includes the steam-carrying
steam
piping and components of the secondary system along with the condenser shells. Depending
upon the duration of the maintenance outage it may be desirable not only to circulate low
humidity air through the system but also to maintain an inert cover gas, typically nitrogen, to
prevent oxidation of the reduced corrosion products present on the piping surfaces. If signifi-
signif
cant humidity is present, a thin water film will develop on the piping surfaces and this will be
fully oxygen saturated, thus promoting general corrosion and pitting on the predominantly
carbon steel surfaces in these systems. It will also convert, at least partially, the reduced
protective oxide formed d on the surfaces during operation (typically magnetite – Fe3O4) to a more
oxidized form such as hematite (Fe2O3). This oxidation process will tend to alter the volume and
thickness of the oxide films leading to poor adhesion and spallation during subsequent plant
start up. Figure 10 shows the relative oxide/steel volume ratios and it is clear that the further
oxidized
ized the corrosion product the larger the volume occupied. Thus, maintaining an oxygen-
oxygen
free, dry environment for all system components during a maintenance outage is seen as the
most prudent step in protecting the plant’s assets.
Figure 10. Oxide/steel volume ratios.
7.2 Wet Lay-up

Some systems may not be drained during a maintenance outage if work is not scheduled to be
performed on the system or if parts of the system in question can be isolated. Thus, at the start
of the lay-up period, the system will have the good chemistry specified for the operational
period but this may be difficult to maintain, especially if the outage is extended or if the project
is large in scope such as an overall mid
mid-life refurbishment. Under these circumstances,
rcumstances, the keys
to maintaining low corrosion rates of all system materials are the same as during operation:
operation
maintain an alkaline pH condition to promote passivity of the corrosion films and maintain
reducing conditions in the fluid by excluding air ingress and, typically, dosing with an oxygen
scavenging chemical such as hydrazine (N2H4). Frequent recirculation of the laid-up laid system

volume during the maintenance outage is recommended to promote good mixing and repre-
sentative sampling.
8 Heavy Water Systems
Heavy water (D2O) is essential to the operation of a CANDU reactor. It is used in place of light
water (H2O) as a moderator and coolant due to its superior neutron moderating properties
(deuterium’s neutron absorption cross-section is orders of magnitude lower than hydrogen’s)
making it feasible to fuel the reactor with uranium containing the natural abundance of the
fissile U-235 isotope (~0.7%). Heavy water or, more specifically, the deuterium isotope, is
naturally abundant in the environment at about 0.015% (atomic percentage) as both the D2O
and the HDO molecule, all in chemical equilibrium as shown in equation 40. The moderator and
PHT systems both require heavy water that is greater than 98% in isotopic for D2O. For reactor
safety purposes in the event of a pressure tube rupture where the PHT water enters and mixes
with the moderator, the PHT isotopic content is typically kept slightly lower than the moderator,
98.6% vs > 99% for example, which will dilute the moderator in the accident scenario lowering
the overall moderation.
H2O + D 2O  2HDO (40)
8.1 Upgrading
Separation techniques are numerous and include fractionation through distillation and various
processes involving atomic exchange through chemical sorbants and equilibrium systems. All of
these processes rely on the fact that the mass of the deuterium atom is twice that of the
hydrogen atom, making properties such as vapour pressure slightly different from those of
compounds containing the hydrogen atom alone. The separation factor for the deuterium-
hydrogen isotopes is defined as the ratio of the deuterium fraction in the desired phase (liquid,
x) over its fraction in the other phase (gas, y), as shown in equation 41. For a light-water/heavy-
water mixture, the separation factor can be estimated through the vapour pressure of each of
the components as shown in equation 42 for the normal boiling point of natural water [Bene-
dict, 1980].
a=
( )
x 1- x
(41)
y (1- y)
pH O
aHD = 2
= 1.026 (42)
pD O
2
The primary method for heavy water production is known as the GS dual-temperature exchange
process. It involves the equilibrium exchange of the deuterium atom between water and
hydrogen sulfide gas in a staged absorption column as shown in equation 43. Two towers are
employed whereby a fresh water feed to the cold tower, which has a higher separation factor for
deuterium ( a c » 2.32 ) since it will be retained preferentially in the liq

liquid
uid phase, is stripped of
some of its hydrogen content to a deuterium
deuterium-rich hydrogen sulfide gas (HDS),, which is effectively
leached of its deuterium content
content. The feed gas to the cold column is produced in the high
temperature column, where the separation factor (ratio of the fraction of D in liquid to D in the
gas) is lower ( ah » 1.80 ),, thereby producing the D
D-rich gas for feed to the primary, cold exchange
column. The basic principles of operation of the GS process are depicted in Figure 11 (from
Benedict et al., 1980). With an appropriate number of stages in the absorption columns, water
can be fairly economically enriched to about 25% - 30% D2O. From this concentration, final
upgrading using conventional distillation is economically feasible.
H2O +HDS  HDO +H2S (43)
Figure 11. Schematic representation of a GS dual temperature absorption column (Benedict et

al., 1980).
The distillation of water to produce a heavy water product is extremely energy intensive,
intensive since
large volumes are required to compensate for the low isotopic content of natural water.
water Thus,
once primary upgrading has been accomplished from pro processes
cesses such as the GS dual- dual
temperature exchange described above, energy requirements are much reduced,reduced particularly
when conducted under vacuum conditions
conditions. All CANDU plants will contain a distillation facility
for upgrading the D2O isotopic content since it can be diluted during operation through addition

of chemicals containing the hydrogen atom (H2, LiOH, N2H4, NH3 etc.) and through neutron
absorption producing tritium (H-3 or T). The overall efficiency of the fission process relies upon
specified isotopic content in the moderator and coolant thus the process water must be periodi-
cally upgraded.
8.2 Clean-up
Like the make-up water that feeds the secondary system and the reactor auxiliaries and is
conditioned to ensure low corrosion rates of the materials of construction and adequate system
lifetimes, the heavy water used in CANDU plants must be very pure. The basic methods of
filtration and ion exchange all apply to heavy water as they do for light water, with the exception
that the isotopic content of the heavy water must also be taken into consideration (the H2O
content in the heavy water may be considered an impurity).
As described above, D2O production and upgrading facilities may produce heavy water at
isotopic concentrations greater than 99.5% quite readily. Impurities introduced during the
upgrading process may include corrosion products from the materials of construction of the
distillation towers and packing columns (Fe, Ni, Cu etc) as well as contaminant cations and
anions (Na+, Ca2+, Cl-, SO42- etc) from ingress of humidity or contamination during transfer
operations. These may be removed through ion-exchange; however, it is imperative that the
ion-exchange resin first be converted to a deuterated form (exchange of the H+ sites with D+) to
minimize the downgrading in isotopic from cation exchange.
Tritium Removal
Some plants, notably the Darlington station in Ontario, include a tritium reduction facility in
order to ensure low environmental releases and to minimize the tritium activity in the operating
plant. Tritium removal plants follow the same isotopic removal principles as upgrading the
deuterium content to produce heavy water, except that the separation factor for D/T is very
close to unity under near atmospheric and/or vacuum conditions. At the Darlington facility,
separation is aided by a series of catalytic exchange columns that facilitate the equilibrium
exchange between tritiated heavy water and the carrier deuterium gas. The catalytic exchange
process follows equation 44 and typically can reduce the tritium concentration in the input
heavy water by a factor of ten [CANTEACH]. The tritium now contained in the deuterium gas
stream is concentrated to >99.9% through a series of cryogenic distillation columns operating at
temperatures of ~ 25 K absolute. A schematic flow sheet of the Darlington facility is shown in
Figure 13. These cryogenic temperatures ensure that any humidity or traces of nitrogen and
oxygen are removed (25 K is well below their boiling points) leaving a pure deuterium/tritium
gas that is easily separated through the density/buoyancy effects utilized in distillation. The
resulting pure tritium gas is encapsulated through reaction with titanium sponge, producing a
titanium hydride and immobilizing the radioactivity. The hydride may then be stored until the
tritium decays to suitable levels or sold for profit.
DTO + D 2  D 2O +DT (44)
Figure 12. Schematic diagram of the Darlington tritium removal facility. (Busigin & Sood)
The chemistry of process systems is controlled in order to optimize the performance of the
variety of materials that make up the systems. There is therefore a strong link to the Chapter 14
on Materials and Corrosion.
10 References
A.V. Bandura & S.N. Lvov, The Ionization Constant of Water over a Wide Range of Temperatures
and Densitites, J. Phys. Chem. Data, vol. 25, pp 15-30, 2006.
D.M. Bartels, The Reaction Set, Rate Constants and g-Values for the Simulation of Radiolysis of
Light Water over the Range of 20o to 350oC Based on Information Available in 2008,
AECL Analyses, 153-127160-450-001, Revision 0, 2009.
M. Benedict, T.H. Pigford and H.W. Levi, Nuclear Chemical Engineering, Second Edition, McGraw-
Hill Book Company, 1981.
A. Busigin and S.K. Sood, Optimization of Darlington Tritium Removal Facility Performance:
Effects of Key Process Variables, Nuclear Journal of Canada, 1:4, pp 368-371.
G.B. Buxton, Radiation Chemistry of the Liquid State: (1) Water and Homogeneous Aqueous
Solutions, in Radiation Chemistry: Principals and Applications, Editors: Farhataziz and
M.A.J. Rodgers, VHC, New York, 1987.
CANTEACH, Introduction to CANDU Process Systems – Heavy Water Production and Manage-
ment, available at www.canteach.ca.
CANTEACH, The Moderator and Auxiliary Systems, Chemistry Course 224, available at
www.canteach.ca.
B. Cox, Mechanisms of Zirconium Alloy Corrosion in High Temperature Water, J. Corr. Sci. & Eng.,
vol.6, 2003.
CRC, Handbook of Chemistry and Physics, 94th Edition, CRC Publishing, 2013-2014.
M.K. Edwards, L. Qui and D.A. Guzonas, Emergency Core Cooling System Sump Chemical Effects
on Strainer Head Loss, Nuclear Plant Chemistry Conference – NPC 2010, Quebec City, Oc-
tober 2010.
A.J. Elliot and C.R. Stuart, Coolant Radiolysis Studies in the High Temperature, Fuelled U-2 Loop
in the NRU Reactor, AECL R&D Report, 153-127160-440-003, Revision 0, 2008.
D.W. Evans et al, Gadolinium Depletion Event in a CANDU® Moderator – Causes and Recovery,
Nuclear Plant Chemistry Conference – NPC 2010, Quebec City, October 2010.
IAPWS, Guideline on the Henry’s Constant and Vapor-Liquid Distribution Constant for Gases in
H2O and D2O at High Temperatures, Kyoto, Japan, available at www.iapws.org, 2004.
IAPWS, Release on the Ionization Constant of H2O, Lucerne, Switzerland, available at

www.iapws.org, 2007.
IAPWS, Technical Guidance Document – 2012 Revision: Instrumentation for Monitoring and
Control of Cycle Chemistry for the Steam-Water Circuits of Fossil-Fired and Combined-
Cycle Power Plants, Boulder, Colorado, available at www.iapws.org, 2012.
IAPWS, Technical Guidance Document: Corrosion Product Sampling and Analysis for Fossil and
Combined-Cycle Power Plants, Moscow, Russia, available at www.iapws.org, 2014.
C. Lertsurasakda, P. Srisukvatananan, L. Liu, D. Lister and J. Mathews The effects of amines on

flow-accelerated corrosion in steam-water systems. Power Plant Chemistry, Vol. 15, No.
3, pp.181-190 (2013)
I.N. Levine, Physical Chemistry, 199??.
D.H. Lister and L.C Lang. A mechanistic model for predicting flow-assisted and general corrosion
of carbon steel in reactor primary coolants. Proc. Chimie 2002; Intern. Conf. Water
Chem. Nucl. Reactor Systems. Avignon, France. SFEN. (2002)
G. Ma et al, Experience of Oil in CANDU® Moderator During A831 Planned Outage at Bruce
Power, Nuclear Plant Chemistry Conference – NPC 2010, Quebec City, October 2010.
J. Matthews, Layup Practices for Fossil Plants, Power, February 2013.
R.E. Mesmer & D.L. Herting, Thermodynamics of Ionization of D2O and D2PO4-, J. of Sol. Chem.
Vol.7, no.12, pp.901-913, 1978.
J.P. Slade and T.S. Gendron, Flow accelerated corrosion and cracking of carbon steel piping in
primary water – operating experience at the Point Lepreau Generating Station, Proc. In-
tern. Conf. on Environmental Degradation of Matls. in Nuclear Power Systems, Salt Lake
City, UT. TMS. (2005).
J.W.T. Spinks & R.J. Woods, An Introduction to Radiation Chemistry, Wiley, New York, 1990.
C.R. Stuart, Mitigation of Hydrogen Production in the Shield Cooling System of CANDU Reactors -
Recent Changes to Chemistry Control, 9th International Workshop on Radiolysis, Elec-
trochemistry and Materials Performance, Paris, France, September 2012.
R.L. Tapping, J. Nickerson, P. Spekkens and C. Maruska, CANDU Steam Generator Life Manage-
ment, Nuc. Eng. & Design, vol. 197, pp. 213-223, 2000.
R.L. Tapping, Chapter 17: Corrosion Issues in Pressurized Heavy Water Reactor (PHWR/CANDU®)
Systems, Nuclear Corrosion Science and Engineering, Woodhead Publishing Limited,
London, 2012.
P.R. Tremaine & J.C. Leblanc, The Solubility of Magnetite and the Hydrolysis and Oxidation of
Fe2+ in Water to 300oC, J. Sol. Chem., vol.9, no.6, pp. 415-442, 1980.
C. Turner & D. Guzonas, Improving Chemistry Performance in CANDU® Plants, Nuclear Plant
Chemistry Conference – NPC 2010, Quebec City, October 2010.
V.F. Urbanic, B. Cox & G.J. Fields, Long Term Corrosion and Dueterium Ingress in CANDU PHW
Pressure Tubes, Zirconium in the Nuclear Industry: Seventh International Symposium,
ASTM STP939, pp. 189-205, 1987.
P.A. Yakabuskie, J.M. Joseph, J.C. Wren & C.R. Stuart, Effect of NO3- and NO2- on Water Radiolysis
Kinetics During -Irradtiation, NPC 2010 – Nuclear Plant Chemistry Conference, Quebec
City, Quebec, 2010.
11 Acknowledgements
Gordon Burton
Dave Guzonas
Craig Stuart
1
CHAPTER 16
Regulatory Requirements and Licensing
prepared by
Dr. Victor G. Snell & Dr. Nikola K. Popov
Summary:
This chapter covers the overall aspects of nuclear reactor regulation, with emphasis on Canada,
but also covering other regimes such as the United States, the United Kingdom, and Europe. It
explains the need for regulation; how regulators work; how they are structured; and the types of
requirements they set. Canada, the United States, and the United Kingdom are used as specific
examples in these areas, with detailed material on Canadian licensing processes, requirements,
and guides.
Disclaimer: This Chapter captures the elements of nuclear power reactor regulation as of ~2013,
when the Chapter was drafted. Although regulatory concepts change slowly, detailed licensing
requirements can change fairly frequently, and some of the details in this Chapter could be
superseded by the time the book is used. The reader is advised to consult the most up-to-date
regulatory documents if the intended use at the time is critical. The information in this Chapter
is current as of late 2012 or early 2013.
Table of Contents
1 Introduction ............................................................................................................................ 4
1.1 Overview ............................................................................................................................. 4
2 Generic Regulatory Requirements.......................................................................................... 5
2.1 Why is a Regulator Needed?............................................................................................... 5
2.2 What a Regulator is Not...................................................................................................... 6
2.3 How a Regulator Exercises its Responsibilities ................................................................... 7
2.4 How a Regulatory Review is Done ...................................................................................... 9
2.5 Regulatory Independence and Social Policy ..................................................................... 10
3 Regulatory Approaches......................................................................................................... 12
4 Case Studies .......................................................................................................................... 13
4.1 Three Mile Island .............................................................................................................. 13
4.2 Fukushima ......................................................................................................................... 14
4.3 Chernobyl.......................................................................................................................... 15
4.4 Olkiluoto............................................................................................................................ 16
5 Pre-Project Licensing Review ................................................................................................ 16
6 International and National Examples.................................................................................... 18
6.1 International Organizations .............................................................................................. 18
6.2 National Regulatory Organizations ................................................................................... 22
7 Nuclear Material Safeguards................................................................................................. 32
7.1 Objectives and Evolution .................................................................................................. 32
7.2 Safeguards Treaty Detection Fundamentals ..................................................................... 34
©UNENE, all rights reserved. For educational use only, no assumed liability. Regulatory Requirements and Licensing – September 2014
7.3 Safeguard Approaches, Concepts, and Measures ............................................................ 34

7.4 Safeguards Design Implications for Nuclear Power Plants ............................................... 36
8 Nuclear Security.................................................................................................................... 38
9 Canadian Regulatory Process................................................................................................ 39
9.1 The CNSC........................................................................................................................... 40
9.2 CNSC Mandate and Organization...................................................................................... 40
9.3 Canadian Regulatory Framework and Regulatory Acts .................................................... 41
9.4 Licences............................................................................................................................. 44
9.5 CNSC Regulatory Documents ............................................................................................ 45
9.6 CNSC Staff Review Procedures.......................................................................................... 56
9.7 CNSC CANDU Safety Issues ............................................................................................... 56
9.8 CNSC Fukushima Action Items .......................................................................................... 57
9.9 Canadian Codes and Standards......................................................................................... 58
10 Reactor Licensing Process in Canada .................................................................................... 60
10.1 Licensing Process .............................................................................................................. 60
10.2 Site Selection and Site Licensing....................................................................................... 64
10.3 Construction Licence......................................................................................................... 65
10.4 Operating Licence ............................................................................................................. 67
10.5 Licence to Decommission and Return Site to Green Site ................................................. 69
11 Problems ............................................................................................................................... 70
12 Appendix A—Examples of U.K. Safety Assessment Principles.............................................. 71
13 Appendix B—Information Required in Licence Applications................................................ 72
14 References............................................................................................................................. 74
15 Glossary................................................................................................................................. 80
List of Figures
Figure 1 Steps in the development of a nuclear power program ................................................. 12
Figure 2 Hierarchy of IAEA safety documents............................................................................... 19
Figure 3 Simplified organization chart of the USNRC ................................................................... 23
Figure 4 Generic hierarchy of regulatory documents ................................................................... 24
Figure 5 Structure of the ONR....................................................................................................... 29
Figure 6 “Farmer” curve................................................................................................................ 30
Figure 7 HSE concept of tolerable risk .......................................................................................... 30
Figure 8 Safeguards state evaluation is a continuous process [IAEA2007b] ................................ 33
Figure 9 Elements of safeguards [IAEA1998a] .............................................................................. 36
Figure 10 Design aspects of safeguards in LWR reactors (Type II) [IAEA1998a] ........................... 37
Figure 11 Design aspects of safeguards in CANDU reactors [IAEA1998a] .................................... 38
Figure 12 CNSC organization ......................................................................................................... 41
Figure 13 Structure of regulatory documents in Canada.............................................................. 42
Figure 14 Design and documentation completeness phases ....................................................... 63
Regulatory Requirements and Licensing 3
List of Tables
Table 1 IAEA guides related to regulation..................................................................................... 19
Table 2 U.K. frequency / dose targets for accidents off-site......................................................... 31
Table 3 Single / dual failure dose limits ........................................................................................ 55
Table 4 Consultative document C-6 dose limits............................................................................ 56
Table 5 New plants in Canada: dose limits.................................................................................... 56
Table 6 Selected key CSA codes and standards............................................................................. 59
1 Introduction
The purpose of this chapter is to:
1. describe the need for, and concepts underlying, nuclear regulation
2. cover specific requirements in Canada in some detail, and examples from other ju-
risdictions where the differences are instructive.
Whereas the first aspect covers “why”, the second aspect covers both “why” and “what”.
1.1 Overview
Regulatory oversight of nuclear power plants is an essential aspect of the application of the
technology, albeit not generally considered a specific engineering discipline. Regulation affects
reactor design and operation by setting requirements at both a high level (safety goals, system
requirements, operations performance metrics, etc.) and in more detail, and through regulatory
review of design and operation.
Section 2 looks at overall regulatory requirements. It begins by explaining the need for regula-
tion. It then looks at generic aspects such as regulatory structure and independence, which
tend to be broadly similar in advanced nuclear-power-capable countries. It discusses how a
good regulator works through drafting standards and guides and through independent review,
but emphasizes that the operator is ultimately responsible for safety. It describes licensing
milestones (site licence, construction licence, operating licence, decommissioning licence, and
licence to abandon), which are common to most countries.
In Section 3, different regulatory approaches are briefly contrasted, varying from risk- or goal-
driven at one end to highly prescriptive at the other.
Section 4 gives a number of case studies of regulatory behaviour (in the same spirit as Chapter
13, of learning from real experience)—not to suggest that the regulator was primarily responsi-
ble for an event, but that weaknesses in the overall regulatory approach led to inattention on
the part of the industry to certain risks.
Section 5 discusses a recent tool for facilitating the licensing of a new design: a pre-project
licensing review.
Section 6 gives some international and national examples of organizations which influence or
apply regulatory requirements. Specifically, we summarize the consensus developed by the
International Atomic Energy Agency (IAEA) as reflected in its guides and touch briefly on re-
gional organizations. Then we discuss examples of prescriptive (United States) and non-
prescriptive (United Kingdom) approaches.
Safeguards are described in Section 7. Security is also covered briefly in Section 8.
Sections 9 and 10 cover the Canadian approach. This can be considered as a balanced ap-
proach, mixing prescriptive requirements and overall goals. The mandate and organization of
the Canadian Nuclear Safety Commission (CNSC) is presented in Section 9. The structure of
Canadian Nuclear Regulations and the CNSC’s Regulatory Documents and Regulatory Guides is
summarized. Certain key Regulatory Documents affecting design and operation are presented
in more detail. The licensing process in Canada for both existing and new nuclear power plants
is described in Section 10; this includes a summary of the environmental assessment process.
Three supplementary sections follow. Section 11 gives some problems for the student to work
through. Section 12 is an Appendix giving some examples of U.K. Safety Assessment Principles
as typical of a non-prescriptive approach. Section 13 is an Appendix listing the information that
is typically required in Canada when applying for construction and operating licences for nuclear
power plants.
Reactor regulation is a very broad field, and we can cover only the essentials in this chapter.
Most sections in the text list references to documents or published papers that the interested
reader can pursue for more detail: see Section 14 for the list of references. Section 15 con-
cludes the chapter with a glossary of terms.

The goal of this chapter is for the student to understand generically:
 The need for regulatory oversight of nuclear technology
 How a regulatory agency is structured
 How it fulfills its responsibilities
 The regulatory process and milestones
 The nature of a regulatory review
 Enforcement of regulatory requirements
 Common threads in international approaches to regulation
 Different approaches to regulation (prescriptive, goal-based, mixed)
 Concepts behind safeguards and security.
In addition, the student will gain a familiarity with the organization and mandate of the Cana-
dian regulatory organization (the Canadian Nuclear Safety Commission, or CNSC). The student
will understand the Canadian licensing process for both new and existing reactors, as well as
acquiring an overall knowledge of CNSC’s most important requirements and guides.
2 Generic Regulatory Requirements

This section covers regulatory requirements and processes in general, supported by selected
examples. Later sections give details for Canada.
2.1 Why is a Regulator Needed?

Nuclear power is an inherently hazardous and complex technology. Society regulates such
technologies, especially if they are relatively new—non-nuclear examples include prescription
drugs and air travel. Regulatory oversight of a technology can encompass most or all of the life
cycle, including site selection, design, construction, commissioning, operation, decommission-
ing, and site abandonment. In addition, fuel fabrication and management of spent fuel (should
it leave the plant site) are also regulated, although these aspects are not covered in this chapter.
Regulators are authorized by the government—usually the national one—and have strong legal
powers to set requirements and to intervene in the technology they regulate, e.g., to order a
drug to be withdrawn from the market, or to order a nuclear power plant to make changes.
They can use sanctions (fines or prison) on both corporations and individuals in case of egre-
gious non-compliance.
An “ideal” regulatory framework balances the risk of the regulated technology against those of
similar technologies, so that social risks are rationalized. This hardly ever happens because the
mandate of regulators is usually restricted to a specific technology. Nuclear power tends to be
regulated more strictly (i.e., lower risks and more risk aversion) than comparable technologies.
There are many reasons for this which we do not have space to explore, but which have been
touched on briefly in Chapter 13, and which include voluntary versus involuntary risk, perceived
risk versus benefit, association of radiation with atomic weapons and cancer (the “dread”
factor), and public unfamiliarity with the technology. The interested reader is referred to the
literature on this subject. [Starr1969] and [Slovic1987] are good places to start.
One of the key functions of a regulator is that through the potential of independent verification
of any aspects of a licensee’s work, the regulator influences all a licensee’s internal processes,
because any piece of safety-related work can be audited. In particular, the simple existence of
regulatory oversight stops the licensee from internally accepting deficient safety cases or
accepting too readily that their assumptions are right and that they have thought of everything.
In other words, oversight encourages the building and maintaining of an adequate safety
culture. Regulatory oversight provides a critical independent challenge to a licensee’s premises,
assumptions, and positions and affects far more than the actual items that a regulator reviews.
2.2 What a Regulator is Not

A good regulator does not remove the onus for safe plant design from the designer or that for
safe plant operation from the operator. This is a key concept. The ultimate responsibility for
safe plant operation lies with the licensed operator. A regulator sets the framework and the
rules to be followed, but these do not guarantee a safe plant. A deficient regulatory structure,
or a regulator that is not sufficiently independent from government, may contribute to an
accident by inadequate oversight, or inadequate requirements, or by being so prescriptive that
the operator relinquishes the responsibility to maintain a questioning attitude. However, an
operator is not absolved by a weak regulator.
Conversely, a good regulatory structure defines a climate of openness, fairness, and high expec-
tations of safe performance that operators will internalize. In addition, a good regulatory
structure ensures that international regulatory trends and approaches are readily adopted and
implemented by a national regulator.
In any case, it is impossible for a regulator to check everything. This has two implications:
 there is an implicit understanding that once a licence has been given, the licensee
will follow the rules and that the regulator is there to verify selected activities from
time to time as it sees fit, usually in some risk-informed manner, as discussed in Sec-
tion 2.5. This is called “trust but verify”1. See Section 2.3.3.
 a regulatory “pass” does not guarantee that the plant is safe, just as a corporate au-
dit does not remove the responsibility of the corporation to exercise fiduciary re-
sponsibility, and just as such an audit may miss mistakes.
As stated in [IAEA2000]: “The prime responsibility for safety shall be assigned to the operator.
The operator shall have the responsibility for ensuring safety in the siting, design, construction,
1 This is a Russian proverb (Доверяй, но проверяй) originally applied to US-USSR relations during the Cold War.
commissioning, operation, decommissioning, close-out or closure of its facilities, including, as

appropriate, rehabilitation of contaminated areas; and for activities in which radioactive mate-
rials are used, transported or handled.” When an operator takes over a plant, it formally as-
sumes most of the responsibility for design safety. In other words, the “design authority”
responsibility is transferred from the technology vendor to the operator. In fact, the operator is
required to satisfy itself that the design is adequate before assuming operating responsibility.
Although technical aspects of the design authority can be delegated in part to the vendor,
informed oversight cannot be delegated, and the operator must remain at least an “intelligent
customer”. In any case, this does not absolve the design organization from its safety responsi-
bilities.
2.3 How a Regulator Exercises its Responsibilities

2.3.1 Scope
Generally, the authority of a regulator is enshrined in national legislation, which gives it the
ability to permit (or deny) siting, construction, commissioning, operation, decommissioning, and
site abandonment of nuclear facilities, and to control the use of radioactive material. Without a
licence for a particular activity, it is illegal to proceed.
Typically regulatory authority covers more than just nuclear power plants (NPPs), including any
significant nuclear activity (uranium mining, radioactive sources, research reactors, medical and
industrial uses of radiation, radioactive waste disposal, etc.). However, this chapter focusses
only on NPPs.
Most nuclear regulators also take on enforcement of national policy on safeguards (Section 7)
and on security (Section 8) as applied to nuclear power and represent the national government
in international fora on nuclear safety.
2.3.2 Nuclear power plant licensing

The steps in the licensing process, through which a regulator can exercise authority, each
involve submission by the applicant of information required by the regulations (Section 10). The
required information is discussed in more detail in Section 10.1 for Canada; this section gives an
overview.
Typically, there are up to eight major approval milestones for a nuclear power plant. In each,
the operator submits the required documentation; the regulator reviews this and (normally)
grants a licence to start the milestone activity2. The regulator continues to review and monitor
progress as the milestone is implemented. Typical milestones are:
 Environmental assessment (including site evaluation)
 Site preparation licence
 Construction licence
 Initial fuel loading
 Low-power commissioning
2 Federal, provincial, and even municipal government bodies may require many other regulatory approvals, which
are outside the scope of this chapter.
 Operation
 Decommissioning
 Site abandonment (i.e., return to unrestricted usage).
In addition, most regulators perform on request a preliminary review of an NPP design even
before a construction licence application, to reduce the risk of licensing issues arising during the
project stage when they are difficult and expensive to address. In Canada, this is called a pre-
licensing vendor design review [CNSC2012]. In the United States, it is called design certification
[USNRC2012], and in the United Kingdom, generic design assessment [HSE2008]. This topic is
discussed in Section 5. Note that there are differences in the depth, scope, and implications of
these approaches in different countries.
A regulatory review is usually performed against a set of published standards, requirements,
and “expectations” (typically through laws, regulations, and guides). Laws are laws of the land,
set by the national government; regulations are requirements set by the government (which in
some countries are also laws of the land); and guides are non-mandatory “good practices”
issued by the regulatory agency. The licence itself is a legal instrument and typically contains
additional specific requirements that must be met. In addition, the regulator may require, or
the operator may volunteer, compliance with other national standards. In Canada, these are
developed by the Canadian Standards Association (CSA). Every country has its own set of
requirements, but now there is more and more international coordination, initially after the
severe accidents at Three Mile Island and Chernobyl, and now post-Fukushima3.
Usually, there is some risk basis for regulatory requirements, which are generally a mix of risk-
informed and prescriptive requirements based on past experience, as discussed in the accident
case studies in Chapter 13.
In many countries, but not all, regulatory deliberations and decisions are made openly, with
many opportunities for public input.
Given the power of the regulator, most regulators set up checks and balances to remain fair.
Most regulators are structured somewhat like a corporation, with a Commission at the top
supported by a small dedicated group of people, and a large separate specialized technical staff
which assesses the safety case in detail and presents its conclusions and recommendations to
the Commission. The Commission makes the final decision, not the staff. The Commission
reports directly to, or is accountable to, the national government in a way that is independent
of the proponent. Some regulators have independent advisory committees (e.g., the United
States Nuclear Regulatory Commission (USNRC) has an Advisory Committee on Reactor Safe-
guards (ACRS) consisting of highly respected technical experts). Such committees are inde-
pendent of the staff, so that technical inputs to the Commission come from more than one
source [USNRC2012a]. This is a “typical” description—there are differences in the organization
of regulators in different countries because the regulatory structure is the responsibility of each
national government.
National and international peer reviews of regulators use the yardstick of international best
practices to encourage improvement, e.g., as implemented by the IAEA through its Integrated
3 Post-Fukushima refers to the changes in nuclear safety regulations and requirements in response to the nuclear
accident in the Fukushima Dai-ichi NPP in Japan in March 2011.
Regulatory Review Service (IRRS).
2.3.3 Ongoing oversight

Granting of a licensing milestone is only part of a regulator's scope. Regulators also monitor,
through documentation review and site inspections, the construction, commissioning, opera-
tion, and decommissioning of nuclear power plants. Trend indicators are used as input to
decisions to take regulatory action. In particular, assessment of the safety culture in different
parts of the nuclear infrastructure is an important indicator of the overall health of the nuclear
safety structure and organization (for a discussion of safety culture, see Chapter 13). An operat-
ing licence renewal or major plant changes provide opportunities (or requirements) for the
operator to perform an in-depth re-assessment of the plant against modern standards and to
close any gaps where it is practical to do so.
Internationally, it is becoming common for periodic safety reviews (PSRs) to be performed,
typically every ten years [IAEA2013]. The purposes of these are to:
 assess the cumulative effects of plant aging and plant modifications, operating expe-
rience, technical developments, and siting aspects, and
 assess plant design and operation against applicable current safety standards and
operating practices.
PSRs are particularly useful if the licence renewal period is long (40 years or more in many
countries).
More routinely, regulators perform audits on key aspects throughout the nuclear-facility life
cycle, effectively “sampling” performance. Negative audit findings in an area will often trigger
broader and more intrusive regulatory audits, until the regulator is satisfied that non-
compliances are isolated, or conversely that regulatory action is warranted to ensure that any
systemic deficiencies are addressed.
2.3.4 Nuclear emergencies

A regulator has a special role in nuclear emergencies. Although practices vary from country to
country, the regulator is not expected to take control of a station during an emergency or to
direct the operators. However regulatory approval may be needed for such things as planned
containment venting. Instead, the regulator is often responsible for, or heavily involved in,
emergency (off-site) planning and in coordinating the various governmental bodies involved in
public notification, sheltering, prophylactic iodine medication, and evacuation. In some cases,
the regulator may be either de jure or de facto the spokesperson to the public.
2.4 How a Regulatory Review is Done

The following are some of the main tools that a (generic) regulator uses to perform an inde-
pendent review of a licence application4 [IAEA2002]. They include:
4 Often the Safety Report is equated with the documentation required for a licensing milestone (called the safety
case in some countries), but this is not correct—many more documents are reviewed, as described in Figure 14.
The Safety Report as a minimum consists of a detailed design description, a deterministic safety analysis, a
summary of the probabilistic safety assessment, a description of programmatic aspects such as quality assurance
 Issuing requirements and guidance for the scope, content, and quality of the safety
case. This ensures that the licensee or proponent is aware beforehand of what is re-
quired and desired. Typical material includes high-level requirements such as safety
goals or dose limits ([HSE2006], [CNSC2008]); lower-level requirements such as the
scope of safety analysis [CNSC2008b]; and guidance on acceptable methods
[CNSC2012b]. Most jurisdictions also issue requirements on design (e.g., [HSE2006],
[CNSC2008], [USNRC2012b] with various levels of prescriptive detail.
 Issuing Review Guides for regulatory staff use so that their reviews are consistent
and complete. Some of these are published ([HSE2002], [USNRC2012c],
[CNSC2013]), while others remain internal.
 Performing an initial acceptance review of the safety case, for completeness.
 Detailed specialist review of the safety case, using such tools as:
o Line-by-line review using the Review Guides;
o Independent calculations using the same computer codes as the licensee, or us-
ing independent codes;
o Expert elicitation, including use of outside panels or individuals and independent
analysis by consultants;
o Assessment against written international standards (e.g., IAEA), industry stan-
dards (e.g., those from the Canadian Standards Association (CSA)), or against the
licensee’s own procedures;
o Quality assurance review and audit, including management and safety culture;
o Review of the completeness and appropriateness of the research and develop-
ment support for any step in NPP implementation, including design, site licens-
ing, etc.;
o Visits to the site;
o Consultation with international bodies and fellow regulators;
o Independent experiments or independent analysis of industry experiments;
o Questions arising from the above to the licensee and discussion and review of
the answers.
o Extensive technical meetings with the proponent.
 The staff review results are consolidated.
 The position is communicated to the Commission in formal meetings, at which the li-
censee has the opportunity to present its position. The meetings are public, along
with the information that is reviewed and discussed at the meetings. It is the Com-
mission that gives the approval to proceed with the relevant milestone.
2.5 Regulatory Independence and Social Policy

Regulatory bodies generally report to a Minister in Cabinet, or to Parliament through a Minister
(as is the case in Canada), or to a similar level of the national government. The reporting is
usually independent of the industry at the government level [IAEA2000]. Personnel movement
between industry and the regulator is fairly common, especially in smaller countries, and is
and human factors, and demonstrations of compliance with regulatory requirements. See [USNRC, 1978] and
[IAEA, 2004] for examples of the format of a Safety Report; the former is far more widely used. Details of the
probabilistic safety assessment are usually submitted along with the Safety Report.
beneficial because it ensures that regulatory staff have deep hands-on knowledge of the tech-
nology they are to regulate. Often personnel movement occurs internationally, which helps
internationalize and harmonize regulatory experience.
At the top level of the regulator, the Commission Chairman, and usually the Commission mem-
bers, are appointed by the government, and therefore the process is political (in the sense of
reflecting the will of the people through their elected representatives). The government
provides ultimate oversight of the regulatory body, but rarely formally intervenes in licensing
decisions.
Regulatory decisions must balance risks, not eliminate them (risk elimination is an impossibility
in almost any technology). Moreover, although most regulators say they evaluate only safety
and do not explicitly acknowledge benefits, in fact they have a responsibility to account for the
fact that an activity may have significant net social benefit. In this sense, regulation sets (or
reflects) social policy. Hence, regulatory approval means that the project poses an acceptable
level of risk. Very few regulators use only risk as a regulatory tool (i.e., based just on the prob-
abilistic safety analysis (PSA)); see Chapter 13 for a discussion on PSA. Many regulators use or
accept a risk framework to sharpen decisions—this is called risk-informed decision making
(RIDM) and is a combination of PSA insights and traditional deterministic requirements. This is
a broad topic—for more detail, see [Apostolakis2004], [Bujor2010], [CSA1997], and
[USNRC2002]. In certain cases, usually when a retrofit is being considered, a cost-benefit case
prepared by the proponent may be considered to ensure that the cost is not disproportionate to
the benefit; see [HSE2001], [CNSC2000].
Finally, regulators must have adequate resources and experience, but they do not have to
replicate all the resources of the designer or operator; cf. our example of a financial audit. To
ensure adequate financial resources, regulators are given strong government support at the
early stage of their development, whereas later most of their daily operations are on a cost-
recovery basis from the licensees.
The above description applies to countries with an established nuclear industry. A country just
embarking on a nuclear program, but with no nuclear infrastructure, faces special challenges in
developing appropriate expertise and organizations ([IAEA2012b], [Popov2012]).
Four mechanisms have been used. These are not exclusive and are often used in combination.
1. The typical mechanism in the past has been to start with a research reactor and to use it
to develop staff for design, operating, and regulatory functions from that base. This
route was followed by Canada, where the power reactor program benefited from staff
trained in the research reactors (NRU, NRX) at Chalk River. Indeed, the Atomic Energy
Control Board (AECB), established in 1946 under the Atomic Energy Control Act, had re-
sponsibility for making regulations and for the Chalk River research project until Atomic
Energy of Canada Limited (AECL) was created in 1952 under the AECB. Two years later,
the Canadian Government formalized the separation of the designer/operator of the
Chalk River site from the regulatory agency [Sims1980].
2. A second mechanism has been to use the vendor of a power reactor to help create the
required infrastructure in the country purchasing the reactor by technology transfer and
training of operating personnel. The regulatory staff in the purchasing country are de-
veloped under an agreement with the regulator of the vendor’s host country. This can
go as far as “importing” experienced regulatory staff.
3. A third mechanism is to use IAEA documents and assistance. This organization has de-
veloped a comprehensive set of requirements and guides covering all aspects of com-
mercial nuclear applications, including the organization and management thereof. The
IAEA also gives assistance to countries developing a nuclear program, such as training
courses in nuclear regulation [IAEA2002c].
4. A fourth mechanism is to hire technical support organizations (TSOs) to provide techni-
cal assistance to the developing regulatory body. These are neutral official organizations
which are either part of the regulatory body or are separate, for example a national
laboratory. See [IAEA2010] for recent developments.
Figure 1 from [IAEA2011] gives an idea of the steps and timescale involved in a new national
nuclear power program.
Figure 1 Steps in the development of a nuclear power program
3 Regulatory Approaches
The approaches taken by regulatory agencies across the world tend to fall into four general
categories (based on the proven approach used by the regulator in the past; of course, this
assertion is oversimplified, and many regulators allow alternative approaches that can be a mix
of the four listed below):
1. A largely risk-based approach (e.g., the early U.K. Farmer curve [Farmer1967], and
Argentina [Barón1998])
2. A risk-informed, largely non-prescriptive approach with specification of goals or re-
sults, not the process (early Canada, some early IAEA, current U.K.);
3. A risk-informed, but more prescriptive approach (current Canada, IAEA);
4. A highly prescriptive approach, with specified detailed requirements (e.g., USNRC
and the U.S. Code of Federal Regulations).
Each approach has its advantages and disadvantages.
A risk-based approach is logical, quantifies and balances risk, and leaves major design and
operating decisions up to the proponent, who decides how to meet the risk goals. It thereby
accommodates and even fosters innovation and evolution. On the other hand, it can lead to
uncertainty in licensing because demonstration of numerical risk can be subjective or limited by
lack of knowledge. It may not account sufficiently for previously unknown failures, which might
be addressed more effectively by a more prescriptive defence-in-depth approach. In particular,
cross-linked or common-cause failures (see Chapter 13) are difficult to identify completely in a
PSA, so that the contribution to risk from these causes can be underestimated.
A mixed approach, which uses both risk-based and prescriptive approaches to various degrees,
captures most of the benefits (and some of the disadvantages) of both extremes and generally
has served Canada and the United Kingdom well. In particular, it is effective in providing de-
fence against accidents not previously identified, without too much misallocation of resources
to truly rare low-risk events.
A highly prescriptive approach in principle leads to well-defined licensing rules and more
regulatory certainty and tends to produce an apparently conservative design. However, it can
shift the burden of proof of safety from the designer/operator to the regulator. If the de-
signer/operator is required mainly to follow detailed rules set by the regulator, and if the
regulator misses something, part of the responsibility for safety has shifted subtly from the
proponent to the regulator. Furthermore, de-emphasis of risk leads to some resource misallo-
cation to severe but rare events, as opposed to more frequent ones with apparently milder
consequences. We shall cover a case study of this situation (Three Mile Island) in the next
section.
Much of the world, especially those countries using designs originating from U.S. light-water
reactor (LWR) technology, follows the U.S. regulatory approach, largely because the regulatory
approach was adopted along with the LWR technology.
4 Case Studies
This section considers a few case studies of regulatory behaviour. Where possible, the regula-
tory agency’s self-evaluation (or that of an independent body) is used, rather than the authors’
opinions. The case studies are presented not as a criticism of a particular organization, but as
lessons learned about effective regulatory behaviour, as was done for design and operational
safety in the case studies in Chapter 13. Some of these examples are also covered in Chapter
13, but with an emphasis on design and operation, rather than regulation.
Success is marked by the lack of accidents and “near misses”, and just as an accident cannot be
laid solely at the feet of a regulator, neither can the lack of accidents be solely ascribed to the
quality of regulation. As Three Mile Island, Chernobyl and Fukushima were studied from the
technical point of view in Chapter 13, we shall now look at these accidents from the regulatory
point of view.
4.1 Three Mile Island

The TMI sequence of events was covered in Chapter 13. A minor event escalated through a
small and unrecognized loss of coolant into a partial core melt which was contained in the
vessel. Two commissions were struck to carry out an independent review.
The President of the United States appointed a Commission headed by John G. Kemeny. Their
report [Kemeny1979] was very broad. Their concentration was on human aspects: “The equip-
ment was sufficiently good that, except for human failures, the major accident at Three Mile
Island would have been a minor incident.” Of interest here are their findings on the U.S. regula-
tory structure at the time, which are summarized as follows:
 The regulations were too voluminous and complex, required immense effort for
compliance, and equated compliance with safety;
 The preoccupation with the most severe accident (the largest-break LOCA) took at-
tention away from more likely, but slower-developing accidents, which were there-
fore not analyzed in depth;
 There was too much preoccupation with equipment performance rather than human
performance;
 There was no requirement to look beyond the single events specified by the USNRC,
for example to multiple failures;
 The role of systems classified as “non-safety-related” in causing accidents was not
recognized;
 There was no systematic way of evaluating prior operating experience or looking for
patterns;
 There were serious deficiencies in internal communication in the USNRC.
An additional review by the Nuclear Regulatory Commission Special Inquiry Group [Rogo-
vin1980] covered very similar ground, albeit with more comments on the USNRC internal
decision-making processes.
Subsequent to TMI, and based on the findings of the Kemeny Commission, severe accidents,
emergency preparedness, plant performance monitoring, and human factors were given far
more prominence on the regulatory side. Insights from PSA were used more widely, not only to
reveal severe accident vulnerabilities, but also to flag requirements that attracted inordinate
resources compared to the risk they mitigated. The industry formed cooperative groups such as
the Institute of Nuclear Power Operations (INPO) to promote high standards of safety and
reliability through plant evaluations, training and accreditation, events analysis, and information
exchange and assistance.
4.2 Fukushima
The Fukushima Dai-ichi sequence of events was covered in Chapter 13. A beyond-design-basis
tsunami caused a loss of all electrical power in four units, resulting in severe core damage or
melting and possible containment failure in the three reactors that contained fuel.
With respect to the Japanese regulator (the Nuclear and Industrial Safety Agency (NISA)), the
National Diet of Japan Nuclear Accident Independent Investigation Commission [Japan2012]
concluded that “root causes were the organizational and regulatory systems that supported
faulty rationales for decisions and actions”, and in particular that the inadequacy of the design
basis for Fukushima was known to both the utility and the regulator, but was not acted on. The
regulator lacked separation from the utility. The Commission recommended formation of a new
regulatory body which would be independent, transparent, professional, consolidated, and
proactive.
The Japanese Government also established the “Investigation Committee on the Accident at
Fukushima Nuclear Power Stations of Tokyo Electric Power Company”, an independent commit-
tee largely composed of academics. Their final report [Japan2012a], although more technical
than the Diet report, raised the same issues and called for reorganization of the regulatory
body. They also noted NISA’s attention to short-term rather than long-term issues. In addition
they focussed on NISA’s performance after the event, in emergency measures off-site, and in
public communication.
As of this writing (early 2014), the Japanese regulatory body has been reorganized and renamed
the Nuclear Regulation Authority (NRA).
Regulators worldwide have applied the technical lessons learned, as outlined in Chapter 13. In
terms of national response, in particular, the radiation threshold levels that were used for
evacuation were based on the Linear No Threshold (LNT) hypothesis, which implies an unrealis-
tically high level of risk for individual doses of radiation <100 mSv, although no effects on human
populations have actually been observed at that level. Although there were no deaths due to
radiation from the accident (and there are not expected to be), the evacuation itself resulted in
many unnecessary deaths [Tanigawa2012].
4.3 Chernobyl
The Chernobyl sequence of events was covered in Chapter 13. A test which involved running
down the main reactor coolant system (RCS) pumps while at low power ended up in a power
transient due partly to the voiding of the core, but mainly to the reverse action of the shutdown
system, destroying the reactor core and bypassing the containment.
The International Safety Advisory Group (INSAG), which provides independent advice to the
IAEA Director-General, performed very extensive fuellings of the Chernobyl accident
[IAEA1992]. This reference includes as an Appendix the Report by a Commission to the USSR
State Committee for the Supervision of Safety in Industry and Nuclear Power, which covered the
technical facts and the role of the regulatory structure. It noted that:
 There were known violations of the safety standards and regulations in force at the
time in the design of Chernobyl Unit 4, but the design was approved and authoriza-
tion given for construction by all the relevant authorities and regulatory bodies.5
 This deficiency resulted from a lack of a well-organized group of experts endowed
with its own resources, rights, and responsibilities for its decisions;
 The USSR State Committee for the Supervision of Nuclear Power Safety could not be
regarded as an independent body because it was part of the same state authority re-
sponsible for the construction of nuclear power plants and for electricity generation.
 The regulatory bodies had no legal basis, no economic methods of control, and no
human and financial resources.
 The operating organization did not have ultimate responsibility and decision-making
authority for safety.
5 Note added by authors: there are always outstanding regulatory issues when a plant is given a licence—what is
important is that the process to determine which issues are acceptable should be an open and robust one.
4.4 Olkiluoto
It is not just accidents than can reveal a regulatory problem. Cost and schedule overruns
because of poor communication between the vendor/operator and the regulator can be very
damaging to a nuclear build project. This has been exemplified by the construction of the third
nuclear power plant unit at Olkiluoto in Finland. The reactor is based on the European pressur-
ized-water reactor (EPR) and is the first of its kind to have started construction. It was meant to
start operation in 2009, but as of this writing (early 2014), it is projected to come into service in
2016. This delay has been accompanied by a cost overrun of almost double the turnkey price of
~3.5 billion euros.
Because the matter is currently in legal dispute, the parties have not published much objective
analysis of what went wrong. However, the Finnish regulator (STUK) did comment publicly on
several occasions, listing lessons learned as follows, from [STUK2006], [Laaksonen2008], [Laak-
sonen2009], [Tiippana2010], [IAEA2012]:
 “too ambitious original schedule for a plant that is first of its kind and larger than any
NPP built earlier;
 inadequate completion of design and engineering work prior to start of construction;
 shortage of experienced designers;
 lack of experience of parties in managing a large construction project;
 world-wide shortage of qualified equipment manufacturers”.
The operator TVO, and particularly the vendor AREVA, have offered their own perspective,
in which one of the causes mentioned is evolution of Finnish regulations during the
construction process, which introduced additional uncertainty into the licensing process.
Of particular interest to the thrust of this chapter is the need for mutual understanding be-
tween the regulator and the operator on how the regulatory practices are to be applied:
“The licensee and the regulator need to discuss early enough on how the national safety
requirements should be best presented in the call for bids.
- just making reference to national requirements and regulatory guides is not adequate to
ensure that requirements are correctly understood by vendors”.
With respect to the last point: In Finland, before a project is committed, the Government makes
a “Decision in Principle”, for part of which the regulator reviews the basic design requirements
and main safety features of each proposed alternative design and gives a formal preliminary
safety assessment. STUK confirmed in this assessment that “no safety issues can be foreseen
that would prevent the proposed plant(s) from meeting Finnish nuclear safety regulations”.
Although the regulator can apply caveats to this statement in terms of changes that would be
required, it is not based on a very detailed review.
5 Pre-Project Licensing Review

The risk to a project from delays associated with uncertainties in the regulatory legal process
and requirements has been recognized for decades, spurred by cost and schedule overruns in
nuclear plant construction in the United States in the 1970s. The solution adopted in many
countries has been a form of pre-project licensing review which is done before project com-
mitment. Typically, a vendor or utility requests a review of a proposed design against the
regulatory requirements of the country in which it is to operate. Such a review ensures that the
vendor understands the regulatory requirements and enables regulatory questions to be
flushed out and answered earlier. The review is very thorough and is similar in many ways to
the construction licence review, but focussing mainly on design details rather than operation.
The format of such a review, and the extent to which it is binding on the regulator, depend on
national legal and cultural practices. We describe briefly three examples.
In Canada, a designer can request a Pre-Licensing Vendor Design Review, which [CNSC2012]
“…is to inform the vendor of the overall acceptability of the reactor design…
A Pre-Licensing Vendor Design Review evaluates whether:
o the vendor understands Canadian regulatory requirements and expectations;
o the design complies with, as applicable, CNSC regulatory documents RD-3376,
Design of New Nuclear Power Plants or RD-367, Design of Small Reactor Facili-
ties, and related regulatory documents and national standards;
o a resolution plan exists for any design issues identified in the review”.
The review is not binding on the CNSC (i.e., the CNSC can request repeated in-depth reviews on
the same topics during construction licence review, and CNSC review results may be subject to
change from pre-licensing to the construction licence review).
In the United States, Standard Design Certification is a legal process covered by Part 52 to Title
10 of the Code of Federal Regulations [USNRC2012]. The design submitted and the review are
both comprehensive, although some detailed design normally performed in the project phase
may be more conceptual (e.g., details of control and instrumentation design). When com-
pleted, the product is a standard reactor design certified by the Commission, independent of a
specific site, through a rulemaking action (Subpart B of Part 52). This rulemaking action can
certify portions of a reactor design for 15 years.
Separately, a licensee can apply for an early site permit for approval of a site for one or more
nuclear power facilities. This is also separate from the filing of an application for a construction
permit or a combined licence (see below) for the facility. A specific design is not required for an
early site permit, but the design eventually chosen must fit within the safety and environmental
parameter envelope requested and specified in the early site permit.
As a third option, a licensee can request a combined licence, which is a combined construction
permit and operating licence for a nuclear power facility.
A licensee can use any combination of these tools. For example, a licensee can obtain an early
site permit based on the environmental and safety performance envelope of a set of designs;
can obtain certification of a specific design that meets the envelope; and can then ask for a
combined construction and operating licence for the certified design on the approved site.
Standard design certification and early site permits are legally binding, subject to conditions
(specified during the certification process) to be met by the detailed project design.
In the United Kingdom, the Office for Nuclear Regulation (ONR) and the Environment Agency
6 Note that the CNSC document titles are in the process of changing; this document has been renamed and revised
as REGDOC 2.5.2. Other CNSC regulatory documents are being revised and renamed as REGDOCs.
(EA) offer a generic design assessment (GDA) of new reactors. The purpose is [HSE2008] an
“assessment of the safety case for a generic design, leading to issue of a design acceptance
confirmation if the outcome is positive”. It consists of four phases: the first is preparation of the
safety case by the proponent, and the next three are reviews by ONR and EA at increasing levels
of detail, namely: 1) fundamental safety overview; 2) overall design safety review; and 3)
detailed design assessment. A design acceptance confirmation is then issued, subject to possi-
ble conditions and exceptions, if the design is considered acceptable. Once completed, a GDA
ensures that within the scope of the design, ONR review results are binding. ONR may extend
the review during the construction licence review, but the results of the prior GDA remain
binding.
6 International and National Examples

6.1 International Organizations
A number of international bodies have a quasi-regulatory role. Regulation is still a national legal
responsibility, and therefore these organizations do not have formal authority. However, they
influence governments, national bodies, and nuclear organizations, and if a country formally
adopts their guidance, it takes on the force of national law or regulation.
As before, we give an overview, and the reader may find out more in the references.
6.1.1 IAEA
The origins and purposes of the International Atomic Energy Agency (IAEA) are described in
Chapter 13, to which the reader is referred.
Its current document structure with respect to nuclear safety and regulation includes a hierar-
chy of Principles, Standards, and Guides, as shown in Figure 2, from [IAEA2012a]. Although
these have no legal or regulatory force unless they are formally adopted on a national level,
they are seen as setting minimum international requirements for safety and regulation. The
structure and nomenclature in Figure 2 are self-explanatory, with Safety Fundamentals and
Safety Requirements being mandatory (if adopted) and Safety Guides being recommendations
and guidance on how to comply with the Requirements. As of this writing (early 2014), they
have been formally adopted in countries such as China and the Netherlands; are used directly to
establish national requirements in Canada, the Czech Republic, Germany, India, Korea, and the
Russian Federation; are used as a reference for review of national standards by most nations;
and are also used as inputs by international organizations such as the Western European
Nuclear Regulators’ Association (WENRA) (Section 6.1.2).
Figure 2 Hierarchy of IAEA safety documents

Of particular relevance to this Chapter, IAEA report GS-R-1 [IAEA2000] deals, among other
things, with overall regulatory structure and organization. It is supported by a number of more
detailed guides.
Table 1 IAEA guides related to regulation
Report Title
Number
SSG-16 Establishing the Safety Infrastructure for a Nuclear Power Programme (2012)
GS-G-1.1 Organization and Staffing of the Regulatory Body for Nuclear Facilities (2002)
GS-G-1.2 Review and Assessment of Nuclear Facilities by the Regulatory Body (2002)
GS-G-1.3 Regulatory Inspection of Nuclear Facilities and Enforcement by the Regulatory
Body (2002)
GS-G-1.4 Documentation for Use in Regulating Nuclear Facilities (2002)
GS-G-1.5 Regulatory Control of Radiation Sources (2004)
GS-G-3.5 The Management System for Nuclear Installations (2009)
SSG-12 Licensing Process for Nuclear Installations (2010)
SSG-16 Establishing the Safety Infrastructure for a Nuclear Power Programme
GS-G-4.1 Format and Content of the Safety Analysis Report for Nuclear Power Plants (2004)
Some of the key principles reflected in these Requirements and Guides are:
 a regulator’s purpose is the verification and assessment of safety in compliance with
regulatory requirements;
 it has sufficient resources to perform its job;
 it is effectively independent from other organizations that could influence its deci-
sions;
 it has qualified and competent staff;
 its management system is aligned with its safety goals;
 it may solicit expert advice, but does not delegate decisions;
 it communicates professionally with interested parties and with the public;
 regulatory control is stable and consistent;
 it authorizes all facilities unless specifically exempt;
 authorization is supported by a demonstration of safety submitted by the applicant;
 regulatory review occurs before authorization and over the facility lifetime;
 review is commensurate with risks;
 it inspects facilities;
 it has legal enforcement powers and can require corrective action;
 it issues, reviews, and promotes regulations and guides;
 it maintains records;
 it promotes a safety culture (see Chapter 13);
 it ensures that emergency preparedness arrangements and emergency plans are in
place.
Finally, there is a growing international consensus on high-level numerical safety goals. Safety
goals are defined and discussed in Chapter 13 (Section 4), and the reader is referred there for
details. Almost all regulators who specify safety goals do so implicitly or explicitly on the
following basis: the goals are set so that the predicted health effects of normal operation and
accidents in nuclear power plants are small compared to other social risks. Although IAEA itself
has not specified numerical safety goals, its senior advisory body (INSAG, in [IAEA1999]) has
proposed the following targets:
 For existing nuclear power plants:
o Frequency of occurrence of severe core damage < 10–4 events per plant operat-
ing year;
o Probability of large off-site releases requiring short term off-site response < 10–5
events per plant operating year.
 For future nuclear power plants:
o Frequency of occurrence of severe core damage < 10–5 events per plant operat-
ing year;
o Practical elimination7 of accident sequences that could lead to large early radio-
active releases.
These targets have been used by many other international and national bodies.
7 i.e., if it is physically impossible for the conditions to occur or if the conditions can be considered with a high
degree of confidence to be extremely unlikely to arise [IAEA, 2004a]
Note also that the frequency of the design basis earthquake (DBE) has recently been changed in
most jurisdictions from 10-3 / year to 10-4 / year, meaning that the DBE that new plants are
required to withstand is significantly more severe.
6.1.2 MDEP and WENRA

There has been an impetus for several years to “harmonize” nuclear regulations, much as
aircraft regulations are common internationally. The ideal is that a reactor design licensed in
one competent jurisdiction would be licensable anywhere because the requirements would be
the same. In practice, licensing requirements are closely linked to design type, and therefore
harmonization has made slow progress because of the wide variety of reactor types requiring
different regulations at the detail level.
An initiative called the “Multinational Design Evaluation Programme” (MDEP) was established in
2006 to:
 enhance multilateral co-operation within existing regulatory frameworks;
 encourage multinational convergence of codes, standards, and safety goals;
 facilitate the licensing of new reactors.
It currently includes nuclear regulatory authorities from 13 countries. Its work includes co-
operation on safety reviews of specific reactor designs. National regulators retain sovereign
authority for all licensing and regulatory decisions.
The Western European Nuclear Regulators’ Association (WENRA) was formed in 1999 to de-
velop a common approach to nuclear safety and to be able to review nuclear safety independ-
ently in countries applying for membership in the European Union (EU).
One of their earliest reports [WENRA2000] surveyed the state of nuclear safety and regulation
in countries applying for EU membership. The report covers the status of the regulatory regime
and regulatory body and of nuclear power plant safety. For the former, issues such as inde-
pendence of the regulatory body, resources, and expertise were priority topics.
Because the predominant type of reactor in Western Europe is the LWR, it is somewhat easier
to propose harmonized requirements. WENRA has adopted common top-level requirements for
any reactor to be built in Western Europe [WENRA2010]. These are largely consistent with the
IAEA Safety Fundamentals [IAEA2006a]. More detailed requirements are given in the WENRA
“Reference Levels” publications, e.g., [WENRA2007], for reactor safety.
Finally, as requested by the European Union (EU) after the accident at Fukushima Dai-ichi in
March 2011, WENRA proposed [WENRA2011] that a series of analytical “stress tests” be done
on all EU nuclear power plants to show their robustness against extreme natural events. Unlike
the traditional approach of analyzing design basis accidents (DBAs) and beyond design basis
accidents (BDBAs) (see Chapter 13), WENRA proposed an analysis assuming:
“sequential loss of the lines of defence in a deterministic approach, irrespective of the
probability of this loss. In addition, measures to manage these situations will be sup-
posed to be progressively defeated.”
In other words, design and accident management measures were to be assumed to fail one
after the other and the consequences at each step tabulated to understand the plant’s robust-
ness. The initiating events conceivable at the plant site included earthquake, flooding and other
extreme natural events. The consequential loss of safety functions included loss of electrical
power, including station black-out, loss of the ultimate heat sink, and a combination of both.
Each EU country (or country applying for EU membership) completed these stress tests for its
own plants and reported the results back to the European Nuclear Safety Regulators’ Group
(ENSREG), an independent expert body formed by the European Commission, for evaluation.
6.2 National Regulatory Organizations

Although each country operating a nuclear power plant has a national regulatory agency, the
overall approach taken generally follows that used in the countries which have designed and
deployed the major types of reactors: Canada for pressurized heavy-water reactors (PHWRs),
the United States for LWRs, and the United Kingdom for gas-cooled reactors. In this section, we
give a brief description of the regulatory approach in the latter two examples; Canada is covered
later and in more detail in Sections 9 and 10.
It is impossible in one section of one chapter to cover the full breadth of regulatory philosophy
and practices in several jurisdictions. Rather, we focus on the overall philosophy rather than the
mechanics of licensing and give references so that the interested reader can find more detail.
6.2.1 USNRC
6.2.1.1 Background
In the United States, the Atomic Energy Act of 1946 established the Atomic Energy Commission
(AEC), transferring atomic energy from military to civilian control (see [Buck1983] for a history).
Its early years were devoted to nuclear weapons production and research, including the military
submarine program, whose reactors became the inspiration for the U.S. PWR commercial power
reactor design. The revised Atomic Energy Act of 1954 gave the Atomic Energy Commission the
responsibility for regulating and licensing commercial atomic activities. Research, military use,
regulation, licensing, and development of the nuclear power industry were all under the same
organization, and although it set up separate divisions, the conflicts of interest became more
and more apparent, until in 1974 the regulatory functions were transferred to the newly-
formed U.S. Nuclear Regulatory Commission (USNRC), which began operations on January 19,
1975.
6.2.1.2 Organization
The USNRC has had enormous influence on regulatory practice in the world because the wide-
spread use of LWR technology was accompanied by the LWR regulatory philosophy. Indeed,
because almost all the reactors licensed in the United States have been PWRs or BWRs8, the
regulatory requirements tend to be specific to these designs.
The USNRC is headed by a five-member Commission appointed by the U.S. President for stag-
gered five-year terms. The President designates one member to serve as Chairman. The
Commission formulates policies and regulations governing nuclear reactor and materials safety,
issues orders to licensees, and adjudicates legal matters brought before it. The Executive
8 The United States have also licensed a fast-breeder power reactor (Fermi 1) and high-temperature gas-cooled
reactors (Peach Bottom and Fort St. Vrain).
Director for Operations (EDO) carries out the policies and decisions of the Commission and
directs the activities of the program offices.
The Commission is supported by a large technical staff, as well as powerful advisory committees
who can challenge both the proponent and the staff, e.g., the Advisory Committee on Reactor
Safeguards, or ACRS. Typically, the ACRS consists of well-known and respected scientists and
experts, often world-renowned professors drawn from universities. It has its own independent
support staff. There is no current equivalent in Canada, although at one point the Reactor
Safety Advisory Committee (RSAC), which later became the Advisory Committee on Nuclear
Safety (ACNS), had a similar role with the then-AECB, albeit on a smaller scale.
The actual trial-level adjudicatory body of the USNRC is the Atomic Safety and Licensing Board
Panel (ASLB). The Panel is composed of administrative judges who are lawyers, engineers, and
scientists, and administrative law judges who are lawyers. It is chaired by the Chief Administra-
tive Judge. The Panel conducts all licensing and other hearings as directed by the Commission,
primarily through individual Atomic Safety and Licensing Boards or single presiding officers
appointed by either the Commission or the Chief Administrative Judge. Individual licensing
boards conduct public hearings on contested issues that arise in the course of licensing and
enforcement proceedings and uncontested hearings on construction of nuclear facilities.
Figure 3 Simplified organization chart of the USNRC

Figure 3 shows a highly simplified organization chart of the USNRC. The five-member Commis-
sion (yellow) is supported by the ACRS, the ASLB, and legal counsel, shown in red. The large
technical staff (turquoise) is headed by the Executive Director for Operations. For nuclear
regulatory purposes, the United States is divided into four regions; the four regional offices
conduct inspection, enforcement, and emergency response programs for licensees within their
borders. USNRC also has a large and powerful R&D Division which conducts experimental and
analytical work of very broad scope, along with development of their own computer tools,
independently of utilities and design organizations.
6.2.1.3 Regulatory approach and documents

As we saw with the IAEA and shall see with other regulators, regulatory documents form a
hierarchy, with legislation at the top, government regulations and licences next (both of which
are enforceable by law), mandatory regulatory requirements set by the regulator next, then
optional regulatory guides provided by the regulator, and finally guidance to regulatory staff
themselves on how to review a submission. Figure 4 shows such a generic structure.
Figure 4 Generic hierarchy of regulatory documents
6.2.1.3.1 Consolidated legislation

In the United States, nuclear energy-related activities are governed by the Atomic Energy Act of
1954, as amended, and the Energy Reorganization Act of 1974, as amended (which created the
USNRC). The interested reader is referred to [USNRC2011] as updated, which consolidates all
nuclear-related legislation and executive orders and runs well over 1000 pages.
The Nuclear Safety and Control Act in Canada [GovCan2013], by contrast, has 54 pages.
[Ahearne1988] is an excellent source of information on the origins of the U.S. regulatory ap-
proach and its differences from that used in Canada. Although the work is dated, it is recom-
mended reading to gain an understanding of the differences. One interesting difference in
practice is the length of a plant licence; typically, in the United States, it is up to 40 years,
whereas in Canada the utility must reapply about every five years.
6.2.1.3.2 Code of Federal Regulations

The technical requirements are covered in the Code of Federal Regulations Title 10 (10CFR),
which corresponds to the second box from the top in Figure 4. The requirements are binding on
all persons and organizations who receive a licence from the NRC to use nuclear materials or to
operate nuclear facilities. Of particular interest are 10CFR Part 50 [USNRC2012d] and 10CFR
Part 52 [USNRC2012].
10CFR Part 50 covers requirements for the traditional domestic licensing of production and
nuclear power facilities. Much of the main text is legal and procedural, and these aspects will
not be discussed further here. Even a detailed technical discussion of 10CFR Part 50 is well
beyond the scope of this chapter. We select a few examples to give an idea of the approach.
Section 50.46 states acceptance criteria for emergency core cooling systems for light-water
nuclear power reactors. The requirements are in the form of numerical analysis values rather
than safety objectives. For example, it is an overall requirement in water reactors to prevent
excessive oxidation of the sheath because if it becomes brittle, it can fragment on rewet, and
the debris can block fuel cooling. Moreover, at high temperatures, the Zircaloy-water reaction
becomes autocatalytic, so that temperature prediction is difficult. Finally the reaction produces
hydrogen, which is a source of energy to containment if it burns. In Canada, these requirements
are stated as objectives, and it is up to the proponent to justify how they are met. However, in
10CFR Part 50, the acceptance criteria can be quite explicit (from 50.46):
“(b)(1) Peak cladding temperature. The calculated maximum fuel element cladding
temperature shall not exceed 2200° F.
(2) Maximum cladding oxidation. The calculated total oxidation of the cladding shall
nowhere exceed 0.17 times the total cladding thickness before oxidation. As used in this
subparagraph, total oxidation means the total thickness of cladding metal that would be
locally converted to oxide if all the oxygen absorbed by and reacted with the cladding lo-
cally were converted to stoichiometric zirconium dioxide. If cladding rupture is calcu-
lated to occur, the inside surfaces of the cladding shall be included in the oxidation, be-
ginning at the calculated time of rupture. Cladding thickness before oxidation means
the radial distance from inside to outside the cladding, after any calculated rupture or
swelling has occurred, but before significant oxidation. Where the calculated conditions
of transient pressure and temperature lead to a prediction of cladding swelling, with or
without cladding rupture, the unoxidized cladding thickness shall be defined as the clad-
ding cross-sectional area, taken at a horizontal plane at the elevation of the rupture, if it
occurs, or at the elevation of the highest cladding temperature if no rupture is calculated
to occur, divided by the average circumference at that elevation. For ruptured cladding,
the circumference does not include the rupture opening.
(3) Maximum hydrogen generation. The calculated total amount of hydrogen generated
from the chemical reaction of the cladding with water or steam shall not exceed 0.01
times the hypothetical amount that would be generated if all of the metal in the clad-
ding cylinders surrounding the fuel, excluding the cladding surrounding the plenum vol-
ume, were to react.”
Although this example illustrates a difference in regulatory philosophy (and not all criteria are
this detailed), it is also related to design differences. In PWRs, if the metal-water reaction
becomes autocatalytic, there is no nearby heat sink to mitigate it, and it can quickly affect a
large part of the core. In CANDU (as discussed in Chapter 13) the nearby moderator can remove
heat from the fuel through the pressure tube and has the effect of tempering the metal-water
reaction rate.
As another example, consider Appendix A to 10CFR 50, the General Design Criteria
[USNRC2012b]. This Appendix consists of 64 basic requirements which underpin much of the
technical approach to safety in the United States. Compliance is mandatory. Although high-
level, many are focussed on LWRs, as stated in the introduction:
“These General Design Criteria establish minimum requirements for the principal design
criteria for water-cooled nuclear power plants similar in design and location to plants for
which construction permits have been issued by the Commission. The General Design
Criteria are also considered to be generally applicable to other types of nuclear power
units and are intended to provide guidance in establishing the principal design criteria
for such other units.”
Many of the GDCs are generic; for example
“Criterion 16—Containment Design. Reactor containment and associated systems shall
be provided to establish an essentially leak-tight barrier against the uncontrolled release
of radioactivity to the environment and to assure that the containment design condi-
tions important to safety are not exceeded for as long as postulated accident conditions
require.”
Some are clearly specific to LWRs, for example:
“Criterion 11—Reactor Inherent Protection. The reactor core and associated coolant sys-
tems shall be designed so that in the power operating range the net effect of the prompt
inherent nuclear feedback characteristics tends to compensate for a rapid increase in re-
activity.”
The underlying purpose of Criterion 11 is that a fast reactivity excursion in an LWR, caused for
example by a rod ejection, must be compensated for (not necessarily terminated, as discussed
in Chapter 13) by negative (Doppler) feedback since (because of the short prompt neutron
lifetime) engineered shutdown systems may not be able to act quickly enough (see Chapter 13).
It is often misinterpreted as a blanket requirement for a negative void reactivity.
At the same level in Figure 4 are plant licences, rules (U.S. term for regulations), and orders.
6.2.1.3.3 Commission Policy Statements

Commission Policy Statements are an example of Regulatory Documents (third box down in
Figure 4). For example, in 2011, the Commission issued a Policy Statement on Safety Culture,
setting forth its expectations for industry to promote a positive safety culture. Such notices are
published in the U.S. Federal Register.
Given the detail in 10CFR, USNRC staff have less need to issue mandatory regulatory docu-
ments.
6.2.1.3.4 Regulatory guides

Regulatory guides (fourth box from the top in Figure 4) provide guidance to proponents on
implementing USNRC regulations, on techniques used by NRC staff in evaluating specific prob-
lems or postulated accidents, and on data needed by staff in its review of applications for
permits or licences. Although nominally optional, exceptions taken to regulatory guides in the
United States (and most other countries) require thorough justification and significant extra
regulatory review.
6.2.1.3.5 Standard review plan

Many regulators now have a standard review plan (SRP, bottom box in Figure 4) written by the
regulator and aimed at regulatory staff performing reviews of proponent applications. This
approach was first developed in the United States and is now being adopted in other countries,
including Canada. The purpose of the SRP is to ensure the quality and uniformity of staff safety
reviews and to make the review process more transparent. For this reason, the SRP is public in
the United States, and proponents can see in advance the criteria against which their submis-
sions will be reviewed and the level of information needed; see [USNRC1987].
6.2.1.4 Licensing process

It is beyond the scope of this chapter to describe the U.S. licensing process in detail. The
following description is highly abbreviated; the reader is referred to [USNRC2009] for more
detail.
We have already described in Section 5 aspects of an alternative licensing approach in the
United States governed by 10 CFR Part 52 that includes options for:
 Early site permits that enable an applicant to obtain approval for a reactor site with-
out specifying the design of the reactor(s) that could be built there;
 Certified standard plant designs which can be used as pre-approved designs;
 A licence which combines a construction permit and an operating licence with condi-
tions for plant operation.
However, at the time this chapter was written (early 2014), all operating plants in the United
States have been licensed under a two-step process defined in 10CFR50:
 A construction licence for a particular plant design at a particular site. This process
includes submission of relevant information on design, safety analysis, intended op-
eration, the site, the environment, and emergency planning; review by NRC staff,
who issue a Safety Evaluation Report summarizing the anticipated effect of the pro-
posed facility on public health and safety; and a public hearing, conducted by the
Atomic Safety and Licensing Board (see Section 6.2.1.2).
 An operating licence. This process includes submission of final design and safety in-
formation by the applicant and review by the USNRC, who prepare a Final Safety
Evaluation Report. A public hearing is not mandatory or automatic, but can be re-
quested by members of the public whose interests might be affected by the issuance
of the licence.
In parallel, the licence applications are reviewed by the ACRS (see Section 6.2.1.2), who provide
the Commission with independent expert advice.
The entire process is remarkably open, with all documents and correspondence related to the
application placed in publicly accessible paper and electronic repositories9. This includes the
plant Safety Report, the NRC Safety Evaluation Reports, letters between the NRC and the
applicant, records of meetings, and other documents. The NRC also holds numerous public
meetings during the licensing process.
In addition, conditions may be placed on the licence, which have the same legal force as the
licence itself. A generic set of conditions is described in 10CFR § 50.54, “Conditions of licenses”.
Many areas are covered, e.g., quality assurance, minimum staffing levels, licence amendment or
revocation, responsibility of the operator, process for making changes, security, and emergency
planning.
Finally, the USNRC has the authority to impose penalties for violations of regulatory require-
ments. It uses three primary enforcement sanctions:
1. Notice of Violation: This identifies a requirement and how it was violated and normally
requires a written response.
2. Civil Penalties: A civil penalty is a monetary fine up to $130,000 per violation per day.
3. Orders: Orders modify, suspend, or revoke licenses or require specific actions by licen-
sees or persons.
6.2.2 United Kingdom

Unlike the United States, the United Kingdom has a very different legal structure, and the
regulator has had to license a wide variety of reactor types (gas-cooled reactors, fast reactors,
heavy-water reactors, and light-water reactors). This task has been accompanied by a less
prescriptive approach.
6.2.2.1 Organization
The Office for Nuclear Regulation (ONR) is responsible for all nuclear regulation in the United
Kingdom. It was formed on April 1, 2011, as an agency of the Health and Safety Executive (HSE).
ONR combined the safety and security functions of HSE’s former Nuclear Directorate, including
Civil Nuclear Security, the U.K. Safeguards Office, and the Radioactive Materials Transport team.
In turn, HSE is a non-departmental public body (NDPB), which is defined as a “body which has a
role in the processes of national Government, but is not a Government Department or part of
one, and which accordingly operates to a greater or lesser extent at arm’s length from Minis-
ters” [UK2009]. It is sponsored by the Department for Work and Pensions. At the top level, it
has a Board, which also contains its Executive Management Team. The work is organized along
program lines, as shown in simplified form in Figure 5.
On April 1, 2014, ONR became a Public Corporation under the 2013 Energy Act. ONR is now the
enforcing authority for licensed sites in Great Britain, as well as the licensing authority.
9 Some records are not disclosed, e.g., records which must be kept secret in the interest of national defence or
foreign policy; internal personnel records; trade secrets and commercial or financial information; records compiled
for law-enforcement purposes. For a comprehensive list, see 10CFR § 9.17, “Agency records exempt from public
disclosure”.
6.2.2.2 Regulatory philosophy

One of the underpinnings of U.K. nuclear safety philosophy is the concept of “so far as is rea-
sonably practicable”, or SFAIRP. This was introduced in the Health and Safety at Work Act of
1974 [UK1974], Section 2-(1), and applies to all industry, not just nuclear:
2.-(1) It shall be the duty of every employer to ensure, so far as is reasonably practicable,
the health, safety, and welfare at work of all his employees.”
For this concept, HSE uses the term As Low as Reasonably Practicable (ALARP) and states that
[HSE2006]:
“for assessment purposes, the terms ALARP and SFAIRP are interchangeable and require
the same tests to be applied. ALARP is also equivalent to the phrase ‘as low as reasona-
bly achievable’ (ALARA) used by other bodies nationally and internationally”.
ALARA is discussed in Chapter 12. However, the implementation and emphasis in regulatory
practice of ALARA and ALARP can be very different. ALARP places the onus on the nuclear-plant
designer and operator to justify why further improvements to all aspects of safety cannot be
practicably achieved, and in the authors’ experience goes considerably beyond ALARA in its
scope.
Figure 5 Structure of the ONR

A second underpinning is the concept of acceptable risk. In 1967, Farmer [Farmer1967] pro-
posed a risk-based criterion for use in siting a nuclear reactor. Figure 6 shows his proposed
iodine-131 release curve in an accident, versus reactor-years between releases (i.e., the inverse
of accident frequency). The basis of the curve was the predicted health effects of an accident,
which (if the criterion was met) would be small compared to other social risks. The flattening at
the top was to minimize the frequency of small releases due to their nuisance value.
Figure 6 “Farmer” curve
6.2.2.3 Tolerability of risk

This approach developed in the United Kingdom into a risk-based approach along with high-
level requirements. These are described in three key documents:
1. The Tolerability of Risk from Nuclear Power Stations [HSE1992];
2. Reducing Risks, Protecting People [HSE2001];
3. Safety Assessment Principles (SAPs) [HSE2006].
The concept evolved from a go/no-go risk line as proposed by Farmer to three distinct regions
as shown in
Figure 7. At the top, the risk of the activity is too high regardless of the benefits. At the bottom,
the risks are broadly acceptable, and no further reduction is usually required unless reasonably
practicable. The middle region of tolerable risk is typical of the risks from activities that people
are prepared to tolerate to secure benefits. ALARP would be applied to outcomes in this region.
Figure 7 HSE concept of tolerable risk

The risk boundaries are set as follows:
Boundary between “tolerable” and “unacceptable” regions for risk entailing fatality:
Worker: 1 in 1000 /y
Member of the public: 1 in 10,000 /y

This is called the Basic Safety Limit (BSL).
Boundary between “broadly acceptable” and “tolerable” regions for risk entailing fatal-
ity:
Worker: 1 in 1,000,000 /y
Member of the public: 1 in 1,000,000 /y
This is called the Basic Safety Objective (BSO).
ALARP is applied in this region between the BSL and the BSO to see if it is reasonably practicable
to reduce the risks below the BSO.
These risk targets are applied to all industries in the United Kingdom. For nuclear power, they
are translated in the SAPs to BSL/BSOs for normal operation and accidents. The latter become
numerical acceptance criteria and targets for the plant PSA. For example, for off-site doses,
Table 2 gives typical frequency/dose targets.
Table 2 U.K. frequency / dose targets for accidents off-site
The reader should consult [HSE2006] for the full set of numerical targets.
6.2.2.4 Safety assessment principles

The SAPs consist of a set of high-level principles, followed by a large number of technical
principles covering the plant life cycle. Although they are expected to be followed by a propo-
nent, they are not legally mandatory (unless they refer to a legal requirement), and are officially
written to guide regulatory staff in their assessment of a safety case. They are generally consis-
tent with IAEA Safety Requirements. Most SAPs are followed by a brief explanation or commen-
tary.
It is outside the scope of this chapter to cover these principles in depth (SAPs consisted of 140
pages as of 2006). Appendix A (Section 12) shows some examples of SAPs.
6.2.2.5 Technical assessment guides

The SAPs are supplemented by a series of Technical Assessment Guides (TAGs), which provide
more detailed guidance to regulatory staff and fill a similar role to the U.S. Standard Review
Plan. Like the latter, they are public.
6.2.2.6 Nuclear site licence conditions

Finally each licensed plant has a series of 36 Standard Conditions attached to its licence, cover-
ing design, construction, operation, and decommissioning [ONR2011]. For example, a number
of conditions cover changes to the plant during construction and operation and the need for
HSE approvals. Because they are attached to the licence, they give HSE a powerful compliance
tool.
7 Nuclear Material Safeguards

IAEA and CNSC material has been used throughout this section and referred to specific docu-
ments where appropriate. Italic fonts have been used to indicate where statements from IAEA
or CNSC documents have been used essentially verbatim.
Previous sections have focussed on the “why” of regulatory approaches and have been rela-
tively high-level. The remainder of this chapter provides more detail on safeguards and security
and on the Canadian regulatory approach. It is intended to be a guide to the practitioner in
finding and understanding relevant requirements, and for Canada, it provides a concise and
unique regulatory roadmap. Because regulatory requirements can affect design and operation
in a significant way, regulators and the IAEA spend much time making each sentence as precise
as possible, much as one does in drafting legislation. For that reason, we have avoided para-
phrasing regulatory requirements as much as practicable, instead quoting directly from the
relevant documents.
Nuclear material safeguards are enforced around the world by the International Atomic Energy
Agency (IAEA). IAEA safeguards are measures through which the IAEA seeks to verify that
nuclear material is not diverted from peaceful uses. States that have signed an agreement with
the IAEA (which are then considered to be nations committed to the peaceful use of nuclear
energy) accept the application of safeguards measures through the conclusion of safeguards
agreements with the IAEA [IAEA2007]. There are various types of safeguards agreements, but
the vast majority of countries have agreed not to produce or acquire nuclear weapons and to
place all their nuclear material and activities under safeguards to enable the IAEA experts to
verify that they are complying with the agreements. The origins of the IAEA and of safeguards
are discussed in Chapter 13.
7.1 Objectives and Evolution

IAEA safeguards implementation started more than 50 years ago and has been subject to
continuous evolution and improvement [IAEA1998], mainly from lessons learned over time.
Safeguards have gone through three major phases [IAEA1998], [IAEA2002d], [IAEA2007b]. The
first phase began in the late 1950s and 1960s as nations started to trade in nuclear plants and
fuel. The safeguards at that time were designed mainly to ensure that this trade did not lead to
the spread of nuclear weapons. The second phase, from 1971 to 1991, reflected a growing
perception that, “pending nuclear disarmament, world security is better served with fewer
rather than more nuclear weapons and nuclear weapon states”. The third and most recent
phase, from 1991 to 1997, consisted of a far-reaching review designed to remedy shortcomings
in the 1971 system. The review culminated in approval of a significantly expanded legal basis of
IAEA safeguards. Figure 8 shows highlights of safeguards evolution from 1991 to the present
time.
Figure 8 Safeguards state evaluation is a continuous process [IAEA2007b]

One should distinguish between the aims that safeguards seek to achieve in the non-nuclear-
weapon and in the declared nuclear-weapon states. The safeguards objectives in the non-
nuclear-weapon states are:
 To provide assurance that non-nuclear-weapon states party to comprehensive safe-
guards agreements are complying with their obligations not to acquire nuclear
weapons or other nuclear explosive devices;
 To deter states that have renounced nuclear weapons from acquiring unsafeguarded
nuclear material10. Safeguards must therefore be effective to detect within a rea-
sonable time the diversion or clandestine production of nuclear material;
 To verify certain facilities and material in some non-nuclear-weapon states with sig-
nificantly developed nuclear programs, but which have not yet concluded compre-
hensive safeguard agreements, and therefore are at high risk of proliferation activi-
ties. These states are placed under safeguards to prevent them from acquiring
nuclear-weapon capability.
The objectives of safeguards related to the formally recognized nuclear-weapon states are:
 To commit not to use nuclear energy to further any military purpose;
 To refrain from any explosive use of nuclear energy;
 To discourage any proliferation of nuclear material that is potentially useful for de-
veloping nuclear weapons;
 To encourage widespread acceptance and implementation of safeguards.
10 Note that “nuclear material” includes all fissionable isotopes that can be used as nuclear fuel or explosive
material, whereas “non-nuclear material” includes all material of strategic importance, such as heavy water,
zirconium tubes, etc.
The technical aim of safeguards is to use technical means to detect and verify safeguards
implementation and to provide for reporting of deviations.
For further background, see [IAEA1983], [IAEA1998a], [IAEA2011a], [IAEA2011b], [IAEA2007b].
7.2 Safeguards Treaty Detection Fundamentals

The Treaty on the Non-Proliferation of Nuclear Weapons (the Non-Proliferation Treaty, or NPT)
is a cornerstone of the nuclear non-proliferation regime and the basis for implementation of
safeguards. It was opened for signature in 1968 and entered into force in 1970. By December
2001, the NPT was in force in 187 states. In 1995, the NPT was extended indefinitely. For
details, see [IAEA2002d].
The key requirements of the NPT are:
 Not to receive and transfer any nuclear weapons or nuclear explosive devices, or control
of such weapons or devices directly or indirectly;
 Not to manufacture or otherwise acquire such weapons or devices;
 Not to seek or receive any assistance in development or manufacture of such devices;
 To accept IAEA safeguards on all fissionable material in peaceful nuclear activities;
 Not to provide source or special fissionable material, or equipment or material especially
designed or prepared for the processing, use or production of special fissionable mate-
rial to any non-nuclear-weapon State;
 To sign safeguard agreements with the IAEA either individually or together with other
States within 18 months of the date on which the State deposits its instruments of ratifi-
cation of or accession to the NPT Treaty
 Allows each State party to the NPT Treaty to develop research, production, and use of
nuclear energy for peaceful purposes and to facilitate and participate in the fullest pos-
sible exchange of equipment, materials, and information for the peaceful uses of nuclear
energy.
7.3 Safeguard Approaches, Concepts, and Measures

The key requirement for a successful safeguards program is to collect sufficient information
from various sources [IAEA2007]. The main sources of safeguards information are:
1. State-supplied information (under the obligation of the safeguards treaties and agree-
ments)
a. Nuclear material accountancy and facility design
b. Voluntary reported information
c. Additional protocol declarations
2. Key results of verification activities
a. Inspection
b. Design information verification (DIV)
3. Other information coming from open and other sources
In the first and second phases of the safeguard evolution process (until the early 1990s) the
focus of the safeguards program was limited to the Comprehensive Safeguards Agreement
protocols, which included:
a. Fuel enrichment facilities
b. Fuel fabrication facilities

c. Nuclear reactors
d. Spent fuel storage
e. Reprocessing facilities.
The Generic safeguards approach specifies the IAEA goals and safeguard activities for a refer-
ence plant. The Facility safeguards approach is prepared for a specific facility, developed and
adjusted to specific facility characteristic features and activities. The State-level safeguards
approach is developed for a specific state, encompassing all nuclear material, nuclear installa-
tions, and nuclear activities.
The integrated safeguards approach, which is a variation of the State-level safeguards approach,
was developed as an optimum combination of all safeguards measures available to the IAEA
under comprehensive safeguard agreements and additional protocols. Starting in the early
1990s, the safeguards program was extended, and is known today as Broad Overview of Ex-
panded Safeguards Coverage Protocols, which in addition to the above targets, include the
following:
a. Equipment manufacturers
b. Mining and milling
c. Conversion facilities
d. R&D centres.
The nuclear material accountancy within the framework of IAEA safeguards begins with the
nuclear-material accounting activities by facility operators and the state system of accounting
for and control of nuclear materials. The IAEA implements nuclear accountancy at various levels
[IAEA2002d]:
1) At the Facility Level by:
a) Dividing operations into nuclear material balance areas (MBAs);
b) Maintaining records on nuclear material quantities in each MBA;
c) Measuring and recording all transfers of nuclear material from each MBA to another
area or facility;
d) Periodically determining quantities of nuclear material in each MBA;
e) Closing the material balance over the period between two successive physical inventory
takings and computing the material unaccounted for (MUF);
f) Providing for a measurement control program to determine the accuracy of calibrations
and measurements;
g) Testing the computed MUF against its limits or errors in the accidental loss of gain;
h) Analysing the collected accounting information to determine the cause and magnitude,
and taking lessons learned for future improvements.
2) At State Authority Level
a) Preparing and submitting nuclear material accounting reports to the IAEA;
b) Ensuring that nuclear material accounting procedures and arrangements are adhered to;
c) Providing for IAEA inspector access and co-ordination arrangements;
d) Verifying facility operators’ nuclear material accountancy performance.
3) At IAEA Level by:
a) Independently verifying nuclear material accounting information in facility records and
State records, and conducting activities as per safeguard agreements signed with the
State;
b) Determining the effectiveness of the State safeguards program;
c) Providing statements to the State about the IAEA verification activities.
7.4 Safeguards Design Implications for Nuclear Power Plants

Most safeguards attention is devoted to nuclear power plants (NPPs) ([IAEA1998a],
[IAEA2014]).. NPPs are subject to a large flow of nuclear materials. At both off-load refueled
LWRs and on-load refueled PHWRs, the entire facility constitutes one MBA. Therefore, all
activities at the NPP site are under the safeguards monitoring program.
Figure 9 shows the elements of safeguards in a generic NPP facility. The top route in Figure 9
shows the real material storage areas and transfer routes, including the possibilities for diver-
sion as dashed lines. The middle route in Figure 9 represents the nuclear-material accounting
by the operator concerning the activities that have occurred. If this declared record of storage
status and transfers accurately reflects the real inventories and transfers, then there can be a
high degree of confidence that diversion activity has not occurred. The bottom route in Figure 9
represents the IAEA activities in monitoring physical activity in the storage and material-transfer
areas for accounting verification purposes. Any activity observed that does not correspond to
that declared will initiate a pre-planned process to re-establish confidence that the declared
record reflects the real inventory. That process might lead to an examination and validation of
the entire inventory, an expensive task which should be avoided.
Figure 9 Elements of safeguards [IAEA1998a]

Safeguards approaches, experience, and the associated guidelines vary between off-load and
on-load refueled reactors:
a) For off-load refueled reactors, refuelling operations and the associated fuel movements
are conducted at low frequency and with large numbers of items moved per refuelling;
b) For on-load refueled reactors, fuel movements and refuelling operations are carried out
at high frequency (daily), but with a small number of items moved per refuelling.
LWRs are refueled off-load, refuelling operations occur with an open vessel, and transfers occur
in channels filled with transparent water shielding. This design combination permits direct
visual observation. Transparent shielding is impractical for refuelling operations in on-load
refueled reactors, and therefore observation by other means is required. In these NPPs, detec-
tion of variations in radiation fields from continuously operating instrumentation together with
optical surveillance provide the information needed on refuelling sequences for safeguard
purposes. Regardless of the NPP design, all fuel is properly labelled by visible and non-visible
marking to be read by modern optical and laser devices.
Figure 10 shows the position of optical surveillance devices in the reactor building and the fuel-
storage pool building in the LWR reactor design.
Application of the general principles for safeguards is common to all pressure-tube PHWRs.
Most IAEA experience in handling PHWRs has been obtained on CANDU reactors.
Figure 11 shows the position of optical surveillance devices in the reactor building and the fuel-
storage pool building in the CANDU reactor design.
CANDU NPPs have horizontal fuel channels and are heavy-water moderated and cooled. The
refuelling system uses two refuelling machines, one at each end of the core, to insert new fuel
at one end and to accept spent fuel at the other. CANDU NPPs require appropriate shielding of
the refuelling machines and operations to protect personnel. Consequently, the positioning of
the fuelling machines in front of the reactor channels, and the refuelling itself, are carried out
by remote control.
Figure 10 Design aspects of safeguards in LWR reactors (Type II) [IAEA1998a]

Figure 11 Design aspects of safeguards in CANDU reactors [IAEA1998a]

Item accountancy and maintaining continuity of knowledge using containment and surveillance
(C/S) measures form the basis of the CANDU safeguards approach. Direct verification of the
operator’s declared nuclear material inventory is implemented for fresh and spent fuel. Al-
though fuel bundles in the core remain inaccessible for verification, continuity of knowledge is
maintained by a monitoring system when they are discharged. C/S covers the discharge of
spent fuel from the reactor and storage in the spent fuel bays to provide continuity of knowl-
edge.
All NPP designer organizations are required to ensure that their designs are safeguards friendly,
i.e., the design allows adequate implementation of the safeguards specific approaches. Some of
these requirements are listed below (not an inclusive list):
 Requirement for early interaction between the designer and the IAEA;
 Design of barriers which form a part of the C/S system;
 Design of seals for moving nuclear material;
 Design of surveillance techniques;
 Design of fuel identification and fuel verification techniques;
 Detection of unrecorded production or use of nuclear material;
 Transfer for spent fuel from the spent fuel pool to storage or reprocessing.
8 Nuclear Security
Nuclear security has become far more important in the last couple of decades, due largely to
changes in the external political environment after the September 11, 2001, terrorist attacks in
the United States. Each utility implements security at its sites, and in Canada, regulation of
nuclear security is part of the mandate of the CNSC.
Top-level requirements are set by the Nuclear Security Regulations [GovCan2010]. This Act
covers the security of both certain nuclear materials and nuclear facilities.
Nuclear materials are divided into three categories related to their security impact: for example,
a quantity of non-irradiated plutonium >2 kg is Category I; between 500 g and 2 kg is Category
II; and between 15 g and 500 g is Category III. Possession in each category requires a licence,
and the security obligations for each case are defined in the regulations.
For nuclear facilities in Canada, it is up to the CNSC Commission to establish and update the
design basis threat which is taken into account by the utility in determining its security provi-
sions. In addition, the utility must conduct an annual facility-specific threat and risk assess-
ment. Nuclear facilities must be located within a protected area enclosed by a barrier. Category
I materials are further protected by an inner-area structure or barrier whose function is to
detect and delay unauthorized access before the on-site nuclear response force can successfully
intervene.
In addition, the operator must identify vital areas. These are areas inside the protected area
containing equipment, systems, devices, or a nuclear substance, the sabotage of which would or
would likely pose an unreasonable risk to the health and safety of persons arising from expo-
sure to radiation. The operator is then required to implement the appropriate physical protec-
tion measures.
The Act also sets out requirements for persons authorized to enter protected and inner areas
and for security personnel. The CNSC has supplemented and expanded the Nuclear Security
Regulations through a number of Regulatory Documents and Guides. Most of those pertaining
to high-security sites such as nuclear power plants contain prescribed information and are not
publicly available, and therefore they are not referenced here.
Further details of security regulations, implementation techniques, and verification methodolo-
gies are beyond the scope of this textbook and are not available to the general public.
9 Canadian Regulatory Process11

The Canadian regulatory process is based on over 60 years of CANDU reactor development and
operation in Canada and internationally. Canada has been in the forefront of nuclear power
development and use since the very beginning, including the development of its nuclear regula-
tory program.
The Parliament of Canada first established legislative control and federal jurisdiction over the
development and use of nuclear energy and nuclear substances in 1946 with the introduction of
the Atomic Energy Control Act (AECA), which also established the Atomic Energy Control Board
(AECB). Fifty years later, the Canadian Government updated its regulatory requirements, and
the Canadian Nuclear Safety Commission was established as a successor to the AECB.
CNSC material has been used throughout this section. We refer to specific documents where
appropriate. As noted earlier, we do not try to paraphrase carefully developed regulatory texts.
Italic fonts have been used to indicate where statements from CNSC documents have been used
essentially verbatim.
11 Some of the descriptive material has been drawn from the Web site of the CNSC over the period from July 2011
to January 2013.
9.1 The CNSC

The Canadian Nuclear Safety Commission regulates the use of nuclear energy and materials to
protect the health, safety, and security of Canadians and the environment, and to implement
Canada’s international commitments on the peaceful use of nuclear energy.
The CNSC develops, maintains, and revises all regulatory requirements and documents under
the authority of the Nuclear Safety and Control Act [GovCan2013], which sets the overall legal
framework for regulation.
9.2 CNSC Mandate and Organization

This section covers the mandate and role of the CNSC in the Canadian NPP regulatory process.
It also covers the key elements of the CNSC organization, starting with the Commission and
continuing with the organization of various departments and their role in Canadian regulatory
processes.
As discussed in Section 2.3, the CNSC consists of an appointed Commission supported by
technical staff.
The CNSC Commission has up to seven appointed permanent part-time members whose
decisions are supported by more than 800 staff. These employees review applications for
licences according to regulatory requirements, make recommendations to the Commission, and
review (and can be delegated by the Commission to enforce) compliance with the Nuclear
Safety and Control Act, regulations, and any licence conditions imposed by the Commission.
Under the NSCA [GovCan2013], CNSC's mandate involves four major areas:
 regulation of the development, production and use of nuclear energy in Canada to
protect health, safety and the environment
 regulation of the production, possession, use and transport of nuclear substances,
and the production, possession and use of prescribed equipment and prescribed in-
formation
 implementation of measures respecting international control of the development,
production, transport and use of nuclear energy and substances, including measures
respecting the non-proliferation of nuclear weapons and nuclear explosive devices
 dissemination of scientific, technical and regulatory information concerning the ac-
tivities of CNSC, and the effects on the environment, on the health and safety of per-
sons, of the development, production, possession, transport and use of nuclear sub-
stances.
The CNSC organization is shown in Figure 12. Note:
 the Regulatory Operations Branch, which is responsible for NPP licensing in Canada;
 the Technical Support Branch, which is home to all the CNSC discipline specialists and
experts who review and assess various reactor designs and all information submitted
in support of a licence application.
President Commission Members
Executive Advisor
Commission Secretariat
Office of Audit,
Evaluation & Ethics
Regulatory Technical Regulatory Corporate

Legal
Operations Support Affairs Services
Services
Branch Branch Branch Branch
Figure 12 CNSC organization
9.3 Canadian Regulatory Framework and Regulatory Acts

The Canadian regulatory documents are arranged in several levels, as shown in Figure 13,
starting at the top level with enabling legislation, then at the middle level with requirements
documents, and at the lowest level with guidance documents. The requirements documents
can be subdivided into regulations, licences, and regulatory documents (defined in Sections
9.5.1 and 9.5.2).
Three Acts are relevant and are described in the following sub-sections.
9.3.1 Nuclear Safety and Control Act

The Nuclear Safety and Control Act [GovCan2013] is the enabling legislation in Canada, as
shown in Figure 13 and as described in Section 9.2.
Figure 13 Structure of regulatory documents in Canada

The Act covers the following areas:
 Establishment and organization of the Canadian Nuclear Safety Commission (CNSC);
 Number and role of CNSC members;
 Role, responsibilities, and duties of the CNSC President;
 Directives and powers of the CNSC;
 Use, establishment, and responsibilities of various panels that may be established by
the CNSC;
 The decision-making process at the CNSC;
 The scope and process for issuing or revoking various licences to the Canadian indus-
try;
 The responsibilities of the CNSC for maintaining various records and reports;
 Appointment of various specialists, analysts, and inspectors of the CNSC;
 Appointment of designated officers (to whom the Commission can delegate some if
its authority to make decisions on its behalf);
 Process for creation of procedures for decisions and orders;
 Process for redetermination and appeal of decisions and orders;
 Scope of regulations and process for their creation and enforcement;
 Provisions for exceptional powers;
 Provisions for offences and punishment;
 Transitional provisions.
Of particular importance is sub-section 24(4) of the Act, which states that:
“No licence shall be issued, renewed, amended or replaced—and no authorization to
transfer one given—unless, in the opinion of the Commission, the applicant or, in the
case of an application for an authorization to transfer the licence, the transferee:
(a) is qualified to carry on the activity that the licence will authorize the licensee to carry
on; and
(b) will, in carrying on that activity, make adequate provision for the protection of the
environment, the health and safety of persons, and the maintenance of national security
and measures required to implement international obligations to which Canada has
agreed.”
The next level in Figure 13 is Regulations, which are published in the Canada Gazette and
include:
 General Nuclear Safety and Control Regulations [GovCan2008]
 Radiation Protection Regulations [GovCan2007]
 Class I Nuclear Facilities Regulations [GovCan2012a]
 Class II Nuclear Facilities and Prescribed Equipment Regulations
 Uranium Mines and Mills Regulations
 Nuclear Substances and Radiation Devices Regulations
 Packaging and Transport of Nuclear Substances Regulations
 Nuclear Security Regulations [GovCan2010]
 Nuclear Non-Proliferation Import and Export Control Regulations [GovCan2010a]
 CNSC Cost Recovery Fees Regulations
 Administrative Monetary Penalties Regulations
 Canadian Nuclear Safety Commission Rules of Procedure
 Canadian Nuclear Safety Commission By-Laws
We have provided references above for some of the key Regulations in the nuclear power-
reactor field.
In addition, when required, the CNSC puts forward proposals for new regulations or proposed
amendments to existing regulations:
 Regulations Amending the Class I Nuclear Facilities Regulations and the Uranium
Mines and Mills Regulation
 Amendments to Nuclear Non-Proliferation Import and Export Control Regulations
 Amendment to Class II Nuclear Facilities and Prescribed Equipment Regulations
 Regulations Amending Certain Regulations made under the Nuclear Safety and Con-
trol Act (Miscellaneous Program)
 Amendments to the Packaging and Transport of Nuclear Substances Regulations.
9.3.2 Canadian Environmental Assessment Act

The second relevant Act is the Canadian Environmental Assessment Act [GovCan2012], which
covers all aspects of the environmental assessment process that must be followed for a new
nuclear energy facility of any purpose and type. It sets out the requirements for addressing all
important environmental issues, radiological and non-radiological.
9.3.3 Canadian Nuclear Liability Act

The third Act is the Canadian Nuclear Liability Act [GovCan1985], which regulates the nuclear
responsibility of various participants in the Canadian nuclear industry. The underlying purpose
of the Act is to require nuclear operators to carry insurance to cover liability should an accident
occur. The operator is absolutely liable for damage irrespective of the cause of the accident:
that is, claimants do not have to prove negligence on the part of the operator—they need only
prove that they have suffered damage. Operators must obtain liability insurance for damages
up to a maximum of $75 million (but see next paragraph); if the consequences exceed that
amount, the federal government appoints an independent Nuclear Damage Claims Commission
that will receive claims, assess damages, and recommend the level of compensation that should
be paid. The responsibility to pay claims exceeding $75 million rests with the federal govern-
ment. The Act does not set a limit on what the government pays.
As of this writing (early 2014), Bill C-22 (the Energy Safety and Security Act) has been tabled in
the House of Commons to replace the Nuclear Liability Act. It raises the absolute liability limit
to $1 billion for both nuclear energy and offshore oil and gas. It is expected that this bill will be
enacted by the House of Commons.
9.3.4 Other acts

Other Acts that pertain to some aspects of nuclear power plants include:
 Canadian Environmental Protection Act
 Fisheries Act
 Species at Risk Act
 Navigable Waters Protection Act
 Transportation of Dangerous Goods Act.
9.4 Licences
The third level in Figure 13 also includes licences. These are mandatory legal documents which
set forth the requirements and limitations for activities undertaken during the life cycle of a
nuclear activity. For example, a Power Reactor Operating Licence typically covers (among other
items):
 Authorization to operate the nuclear power plant for a specified period;
 Reporting of events to the CNSC;
 Control of the exclusion zone within which no permanent dwelling is permitted and
temporary access is controlled;
 References to some of the CNSC Regulatory Documents listed in Section 9.5, which
therefore become mandatory;
 References to some of the CSA Standards listed in Table 6, which therefore become
mandatory;
 Maintaining minimum operating-crew size and composition, including obtaining op-
erating licenses for key staff;
 Complying with the requirement to use a set of Operating Policies and Principles, in-
cluding Operating Limits;
 Complying with reactor power limits, including bundle and channel power limits
 Limits on permissible changes without CNSC prior consent. Specifically, this forbids
changes which “introduce hazards different in nature or greater in probability than
those considered by the Final Safety Analysis Report and Probabilistic Safety Assess-
ment”. Such a clause binds the results of the safety assessment performed by the
operator to the licence;
 Using only those fuel-bundle designs approved by the CNSC;
 Complying with requirements for an emergency plan for the facility;
 Complying with the limits on releases of radioactive material in normal operation
(derived release limits);
 Following requirements for a maintenance program; and
 Complying with requirements for radiation protection, site security, and safeguards.
The licence is accompanied by, and references, a rather lengthy “Licence Conditions Handbook”,
which provides:
 compliance verification criteria to meet the conditions listed in the licence;
 information regarding delegation of authority and current versions of documents;
 implementation timelines for specific licence conditions; and
 an explanation of each regulatory requirement specified in the licence.
In effect, the Licence Conditions Handbook extends the scope of the licence.
9.5 CNSC Regulatory Documents

The fourth level in Figure 13 contains the Requirements Documents. They are not legally
mandatory until and unless referenced in a licence. They are effectively mandatory for new
designs and are used as an evaluation tool for existing designs.
The fifth level in Figure 13 contains the Guidance Documents. These documents provide
direction to licensees and applicants on how to meet the requirements set out in the CNSC’s
Regulations, regulatory documents, and licences, and the techniques used by CNSC staff to
evaluate specific problems or data needed in the review of applications for permits or licences.
The following sub-sections provide a basic explanation of key CNSC regulatory and guidance
documents.
9.5.1 Historical classification schemes

The CNSC documents have followed several labelling and numbering schemes that reflect the
different periods when these documents were created. We review the historical schemes
briefly because many of these documents are still referenced under their original names in
licences.
Originally the AECB issued all requirements documents as “R” (“regulatory”) documents.
This system was superseded by the following classification:
 “S” (“Standard”) documents, which were mandatory
 “G” (“Guidance”) documents, which provided guidance
 “P” (“Policy”) documents, which provided direction to CNSC staff and information for
stakeholders.
More recently, CNSC has used the labels “RD” (Regulatory Document) and “GD” (Guidance
Document). Occasionally, CNSC has issued documents as “RD/GD”, which means that some
parts of the document contain (effectively) requirements and other parts contain guidance. In
such cases, CNSC staff clarifies when information is mandatory.

The “RD” and “GD” labels have now been replaced by “REGDOC”, as discussed in Section 9.5.2.
In CNSC regulatory documents, “shall” is used to express a requirement, i.e., a provision that a
licensee or licence applicant is obliged to satisfy to comply with the requirements of the docu-
ment. “Should” is used to express guidance, or that which is advised, but not required. “May”
is used to express an option or that which is permissible within the limits of the document.
“Can” is used to express possibility or capability.
9.5.2 Current classification scheme

The CNSC has recently developed a classification scheme (by subject) for its existing and
planned regulatory documents. As of April 2013, all these documents are now organized under
three key categories and twenty-five series as follows:
1.0 Regulated facilities and activities

1.1 Reactor facilities
1.2 Class IB facilities
1.3 Uranium mines and mills
1.4 Class II facilities
1.5 Certification of prescribed equipment
1.6 Nuclear substances and radiation devices
2.0 Safety and control areas

2.1 Management system
2.2 Human performance management
2.3 Operating performance
2.4 Safety analysis
2.5 Physical design
2.6 Fitness for service
2.7 Radiation protection
2.8 Conventional health and safety
2.9 Environmental protection
2.10 Emergency management and fire protection
2.11 Waste management
2.12 Security
2.13 Safeguards and non-proliferation
2.14 Packaging and transport
3.0 Other regulatory areas

3.1 Reporting requirements
3.2 Public and aboriginal engagement
3.3 Financial guarantees
3.4 Commission proceedings
3.5 Information dissemination
The numbering scheme used in this categorization is “REGDOC-x.yy.zz”, where “x” is the number
of the key category, “yy” is the number of the series, and “zz” is a number which identifies the
document uniquely within that category and series. For example, the design requirements for
new reactor designs are documented in REGDOC-2.5.2, where the first “2” means the document
is in the “safety and control area”; “5” means it is under the “physical design” series; and the
last “2” is a sequence number identifying the specific document.
9.5.3 Licensing basis

Many of these documents form the licensing basis of a regulated facility or activity. The licens-
ing basis [CNSC2010a] is a set of requirements and documents consisting of:
a) Regulatory requirements set out in the applicable laws and regulations.
 Examples: the Nuclear Safety and Control Act, the Canadian Environmental Assess-
ment Act, the Canadian Environmental Protection Act, the Nuclear Liability Act, the
Transportation of Dangerous Goods Act, the Radiation Emitting Devices Act, the Ac-
cess to Information Act, and the Canada/IAEA Safeguards Agreement.
b) Conditions and safety and control measures described in the facility licence and the
documents directly referenced in that licence.
 Examples: the Nuclear Security Regulations, the Nuclear Non-Proliferation Import
and Export Control Regulations, the General Nuclear Safety and Control Regulations,
the Radiation Protection Regulations, the Uranium Mines and Mills Regulations, the
Class I Nuclear Facilities Regulations, the Nuclear Substances and Radiation Devices
Regulations, the Class II Nuclear Facilities and Prescribed Equipment Regulations,
and the Packaging and Transport of Nuclear Substances Regulations.
c) Safety and control measures described in documents directly referenced in the licence.
 Examples: regulatory documents (such as RD-204, RD/GD-210, S-294, plus others),
industry codes and standards (such as CSA N286-05, CSA N285.4, CSA N290.13), pro-
ponent- or licensee-produced documents, and any subsequent changes made to
these documents in accordance with a CNSC-approved change control process (such
as Management System Manual, Limits and Conditions of Operation).
d) Documents needed to support the licence application.
 Examples are documents describing the design, safety fuellings, and all aspects of
operation to which the licensee makes reference; conduct of operations; and con-
duct of maintenance.
The licensing basis sets the boundary conditions for acceptable performance at a regulated
facility or activity and thus establishes the basis for the CNSC’s compliance program with
respect to that regulated facility or activity.
We now summarize selected regulatory documents. Note that document numbers and titles
are shown in terms of the currently issued documents (and not in terms of the new classifica-
tion scheme shown in Section 9.5.2).
9.5.4 Design of new NPPs in Canada: RD-337

RD-337 [CNSC2008] sets out comprehensive CNSC requirements with respect to the design of
new water-cooled NPPs and provides examples of optimal design characteristics12. Excerpts
follow:
The information provided in this document is intended to facilitate high quality design,
and consistency with modern international codes and standards, for new water-cooled
NPPs. It is recognized that specific technologies may use alternative approaches. If a
design other than a water-cooled reactor is to be considered for licensing in Canada, the
design is subject to the safety objectives, high level safety concepts and safety manage-
ment expectations associated with this regulatory document.
To a large degree, this document represents the CNSC’s adoption of the principles set
forth in International Atomic Energy Agency (IAEA) document NS-R-1, Safety of Nuclear
Plants: Design, and the adaptation of those principles to align with Canadian practices.
The scope of NS-R-1 has been expanded to address the interfaces between NPP design
and other topics, such as environmental protection, radiation protection, ageing, human
factors, security, safeguards, transportation, and accident and emergency response
planning.
The main components of RD-337 are:
i. Safety objectives and concepts
This section covers safety objectives in Canada (general, radiation protection, and tech-
nical), safety goals, radiation dose criteria, accident mitigation and management, and
safety concepts (defence-in-depth, physical barriers, and operating limits and condi-
tions).
ii. Safety management during design
This section discusses safety during the design process. It includes all important aspects
of the design process, including design authority, design management, QA, proven engi-
neering practices, design documentation, etc.
iii. Safety considerations
This section covers all aspects of safety applications in a design, including the application
of defence-in-depth, safety functions, accident prevention and plant safety characteris-
12 While this Chapter was being finalized, CNSC issued an update to RD-337 – called REGDOC 2.5.2 – which adds
lessons learned from Fukushima. See [CNSC2014].
tics, radiation protection principles and acceptance criteria, exclusion zones, and facility
layout.
iv. General design considerations
This section contains general requirements for all technical disciplines and reactor engi-
neering areas. It covers:
 structure, system, and component classifications;
 plant design envelope and operating plant states;
 guidance for consideration of all important hazards and initiating events that need to
be considered in the design;
 design rules, limits, and reliability for safety (such as common-cause failures, single
failures, fail-safe requirements, etc.);
 pressure-retaining systems and components;
 equipment classification (seismic, fire, environmental, etc.);
 instrumentation and control;
 safety support systems;
 civil structures;
 human factors;
 commissioning;
 security;
 safeguards;
 decommissioning.
v. System-specific requirements
This section covers requirements specific to most important systems and functions, in-
cluding the reactor core, heat-transport system, steam-supply system, means of shut-
down, emergency core cooling system, containment, emergency power supply, heat
transfer to an ultimate heat sink, emergency heat-removal system, control facilities,
waste management and control, fuel handling and storage, and radiation protection.
vi. Safety analysis
The safety analysis sections cover the most important aspects of safety analysis, namely
analysis objectives, hazards analysis, deterministic analysis, and probabilistic analysis.
The topic is covered in RD-310 in more detail [CNSC2008b].
vii. Environmental protection and mitigation
This section covers environmental protection design and requirements, as well as re-
quirements for released radioactive and hazardous substances.
viii. Alternative approaches
RD-337 is intended to be technology-neutral for water-cooled reactor designs. It is rec-
ognized that specific technologies may use alternative approaches.
The CNSC considers alternative approaches to the expectations in this document where:
 The alternative approach would result in an equivalent or superior level of safety;
 Application of the expectations in this document conflicts with other rules or re-
quirements;
 Application of the expectations in this document would not serve the underlying
purpose, or is not necessary to achieve the underlying purpose; or
 Application of the expectations in this document would result in undue hardship
or other costs that significantly exceed those contemplated when the regulatory
document was adopted.
One example of an alternative approach is consideration of best estimate and uncertainty
analysis instead of a conservative analysis; see Chapter 13.
9.5.5 NPP site evaluation: RD-346

This regulatory document [CNSC2008c] sets out requirements for the evaluation of sites for new
NPPs before an application is made for a Licence to Prepare Site and before an environmental
assessment determination is initiated.
RD-346 represents the CNSC staff’s adoption, or where applicable, adaptation of the principles
set forth by the International Atomic Energy Agency (IAEA) in NS-R-3, Site Evaluation for Nuclear
Installations. The scope of RD-346 goes beyond NS-R-3 in several aspects such as the protection
of the environment, security of the site, and protection of prescribed information and equip-
ment, which are not addressed in IAEA’s NS-R-3.
Site evaluation is a process that should precede the submission of an application to prepare a
site for the construction of a new NPP. RD-346 is written to serve the broader licensing needs
under the Nuclear Safety and Control Act and the Canadian Environmental Assessment Act, and
will facilitate a more effective and efficient regulatory review.
The document covers criteria for site evaluation; type and scope of baseline data required in the
site evaluation process; evaluation of natural external events; evaluation of external, non-
malevolent, human-induced events; security considerations; decommissioning; quality assur-
ance; and the consultation process.
9.5.6 NPP life extension: RD-360

The regulatory document RD-360 [CNSC2008d] covers the regulatory requirements for NPP life-
extension programs. The purpose of this regulatory document is to inform licensees about the
steps and phases to consider when undertaking a project to extend the life of a nuclear power
plant. The main topics are:
 Key elements to consider when establishing the scope of the life-extension project;
and
 Considerations to be taken into account in planning and executing a life-extension
project.
The document covers the initiation of the life-extension project, preparation of the integrated
project plan (which includes environmental assessment, safety assessment, integrated imple-
mentation plan and its adequacy, etc.), project execution, and return to service.
A major part of the assessment for life extension requires reviews of the existing plant against
modern codes and standards, including, but not limited to, CNSC documents more recent than
were in place at the time the plant was originally licensed. Although compliance is not manda-
tory, gaps must be identified, assessed, and remedied if practical. In addition, the condition of
systems, structures, and components (SSC) is assessed, largely to capture the effects of aging
and degradation. These assessments may also identify gaps. The entire assessment is collected
in a Global Assessment Report (GAR), from which an Integrated Implementation Plan (IIP) is
developed. The IIP is a work plan and schedule to fix the identified gaps insofar as is practical
and risk-effective. Such life-extension assessments are similar to, but much broader in scope
than, periodic safety reviews.
9.5.7 Safety analysis for NPPs: RD-310 and GD-310

Safety analysis is one of the most important topics in NPP licensing because it provides the
assurance that the plant can be operated safely and that the postulated events can be pre-
vented, mitigated, and managed. The CNSC has issued both the requirements document (RD-
310 [CNSC2008b]) and the guidance document (GD-310 [CNSC2012b]).
RD-310 identifies high-level regulatory information for a nuclear power plant licence ap-
plicant’s preparation and presentation of a safety analysis. In particular, it establishes a
more modern risk-informed approach to the categorization of accidents, one that con-
siders a full spectrum of possible events, including the events of greatest consequence
to the public. The CNSC expects proponents and applicants for new reactor licences to
immediately apply this regulatory document in new-build submissions. In the context of
existing reactors, CNSC expects the licensees to apply this document, in a graduated
manner, to all relevant programs in future submissions.
RD-310 explains the objectives of deterministic safety analysis, and in detail covers the
requirements for safety analysis. In particular, it covers identification and classification
of events following the international practice, i.e., groups them into Normal Operation
(NO), Anticipated Operational Occurrences (AOOs), Design Basis Accidents (DBAs), and
Beyond Design Basis Accidents (BDBAs), and explains the acceptance criteria. However,
most of the details in the document are focussed on the safety analysis methods and as-
sumptions, including methods, data, assumptions, computer code requirements, and
required conservatisms in the analysis. The document also covers the required docu-
mentation that a licence applicant needs to prepare, and the time line for updating this
documentation.
9.5.8 Reliability programs for NPPs: RD/GD-98

Regulatory document RD/GD-98, Reliability Programs for Nuclear Power Plants [CNSC2012a],
sets out the requirements and guidance of the CNSC for the development and implementation
of a reliability program for nuclear power plants in Canada. One of the important aspects of this
document is to provide a clear definition of the Systems Important to Safety (SIS) and how these
are treated and classified in NPP design.
9.5.9 Reporting requirements for operating nuclear power plants: S-99

The purposes of this Regulatory Standard [CNSC2003] are:
 To help the Canadian Nuclear Safety Commission collect the information that it
needs to ensure that a nuclear power plant is operating safely and to verify that the
licensee is complying with regulatory requirements;
 To help applicants for operating licences for nuclear power plants design programs
for collecting and reporting information in accordance with regulatory requirements;
and
 To facilitate CNSC evaluations of the appropriateness, completeness, and timeliness
of information reported to the CNSC by operators of nuclear power plants.
This document is a key document for utilities that operate nuclear power plants because it
states the requirements for reporting of operational issues. The document groups the require-
ments into two groups: reporting unscheduled events, and reporting scheduled events.
9.5.10 Pre-licensing review of a vendor’s reactor design: GD-385

Guidance document GD-385, Pre-Licensing Review of a Vendor’s Reactor Design [CNSC2012],
describes the pre-licensing review process provided by the CNSC for assessing a vendor’s design
for a nuclear power plant or small reactor. The review considers the areas of design that relate
to reactor safety, security and safeguards. For this topic, a licensee or licence applicant is not
obliged to satisfy any provisions through regulations or licence conditions, and therefore no
regulatory document (RD) accompanies this guidance document.
A pre-licensing review is an optional service provided by the CNSC. The objective of a pre-
licensing review is to increase regulatory certainty while ensuring public safety. The review can
be undertaken by a reactor vendor prior to an applicant’s submission of a licence application to
the CNSC and can provide early identification and resolution of potential regulatory or technical
issues in the design, particularly those that could result in significant changes to the design or
safety analysis.
This review does not certify a reactor design and does not involve the issuance of a licence
under the Nuclear Safety and Control Act. It is not required as part of the licensing process for a
new nuclear power plant or small reactor. The conclusions of a design review do not bind or
otherwise influence decisions made by the Commission, with whom the authority resides to
issue licences for nuclear power plants and small reactors. It is different in these respects from
the U.S. Standard Design Certification.
9.5.11 Licence to construct a nuclear power plant: RD/GD-369

To request CNSC assessment of a request for a construction licence, a formal application must
be submitted to the Canadian Nuclear Safety Commission (CNSC) along with appropriate
documentation. This CNSC guidance identifies the information that should be submitted to
support such an application.
RD/GD-369 [CNSC2012c] applies to applications for a licence to construct a water-cooled
nuclear power plant. It does not presuppose or limit an applicant’s intention to follow any
particular kind of water-cooled reactor technology. This document follows the format of the
IAEA Safety Guide No. GS-G-4.1, Format and Content of the Safety Analysis Report for Nuclear
Power Plants, but is more specific to Canadian regulatory practice. In following these guide-
lines, an applicant must submit the appropriate information to demonstrate that it is qualified
and will make adequate and reasonable provisions to undertake the activity to be licensed,
pursuant to sub-section 24(4) of the Nuclear Safety and Control Act and associated regulations.
RD/GD-369 is a combined document that lists and explains the requirements, but also provides
guidance on how to address them. The document covers plant description; management of
safety; site evaluation; general design aspects and support programs; design of plant structures,
systems, and components; safety analysis; construction and commissioning; operational as-
pects; operational limits and conditions; radiation protection; environmental protection; emer-
gency preparedness; radioactive and hazardous waste management; decommissioning and end-
of life-aspects; and safeguards.
9.5.12 Severe accident management programs for nuclear reactors: G-306

The CNSC Guide on Severe Accident Management Programs for Nuclear Reactors [CNSC2006]
describes a typical severe accident management (SAM) program for a nuclear reactor. A person
who applies for, or holds, a licence to construct or operate a nuclear reactor can use this guide
when developing and implementing measures to help:
a) Prevent the escalation of a reactor accident into an event involving severe damage to
the reactor core;
b) Mitigate the consequences of an accident involving severe damage to the reactor core;
and
c) Achieve a safe, stable state of the reactor and plant over the long term.
Some severe accidents (in particular dual failures, see Chapter 13) have historically been part of
the CANDU design basis. However the spectrum of severe accidents that must be evaluated has
broadened in the past few decades due to the maturation of probabilistic safety assessment as
an evaluation tool and to the real events at Three Mile Island, Chernobyl, and Fukushima.
Because operating plants may not have been explicitly designed against extreme events, in the
past, severe accident assessment and management looked at the capabilities of existing plant
systems to mitigate such events and developed severe-accident management procedures to
take advantage of these capabilities. However, especially post-Fukushima, design measures
(retrofits) are being implemented specifically to mitigate a spectrum of severe accidents, for
example, filtered containment venting.
A Severe Accident Management (SAM) program provides an additional defence against the
consequences of those accidents that fall beyond the scope of events considered in the reactor
design basis. The establishment of a SAM program should ensure that personnel involved in
managing an accident have the information, procedures, and resources necessary to carry out
effective on-site actions.
G-306 provides guidance on the goals and principles of severe-accident management, consid-
erations for program development (risk assessment and accident analysis), determination of
high-level accident response (preventive and mitigating actions, evaluation of systems and
equipment, and assessment of material resources), and SAM procedures and guidelines.
9.5.13 Containment systems for CANDU nuclear power plants: R-7

This regulatory document [CNSC1991] is one of the original series of AECB documents. It was
superseded by RD-337, except in areas in which it is consistent with RD-337, but provides more
information.
The document describes the design, operating, and testing requirements for the containment
system. It includes definition of the containment envelope, dose limits under accident condi-
tions (see Section 9.5.17), structural integrity, leakage criteria, and requirements in several
important areas, such as availability, environmental qualification, shielding, status monitoring,
and seismic provisions.
9.5.14 Shutdown systems for CANDU nuclear power plants: R-8

This regulatory document [CNSC1991a] is one of the original series of AECB documents. It was
superseded by RD-337 except in areas in which it is consistent with RD-337, but provides more
information.
The document describes the design, operating, and testing requirements for the shutdown
systems. It includes definition of the minimum allowable performance standards and require-
ments in several important areas, such as performance, environmental qualification, availability,
separation and independence, actuation instrumentation, status monitoring, and seismic
provisions.
9.5.15 Emergency core cooling systems for CANDU nuclear power plants: R-9
This regulatory document [CNSC1991b] is one of the original series of AECB documents. It was
superseded by RD-337, except in areas in which it is consistent with RD-337, but provides more
information.
The document describes the design, operating, and testing requirements for the emergency
core cooling system. It includes definition of the minimum allowable performance standards
and requirements in several important areas, such as core cooling, environmental qualification,
availability, separation and independence, actuation instrumentation, leakage control, inadver-
tent operation, shielding, status monitoring, and seismic provisions.
9.5.16 Use of two shutdown systems in reactors: R-10

This regulatory document [CNSC1991b] is one of the original series of AECB documents. It was
superseded by RD-337.
The document formalized the requirement for two independent shutdown systems:
“All nuclear power reactors licensed for construction in Canada after January 1, 1977, shall
incorporate two independent protective shutdown systems unless otherwise approved by the
Board.”
It required that the two shutdown systems be independent, reliable, diverse, and individually
capable of meeting regulatory dose limits13. Therefore, it could be assumed that at least one
system would operate as designed when required (i.e., there was no need to postulate simulta-
neous unavailability of both systems after an accident).
9.5.17 Classification of systems, design basis accidents, and dose limits

The selection and classification of accidents for safety analysis, or in other words the definition
of design basis accidents, is covered extensively in Chapter 13 and is summarized here only for
convenience:
The deterministic safety analysis limits under which all large operating CANDU plants up to but
excluding Darlington have been licensed can be found in [Hurst1972]. The spectrum of possible
design basis accidents was collapsed into two broad categories: single failures, or the failure of
any one process system in the plant, and dual failures, a much less likely event defined as a
single failure coupled with the unavailability of either a shutdown system, or containment, or
the emergency core cooling system, these constituting the so-called special safety systems. For
each category, a frequency and a consequence limit was chosen that had to be satisfied. In
addition, to deal with the siting of a reactor (Pickering A) next to a major population centre
(Toronto), population dose limits were defined for each category of accident.
The limits were as follows (sometimes this is called the Siting Guide):
Table 3 Single / dual failure dose limits
Accident Maximum Fre- Individual Dose Limit Population Dose

quency Limit
Single Failure 1 per 3 years 0.005 Sv 102 Sv

Dual Failure 1 per 3000 years 0.25 Sv 104 Sv

For the licensing of Darlington, as discussed in Chapter 13, Consultative Document C-6 was
used. This expanded the number of accident classes to five, as follows:
13 This is an oversimplification of R-10. R-10 actually specifies that for single failures, “at least” one shutdown
system had to meet dose limits, whereas for dual failures, each shutdown system had to be individually capable of
so doing. The distinction was removed in later practice, so that each shutdown system alone had to meet require-
ments for all design basis accidents.
Table 4 Consultative document C-6 dose limits
Event Expected Frequency Whole-Body Dose (Sv) Thyroid Dose (Sv)

Class per reactor-y
[Charak, 1995]14
1 > 10-2 0.0005 0.005

2 10-2 to 10-3 0.005 0.05
3 10-3 to 10-4 0.03 0.3
4 10-4 to 10-5 0.1 1
5 < 10-5 0.25 2.5
Finally, for new plants (and as an evaluation tool for existing plants), a classification scheme
based more on international practice has been used (from [CNSC2008] and [CNSC2008b]):
Table 5 New plants in Canada: dose limits
Event Frequency Dose Limit (Sv)
Anticipated Operational Occurrence  10-2 / reactor-year 0.0005
Design Basis Accident 10-2 to 10-5 / reactor-year 0.020
Plant licences refer to the Final Safety Analysis Report, which in turn references the appropriate
classification scheme.
9.6 CNSC Staff Review Procedures

CNSC staff have prepared a set of Staff Review Procedures (SRPs) used during the technical
assessment stage of the environmental assessment and site licensing process. They provide
instructions to CNSC staff on the conduct of an assessment. They also inform potential appli-
cants and the public about the criteria used to assess environmental impact statements and
licence applications for new nuclear power plants. SRPs are not regulatory documents. They
are available on the CNSC Web site. Other SRPs are under development.
9.7 CNSC CANDU Safety Issues

All regulators track safety-related issues which are not so pressing as to require immediate
regulatory action, but have not been addressed convincingly enough by the operator. Typically,
these include knowledge gaps, safety analysis assumptions, and methodologies. Such issues in
Canada were formerly called “Generic Action Items” (GAIs); these have now been phased out as
a regulatory tool and replaced by the CANDU Safety Issues. Generally, they pertain to more
than one operating plant.
14 Expected frequency ranges are not part of C-6; they were used by Ontario Hydro in the licensing of Darlington
to classify events not listed in C-6.
The list of CANDU Safety Issues was developed from:

 The IAEA document, Generic Safety Issues for Nuclear Power Plants with Pressurized
Heavy Water Reactors and Measures for their Resolution [IAEA2007a];
 Regulatory oversight of currently operating reactors;
 Results of life-extension assessments;
 Safety issues identified in pre-licensing reviews of new CANDU designs.
They are classified into three categories:
 Category 1 means issues that have been satisfactorily addressed in Canada;
 Category 2 identifies issues that are a concern in Canada, but have appropriate
measures in place to maintain safety margins;
 Category 3 issues are a concern in Canada and measures are in place to maintain
safety margins, but the adequacy of these measures needs to be confirmed.
For the last, the industry and the CNSC have used a risk-informed decision-making (RIDM)
process to identify, estimate, and evaluate the risks associated with each and to recommend
measures to control these risks. None of these issues has yielded a risk significance level that
required immediate corrective action.
As of August 2013 [GovCan2013a], there are 16 Category 3 issues outstanding, including six
associated with large-break loss-of-coolant accidents (LBLOCA) and ten other safety issues not
associated with LBLOCA.
The current status of each CANDU Safety Issue is reported publicly at International Nuclear
Safety Convention meetings [GovCan2013a].
9.8 CNSC Fukushima Action Items

On March 11, 2011, a magnitude 9.0 earthquake, followed by a devastating tsunami, struck
Japan. The combined impact of the earthquake and tsunami on the Fukushima Dai-ichi nuclear
power plant caused a severe nuclear accident. Details are presented in Chapter 13.
The CNSC established the CNSC Fukushima Task Force to evaluate the operational, technical,
and regulatory implications for CANDU nuclear power plants and the capability of CANDU NPPs
to withstand conditions similar to those that triggered the Fukushima accident. It issued its
recommendations in the Fukushima Task Force Report [CNSC2011]. In response, the CNSC Staff
developed an Action Plan [CNSC2011a], which mandates actions both for the utilities and for
the CNSC itself. In these documents, CNSC staff requested utilities to conduct certain activities
in the following areas:
a) Safety-review criteria, including identification and magnitudes of external events,
adequacy of design-basis-accident analysis, consideration of beyond-design-basis ac-
cidents, implementation of severe-accident management, licensees’ emergency re-
sponse plans, nuclear emergency management in Canada, and CNSC regulatory
frameworks and processes;
b) Strengthening reactor defence-in-depth;
c) Enhancing emergency response; and
d) Improving regulatory framework and licensing.
This work resulted in a number of Fukushima Action Items that the Canadian industry is putting
in place.
9.9 Canadian Codes and Standards

The Canadian Standards Association (CSA) is a not-for-profit membership-based association
serving business, industry, government, and consumers in Canada and the global marketplace.
CSA develops standards that enhance public safety and health, advance the quality of life, and
help to preserve the environment.
The objective of the CSA Nuclear Standards Program is to help promote a safe and reliable
nuclear power industry in Canada and to have a positive influence on the international nuclear
power industry. Although focussing on nuclear power plants, the program provides guidance
for other types of nuclear facilities on selected topics such as radioactive waste management
and environmental releases. Compliance with CSA nuclear standards is mandatory for those
that are referenced in a plant licence and is de facto mandatory for other organizations which
contribute to design, analysis, and other aspects of plant construction and operation. Table 6
shows selected CSA Standards relevant to nuclear power plants.
Table 6 Selected key CSA codes and standards
Document Title Creation /

number Revision or
Reaffirmed
date
CSA N285.0 General Requirements for Pressure-Retaining Systems and 2008

Components in CANDU Nuclear Power Plants/Material Stan-
dards for Reactor Components for CANDU Nuclear Power Plants
CSA N285.4 Periodic Inspection of CANDU Nuclear Power Plant Components 2009
CSA N285.5-13 Periodic Inspection of CANDU Nuclear Power Plant Containment 1998
Components (2013)
CSA N285.8 Technical Requirements for In-Service Evaluation of Zirconium 2010

Alloy Pressure Tubes in CANDU Reactors
CSA N286-05 Management System Requirements for Nuclear Power Plants 2005
(2010)
CSA N286.7 Quality Assurance of Analytical, Scientific, and Design Computer 1999
Programs for Nuclear Power Plants (2007)
CSA N287.1 General Requirements for Concrete Containment Structures for July 1993
CANDU Nuclear Power Plants (2009)
CSA N287.7-08 In-Service Examination and Testing Requirements for Concrete 2008
(R2013) Containment Structures for CANDU Nuclear Power Plants (2013)
CSA N288.4-10 Environmental Monitoring Programs at Class I Nuclear Facilities 2010

and Uranium Mines and Mills (1990)
CSA N288.5-11 Effluent Monitoring Programs at Class I Nuclear Facilities and 2011
Uranium Mines and Mills
CSA N288.6-12 Environmental Risk Assessments at Class I Nuclear Facilities and 2012
Uranium Mines and Mills
CSA N288.1 Guidelines for Calculating Derived Release Limits for Radioactive 2008
Material in Airborne and Liquid Effluents for Normal Operation
of Nuclear Facilities
CSA N289.1 General Requirements for Seismic Qualification of CANDU 2008

Nuclear Power Plants
Document Title Creation /

number Revision or
Reaffirmed
date
CSA N290.1 Requirements for the Shutdown Systems of CANDU Nuclear 1980
Power Plants (2006)
CSA N290.15- Requirements for the Safe Operating Envelope of Nuclear Power 2010
10 Plants
CSA N291 Requirements for Safety-Related Structures for CANDU Nuclear 2008
Power Plants
CSA N293 Fire Protection for CANDU Nuclear Power Plants 2007
CSA N294 Decommissioning of Facilities Containing Nuclear Substances 2009
10 Reactor Licensing Process in Canada15

CNSC material has been used throughout this section and referred to in specific documents
where appropriate. Italic fonts have been used to indicate where statements from CNSC
documents have been used essentially verbatim.
This section covers the licensing process of both new-build NPPs and operating reactors. The
emphasis is on new-build NPPs. However, certain aspects of the licensing (relicensing) of
operating reactors are also covered. See [CNSC2008a] and [Andrews2012].
The focus is on the Canadian licensing process, but relevant and important aspects of, and
differences from, the licensing process in other countries are also covered at a high level.
10.1 Licensing Process

Nuclear power plants are defined as Class I nuclear facilities, and the regulatory requirements
for these facilities are found in the Class I Nuclear Facilities Regulations. The regulations also
require separate licences for each of the five phases in the lifecycle of a nuclear power plant:
i. licence to prepare a site, including the environmental assessment;
ii. licence to construct;
iii. licence to operate;
iv. licence to decommission; and
v. licence to abandon.
There are four major steps in the licensing process:
1. Applicant submits a licence application
2. Environmental assessment
15 Some of the descriptive material has been drawn from the Web site of the CNSC.
3. Licensing technical assessment

4. CNSC renders its decision.
We summarize these in turn.
10.1.1 Licence application

In Canada, the licensing process begins when an application is received by CNSC. All new
licence applications or amendments to existing licences require the approval of the Commission
or a CNSC Designated Officer. The Commission is notified when an application that requires a
decision from them has been filed.
The preparation of a licence application needs to consider all regulatory criteria as defined by
the Nuclear Safety and Control Act, and relevant regulations, CNSC requirements and expecta-
tions, international and domestic standards, and applicable international obligations.
For major resource projects such as nuclear power plants, uranium mines, or fuel processing
facilities, a project description is sent to Natural Resources Canada’s Major Projects Manage-
ment Office (MPMO). MPMO is responsible for coordinating the work of all the federal de-
partments and agencies that have a role to play in the regulatory process for major resource
projects. MPMO offers licensees a single entry point into the federal regulatory system.
For new nuclear power plants in Canada, three licences are initially required:
1. licence to prepare site, including an environmental assessment
2. licence to construct the nuclear power plant, and
3. licence to operate.
The regulatory process in Canada in principle makes it possible to combine these processes, but
the normal route is for an applicant to request them sequentially, i.e., each process is com-
pleted when the appropriate licence is issued.
Very often, before a formal licensing process is initiated through an application by the NPP
owner or operator, a pre-licensing process is performed by the NPP design organization, as
explained in detail in Section 5.
The key element of a successful licensing process is to prepare adequate documentation that
meets regulatory requirements and expectations. Figure 14 provides a general idea of the level
of detail required during different phases of the licensing process. On the left-hand side, the
reactor design process is shown, and on the right-hand side, the level of documentation com-
pleteness. From top to bottom, the design and documentation move from the initial stage to
completeness. The middle of the diagram provides an indication of the percentage complete-
ness of the design. However, the design and documentation completeness level is discussed
and agreed in detail with CNSC at the beginning of a licensing process.
The technology vendor must complete a conceptual design and appropriate documentation
before engaging in any licensing activity. After the conceptual design has been firmed up, which
means that key design options and characteristics have been decided, a basic engineering
program begins to develop a preliminary design. During the basic engineering process, at some
point, the technology vendor may decide to engage the regulator in performing a non-binding
pre-licensing review. This is not a mandatory process, and technology vendors may or may not
decide to request it.
As shown in Figure 14, from [CNSC2012], following the preliminary design, detailed design is
conducted. After the detailed design and appropriate documentation have been prepared, the
licensee applies for a construction licence. During the regulatory review for the construction
licence, further detailed design activities related to the construction phase may be conducted.
After the regulatory body issues an operating licence, the licensee will continue to provide
ongoing design support activities.
Section 13 gives as an example the information that must be submitted in Canada for construc-
tion and operating licences.
At any time during operation, if a design or procedural change is considered for implementation
which deviates from the safety case or from any prior licence issued by CNSC, an approval
process for the particular change must be conducted and CNSC approval obtained before
implementing the change.
Figure 14 Design and documentation completeness phases
10.1.2 Environmental assessment

CNSC has obligations and responsibilities under the Canadian Environmental Assessment Act
(CEAA), which is the basis for federal Environmental Assessments in Canada. See [CEAA2012].
During the application review process, CNSC determines whether an environmental assessment
(EA) is required for a proposed project. EAs are used to predict the potential environmental
effects of a specific project and to determine whether these effects should be mitigated, and by
what means, before a project is carried out. EAs consider environmental components such as:
 air, water, and soil quality;
 noise;
 human health;
 aboriginal interest, physical and cultural heritage; and
 use of land and resources.
EAs provide opportunities for public participation in activities undertaken by the potential
licensees, the CNSC, or both, including aboriginal consultations.
Under the CEAA [GovCan2012], there are two levels of environmental assessment: a standard
environmental assessment and a panel review. The standard environmental assessment is the
default level of assessment applicable to a designated project; the second, higher level of
environmental assessment is a panel review. The CEAA imposes time limits for environmental
assessments. Most standard environmental assessments must be completed within one year,
while panel reviews are limited to two years.
At the beginning of an environmental assessment, CNSC provides project-specific advice and
guidance about what should be assessed, such as the scope of the project and the factors that
apply to each individual project. Further guidance is available from CNSC for conducting these
technical studies.
All applications for new nuclear facilities and some amendments to existing licences require the
applicant to undertake an EA and technical studies. After technical studies are completed, the
applicant can submit an Environmental Impact Statement (EIS) to CNSC, who will assess the EIS
and prepare an EA Report. For major new facilities, the assessment of an applicant’s EIS,
referred to as an EA technical assessment, and is guided by an Assessment Plan and EIS-related
Staff Review Procedures. The EA report contains recommendations for an EA decision, including
all appropriate mitigation measures and the requirements of a follow-up program. Pursuant to
the CEAA, the Responsible Authority (the Commission or a delegated officer) must make its EA
decision before any licensing action can be considered for the project.
Only one EA is required for a project, even if that project requires separate licences for different
phases (i.e., licence to prepare site, licence to construct, etc.). However, additional EAs may be
required if a project changes over time. For a new nuclear power plant, if construction is to be
performed at a new site, the EA is conducted before a site licence is issued. For a new nuclear
power plant where construction is intended on a site that already has other NPP units, the EA is
conducted before site preparation can begin for the new NPP addition.
10.1.3 Licensing technical assessment

After receipt of a licence application, CNSC staff undertakes a variety of technical assessments
according to a prescribed Assessment Plan and licensing-related Staff Review Procedures to
ensure that each application complies with its corresponding regulatory requirements. Peer
reviews are sometimes used, when additional rigour is required.
The Licensing Technical Assessment considers all regulatory criteria as defined by the Nuclear
Safety and Control Act, relevant regulations, CNSC requirements and expectations, international
and domestic standards, and applicable international obligations.
At the end of this process, CNSC staff make a recommendation for a decision on the licence
application through an integrated assessment report. A recommended compliance plan for
each licence is also developed, and the mitigation measures included in the follow-up program,
if applicable, are included in the licence.
10.1.4 CNSC decision

The final decision on a licence is made by either the Commission (e.g., for licences for nuclear
power plants) or a CNSC designated officer.
When the decision is to be made by the Commission, public hearings may be held to take into
account the views, concerns, and opinions of interested parties, including the applicant and
interveners. This is an important part of the process of establishing regulatory policy, making
licensing decisions, and implementing programs. CNSC staff recommendations are presented,
and the applicant and the public are normally given a chance to make statements.
Most decisions involving major nuclear facilities are made through a two-session16 public
hearing process. The first hearing is for presentations by the applicant and CNSC staff, and the
second for intervener presentations. Commission members may ask questions in both cases.
The applicant and CNSC staff also attend the second hearing and respond to further questioning
by the Commission.
Whether made by the Commission or a CNSC designated officer, the decision will take the form
of the granting of a licence or certificate, or else a letter of refusal.
The next section provides more information about the EA process for NPP construction on a
new site.
10.2 Site Selection and Site Licensing

This section covers the criteria used for appropriate site selection. A number of regulatory
documents in Canada and internationally cover this topic, e.g., [CNSC2008c].
The CNSC must be satisfied that it is feasible to perform the site preparation activities in a
manner that will satisfy all health, safety, security, and environmental protection requirements.
In addition, the Commission cannot issue a site preparation licence unless a decision as a result
of the EA has been made, indicating that the project may proceed; however, these approval
16 The first hearing is known as “Day 1” and the second as “Day 2”, although they may last more than one day
each.
processes may proceed in parallel.

The following aspects are considered when evaluating the suitability of a site over the life of a
nuclear power plant:
 the potential effects of external events (such as seismic events and floods) and hu-
man activity on the site;
 the characteristics of the site and its environment which could influence the path-
ways of radioactive and hazardous material that may be released; and
 the population density, population distribution, and other characteristics of the re-
gion, insofar as they may affect the implementation of emergency measures and the
evaluation of risks to individuals, the surrounding population, and the environment.
Under the regulations, an applicant must submit, for any licence, a Project Description of the
facility and plans showing the location, perimeter, areas, structures and systems of the facility.
An application for a Licence to Prepare Site does not require detailed information or determina-
tion of a reactor design; however, high-level design information is required for the environ-
mental assessment that precedes the licensing decision for a Licence to Prepare Site. An
application for a Licence to Construct must contain more detailed information about the reactor
design and a supporting safety case.
The goal of the CNSC, during the site preparation stage, is to ensure that the site characteristics
which may have an impact on health, safety, security, and the environment have been identi-
fied, and that these characteristics can and will be taken into consideration in the design,
operation, and decommissioning of the proposed nuclear power plant. The technical informa-
tion arising from the consideration of external events, site-specific characteristics, and support-
ing assessments is used as input into the design of the nuclear power plant and must be in-
cluded in the application.
The Canadian Environment Assessment Act, as amended in 2012 [GovCan2012], gives the CNSC
decision-making authority on all environmental assessments (EAs) for nuclear projects. There-
fore, both environmental assessment and site licensing processes occur concurrently under the
authority of the CNSC Commission. This ensures that the information submitted by the propo-
nent can be considered by public and government agencies through a single process and that
any appropriate decisions under EA and site licensing can be made by a single body—the CNSC
Commission. The Commission process establishes the level of an environmental assessment
based on risk and other factors; for example, an application for a site for a new nuclear power
plant will almost certainly require a panel review.
The CNSC Commission makes a decision on the EA within the 24-month timeline established for
site licensing.
10.3 Construction Licence

During this stage, CNSC undertakes a variety of technical assessments according to a prescribed
assessment plan and licensing-related staff review procedures to ensure that each application
complies with its corresponding regulatory requirements. The scope and duration of each
assessment will vary depending on the type of licence requested. Peer reviews are sometimes
used when additional rigour is required.
The licensing technical assessment considers all regulatory criteria as defined by the Nuclear
Safety and Control Act, relevant regulations, CNSC requirements and expectations, international
and domestic standards, and applicable international obligations.
The CNSC reviews the submitted application and judges whether the information contained in
the application is acceptable and whether the licence applicant is qualified and will make
adequate provision for safety, etc., in carrying out the licensed activity (in accordance with 24(4)
of the NSCA). If the CNSC accepts this information, it becomes the reference safety case for the
plant and will form part of the licensing basis at the construction licence stage. The information
provided with the licence application, including the documents to which the application makes
reference, constitutes the construction safety case. The requirements for the safety case are
listed in detail in [CNSC2012c].
The information required in support of the application to construct a nuclear power plant
includes, for example:
 a description of the proposed design for the nuclear power plant, taking into consid-
eration the physical and environmental characteristics of the site;
 environmental baseline data on the site and surrounding area;
 a preliminary safety analysis report showing the adequacy of the design;
 measures to mitigate the effects on the environment and the health and safety of
persons that may arise from construction, operation, or decommissioning of the fa-
cility;
 information on potential releases of nuclear substances and hazardous materials and
proposed measures to control them;
 programs and schedules for recruiting and training operations and maintenance
staff;
 and in general, those safety control areas listed in Section 9.5.2.
After the construction licence application has been received, the CNSC performs a comprehen-
sive assessment of the design documentation, the preliminary safety analysis report, the
construction program, and any other information required by the regulations. The assessment
focusses on determining whether the proposed design and safety analysis, along with other
required information, meet regulatory requirements.
Specifically, the evaluation involves rigorous engineering, scientific analysis, and engineering
judgment, taking into consideration the CNSC’s experience and knowledge of the best practices
in nuclear plant design and operation, as gained from existing power plants in Canada and
around the world. This review may take place in parallel with the Environmental Assessment
and site preparation licensing process. At the end of this process, CNSC staff makes a recom-
mendation for a decision on the licence application, through an integrated assessment report.
Occasionally, the recommendations include proposed changes to the regulatory framework to
accommodate evolving nuclear technologies. A recommended compliance plan for each licence
is also developed, and the mitigation measures included in the follow-up program, if applicable,
are included in the licence.
In addition to reviewing the information included in the application, the CNSC also verifies that
any outstanding issues from the site preparation stage have been resolved. The CNSC staff’s
conclusions and recommendations from these reviews are documented in reports submitted to
the Commission; the Commission then makes the final decision on the issuance of the construc-
tion licence. As noted earlier, the Commission will not issue a licence unless it is satisfied that
the applicant will make adequate provisions to protect health, safety, security and the environ-
ment, and to respect the international obligations to which Canada has agreed. As such, it is
the responsibility of the applicant to show that there are no major safety issues outstanding at
the time the Commission considers the application for a construction licence.
During the construction phase, the CNSC carries out compliance activities to verify that the
licensee is complying with the NSCA, with associated regulations, and with its licence. Such
compliance activities focus on confirming that plant construction is consistent with the design,
that the licensee is demonstrating adequate project oversight and confirming that quality
assurance requirements are met, and that the licensee is respecting any requirements of the EA
follow-up program.
In the latter part of construction, regulatory attention turns towards the inactive commissioning
program (without fuel loaded) and associated activities, whose purpose is to demonstrate to
the extent practicable that all systems, structures, and components function in accordance with
their design specifications.
10.4 Operating Licence

When applying for a Licence to Operate a nuclear power plant, it is the responsibility of the
applicant to demonstrate to the CNSC that it has established the safety management systems,
plans, and programs that are appropriate to ensure safe and secure operation. Information
required by the CNSC in support of the application for a licence to operate includes, for exam-
ple:
 a description of the structures, systems, and equipment at the nuclear power plant,
including their design and operating conditions;
 the final safety analysis report; and
 proposed measures, policies, methods, and procedures for:
o commissioning systems and equipment;
o operating and maintaining the nuclear facility;
o handling nuclear substances and hazardous materials;
o controlling the release of nuclear substances and hazardous materials into the
environment;
o preventing and mitigating the effects on the environment and on health and
safety resulting from the operation and subsequent decommissioning of the
plant;
o assisting off-site authorities in emergency preparedness activities, including as-
sistance to deal with an accidental off-site release; and
o nuclear security.
A typical application [OPG2013] would also cover:
 Management system, including safety culture;
 Human performance management;
 Operating performance and corrective action program;
 Fitness for service;
 Probabilistic safety analysis;
 Radiation protection;
 Conventional health and safety;
 Emergency management and fire protection;
 Waste management;
 Safeguards;
 Packaging and transport;
 Community relations and public information;
 Financial guarantees;
 Nuclear liability insurance;
 Open action items.
In addition to assessing the information included in the application to operate the nuclear
power plant, the CNSC also verifies that any outstanding issues from the construction licensing
stage have been resolved.
The CNSC staff’s conclusions and recommendations from these reviews are documented in
reports submitted to the Commission, which then makes the final decision on issuance of the
operating licence.
The Licence to Operate will enable the operator to begin active commissioning. The purpose of
the commissioning activities is to demonstrate that the plant has been constructed in accor-
dance with the design and that the systems, structures, and components important to safety
are functioning in accordance with their design specifications. The initial operating licence is
typically issued with conditions (hold points) to load nuclear fuel, permit reactor start-up, and
operation at power in steps up to the design rating of the plant. All the relevant commissioning
tests must be satisfactorily completed before the hold points can be released.
During the subsequent long-term operation of the plant, the CNSC carries out compliance
activities in order to verify that the licensee is complying with the NSCA, associated regulations,
and its licence terms. If the compliance activities identify any non-compliance or adverse trend,
there is a range of possible actions that the CNSC can take, from a request for licensee action to
prosecutions.
10.4.1 Ongoing oversight

Operating NPPs are subject to ongoing regulatory monitoring, screening, and inspections.
Canadian CANDU plants have resident full-time CNSC inspectors on-site to provide immediate
interaction with the operator and to view the station and equipment first-hand.
Aspects of continuing CNSC oversight include:
 Informal information exchange;
 Attending industry working groups as observers;
 Periodic formal audits and inspections of plant operation and management;
 Planned routine assessments and evaluations;
 Reactive inspections following events;
 Formal correspondence;
 Monitoring of progress on CANDU Safety Issues and on specific station action items;
 Conducting an annual review of the safety performance of NPPs;
 Revision to the Licence Condition Handbooks;
 Enforcement actions;
 Orders (legal instruments);
 Licence amendments initiated by CNSC staff;
 Licence amendments or renewals initiated by licensees; and
 Commission meetings and hearings.
10.4.2 Operating licence renewal and refurbishment

There are two longer-term activities which provide an opportunity to ensure that a plant
continues to meet regulatory requirements.
Operating licences are typically renewed about every five years in Canada. This requires a
review, update, and resubmission of the safety case to reflect changes in plant design and to
assess the plant against evolving regulatory requirements. The status of Action Items, CANDU
Safety issues, station-specific items, station safety performance, operating experience, quality
assurance, and other items are all reviewed at the time of renewal of an operating licence.
CANDU plants can be refurbished by replacement of the fuel channels, typically after about 25
years of operation. This action triggers a much more extensive comparison of the plant against
modern standards, as described in Section 9.5.6.
10.5 Licence to Decommission and Return Site to Green Site

At the end of a nuclear power plant’s useful life, it will be necessary to decommission the
facility. This will require a separate licence from the Commission. Information on decommis-
sioning plans and financial guarantees will, in practice, have been taken into account at all
stages of licensing (site preparation, construction, and operation) [CNSC2008a]. Factors taken
into account when evaluating an application to decommission a nuclear power plant include,
but are not limited to:
 the major components and systems within the facility, which must be properly con-
sidered during decommissioning planning;
 the design features that will facilitate decommissioning activities and reduce the
spread of contamination during operation;
 the expected levels of activation and contamination within the facility following the
end of operation;
 an assessment of structures to ensure that they are capable of being maintained for
the proposed period of storage and monitoring;
 the disposal of some of the nuclear materials and radiation devices (e.g., fresh fuel,
spent fuel, heavy water, water contaminated with tritium, and other prescribed nuc-
lear materials); and
 the quantities or volumes of wastes of all types (radioactive and hazardous) ex-
pected during decommissioning activities.
In addition, the licensees must show that they have sufficient funds to decommission the plant,
provide for the long-term management of spent nuclear fuel, and provide continuing environ-
mental monitoring and maintenance of the site for the duration of the licence.
11 Problems
1. Tokai-Mura, 1999: Locate the following material: International Atomic Energy Agency,
Report on the Preliminary Fact Finding Mission Following the Accident at the Nuclear
Fuel Processing Facility in Tokaimura, Japan; IAEA report, 1999. You may also find or use
other documents as well. Assess the accident in terms of regulatory effectiveness, using
some of the ideas in this chapter. Propose lessons learned. Which aspects of regulatory
performance were also apparent after the Fukushima accident?
2. Compare the approaches taken by the USNRC and the CNSC with respect to large Loss of
Coolant Accident (LOCA) regulatory acceptance criteria.
a. Find out what the acceptance criteria are and where they are documented.
Note: not all Canadian criteria are set by the CNSC; you may have to look at
CANDU safety reports.
b. Both Canadian and U.S. practices set limits on sheath oxidation in a LOCA to
prevent embrittlement on rewet by Emergency Core Cooling (ECC). Compare the
two criteria and explain the rationale in each case.
c. Find the requirements on shutdown systems for both the United States and
Canada. What is the typical reliability of the shutdown system in each case?
How are postulated impairments of the shutdown system handled in each case?
3. Review the U.S. General Design Criteria (GDC) in 10CFR, Appendix A to part 50. Which
design criteria (if any) would be intrinsically difficult for CANDU to meet? If you find any
such criteria, what is the safety concern that these GDCs address, and how is that
concern addressed in CANDU design?
4. Locate the U.K. Safety Assessment Principles issued by HSE. Compare and contrast the
requirements for safety systems with those in CNSC document RD-337. Which ideas are
common? Which are different? (Compare the underlying requirement, not the exact
wording).
12 Appendix A—Examples of U.K. Safety Assessment Principles

This Appendix lists some examples of U.K. Safety Assessment Principles which illustrate the non-
prescriptive technology-neutral approach. The text of the SAP is in bold and the commentary in
normal typeface.
=====================================================================
“Engineering principles: design for reliabilitySingle failure criterion EDR.4
During any normally permissible state of plant availability no single random failure, assumed to
occur anywhere within the systems provided to secure a safety function, should prevent the
performance of that safety function.
175 Consequential failures resulting from the assumed single failure should be considered as an
integral part of the single failure. Further discussion of the single-failure criterion is given in
IAEA Safety Standard NS-G-1.2”
=====================================================================
“Engineering principles: safety systems: Diversity in the detection of fault sequences ESS.7
The protection system should employ diversity in the detection of fault sequences, preferably
by the use of different variables, and in the initiation of the safety system action to terminate
the sequences.
342 This principle applies in particular to U.K. civil nuclear power reactor safety systems and in
particular to high-integrity safety systems.”
=====================================================================
“Engineering principles: reactor core Shutdown systems ERC.2
At least two diverse systems should be provided for shutting down a civil reactor.
444 Where a shutdown system is also used for the control of reactivity, a suitable and sufficient
shutdown margin should be maintained at all times.
445 Reactor shutdown and subsequent hold-down should not be inhibited by mechanical
failure, distortion, erosion, corrosion, etc., of plant components, or by the physical behaviour of
the reactor coolant, under normal operation or design basis fault conditions.”
=====================================================================
This last SAP was part of the basis for fitting a second shutdown mechanism (large-volume
boron injection tanks) to the Sizewell-B PWR.
13 Appendix B—Information Required in Licence Applications

(Construction and Operating Licences in Canada, as Examples)
From [GovCan2012a]
Licence to Construct
An application for a licence to construct a Class I nuclear facility shall contain the following
information in addition to the information required by section 3:
(a) a description of the proposed design of the nuclear facility, including the manner in which
the physical and environmental characteristics of the site are taken into account in the de-
sign;
(b) a description of the environmental baseline characteristics of the site and the surrounding
area;
(c) the proposed construction program, including its schedule;
(d) a description of the structures proposed to be built as part of the nuclear facility, including
their design and their design characteristics;
(e) a description of the systems and equipment proposed to be installed at the nuclear facility,
including their design and their design operating conditions;
(f) a preliminary safety analysis report demonstrating the adequacy of the design of the nuclear
facility;
(g) the proposed quality assurance program for the design of the nuclear facility;
(h) the proposed measures to facilitate Canada's compliance with any applicable safeguards
agreement;
(i) the effects on the environment and the health and safety of persons that may result from
the construction, operation and decommissioning of the nuclear facility, and the measures
that will be taken to prevent or mitigate those effects;
(j) the proposed location of points of release, the proposed maximum quantities and concen-
trations, and the anticipated volume and flow rate of releases of nuclear substances and
hazardous substances into the environment, including their physical, chemical and radiologi-
cal characteristics;
(k) the proposed measures to control releases of nuclear substances and hazardous substances
into the environment;
(l) the proposed program and schedule for recruiting, training and qualifying workers in respect
of the operation and maintenance of the nuclear facility; and
(m) a description of any proposed full-scope training simulator for the nuclear facility.
Licence to Operate
An application for a licence to operate a Class I nuclear facility shall contain the following
information in addition to the information required by section 3:
(a) a description of the structures at the nuclear facility, including their design and their de-
sign operating conditions;
(b) a description of the systems and equipment at the nuclear facility, including their design
and their design operating conditions;
(c) a final safety analysis report demonstrating the adequacy of the design of the nuclear
facility;
(d) the proposed measures, policies, methods and procedures for operating and maintain-
ing the nuclear facility;
(e) the proposed procedures for handling, storing, loading and transporting nuclear sub-
stances and hazardous substances;
(f) the proposed measures to facilitate Canada's compliance with any applicable safeguards
agreement;
(g) the proposed commissioning program for the systems and equipment that will be used
at the nuclear facility;
(h) the effects on the environment and the health and safety of persons that may result
from the operation and decommissioning of the nuclear facility, and the measures that
will be taken to prevent or mitigate those effects;
(i) the proposed location of points of release, the proposed maximum quantities and con-
centrations, and the anticipated volume and flow rate of releases of nuclear substances
and hazardous substances into the environment, including their physical, chemical and
radiological characteristics;
(j) the proposed measures to control releases of nuclear substances and hazardous sub-
stances into the environment;
(k) the proposed measures to prevent or mitigate the effects of accidental releases of nu-
clear substances and hazardous substances on the environment, the health and safety of
persons and the maintenance of security, including measures to
(i) assist off-site authorities in planning and preparing to limit the effects of an acciden-
tal release,
(ii) notify off-site authorities of an accidental release or the imminence of an accidental
release,
(iii) report information to off-site authorities during and after an accidental release,
(iv) assist off-site authorities in dealing with the effects of an accidental release, and
(v) test the implementation of the measures to prevent or mitigate the effects of an ac-
cidental release;
(l) the proposed measures to prevent acts of sabotage or attempted sabotage at the nu-
clear facility, including measures to alert the licensee to such acts;
(m) the proposed responsibilities of and qualification requirements and training program for
workers, including the procedures for the requalification of workers; and
(n) the results that have been achieved in implementing the program for recruiting, training
and qualifying workers in respect of the operation and maintenance of the nuclear facil-
ity.
14 References
[Ahearne1988] J. F. Ahearne, “A Comparison between Regulation of Nuclear Power in Canada
and the United States”, Progress in Nuclear Energy 22(3), 215-229 (1988).
[Andrews2012] L. Andrews, “Licensing New Major Facilities”, Presentation at the New Energy
Forum, Toronto, Canada, March 26-27, 2012.
[Apostolakis2004] G. E. Apostolakis, “How Useful is Quantitative Risk Assessment?” Risk Analy-
sis 24(3), 515-520 (2004).
[Barón1998] J. H. Barón, C. E. Chiossi, and H. R. Vallerga, “PSA-Based Regulatory Approach in
Argentina”, Presented to International Conference on Probabilistic Safety Assessment
and Management (PSAM 4), New York, USA, September 13-18, 1998.
[Buck1983] A. L. Buck, A History of the Atomic Energy Commission, U.S. Department of Energy
report DOE/ES-0003/l, July 1983.
[Bujor2010] A. Bujor, G. Rzentkowski, and D. Miller, Development of Risk-Informed Regulatory
Positions on CANDU Safety Issues, Part I: Methodology Development for Risk Estimation
and Evaluation, International Atomic Energy Agency report IAEA-TECDOC-1650, June
2010.
[CEAA2012] First Session, Forty-First Parliament, 60-61 Elizabeth II, 2011-2012; Statutes of
Canada 2012, Chapter 19: An Act to Implement Certain Provisions of the Budget Tabled
in Parliament on March 29, 2012 and Other Measures, assented to June 29, 2012, Bill C-
38.
[Charak, 1995] I. Charak and P. H. Kier, CANDU Reactors, Their Regulation in Canada, and the
Identification of Relevant NRC Safety Issues, Argonne National Laboratory report
NUREG/CR-63 15 and ANL-9515, April 1995 (prepared for U.S. Nuclear Regulatory Com-
mission).
[CNSC1977] Canadian Nuclear Safety Commission, The Use of Two Shutdown Systems in Reac-
tors, AECB report R-10, January 1977.
[CNSC1991] Canadian Nuclear Safety Commission, Requirements for Containment Systems for
CANDU Nuclear Power Plants, AECB report R-7, February 1991.
[CNSC1991a] Canadian Nuclear Safety Commission, Requirements for Shutdown Systems for
CANDU Nuclear Power Plants, AECB report R-8, February 1991.
[CNSC1991b] Canadian Nuclear Safety Commission, Requirements for Emergency Core Cooling
Systems for CANDU Nuclear Power Plants, AECB report R-9, February 1991.
[CNSC2000] Canadian Nuclear Safety Commission, Considering Cost-Benefit Information, CNSC
Regulatory Policy P-242, October 2000.
[CNSC2003] Canadian Nuclear Safety Commission, Reporting Requirements for Operating
Nuclear Power Plants, CNSC report S-99, March 2003.
[CNSC2006] Canadian Nuclear Safety Commission, Severe Accident Management Programs for
Nuclear Reactors, CNSC report G-306, May 2006.
[CNSC2008] Canadian Nuclear Safety Commission, Design of New Nuclear Power Plants, CNSC
report RD-337, November 2008. Updated in 2014: see [CNSC2014]

[CNSC2008a] Canadian Nuclear Safety Commission, Licensing Process for New Nuclear Power
Plants in Canada, CNSC report INFO-0756, Rev. 1, May 2008.
[CNSC2008b] Canadian Nuclear Safety Commission, Safety Analysis for Nuclear Power Plants,
CNSC report RD-310, February 2008.
[CNSC2008c] Canadian Nuclear Safety Commission, Site Evaluation for New Nuclear Power
Plants, CNSC report RD-346, November 2008.
[CNSC2008d] Canadian Nuclear Safety Commission, Life Extension of Nuclear Power Plants,
CNSC report RD-360, February 2008.
[CNSC2010] Government of Canada, Canadian National Report for the Convention on Nuclear
Safety, Fifth Report, Canadian Nuclear Safety Commission report INFO-0805, September
2010.
[CNSC2010a] Canadian Nuclear Safety Commission, Licensing Basis Objective and Definition,
CNSC report INFO-0795, January 2010.
[CNSC2011] Canadian Nuclear Safety Commission, CNSC Fukushima Task Force Report, CNSC
report INFO-0824, October 2011.
[CNSC2011a] Canadian Nuclear Safety Commission, CNSC Staff Action Plan on the CNSC Fuku-
shima Task Force Recommendations, CNSC report INFO-0828, December 2011.
[CNSC2012] Canadian Nuclear Safety Commission, Pre-Licensing Review of a Vendor’s Reactor
Design, CNSC report GD-385, May 2012.
[CNSC2012a] Canadian Nuclear Safety Commission, Reliability Programs for Nuclear Power
Plants, Licence Application Guide, CNSC report RD/GD-98, August 2012.
[CNSC2012b] Canadian Nuclear Safety Commission, Guidance on Safety Analysis for Nuclear
Power Plants, CNSC report GD-310, March 2012.
[CNSC2012c] Canadian Nuclear Safety Commission, Licence to Construct a Nuclear Power Plant,
Licence Application Guide, CNSC report RD/GD-369, August 2012.
[CNSC2013] Canadian Nuclear Safety Commission, CNSC Licensing Process, CNSC web site,
January 2013.
[CNSC2014] Canadian Nuclear Safety Commission, Design of Reactor Facilities: Nuclear Power
Plants, CNSC report REGDOC-2.5.2, May 2014.
[CSA1997] Canadian Standards Association, Risk Management: Guideline for Decision-Makers,
CSA report CAN/CSA-Q850-97, October 1997.
[Farmer1967] F. R. Farmer, “Reactor Safety and Siting: A Proposed Risk Criterion”, Nuclear
Safety, 8(6), Nov.-Dec. 1967.
[GovCan1985] Government of Canada, Nuclear Liability Act, R.S.C., 1985, c. N-28, published by
the Minister of Justice.
[GovCan2007] Government of Canada, Radiation Protection Regulations, published by the
Minister of Justice, SOR/2000-203, last amended on September 18, 2007.
[GovCan2008] Government of Canada, General Nuclear Safety and Control Regulations, pub-
lished by the Minister of Justice, SOR/2000-202, last amended on April 17, 2008.
[GovCan2010] Government of Canada, Nuclear Security Regulations, SOR/2000-209, published
by the Minister of Justice, last amended on May 13, 2010.
[GovCan2010a] Government of Canada, Nuclear Non-Proliferation Import and Export Control
Regulations, SOR/2000-210, published by the Minister of Justice, last amended on May
13, 2010.
[GovCan2012] Government of Canada, Canadian Environmental Assessment Act, 2012, S.C.
2012, c. 19, s. 52, published by the Minister of Justice, last amended on November 25,
2013.
[GovCan2012a] Government of Canada, Class I Nuclear Facilities Regulations, SOR/2000-204,
last amended on December 14, 2012.
[GovCan2013] Government of Canada, Nuclear Safety and Control Act, S.C. 1997, c. 9, published
by the Minister of Justice, last amended on July 3, 2013.
[GovCan2013a] Canadian National Report for the Convention on Nuclear Safety, Sixth Report,
August 2013.
[HSE1992] Health and Safety Executive, The Tolerability of Risk from Nuclear Power Stations, The
Stationery Office, 1992 ISBN 0 11 886368 1.
[HSE2001] Health and Safety Executive, Reducing Risks, Protecting People: HSE’s Decision-
Making Process, HSE report, 2001.
[HSE2002] Health and Safety Executive, Nuclear Safety Directorate, Technical Assessment Guide:
Foreword, Issue 002, HSE report T/AST/FWD, July 2002.
[HSE2006] Health and Safety Executive, Safety Assessment Principles for Nuclear Facilities, HSE
report, 2006 edition, Revision 1.
[HSE2008] Health and Safety Executive, New Nuclear Power Stations, Generic Design Assess-
ment—Guidance to Requesting Parties, HSE report, Version 3, August 2008.
[Hurst1972] D. G. Hurst and F. C. Boyd “Reactor Licensing and Safety Requirements”, Paper 72-
CNA-102, 12th Annual Conference of the Canadian Nuclear Association, Ottawa, Ontario,
June 1972.
[IAEA1983] International Atomic Energy Agency, IAEA Safeguards: Aims, Limitations, Achieve-
ments, IAEA report IAEA/SG/INF/4, 1983.
[IAEA1992] International Nuclear Safety Advisory Group, The Chernobyl Accident: Updating of
INSAG-1, IAEA report INSAG-7, November 1992.
[IAEA1998] International Atomic Energy Agency, The Evolution of IAEA Safeguards, International
Verification Series No. 2, IAEA report, 1998.
[IAEA1998a] International Atomic Energy Agency, Design Measures to Facilitate Implementation
of Safeguards at Future Water Cooled Nuclear Power Plants, IAEA Technical report Series
No. 392, IAEA, 1998.
[IAEA1999] International Nuclear Safety Advisory Group, Basic Safety Principles for Nuclear
Power Plants, IAEA report INSAG-12, October 1999.
[IAEA2000] International Atomic Energy Agency, Legal and Governmental Infrastructure for
Nuclear, Radiation, Radioactive Waste and Transport Safety, IAEA Safety Standard GS-
R-1, September 2000.
[IAEA2002] International Atomic Energy Agency, Review and Assessment of Nuclear Facilities by
the Regulatory Body, IAEA Safety Guide GS-G-1.2, August 2002.
[IAEA2002a] International Atomic Energy Agency, Organization and Staffing of the Regulatory
Body for Nuclear Facilities, IAEA Safety Guide GS-G-1.1, September 2002.
[IAEA2002b] International Atomic Energy Agency, Regulatory Inspection of Nuclear Facilities and
Enforcement by the Regulatory Body, IAEA Safety Guide GS-G-1.3, September 2002.
[IAEA2002c] International Atomic Energy Agency, Regulatory Control of Nuclear Power Plants,
IAEA Training Course Series #15, 2002.
[IAEA2002d] International Atomic Energy Agency, IAEA Safeguards Glossary – 2001 Edition,
International Verification Series No. 3, IAEA report, 2002.
[IAEA2004] International Atomic Energy Agency, Format and Content of the Safety Analysis
Report for Nuclear Power Plants, IAEA Safety Guide GS-G-4.1, May 2004.
[IAEA2004a] International Atomic Energy Agency, Design of Reactor Containment Systems for
Nuclear Power Plants, IAEA Safety Guide NS-G-1.10, 2004.
[IAEA2006] International Atomic Energy Agency, The Management System for Facilities and
Activities Safety Requirements, IAEA report GS-R-3, July 2006.
[IAEA2006a] International Atomic Energy Agency, Fundamental Safety Principles, IAEA report SF-
1, November 2006.
[IAEA2007] International Atomic Energy Agency, IAEA Safeguards: Staying Ahead of the Game,
IAEA report, 2007.
[IAEA2007a] International Atomic Energy Agency, Generic Safety Issues for Nuclear Power Plants
with Pressurized Heavy Water Reactors and Measures for their Resolution, IAEA TECDOC-
1554, June 2007.
[IAEA2010] International Atomic Energy Agency, International Conference on Challenges Faced
by Technical and Scientific Support Organizations in Enhancing Nuclear Safety and Secu-
rity, Tokyo, Japan, 25–29 October 2010; IAEA report STI/PUB/1519, September 2011.
[IAEA2011] International Atomic Energy Agency, Establishing the Safety Infrastructure for a
Nuclear Power Programme - Specific Safety Guide, IAEA report SSG-16, STI/PUB/1507,
Vienna, December 2011.
[IAEA2011a] International Atomic Energy Agency, Verifying Compliance with Nuclear Non-
Proliferation Undertakings, IAEA Safeguards and Additional Protocols, IAEA report, Sep-
tember 2011.
[IAEA2011b] International Atomic Energy Agency, Non-Proliferation of Nuclear Weapons and
Nuclear Security—Overview of Requirements for States with Limited Nuclear Material
and Activities, IAEA report, September 2011.
[IAEA2012] International Atomic Energy Agency, Project Management in Nuclear Power Plant
Construction: Guidelines and Experience, IAEA report STI/PUB/1537, 2012.
[IAEA2012a] International Atomic Energy Agency, Long-Term Structure of the IAEA Safety
Standards and Current Status, IAEA report from http://www-ns.iaea.org/committees/
files/CSS/205/status.pdf, December 2012.
[IAEA2012b] International Atomic Energy Agency, Licensing the First Nuclear Power Plant, IAEA
report INSAG-26, Vienna, 2012.
[IAEA2013] International Atomic Energy Agency, Periodic Safety Review for Nuclear Power
Plants, IAEA report SSG-25, Vienna, 2013.
[IAEA2014] International Atomic Energy Agency, International Safeguards in the Design of
Nuclear Reactors, IAEA Report NP-T-2.9, Vienna, 2014.
[Japan1948] The National Government Organization Law (Law No. 120 of 1948, as amended).
[Japan2012] The National Diet of Japan, The Fukushima Nuclear Accident Independent Investiga-
tion Commission, Executive Summary, report, 2012.
[Japan2012a] Investigation Committee on the Accident at the Fukushima Nuclear Power Sta-
tions, Final Report, July 2012.
[Kemeny1979] J. G. Kemeny, Report of the President’s Commission on the Accident at TMI, 1979.
[Laaksonen2008] J. Laaksonen, “General Overview of Licensing Process and Regulatory Over-
sight in the Design and Construction Stage of a Nuclear Power Plant”, presented at the
Workshop on Licensing and Regulatory Oversight of New Nuclear Builds, Helsinki and
Olkiluoto, September 1–4, 2008.
[Laaksonen2009] J. Laaksonen, “Regulatory Oversight of Olkiluoto 3 (EPR) Construction: Lessons
Learnt”, presented at 20th International Conference on Structural Mechanics in Reactor
Technology, Otaniemi, Finland, August 9–14, 2009.
[ONR2011] Office for Nuclear Regulation, Licence Condition Handbook, October 2011.
[OPG2013] Ontario Power Generation, Darlington Nuclear Generating Station: Application for
Licence Renewal, December 2013.
[Popov2012] N. Popov, “Licensing Challenges for New Countries Entering Nuclear Power Pro-
gram”, paper ICONE20POWER2012-55244, Proceedings, 20th International Conference
on Nuclear Engineering and ASME 2012 Power Conference (ICONE20 & POWER2012),
Anaheim, California, USA, July 30–August 3, 2012.
[Rogovin1980] M. Rogovin and G. T. Frampton, Three Mile Island: A Report to the Commissioners
and to the Public, Nuclear Regulatory Commission Special Inquiry Group, January 1980.
[Sims1980] G. H. E. Sims, A History of the Atomic Energy Control Board, AECB Publication INFO-
0026, July 1980; Canadian Government Publishing Centre, Catalog CC172-3/1981E.
[Slovic1987] P. Slovic, “Perception of Risk”, Science, New Series, 236(4799), pp. 280-285 (April
17, 1987).
[Starr1969] C. Starr, “Social Benefit versus Technological Risk”, Science, New Series, 165(3899),
pp. 1232-1238 (Sep. 19, 1969).
[STUK2006] Säteilyturvakeskus, Management of Safety Requirements in Subcontracting During
the Olkiluoto 3 Nuclear Power Plant Construction Phase, STUK Investigation Report 1/06,
Translation 1.9.2006, July 2006.

[Tanigawa2012] K. Tanigawa, Y. Hosoi, N. Hirohashi, Y. Iwasaki, and K. Kamiya, “Loss of Life after
Evacuation: Lessons Learned from the Fukushima Accident”, The Lancet, 379(9819), pp.
889–891 (2012).
[Tiippana2010] P. Tiippana, “STUK Experience in Licensing and Oversight of New Reactors”,
presented at the ELFORSK Conference, January 21, 2010.
[UK1974] United Kingdom, Health and Safety at Work, etc., Act 1974, Legislation.
[UK2009] United Kingdom Cabinet Office, Public Bodies 2009. See
http://www.civilservice.gov.uk/about/resources/ndpb.
[USNRC1978] United States Nuclear Regulatory Commission, Standard Format and Content of
Safety Analysis Reports for Nuclear Power Plants – LWR Edition, Regulatory Guide 1.70,
Revision 3, November 1978.
[USNRC1987] United States Nuclear Regulatory Commission, Standard Review Plan for the
Review of Safety Analysis Reports for Nuclear Power Plants - LWR Edition, USNRC report
NUREG-0800, June 1987.
[USNRC2002] United States Nuclear Regulatory Commission, Regulatory Guide 1.174 - An
Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-
Specific Changes to the Licensing Basis, November 2002.
[USNRC2009] United States Nuclear Regulatory Commission, Nuclear Power Plant Licensing
Process, USNRC report NUREG/BR-0298, Rev. 2, July 2009.
[USNRC2011] United States Nuclear Regulatory Commission, Nuclear Regulatory Legislation -
111th Congress; 2nd Session, USNRC report NUREG-0980, January 2011.
[USNRC2012] United States Nuclear Regulatory Commission, Licenses, Certifications, And
Approvals For Nuclear Power Plants, Code of Federal Regulations 10 CFR Part 52, 2012.
[USNRC2012a] United States Nuclear Regulatory Commission, Advisory Committee on Reactor
Safeguards, Code of Federal Regulations 10 CFR Part 1.13, 2012.
[USNRC2012b] United States Nuclear Regulatory Commission, General Design Criteria for
Nuclear Power Plants, Code of Federal Regulations 10 CFR Appendix A to Part 50, 2012.
[USNRC2012c] United States Nuclear Regulatory Commission, Standard Review Plan for the
Review of Safety Analysis Reports for Nuclear Power Plants: LWR Edition, USNRC report
NUREG-0800, formerly issued as NUREG-75/087, 2012.
[USNRC2012d] United States Nuclear Regulatory Commission, Domestic Licensing of Production
And Utilization Facilities, Code of Federal Regulations 10 CFR Part 50, 2012.
[WENRA2000] Western European Nuclear Regulators’ Association, Nuclear Safety in EU Candi-
date Countries, WENRA report, October 2000.
[WENRA2007] Western European Nuclear Regulators’ Association, Reactor Harmonization
Working Group, WENRA Reactor Safety Reference Levels, WENRA report, January 2007.
[WENRA2010] Western European Nuclear Regulators’ Association, WENRA Statement on Safety
Objectives for New Nuclear Power Plants, November 2010.
[WENRA2011] Western European Nuclear Regulators’ Association, “Stress Test” Specifica-

tions—Proposal by the WENRA Task Force, April 2011.
15 Glossary
ACNS Advisory Committee on Nuclear Safety (Canada, historical)
ACRS Advisory Committee on Reactor Safeguards (USA)
AEC Atomic Energy Commission (USA)
AECA Atomic Energy Control Act (Canada)
AECB Atomic Energy Control Board (Canada)
ALARA As Low as Reasonably Achievable
ALARP As Low as Reasonably Practical (UK)
ASLB Atomic Safety and Licensing Board (USA)
BSL Basic Safety Limit (UK)
BSO Basic Safety Objective (UK)
CEAA Canadian Environmental Assessment Act
CFR Code of Federal Regulations (USA)
C/S Containment and Surveillance
CSA Canadian Standards Association
DIV Design Information Verification
EA Environment Agency (UK)
EA Environmental Assessment (Canada)
EIS Environmental Impact Statement
ENSREG European Nuclear Safety Regulators Group
EU European Union
GAR Global Assessment Report
GDA Generic Design Assessment (UK)
GDC General Design Criteria (US)
GAI Generic Action Item
HSE Health and Safety Executive (UK)

IIP Integrated Implementation Plan
INSAG International Safety Advisory Group (IAEA)
IRRS Integrated Regulatory Review Service (IAEA)
LBLOCA Large Break Loss of Coolant Accident
LNT Linear No Threshold (dose response hypothesis)
LOCA Loss of Coolant Accident
LWR Light Water Reactor
MBA Material Balance Area
MDEP Multinational Design Evaluation Programme
MPMO Major Projects Management Office
MUF Material Unaccounted For
NDPB Non Departmental Public Body (UK)
NISA Nuclear and Industrial Safety Agency (Japan)
NPP Nuclear Power Plant
NRA Nuclear Regulation Authority (Japan)
NRU National Research Universal (reactor)
NRX National Research Experimental (reactor)
ONR Office for Nuclear Regulation (UK)
PHWR Pressurized Heavy Water Reactor
PWR Pressurized Water Reactor
PSR Periodic Safety Review
RCS Reactor Coolant System
RIDM Risk-Informed Decision Making
RSAC Reactor Safety Advisory Committee (Canada, historical)
SAPs Safety Assessment Principles (UK)
SAM Severe Accident Management
SER State Evaluation Report
SFAIRP So Far As Is Reasonably Practicable (UK)
SIR Safeguards Implementation Report

SIS Systems Important to Safety
SRP Standard Review Plan (US)
SRP Staff Review Procedure (Canada)
SSC Systems, Structures and Components
STUK Säteilyturvakeskus (Radiation and Nuclear Safety Authority, Finland)
TAG Technical Assessment Guide (UK)
TOR Tolerability of Risk (UK)
TSO Technical Support Organization
WENRA Western European Nuclear Regulators’ Association
16 Acknowledgements
We gratefully thank the following reviewers for their hard work and excellent comments during the
development of this Chapter. Their feedback has much improved it. Of course the responsibility for any
errors or omissions is entirely the authors’.
Glenn Archinoff
Joe Boyadjian
Dave Newland
Terry Rogers
We also thank Diana Bouchard for expertly editing and assembling the final copy.
1
CHAPTER 17
Fuel
Prepared by
Mukesh Tayal and Milan Gacesa – Independent Consultants
Learning Objectives
The principal learning objectives of this chapter are:
To gain a general understanding of the major mechanisms affecting fuel behaviour under
normal operating conditions;
To understand the underlying sciences of the relatively more important phenomena involving
fuel, including fission gas release, internal gas pressure, thermal stresses, element defor-
mation, environmentally assisted cracking, lateral vibrations of fuel elements, and fatigue
of the bundle assembly weld; and
To be able to perform calculations in the above subject areas to enable predictions.
Summary
Nuclear fuel, like conventional fuel, is responsible for generating heat and transferring the heat
to a cooling medium. Unlike conventional fuel, nuclear fuel presents the additional challenge of
retaining all the by-products of the heat-generating reaction within its matrix. Conventional fuel
releases almost all its combustion by-products into the environment. Experience has shown that
at times the integrity of CANDU fuel can be challenged while performing these roles. Eighteen
failure mechanisms have been identified, some of which cause fission by-products to be released
out of the fuel matrix. Other chapters in this book deal with preventing fission by-products
which escape from the fuel matrix from reaching the public. CANDU reactors can locate and
discharge fuel assemblies that release fission by-products into the coolant at power to minimize
the effect of fuel defects on plant operation and the public. Acceptance criteria are established
against which a fuel design is assessed to verify its ability to fulfill the design requirements
without failing. A combination of analyzes and tests is used to complete the verification assess-
ments.
©UNENE, all rights reserved. For educational use only, no assumed liability. Fuel – December 2014
Table of Contents
1 Introduction ............................................................................................................................ 4
2 Overview ................................................................................................................................. 5
2.1 Operating Environment .................................................................................................... 5
2.2 Fuel Design ....................................................................................................................... 6
2.3 In-Reactor Challenges....................................................................................................... 8
3 Power and Burnup ................................................................................................................ 13
4 Collapsible Sheaths ............................................................................................................... 14
4.1 Diametric Collapse.......................................................................................................... 15
4.2 Longitudinal Ridging ....................................................................................................... 16
4.3 Collapse into the Axial Gap............................................................................................. 17
5 Thermal Performance of Fuel ............................................................................................... 18
6 Fission Gas and Internal Pressure ......................................................................................... 19
6.1 Overview......................................................................................................................... 19
6.2 Recoil .............................................................................................................................. 24
6.3 Knockout ......................................................................................................................... 24
6.4 Diffusion of Fission Gas to Grain Boundaries ................................................................. 25
6.5 Internal Gas Pressure...................................................................................................... 25
7 Stresses and Deformations ................................................................................................... 26
7.1 Thermal Stresses in Pellets ............................................................................................. 27
7.2 Changes in Element Diameter and Length ..................................................................... 28
8 Environmentally Assisted Cracking ....................................................................................... 31
8.1 EAC Processes ................................................................................................................. 33
8.2 Defect Threshold ............................................................................................................ 35
8.3 Defect Probability ........................................................................................................... 36
8.4 Mitigation ....................................................................................................................... 38
9 Vibration and Fatigue............................................................................................................ 39
9.1 Alternating Stresses........................................................................................................ 39
9.2 Fatigue of the Endplate and the Assembly Weld ........................................................... 45
10 Fuel Design Verification ........................................................................................................ 46
10.1 Damage Mechanisms ..................................................................................................... 51
10.2 Acceptance Criteria and Integration............................................................................... 54
11 Operating Constraints and Inputs to the Safe Operating Envelope...................................... 59
12 Removal of Fuel Bundles Containing Defects ....................................................................... 60
12.1 Deterioration of Primary Defects in Fuel Elements........................................................ 60
12.2 Defect Removal Systems................................................................................................. 62
12.3 Confirmation of Defect Removal .................................................................................... 63
13 Closure .................................................................................................................................. 64
14 Problems ............................................................................................................................... 65
15 References............................................................................................................................. 68
16 Further Reading .................................................................................................................... 70
17 Relationships with Other Chapters ....................................................................................... 71
Fuel 3
List of Figures
Figure 1 37-element CANDU fuel bundle........................................................................................ 4

Figure 2 Causes of defects in CANDU fuel (1967–1996)............................................................... 10
Figure 3 Illustrative power histories.............................................................................................. 14
Figure 4 Sheath collapse: forms and stages.................................................................................. 15
Figure 5 Illustrative permanent strain at the tip of a longitudinal ridge ...................................... 16
Figure 6 Collapse into axial gap..................................................................................................... 17
Figure 7 Illustrative heat transfer and pellet temperature in a fuel element............................... 18
Figure 8 Terms for fission gas release ........................................................................................... 20
Figure 9 Steady-state release of 88Kr from a single crystal of UO2................................................ 20
Figure 10 Grain boundary bubbles and tunnels ........................................................................... 21
Figure 11 Changes in sizes and shapes of UO2 grains during irradiation...................................... 22
Figure 12 Equiaxial grain growth................................................................................................... 22
Figure 13 Columnar grain growth ................................................................................................. 23
Figure 14 Variation of internal gas pressure during irradiation.................................................... 24
Figure 15 Typical cracks in a high-power pellet ............................................................................ 27
Figure 16 Thermal stresses in a pellet .......................................................................................... 28
Figure 17 Removal of as-fabricated pores during irradiation ....................................................... 29
Figure 18 Fuel rod before and after startup ................................................................................. 30
Figure 19 Causes of hourglassing in pellets .................................................................................. 31
Figure 20 Power ramp cracks originating at a circumferential ridge ............................................ 32
Figure 21 Power ramp cracks near a sheath/endcap junction ..................................................... 33
Figure 22 Key terms for environmentally assisted cracking during power ramps........................ 33
Figure 23 Mechanism for power ramp failures............................................................................. 34
Figure 24 Defect threshold due to power ramp in non-CANLUB fuel at 140±20 MWh/kgU........ 36
Figure 25 Defect probability.......................................................................................................... 37
Figure 26 Lateral vibration of a fuel element................................................................................ 40
Figure 27 Endplate as a beam on elastic foundations .................................................................. 41
Figure 28 Stiffening of a fuel element at operating power........................................................... 43
Figure 29 Margins in criteria and analyzes ................................................................................... 57
Figure 30 Secondary hydriding...................................................................................................... 62
Figure 31 Schematic diagram of a delayed neutron monitoring system in a CANDU-6 reactor .. 63
List of Tables
Table 1 Illustrative design data for C6 fuel.................................................................................... 11

Table 2 Illustrative data for fuel assessments ............................................................................... 12
Table 3 Illustrative list of tests on prototype fuel bundles............................................................ 47
Table 4 Fuel design acceptance criteria for normal operating conditions.................................... 54
Table 5 Illustrative list of computer codes used at AECL for fuel analyzes ................................... 58
1 Introduction
The principal roles of fuel are to generate heat and to maintain a reasonable geometry so that
the coolant can carry that heat away. In addition, because fuel is a significant source of radio-
logical hazard, appropriate fuel integrity needs to be ensured at all times—including during
fabrication, operation, postulated accidents, and storage.
The science of the first two of these aspects, the generation and removal of heat, has been
discussed in earlier chapters. In the current chapter, we focus on the science of those aspects of
fuel that enable its reasonable integrity during normal operating conditions. Fuel integrity
during accidents and during abnormal occurrences is covered in another chapter.
This chapter uses CANDU-6 (also called C6) fuel for illustration, as shown in Figure 1. The figure
shows a fuel bundle in which fissile uranium is contained in 37 sealed fuel elements which are
held together by two welded endplates. Separation of the hot fuel elements from each other,
and from the pressure tube, is maintained by pads that are brazed to the fuel sheaths [Page,
1976].
Figure 1 37-element CANDU fuel bundle

Broad overviews of fuel design have been provided in earlier chapters. The focus in this chapter
is on the natural-uranium fuel used in current CANDU reactors, using the CANDU-6 reactor for
illustration. Chapter 18 covers alternative fuel cycles that have also been considered for appli-
cation in the CANDU reactor. Chapter 18 also covers, albeit briefly, some aspects of fuel manu-
facturing, including uranium mining, as part of a discussion on the natural uranium fuel cycle.
Chapter 19 describes long-term storage and disposal of irradiated CANDU fuel. Discussion of
Fuel 5
radiological hazards is beyond the scope of this chapter and is addressed in another chapter.
2 Overview
Before we can understand the science that underlies fuel integrity and therefore the chosen
design, it is important to obtain sufficient background about the operating conditions for which
nuclear fuel must be designed and the specific design that has been chosen to operate in them.
These factors, in turn, determine which specific scientific disciplines are pertinent to fuel
integrity.
2.1 Operating Environment

This section describes certain key illustrative (but not necessarily comprehensive) operating
conditions that have major impacts on design choices for fuel and the science that underpins its
performance in the reactor.
 Power Intensity: CANDU fuel produces a large amount of power in a small space. In a
length of approximately 0.5 m, a fuel element some 13 mm in diameter produces, at its
peak, some 30 kW. In comparison, a typical household electrical heater produces ap-
proximately 1 kW of heat when it is red hot. Therefore, nuclear fuel is 30 times more in-
tense than the household electrical heater. This high heat intensity is a principal source
of thermal and related challenges in nuclear fuel.
 Coolant Temperature: Reactor coolant is at high temperature—about 300° Celsius. This
leads to concerns such as corrosion, hydriding, and creep in metals. Corrosion is also in-
fluenced by coolant chemistry. Corrosion and hydriding reduce the capability of Zircaloy
to carry loads.
 Coolant Velocity: Coolant flows in the pressure tube at high velocity—approximately 10
m/s, or about 36 km/h—and at times this flow is turbulent. This leads to high drag loads
and also to a potential for vibration, fatigue, and fretting.
 Embrittlement: Processes such as fast neutrons and hydrogen/deuterium picked up from
corrosion tend to embrittle many metals and alloys. Therefore, shortly after insertion in
the reactor core, fuel’s structural materials have a much diminished capacity to resist
tensile strains. These phenomena are important sources of microstructural contribu-
tions to structural challenges.
 Fission Products: Fission reaction results in solid and gaseous fission products. These
fission products have a larger volume than the original uranium, and their higher volume
must be accommodated. Gaseous fission products increase the internal pressure on Zir-
caloy, and if excessive, this pressure can crack the Zircaloy. In addition, some fission
products are corrosive and can aggravate the structural challenges from mechanical
loads, especially in materials that become embrittled during irradiation.
 Fuel Management: Natural uranium fuel results in relatively low excess neutronic reac-
tivity; therefore, a CANDU reactor requires frequent addition of new fuel to maintain re-
activity (see the earlier chapters on physics). For this reason, a CANDU reactor is re-
fueled on an almost daily basis. Shutting down the reactor every day for refuelling
would have a major impact on the total power produced by the reactor. Therefore,
CANDU reactors are refueled while they are operating at full power. During on-power
fuelling, the reactor coolant carries the new fuel bundle into the C6 fuel channel until it
hits an existing bundle in the channel. Therefore, the new as well as the old fuel bun-
dles—the latter potentially more brittle—need to withstand impact loads. On-power

fuelling also means that some irradiated fuel bundles are moved from a low-power axial
position in the channel to (or through) a higher-powered axial location. This can lead to
power-ramp challenges, as described in Section 8.
 Refuelling: The fuel bundle is loaded and unloaded on-power, remotely, through tight
spaces, through bends, and sometimes through interrupted (discontinuous) supports.
These aspects require tight dimensional control, good dimensional stability, reasonable
flexibility, and allowances for wear.
 Power Manoeuvring: The reactor may be required to change its power level periodically
to respond to grid demand. Periodic changes in the power extracted from the fuel add
another dimension to fatigue considerations.
 Long Duration: A C6 fuel bundle stays in the reactor core for six months to two years.
Therefore, the challenging conditions described above must be endured continually for
long periods.
2.2 Fuel Design

Fuel is designed to address the operating conditions described above. Each fuel channel of a C6
reactor contains 12 fuel bundles.
Figure 1 shows a typical C6 fuel bundle. Each fuel bundle is made up of 37 fuel elements. Each
fuel element consists of a thin cylindrical tube called a sheath, which is made of Zircaloy-4. The
sheath encloses about 30 cylindrical pellets of UO2. A thin coating of a graphite-based com-
pound (called CANLUB) is applied to the inside surface of the sheath to mitigate the harmful
effects of power ramps, as described later in Section 8. Each fuel element is sealed at both ends
by endcaps that are welded to the sheaths; these welds are called "closure welds". Endplates
hold the fuel elements together in the geometric configuration of a fuel bundle. The endplates
are welded to the fuel elements; these welds are called "assembly welds". Spacing between
individual fuel elements is maintained using spacer pads. Spacing between the fuel bundle and
the pressure tube is maintained by bearing pads. The bearing pads and the spacer pads are
brazed to the fuel sheaths [Page, 1976; Gacesa et al., 1983].
Figure 1 also shows some key aspects of the internal design of a fuel element. The axial profile
of the pellet contains a dish to accommodate thermal expansion of the pellet that occurs
because of high temperature during irradiation. To make it easier to insert the pellets into the
sheath, (a) a diametric clearance is provided between the pellets and the sheath, and (b)
chamfers are provided in the pellets. Axial clearance between the pellet stack and the two
endcaps limits or even prevents axial interference between the two. A cavity in the endcap
provides space to store fission gas produced during irradiation.
The following are key rationales for some of the chosen features:
 Subdivision of the Bundle: As one illustrative example, a choice was made to subdivide
the uranium in the fuel bundle into a number of fuel elements. This permits the coolant
to be near the source of the heat at many locations, leading to lower UO2 temperature
and reduced threats from many thermally driven damage mechanisms. At the same
time, however, we should avoid excessive subdivision in the form of significantly more
numerous but much smaller fuel elements along with additional spacer pads and bear-
ing pads. The latter approach would add zirconium to the bundle, which would, among
Fuel 7
other effects, increase parasitic absorption of neutrons; increase fabrication cost; pro-
vide additional resistance to flow and hence require bigger and costlier pumps to gener-
ate the same flow of reactor coolant; lead to greater potential for fatigue and fretting
due to lower resistance of the smaller fuel elements to flow-induced vibrations; and ex-
hibit higher propensity for creep sag of the smaller fuel elements, which could poten-
tially lead to jamming of the fuel bundle in the channel. Increased subdivision also ad-
versely impacts the neutronic-thermal-hydraulic interaction during accidents (coolant
void reactivity), and, taken to the extreme, a much smaller element diameter would re-
duce the available surface area and decrease the efficiency of heat removal by the cool-
ant. Hence, a judicious balance must be struck regarding the degree of subdivision.
 Circular Cross Section of Fuel Elements: As a second illustrative example, the circular
cross section of the fuel elements reduces the stress concentration that would other-
wise be introduced by non-circular cross sections. This increases the mechanical
strength of the fuel elements to resist in-reactor loads.
 Ceramic, High-Density UO2: A large amount of heat is generated in the uranium fuel un-
der high external coolant pressure; therefore, the uranium must be in a form that will
transmit heat efficiently and maintain structural integrity. Ideally, uranium would be
used in a metallic form (e.g., uranium-aluminum alloy) to promote better heat transfer.
However, uranium metal has two undesirable characteristics: (1) if a fuel sheath were to
rupture and hot water from the coolant were to penetrate the fuel element, uranium
metal would oxidize rapidly and lose its strength; and (2) uranium metal is also dimen-
sionally unstable because of its anisotropy. Hence, uranium in the form of UO2 is a good
compromise. Although the low thermal conductivity of UO2 leads to high temperatures
in the fuel, its high melting point (about 2840°C) provides a large tolerance to melting.
Maximizing the density of UO2 leads to (1) higher fissile content in the element, hence
more energy; (2) improved thermal conductivity and hence lower temperature; and (3)
reduced in-reactor densification and hence higher dimensional stability.
 Zircaloy-4: Unlike the enriched uranium used in other types of reactors, CANDU reactors
use natural uranium. The latter has much less 235U to undergo fission. Therefore, to sus-
tain the chain reaction, it is essential that neutrons be conserved and that absorption of
neutrons by structural materials be minimized. For this reason, the structural compo-
nents are made from Zircaloy-4, which has a relatively small cross section for neutron
absorption. In addition, the structural components are kept as thin as possible, includ-
ing sheaths that are allowed to collapse if needed under operating pressure and tem-
perature. Zircaloy-4 also has high corrosion resistance, which is important considering
the high temperatures and the long residence periods in the reactor. Finally, Zircaloy-4
also has suitable mechanical properties.
 Thin Collapsible Sheaths: As a fifth illustrative example, the sheaths on CANDU fuel are
purposely kept very thin, mainly to conserve neutrons. Another benefit of a thin sheath
is that it can collapse diametrically under the operating pressure and temperature of the
coolant. This improves heat transfer from the pellet to the coolant, which in turn lowers
pellet temperature and reduces fission gas release. At the same time, the sheath must
be strong enough to carry the in-reactor loads and to avoid forming overly sharp ridges
(called longitudinal ridges, described in more detail in a later section of this chapter)
which can fail by overstrain or by fatigue.
 Helium Filling Gas: To minimize damage from several mechanisms such as melting, fis-
sion gas pressure, and environmentally assisted cracking, the operating temperature of
the pellet should be kept as low as reasonably achievable. Towards that end, empty
spaces within a fuel element are filled with gas that has high thermal conductivity. This
is called filling gas, and helium, or helium mixed with argon and air, is generally used for
this purpose. Another advantage of helium is that it has very small atoms that can es-
cape through small faults in joints: if a fuel element is not perfectly sealed by the end-
caps during manufacturing, escaping helium can be detected by appropriate sensors
during fuel fabrication.
 Internal Shapes and Spaces: Many internal shapes and spaces are carefully chosen to
balance a number of conflicting objectives, e.g., to reduce stresses, to provide space for
expansion during irradiation, to provide space to store fission gases produced during ir-
radiation, and to promote large-volume, automated, low-cost production. Some such
features include the pellet profile, the profile of the sheath/endcap junction, the endcap
profile, the diametric clearance between the pellets and the sheath, and the axial clear-
ance between the pellet stack and the endcaps. Several of these aspects are further
elaborated upon in later sections of this chapter.
 Manufacturing: Some features are driven primarily by considerations of automation in
manufacturing, e.g., resistance welding of endcaps and sheaths and brazing of pads.
Both processes enable fast throughputs. Chapter 18 provides additional details of some
aspects of fuel manufacturing.
 Ease of Handling: The length of the bundle was chosen primarily to facilitate on-power
refuelling, although both the length and weight of the bundle have ergonomic advan-
tages; one individual can lift and carry the bundle by hand.
 Fuel Management: Some important considerations in fuel management are: (a) to main-
tain neutronic reactivity at the desired level; (b) to achieve reasonable burnup; and (c) to
avoid defects due to power ramps. The last aspect is discussed in more detail in a later
section.
These configurations were chosen to facilitate the primary objectives of the fuel bundle: to
produce heat, to maintain its geometry so that the coolant can carry the heat away, to ensure
in-reactor integrity of the fuel elements and the fuel bundle, and to do all this consistently over
a lengthy period.
In several other ways as well, the chosen features reflect careful balances among many conflict-
ing requirements, processes, and forces. Fuel design is a complex multidisciplinary optimization
process involving reactor physics, thermal-hydraulics, heat transfer, structural mechanics,
microstructural processes, geometric stability, and interfacing systems, all under normal, off-
normal, and accident conditions.
The choices described above—thin structural materials, high operating temperatures, axial
movement of the fuel bundle during on-power fuelling, large in-reactor mechanical forces, and
material degradation due to irradiation—all combine to pose unique and significant challenges
to the in-reactor integrity of CANDU nuclear fuel under normal operating conditions. An
overview of these challenges is given in the next section.
2.3 In-Reactor Challenges

The major challenges to fuel’s in-reactor integrity can be grouped into three categories: (1)
Fuel 9
thermal challenges, (2) mechanical challenges, and (3) compatibility challenges.

Thermal challenges, if excessive, can potentially cause melting in the pellet or in the structural
materials (mainly Zircaloy). Pellet melting can result in (a) escape of additional fission gas from
the UO2 matrix into the pellet-to-sheath gap and (b) thermal expansion of the pellet, causing
the sheath to rupture. This combination can send additional radioactive fission products into
the coolant. Excessive amounts of radioactive fission products in the coolant could increase the
radiological hazard to the health of station staff.
Excessive mechanical damage can potentially cause holes, cracks, or breaks in structural materi-
als, with consequences similar to the thermal challenges described above.
Incompatibility can potentially cause the fuel bundle to jam inside the reactor or in its accesso-
ries such as fuel handling equipment, interfering with its proper insertion or eventual removal.
Incompatibility can also potentially harm neighbouring or interfacing components through
processes such as crevice corrosion.
For these reasons, a major objective in fuel design is to limit the extent of damage from the
mechanisms just described. Doing so requires, among other things, assessments of non-linear
heat transfer, non-linear stress analyzes, and microstructural processes.
All damage mechanisms are under sufficient control in current CANDU fuels. In most mature
CANDU cores, fuel defects typically occur in less than one bundle per year and tend to be
caused mostly by fretting due to debris.
The defect rate of CANDU fuel is remarkably low. During 2007–2008, the International Atomic
Energy Agency (IAEA) conducted a world-wide survey of fuel defect rates from 1994 to 2006
[IAEA, 2010]. World-wide, water-water energy reactors (WWERs) were found to have a fuel
defect rate of 94 defective fuel elements per million discharged (also called ppm, or parts per
million). The fuel defect rate was 87 ppm in pressurized water reactors and 65 ppm in boiling
water reactors. In comparison, fuel in Canada had a defect rate of 3.5 ppm. The Canadian fuel
defect rate is very acceptable—almost in the range of impurities found in most substances—and
much lower than for other types of nuclear fuels. This low rate confirms the basic soundness of
the practices built into all major aspects of Canadian fuel, from research to development, to
design, to fabrication, to operation, and to feedbacks among them.
Nevertheless, several defect excursions have occurred in the past, and it is illustrative to exam-
ine the historical causes of significant fuel defects. Figure 2 shows readily available illustrative
data on fuel bundles which have been damaged, mostly in defect excursions in power reactors
during the first three decades of commercial nuclear power, although the figure also includes
some data that would not strictly be considered to represent a major defect excursion.
Note: Above data exclude failures due to debris

Figure 2 Causes of defects in CANDU fuel (1967–1996)
The International Atomic Energy Agency has defined a “defect excursion” as one in which 10 or
more fuel bundles fail within a 12-month period [IAEA, 2010]. Figure 2 does have limitations: it
focusses on defects related to design and/or operation, mostly in Canada, and it excludes
defects due to fretting by debris. One can postulate that for any given damage mechanism, the
amount of damage is an indicator of how close its failure threshold is to the operating condi-
tions—the closer one operates the fuel to the defect threshold of a specific damage mecha-
nism, the higher will be the likelihood of damage from that mechanism. Figure 2 demonstrates
that in the past, power ramps have been a dominant damage mechanism in CANDU fuels.
The underlying science of the relatively more important challenges to fuel integrity is described
in later sections of this chapter. Table 1 provides some illustrative design data for C6 fuel.
Likewise, Table 2 provides some illustrative data for the physical, thermal, and mechanical
properties of CANDU fuel; these can be used for illustrative calculations of fuel performance.
Fuel 11
Table 1 Illustrative design data for C6 fuel

[Major Source: Nuclear Engineering International, September 2006]
Parameter Value Parameter Value
Number of elements (rods) per bundle (assem- 37 Element length (mm) 493
bly)
Peak linear fuel element rating (kW/m) 57 Element outside diameter 13

(mm)
Average discharge burnup of fuel bundle 170 Sheath (clad) thickness (mm) 0.4
(MWh/kgU)
Maximum residence period in the reactor 700 Pellet outside diameter (mm) 12
(days)
U weight per bundle (kg) 19.2 Pellet length (mm) 16
Zr weight per bundle (kg) 2.2 Pellet density (g/cm3) 10.6
Overall bundle length (mm) 495 Endplate width, outer ring 4.9
(mm)
Overall maximum bundle diameter (mm) 102 Endplate thickness (mm) 1.6
Average sheath temperature (°C) 340 Diameter of assembly weld 4

(mm)
Table 2 Illustrative data for fuel assessments
Parameter Value Parameter Value
UO2 Zr-4
Thermo-Physical Data
Theoretical density (g/cm3) 10.97 Density (g/cm3) 6.56
As-fabricated density (g/cm3) 10.6 Thermal conductivity at 300°C 16.4

(W/(m.k))
Thermal conductivity at 1000°C 2.8 Specific heat capacity at 300°C 327

(W/(m.K) (J/(kg.K))
Specific heat capacity at 1000°C 328 Melting point (°C) 1850

(J/(kg.K))
Melting point at zero burnup (°C) 2840 Sheath-to-coolant heat transfer 50

coefficient (kW/m2K)
Pellet-to-endcap heat transfer 1 Pellet-to-sheath heat transfer 80

coefficient (kW/m2K) coefficient (kW/m2K)
(at high power, hard contact, low (at high power, hard contact, low
burnup) burnup)
Mechanical Data
Young’s modulus at 1000°C (GPa) 190 Young’s modulus at 300°C (GPa) 80
(at 98% theoretical density) Shear modulus (GPa) 27
Poisson’s ratio 0.316 Poisson’s ratio 0.37
Yield strength at 1000°C (MPa) 180 Yield strength at 300°C 274

(for grain size of 25 µm) (78% cold worked and stress re-
lieved at 510°C)
Coefficient of linear thermal expan- 12.5 Coefficient of linear thermal expan- 6.72
sion (µm/(m.K)) sion (µm/(m.K))
(at 1000°C) (diametric, at 300°C)
Fuel 13
3 Power and Burnup

Recall from the earlier chapters on physics that the heat output of a fuel element is usually
expressed in kW/m. The energy produced by a fuel element per unit mass is called “burnup”. It
is also customary to use burnup as an indicator of irradiation time, although strictly speaking,
time is equal to energy divided by power. In the CANDU fuel community, burnup is usually
measured in MWh/kgU. In other types of fuel, e.g., in PWRs and BWRs, fuel burnup is usually
measured in GWd/tU. 1 GWd/tU equals 24 MWh/kgU.
Sometimes "atomic percent" is also used as a unit of burnup. It measures the ratio of the
number of fissions to the initial number of uranium atoms. As a working rule, 1 atomic percent
equals 240 MWh/kg U.
Exercise: A fuel element is operating at 60 kW/m and contains UO2 cylinders of 12 mm diame-
ter. What will be the fuel burnup after 30 days?
One meter of this fuel element contains a UO2 volume of 113 ml. Assuming a UO2 density of
10.6 g/cm3, the mass of UO2 is 1.2 kg. In UO2, 88% is U. Therefore, one meter of this fuel
element contains (1.2 * 0.88) = 1.06 kg of U. This fuel element is producing 60 kW; therefore, in
30 days, it produces (60* 720/1000) = 43 MW•h. Therefore, its burnup is (43/1.06) = 41 MW•h
/kgU.
This example illustrates that a typical CANDU fuel element operating at 60 kW/m accumulates
about 41 MW•h/kgU of burnup every month.
Recall from the earlier physics chapters that in a CANDU channel, the flux can change signifi-
cantly from one channel to another and from one axial location in a channel to another. Also
recall from the chapters on physics that to maintain reactivity, some fuel bundles are periodi-
cally moved from one axial location in the channel to another. These factors produce a variety
of power histories in any given bundle.
Figure 3 shows two illustrative power histories for CANDU fuel. The first is called a “declining”
power history and would be experienced by a fuel bundle that is loaded in a central axial
location of a C6 fuel channel. The second includes a “power ramp” and is experienced by
CANDU fuel which is initially loaded in a low-flux axial location of the channel and later shifted
to a location of higher flux.
Different types of power histories have different impacts on fuel performance. For example, an
element with a declining history generally experiences high power for a long period, which
usually results in relatively high gas pressure in the fuel element. On the other hand, a ramped
history exposes the fuel element to relatively high risk of “environmentally assisted cracking”.
These aspects are illustrated further in later sections.
Figure 3 Illustrative power histories
4 Collapsible Sheaths
To minimize unnecessary parasitic absorption of neutrons, CANDU fuel sheaths are kept very
thin—so thin that coolant pressure induces high compressive hoop stresses in them. At suffi-
ciently high compressive hoop stress, the sheath will lose its elastic stability and collapse
diametrically inwards into the diametric gap between the sheath and the pellet until it is sup-
ported by pellets. Figure 4(a) illustrates this phenomenon.
This collapse improves contact between the pellet and the sheath and therefore improves heat
transfer between them. This, in turn, lowers pellet temperature and reduces thermally driven
processes. Some such processes, such as diffusion and release of fission gas, decrease expo-
nentially with reduced local temperature. Therefore, even a modest decrease in local pellet
temperature leads to a disproportionately large reduction in these processes. For this reason,
designers of CANDU fuel prefer to promote diametric collapse of the sheath.
At the same time, excessive diametric or axial collapse can risk sheath failure from overly high
local strains. These conditions are called longitudinal ridging and axial collapse respectively (see
Figures 4(b) and 4(c)) and must be avoided. Therefore, a delicate balance must be struck
between promoting sufficient collapse to improve pellet-to-sheath heat transfer and avoiding
overstrain due to excessive collapse. This trade-off is explained in the following sections.
Fuel 15
Note: Shapes have been exaggerated for clarity

Figure 4 Sheath collapse: forms and stages
4.1 Diametric Collapse

Elastic instability in a thin circular cylinder can be calculated from Bryan’s equation [Bryan,
1888]:
P = E (t/r)3/[4(1-ν2)],
where P is the differential pressure ( = coolant minus internal), E is Young’s modulus, t is thick-
ness, r is radius, and ν is Poisson’s ratio.
Exercise: Consider a fuel sheath with the following properties: diameter, 13 mm; thickness, 0.4
mm; Young’s modulus, 80 GPa; and a Poisson’s ratio of 0.37. Its collapse pressure is 80,000 *
(0.4/6.5)3 / (4 * (1 - 0.372)) = 5.4 MPa.
In other words, the sheath just described will become elastically unstable when the differential
pressure (coolant minus internal) exceeds 5.4 MPa. By comparison, the coolant pressure is 10
MPa. Therefore, when a C6 fuel element is first loaded into the reactor, the sheath will collapse
elastically into the diametric gap.
Internal gas pressure builds up during irradiation (see Section 6). If the internal pressure in the
fuel element exceeds about (10 – 5.4) = 4.6 MPa, the differential pressure will drop below the
collapse pressure. The sheath will then no longer have much circumferential contact with the
pellet, unless by then creep has converted sufficient elastic strain into permanent plastic defor-
mation.
Improved heat transfer from diametric collapse is comparatively more important at high power
because fission gas diffusion increases exponentially with higher temperature. At the beginning
of life, the internal pressure is low, the sheath is elastically unstable, the sheath is in contact
with the pellet, and pellet-to-sheath heat transfer is good. This helps reduce the diffusion and
release of fission gas. By the time internal gas pressure builds up sufficiently to provide elastic
stability to the sheath, the element rating has usually decreased with burnup. Therefore, by
then, good heat transfer is not as critically important.
4.2 Longitudinal Ridging

If the coolant pressure increases well past the collapse pressure, eventually the sheath starts to
wrap itself around the pellet, as shown in Figure 4(b). This eventually creates a sharp ridge that
runs along the length of the sheath, called a longitudinal ridge. Depending on how sharp the
ridge is, the tip of the ridge can experience high bending strain. In an experimental irradiation,
fatigue cracks eventually developed at longitudinal ridges. Some experts have also postulated
that hydrogen in the sheath can migrate preferentially to longitudinal ridges, potentially hydrid-
ing and thereby embrittling the sheath at that location. For these reasons, strain at longitudinal
ridges must be controlled to reasonable levels.
Figure 5 shows a typical development of permanent strain at the tip of a longitudinal ridge
when pressure is increased. The two curves in this figure illustrate typical scatter in two differ-
ent test specimens in this specific experiment.
[Courtesy of COG]
Figure 5 Illustrative permanent strain at the tip of a longitudinal ridge
Initially, the strain increases rather gradually. Beyond a critical strain, however, a further in-
crease in pressure increases the sheath strain fairly rapidly. The point at which this occurs is
called the critical strain and has been experimentally found to be 0.5%. It is good practice to
design CANDU fuel to stay below this strain value at the longitudinal ridge. Fuel designed in this
way has never exhibited sheath failures at longitudinal ridges.
Fuel 17
Experiments have established that the following parameters are relatively more important in
influencing ridge strain: differential pressure, diametric clearance, sheath thickness, sheath
diameter, yield strength, and Young’s modulus.
4.3 Collapse into the Axial Gap

If the sheath is too thin, or if the axial gap is too long, the sheath can potentially collapse into
the axial gap, as shown in Figures 4(c), 6(a), and 6(b). Such a collapse has the potential to crack
the sheath, and therefore it is good engineering practice to avoid it.
[Figure (c) courtesy of COG]

Figure 6 Collapse into axial gap
The following major parameters influence axial collapse: differential pressure, length of the axial
gap, radius and thickness of the sheath, and the yield strength, Young’s modulus, and Poisson’s
ratio of the sheath material.
Figure 6(c) shows an illustrative calculation for instantaneous axial collapse in 37-element fuel.
Creep collapse can potentially occur at a lower pressure than the instantaneous collapse pres-
sure.
In CANDU fuel, axial collapse has been observed only under unusual situations such as absence
of a large part of a pellet or an excursion to unusually high temperature in experimental fuel.
5 Thermal Performance of Fuel

High temperatures in the UO2 pellet and in the Zircaloy sheath are important drivers for many
damage mechanisms. Other chapters discuss the equations that can be used to calculate
temperature distributions inside the fuel element.
Table 2 and Figure 7 provide some illustrative data that can be used to assess the thermal
performance of fuel.
[Source: Tayal et al., 1987]

Figure 7 Illustrative heat transfer and pellet temperature in a fuel element
Some parameters are not constant, but depend on other parameters. For example, Figure 7(a)
shows that the thermal conductivity of UO2 depends on local temperature. Likewise, the heat
transfer coefficient between the pellet and the sheath changes significantly with time, as shown
in Figure 7(b), because of changes during irradiation in factors such as the interface pressure
between the pellet and the sheath and the volume and composition of fission and filling gases
within the fuel element. In such cases, Table 2 lists an average representative value that can be
used for illustrative calculations.
Figure 7(c) illustrates a typical local temperature profile across the pellet radius. Note that (a)
the temperature distribution is approximately parabolic, (b) the temperature at the centre of
the pellet is about 1,800°C, and (c) the temperature at the surface of the pellet is about 400°C.
Needless to say, the absolute values of temperature depend on many parameters, including
Fuel 19
element power; Figure 7(c) is merely an illustrative example.

The above absolute temperatures, as well the temperature gradient across the pellet radius,
have significant consequences for fuel performance, as discussed in more detail later in this
chapter.
6 Fission Gas and Internal Pressure

6.1 Overview
Each fission reaction produces fission products (fission fragments), as described in earlier
chapters on physics. Some fission products are solid, whereas others are gaseous. These fission
products accumulate over time. The gaseous fission products lead to internal pressure which, if
excessive, can potentially create a hole in the sheath, either through mechanical overstrain or
through environmentally assisted cracking. A hole can provide a path for radioactive fission
products to leach into the coolant and to pose a radiological hazard to operating staff. For this
reason, fission gas pressure must be controlled and limited to acceptably low values.
As noted earlier, CANDU fuel is designed to have collapsible sheaths to improve heat transfer to
the coolant, which reduces pellet temperature, which in turn reduces the amount of fission gas
that can escape from the pellet and contribute to gas pressure in the fuel element. In addition,
through appropriate pellet shapes and clearances, empty space—called “open void”—is pro-
vided inside a fuel element to keep internal gas pressure below acceptable levels.
Over the years, CANDU reactors have maintained good control of fission gas pressure. In a rare
instance, however, control of fission gas pressure was lost for a very brief period in the 1980s in
a Bruce reactor in Ontario when unavailability of a fuelling machine prevented timely refuelling
of a fuel channel. As a result, some fuel bundles resided in the core for longer durations than
designed, resulting in excessive accumulation of fission gas and in high internal pressure. This in
turn caused cracks in the sheaths of a few fuel elements. The internal gas pressure can be
calculated using a model summarized below [Notley et al., 1979].
Figures 8–14 show some processes pertinent to fission gas pressure. Fission produces gas
uniformly within the grains of UO2. However, to exert pressure on the sheath, the gas must first
reach the “open space” within the fuel element (Figure 8a). This is the space in the axial and
diametric gaps, in the dishes, in the chamfers, in the cracks within the pellets, in the surface
roughnesses of the sheath and of the pellets, and in the cavity of the endcap. The movement of
gas from within grains to the open space is described below.
Figure 8 Terms for fission gas release

A number of processes are involved in transporting fission gas from within grains to the open
voidage. They include knock-out, recoil, diffusion, grain boundary sweep, storage in grain
boundary bubbles, growth of grain boundary bubbles, interlinkage of grain boundary bubbles,
and venting of grain boundary bubbles. These are described in the following paragraphs.
Figure 9 is an illustrative example of the steady-state release rate of 88Kr from a single crystal of
UO2 [Carroll et al., 1965].
[After Carroll et al., 1965]

Figure 9 Steady-state release of 88Kr from a single crystal of UO2
Below about 625°C, two thermally-independent processes—recoil and knockout—make signifi-
cant contributions to total release of fission gas. They are also sometimes called “athermal”
release mechanisms.
Above about 625°C, the emission rate increases rapidly with temperature. This is due to diffu-
sion, because diffusivity increases exponentially with temperature. This phenomenon is also
sometimes called “thermal” release. If high fission gas pressure does occur in CANDU fuel, it is
almost always driven primarily by thermal diffusion.
Fuel 21
Fission gas is produced within grains. If the grain boundary acts as a sink, a gradient develops in
the concentration of gas across the grain's radius. This concentration gradient drives the rate of
diffusion from the grain to the boundary.
At the grain boundary, the gas accumulates in bubbles (Figures 8(b) and 10), which become
progressively bigger as more gas reaches the grain boundaries.
B: Bubbles; T: Tunnels
[Source: Hastings, 1983]
Figure 10 Grain boundary bubbles and tunnels
In parallel, high temperature can also cause the grains to grow (Figures 8(c) and 11), either
through equi-axial grain growth (Figure 12) or, at even higher temperatures, through columnar
grain growth (Figure 13). The boundaries of the growing grains collect the gas that was previ-
ously contained in the areas swept by grain growth. This gas is added to the inventory of gas at
grain boundaries. This process is called grain boundary sweep.
Figure 11 Changes in sizes and shapes of UO2 grains during irradiation
Figure (b) also shows fission gas bubbles at grain boundaries

Figure 12 Equiaxial grain growth
Fuel 23

Figure 13 Columnar grain growth
As additional gas reaches grain boundaries, the grain boundary bubbles eventually become big
enough to touch each other. This is called interlinkage. The volume contained within inter-
linked bubbles reflects the storage capacity of the grain boundary bubbles. Any gas that
reaches the grain boundaries over and above the bubble capacity is called excess gas.
When fuel power changes, thermal expansions or contractions in the pellet create micro-cracks
in the UO2 ceramic. These provide a pathway for the excess gas in interlinked bubbles to vent to
the open space between the pellet and the sheath. This can be deduced from Figure 14, which
shows measurements of internal pressure during irradiation [Notley et al., 1979].
Figure 14 shows that at high power, internal pressure does not change during constant-power
periods, even though fission gas must be produced continually during these periods. Internal
pressure increases only immediately after changes in power, suggesting that grain boundary gas
escapes to the open space through interconnected micro-cracks formed upon changes in power.
Under the usual operating conditions in CANDU power reactors, only a small fraction (<10%) of
the gas produced vents to the open space. Most of the gas stays tied up within the grains of
UO2 and in grain boundary bubbles.
[Source: Notley et al., 1979]

Figure 14 Variation of internal gas pressure during irradiation
The process by which fission gas reaches the open space is called fission gas release. It in-
creases rapidly with higher local temperature, and for a given temperature, it increases with
burnup and with smaller grain size.
Later sections of this chapter will describe a simplified model to quantify the amount of fission
gas that reaches the grain boundaries through diffusion.
6.2 Recoil
Recoil is one of the mechanisms that releases fission fragments outside the pellet. To a first
approximation, fission fragments travel through UO2 in a straight line. They lose energy as they
encounter (primarily) electrons in the material. When the initial kinetic energy of the fission
fragment has been fully dissipated in this process, the fission fragment stops as a fission prod-
uct. This occurs in about 10 µm. However, if the fission fragment reaches a surface of the pellet
before expending all its energy, it escapes the pellet. Therefore, recoil is a contributing factor in
fission product release from near a surface of the pellet (~10 µm), especially at low temperature
[Olander, 1976].
6.3 Knockout
Knockout is another athermal mechanism for fission product release that occurs at low temper-
ature. As a fission fragment travels through the pellet, it may occasionally collide with nuclei of
other atoms in the lattice. The latter are called “knock-ons”. Knock-ons are usually uranium or
oxygen atoms in the fuel, but occasionally a knock-on can be a fission gas atom lodged in the
lattice. A knock-on acquires kinetic energy from the collision, and it also travels in a straight line
Fuel 25
and loses energy as it travels. A primary knock-on generally travels a distance of about 200 Å
(Angstroms) before coming to rest. If the knock-on reaches a surface of the pellet before
stopping, it escapes the pellet. If the knock-on was a fission fragment, this process results in
release of fission products [Olander, 1976].
6.4 Diffusion of Fission Gas to Grain Boundaries

As noted earlier, diffusion is a key process that drives the fraction of produced gas that eventu-
ally reaches the open space at high operating power. This section discusses the rate of diffusion
of stable gases to grain boundaries after a steady-state profile of gas concentration has been
built across the grain. This can be calculated from Booth’s model [Booth, 1957].
Grains of UO2 are of irregular shape and size. Nevertheless, this calculation model postulates
that gas diffusion out of grains can be approximated by assuming that the pellet consists of a
collection of spheres of equal radii. The radius “a” of the equivalent sphere is equal to the
average grain radius.
To calculate the leak rate at the grain boundary, we first need to know the gas concentration
profile across the grain radius. Booth solved Fick’s equation for diffusion in a sphere [Fick, 1855]
assuming that the surface of the sphere was a sink. This results in a parabolic concentration
profile:
c (r ) 
f
6D
a  r  ,
2 2
(1)
where c is the local concentration, r is the radial distance from the grain centre, a is the grain
radius, f is the fission rate,  is the yield, and D is the diffusivity. Concentration means the
number of nuclides per unit volume of the grain. By integrating the above equation, we can
obtain the volume-average concentration cav.
Diffusion due to a concentration gradient results in the following rate of nuclide flow per unit
area of grain surface, q:
c
q  D at r  a . (2)
r
The leak rate LEAK is defined as the rate of nuclide diffusion from the grains per unit grain
volume. Therefore [Tayal et al., 1989],
3 c
LEAK   D at r  a . (3)
a r
By differentiating Equation (1), substituting the results into Equation (3), and then substituting
the expression for cav into Equation (3), we get:
15 Dcav
LEAK  . (4)
a2
Equation (4) describes the rate at which fission gas diffuses to the grain boundary during steady
state.
6.5 Internal Gas Pressure

As noted earlier, internal gas pressure must be kept within acceptable levels to avoid overstrain
failures. Therefore, internal gas pressure is a consideration in selecting some details of the
internal design of a fuel element, such as pellet shape and axial gap.
Internal gas pressure is driven by the combined volumes of filling gas and fission gas. These
gases are stored in a variety of locations within a fuel element, e.g., in dishes, in pellet/sheath
radial gaps, in axial gaps, in pellet cracks, and in surface roughnesses of the pellet. Furthermore,
the local temperature at each storage location differs significantly because of the steep tem-
perature gradients within the fuel element.
To calculate the internal pressure, one usually divides the fuel element into many regions, each
with a nearly constant temperature. Then the ideal gas law is used to integrate the impacts on
internal gas pressure:
p1v1 T1  p2v2 T2 , (5)
where p, v, and T represent pressure, volume, and temperature respectively and subscripts 1
and 2 refer to two different states.
Exercise:
Consider a fuel element which was initially filled with 2 ml of helium at standard temperature
and pressure (STP). During irradiation, 10 ml (at STP) of fission gas was released to the open
space. Let us assume for simplicity that at operating power, the open space in the fuel can be
divided into three zones of essentially constant temperature, each with a volume:
 0.4 ml at 1,900 K;
 1 ml at 1,500 K; and
 0.6 ml at 1,000 K.
This fuel element’s internal gas pressure can be calculated as follows. Recall that at STP, the
temperature is 273 K and the pressure is 0.1 MPa. Therefore, corresponding to the filling gas
volume of 2 ml at STP, the “voidage” (V/T) is (2/273) = 0.00733 ml/K.
On-power, the same gas is now stored at an effective voidage of (0.4/1900 + 1/1500 + 0.6/1000)
= 0.00148 ml/K. In other words, the on-power voidage is about (0.00148/0.00733), or one-fifth
the off-power voidage. Therefore, from the ideal gas law, the on-power pressure of the filling
gas is (0.1 * 0.00733/0.00148) = 0.5 MPa.
Including the fission gas, the total STP volume of gas is (2+10) = 12 ml. On-power, because 2 ml
(STP) of filling gas develops a pressure of 0.5 MPa, 12 ml (STP) of filling gas plus fission gas will
develop a pressure of (0.5 * 12/2) = 3 MPa.
7 Stresses and Deformations

A variety of mechanical and thermal loads and microstructural changes lead to in-reactor
stresses, strains, and deformations in fuel. Excessive stresses and strains can in turn potentially
result in a hole in the sheath or in the endcap, which can release radioactive fission products to
the coolant. Excessive stresses or strains in the endplate can break it also, which may pose
difficulties in discharging the fuel from the reactor. Excessive deformations can potentially jam
the fuel in the channel, reduce local cooling, or damage surrounding components through, for
example, crevice corrosion.
It would be unrealistic to operate fuel with zero stresses, strains, and deformations, and it
Fuel 27
would be very expensive to limit these always to trivial values through massively over-
conservative designs. At the same time, it is essential to keep them reasonably below harmful
levels. Some such considerations are described in this section.
7.1 Thermal Stresses in Pellets

As noted in an earlier chapter, the radial temperature distribution in an operating fuel pellet is,
to a first approximation, parabolic. The temperature gradient is steep, about 1800°C at the
centre and about 400°C at the surface, over a radius of only about 6 mm.
Think of a pellet as being made up of many concentric rings (or annuli). Thermal expansion is
higher in the inner hotter rings and comparatively lower in the neighbouring, relatively cooler
outer rings. Differential thermal expansions lead to dimensional misfits between neighbouring
annuli, which in turn lead to stresses. For this reason, operating UO2 pellets are often found
with extensive radial cracks in their outer regions, as shown in Figure 15.

Figure 15 Typical cracks in a high-power pellet
Textbooks in mechanics provide a framework of equations which can be used to calculate
thermal stresses in a cylinder experiencing parabolic temperature profile across its radius [e.g.,
see Timoshenko et al., 1970]. Figure 16 illustrates a typical result.
[Courtesy COG]
Figure 16 Thermal stresses in a pellet
Note that the hoop stress is compressive in the inner part of the pellet and tensile in the outer
part of the pellet. The tensile hoop stresses lead to radial cracking in the outer parts of the fuel
pellet, as can be observed in Figure 15. The radial cracks add to the storage space for fission
gas, reducing internal gas pressure. They also increase pellet expansion and hence sheath
stress, encouraging detrimental environmentally assisted cracking (described in Section 8).
All three major components of stress are compressive in the central part of the pellet. This
leads to compressive hydrostatic stresses, which reduce the size of fission gas bubbles at the
grain boundaries. Moreover, the stresses are well into the plastic range. This can potentially
lead to permanent deformations in the pellet. High local temperatures plus high local stresses
can also lead to high rates of creep, leading to relatively rapid relaxation of stress. High local
creep plus local plasticity are also believed to cause "dish filling" at high local temperatures.
7.2 Changes in Element Diameter and Length

During irradiation, several processes affect the diameter and length of the pellet and sheath.
Notable among these are densification, swelling, creep, dish filling, thermal expansion, and
hourglassing. Brief descriptions of these are given below.
Densification of pellets is caused primarily by irradiation-induced sintering of UO2 in the reactor.
Figure 17 shows an illustrative example of initial as-fabricated pores that have been removed
during irradiation, resulting in densification. Although research has identified many factors that
can affect densification, the relatively more important primary influences are believed to be
temperature, initial density, and duration of high temperature.
Fuel 29

Figure 17 Removal of as-fabricated pores during irradiation
Swelling of pellets is caused by fission products—solid as well as gaseous. The combined
volume of solid fission products is larger than that of the parent material, resulting in swelling of
the pellet. Unreleased fission gases also contribute to pellet swelling. For each percent atomic
burnup, solid fission product swelling increases pellet volume by approximately 0.32% [Olander,
1976]. However, this number is subject to large uncertainties. [Note: 1 atomic percent = 240
MWh/kgU].
Creep due to external coolant pressure decreases sheath diameter. However, outward creep of
the sheath can be expected if the internal gas pressure exceeds the coolant pressure.
Creep within the pellet is believed to cause axial expansion of the hot inner core of the pellet.
This is also the region where dishes are located, as shown in Figure 1. In fuel that is operated at
high local temperature, dishes are usually found to have become smaller after irradiation than
before. This process is called dish filling.
The operating temperatures cause thermal expansion in the pellet as well as in the sheath. The
pellet operates at much higher temperature than the sheath—for example, the average pellet
temperature in the illustration in Figure 7 is about 1000°C, whereas the average illustrative
sheath temperature in Table 1 is about 340°C. As well, the coefficient of thermal expansion of
UO2 is about twice that of Zr-4, as can be seen in Table 2. Therefore, the pellet usually expands
much more than the sheath. At times, this results in the pellet filling the diametric gap and
“pushing” the sheath outward. This in turn creates high tensile stresses in the sheath, often in
the plastic range.
Diametric expansion of the pellet is usually much larger at the two ends of the pellet length
than at the mid-point, as shown in Figure 18. This is called “hourglassing” and is caused by
three factors: non-uniform temperature across the radius of the pellet, axial compression, and
variations in initial density. These are illustrated in Figure 19 and explained below.
Figure 18 Fuel rod before and after startup

 Non-uniform local stiffness: As noted in Section 7.1, the parabolic temperature profile in
the pellet leads to thermal stresses. The net diametric expansion of the pellet is influ-
enced by these thermal stresses. However, these vary from one transverse cross section
to the next. At the pellet mid-plane, the deformation of a transverse cross section is re-
sisted, not only by the stiffness of that cross section, but also by the stiffness of the sur-
rounding UO2. However, at and near the end of the pellet, a transverse cross section is
surrounded by less UO2. Therefore, it can and does expand more than the cross section
at the mid-plane of the pellet. This is one reason for pellet hourglassing.
 Axial compression: When the pellet expands axially, there may be axial interference be-
tween neighbouring pellets, between the stack of pellets and the endcap, or both. This
results in compressive axial forces, which add to hourglassing. Axial compression has a
larger influence on hourglassing if the force is applied near the surface of the pellet
rather than near the centre.
 Initial variation in pellet density: One step during pellet fabrication is pressing of the pel-
let at its end (see Chapter 18). This causes the as-pressed densities to be higher near
the end than at the centre. The pellet subsequently sinters during fabrication and dur-
ing in-reactor densification. The resulting densification is higher at the mid-plane than
at the ends because of the lower initial density at the mid-plane. This also contributes
to hourglassing.
Fuel 31
[Source: Tayal, 1987]

Figure 19 Causes of hourglassing in pellets
When an expanded hourglassed pellet pushes against the sheath, it causes “circumferential
ridges” to be formed in the sheath at the ends of the pellets, giving the sheath the appearance
that resembles a bamboo stalk, as shown in Figure 18. The circumferential ridges are locations
of high stresses and strains, often well into the plastic range, and are a main location for envi-
ronmentally assisted cracking, as discussed in the next section.
The pellet and sheath diameters change continually during irradiation in response to the proc-
esses described above and to continual changes in element power. This means that the pellet-
to-sheath gap and the interface pressure between the pellet and the sheath also change con-
tinually during irradiation. These changes feed back to pellet temperature and to fission gas
release, which in turn feed back to pellet expansion. Thus, a complex feedback loop is set up.
8 Environmentally Assisted Cracking

Environmentally assisted cracking (EAC) occurs in nuclear fuel when irradiation- and/or hydride-
embrittled Zircaloy experiences high tensile stresses and strains in the presence of corrosive
agents. These high stresses and strains are usually caused either by changes in power (called
power ramps) or by excessive internal gas pressure. Figure 2 demonstrates that in the past, EAC
caused by power ramps has been a dominant damage mechanism in CANDU fuels and suggests
that CANDU fuels might have relatively low margins to defects from this mechanism.
Furthermore, our experience has been that when conditions are right for power-ramp EAC in a
CANDU reactor, a number of fuel bundles tend to fail within a short period of time. As well,
multiple fuel elements can often fail in a given fuel bundle. This combination can cause iodine
levels in the coolant to increase quickly. For the above reasons, it is particularly important to
have a good understanding of this damage mechanism and of ways to quantify and control it.
To date, EAC has manifested in CANDU fuels through three main routes:
 EAC failures in the sheaths at interfaces between neighbouring pellets (called circumfer-
ential ridges) as a result of excessive pellet expansion during increases in power (called
power ramps). Figure 20 shows an illustrative example [Penn et al., 1977].
 EAC failures in sheaths near their welds with endcaps as a result of excessive pellet ex-
pansion during power ramps. Figure 21 shows an illustrative example of a typical crack
near the sheath/endcap junction [Floyd et al., 1992].
 EAC failures in the sheaths near their junctions with endcap welds as a result of exces-
sive gas pressure accumulated during abnormally long residences of fuel in the reactor
[Floyd et al., 1992]. Such defects appear similar to those shown in Figure 21.
EAC is also called by other names in the scientific literature. In Canada, it is also called “stress
corrosion cracking” (SCC). In the LWR community, it is usually called “pellet/clad interaction”
(PCI). Both these names were coined shortly after the initial rash of power-ramp defects and
were based on initial understanding of their mechanisms. Subsequently, much additional
knowledge about the mechanism has been uncovered in a myriad of tests around the world. As
a result, several labels very similar to EAC were proposed in later literature. In light of this
additional information, a re-think in the early 2000s concluded that EAC is an appropriate label
for this mechanism.
Figure 22 illustrates some key terms associated with EAC during power ramps.
[Courtesy COG]
Figure 20 Power ramp cracks originating at a circumferential ridge
Fuel 33
[Courtesy COG]
Figure 21 Power ramp cracks near a sheath/endcap junction
Figure 22 Key terms for environmentally assisted cracking during power ramps
8.1 EAC Processes

To control a CANDU reactor’s relative balances of neutronic powers in different parts of the
reactor, CANDU fuel bundles can be repositioned in more than one axial location within a fuel
channel. This is achieved by on-power refuelling, which is an important and unique feature of
CANDU reactors.
Different parts of a fuel channel usually have different neutron fluxes, generally in a cosine
shape over the channel length. Therefore, during on-power refuelling, a previously irradiated
fuel bundle can go from a relatively lower initial power to (or through) a relatively higher final or
intermediate power. This causes power ramps in irradiated fuel elements.
The resultant pellet expansion strains and stresses the sheath, in the process breaking any
protective layer of ZrO2 that might have built up during initial irradiation. By then, the Zircaloy
has also been embrittled by two processes: (a) by fast neutrons experienced during many
months of irradiation at the low-power axial position, and (b) by hydrides, which can either
promote or hinder initiation and growth of cracks depending on their orientation. In addition,
by then, previous irradiation has also released fission products inside the fuel element, some of
which are corrosive (such as iodine, cesium, and cadmium). Some fraction of these corrosive
fission products is believed to react chemically with the CANLUB layer that is applied on the
inside of the sheath, reducing the amount of fission products that are available to attack Zir-
caloy. The balance of the fission products is available for chemical attack on Zircaloy while the
Zircaloy is also experiencing high stresses and strains due to the power ramp. These situations
create a potentially potent combination for EAC: high stresses and strains, concurrent with
embrittled Zircaloy, and also concurrent with a corrosive internal environment.
Figure 23 provides a specific example of some of the processes described above. It is based on
in-reactor measurements of ridge strains during a power ramp to about 55 kW/m [Smith et al.,
1985]. Specifically, this fuel element was first irradiated in a base irradiation at about 30 kW/m
to about 4 MW•d/kgU. Then it was inserted into an experimental rig that could measure sheath
diameters during irradiation. The power in the experimental rig was first increased from 0 to
about 30 kW/m, where it was held constant for about 30 days. Then the power was increased
again, this time to about 55 kW/m.
[Source: Tayal et al., 2008-2]

Figure 23 Mechanism for power ramp failures
Fuel 35
On the plot in Figure 23 for ridge strain vs. element power, we have also superposed the strain
at which ZrO2 breaks and the strain at which Zr-4 becomes plastic [Tayal et al., 2008-2]. One can
see that ZrO2 breaks very early during the ramp and therefore is unlikely to play a significant
role in protecting the sheath from chemical attack by fission products. We can also see that
defects start much later than when the sheath has become plastic. During the plastic part of
loading, stresses increase but little, yet significant plasticity is required before EAC failures occur
in CANDU fuel. Therefore, the primary mechanical driver of EAC is not stress alone, but rather a
combination of stress and strain.
To maintain fuel integrity, the above combination needs to be managed to acceptable levels.
This is done in part by designing and operating fuel below defect thresholds for EAC, as de-
scribed in the next section.
8.2 Defect Threshold

Like many processes, EAC contains some inherent variability. A defect threshold delineates an
area below which no fuel defects are expected from EAC. Above the defect threshold, failures
may or may not occur. Generally, the farther above a defect threshold the process is, the higher
the probability of fuel defects.
An extensive literature of well over a thousand papers suggests that a very large number of
parameters can potentially influence EAC. It is not practical to provide here an exhaustive
mechanistic treatment of such a large list of parameters. Therefore, the authors have arbitrarily
chosen to illustrate key concepts through a relatively recent insight [Tayal et al., 2000]. It is
acknowledged that other perspectives are also available in the literature, but they are not
discussed here for reasons of space.
In the formulation described here, stresses and strains in Zircaloy are represented by the size of
the power ramp (ΔP). Corrosive agent concentration is represented by a combination of pre-
ramp power (Pi) and burnup (ω) at the time of the ramp. Material degradation is represented
by the burnup at the time of the ramp (ω).
Figure 24 shows the empirically obtained defect threshold at 140±20 MWh/kgU for non-
CANLUB fuel.

Figure 24 Defect threshold due to power ramp in non-CANLUB fuel at 140±20 MWh/kgU
The figure contains two lines. The lower line is the defect threshold for non-CANLUB fuel.
Based on empirically derived curves, the defect threshold for CANLUB fuel is located about 12
kW/m higher than that shown in Figure 24 [Tayal et al., 2000].
Exercises:
Consider a non-CANLUB fuel element for which the operator is contemplating a power ramp
from 30 kW/m to 50 kW/m for 2 hours at 140 MWh/kgU. Its pre-ramp power, Pi, is 30 kW/m.
Its power ramp, ΔP, is (50 – 30) = 20 kW/m. From Figure 24, the fuel element can withstand a
power ramp of about 14 kW/m at 140 MWh/kgU. Therefore, the applied power ramp will
exceed the defect threshold and expose the fuel element to risk of an EAC defect.
As a second exercise, consider the same ramp being applied to a CANLUB fuel element. Its
defect threshold is at (14 + 12) = 26 kW/m. The applied ramp of 20 kW/m is less than its defect
threshold, and therefore the fuel can be expected to survive the ramp.
8.3 Defect Probability

The basic rule is always to operate fuel in a region which is safe from EAC defects, i.e., below the
defect threshold. Nevertheless, on very rare occasions it may become necessary to operate a
small number of fuel bundles slightly above the defect threshold in a one-off situation that
might have become unavoidable due to an unusual combination of circumstances, such as
during restart under a distorted flux shape after a shutdown. One then needs to consider taking
a calculated risk. To develop an effective strategy for such rare occasions, one needs to quantify
the risk of fuel failure above the defect threshold. This is done through knowledge of defect
probability beyond the defect threshold, as described below.
Fuel 37
Figure 24 contains not one, but two lines. As noted earlier, the lower line is called the defect
threshold; it is the line below which no defect occurs. The upper line is one above which no fuel
bundle stays intact. The region in between contains some scatter: some bundles in that region
fail, whereas others stay intact.
Empirical evidence suggests that for all practical purposes, the defect threshold represents a
trivial likelihood of defect, essentially zero. Nevertheless, in statistical theory, probabilities
usually have asymptotic ends, that is, a probability is never zero, nor is it ever 100%. Therefore,
the defect threshold shown in Figure 24 is usually assigned a low arbitrary probability such as
1%. Likewise, the upper line in Figure 24 is usually assigned an arbitrarily high defect probability
close to 100%.
Figure 25 shows an illustrative increase in defect probability as a fuel bundle experiences power
ramps that exceed the defect threshold [Penn et al., 1977]. Theoretically, the probability should
be an “S”-shaped curve with asymptotic ends. For practical purposes, however, empirical
evidence shows that the asymptotic parts near the two ends are small and that for the most
part, the defect probability increases essentially linearly with increasing distance from the
defect threshold (see Figure 25).
[Source: Penn et al., 1977]

Figure 25 Defect probability
The distance between the two lines shown in Figure 24 is about 5 kW/m; however, when the
data from other burnups are also considered, the overall separation between the two lines for
all burnups is closer to 10 kW/m. Therefore, to a first approximation, and if one ignores small
asymptotic regions near the two extremes, for every increment of one kW/m above the defect
threshold, the defect probability increases by about 10 percentage points until it reaches 100%.
The portion of the ramp that is above the defect threshold is sometimes also called the “excess
ramp”. In the illustrative example below, the “excess ramp” is 8 kW/m.
Exercise: Consider a non-CANLUB fuel bundle where the steady-state power in the outer
element changes from 35 kW/m to 50 kW/m at 140 MWh/kgU. This gives it a ramp of (50 – 35)
= 15 kW/m. From Figure 24, the defect threshold at this initial power is 7 kW/m. Therefore, the
actual ramp is (15 – 7) = 8 kW/m above the defect threshold. Therefore, this fuel bundle has a
(8*10%) = 80% chance of fuel defects through EAC.
8.4 Mitigation
Based on past experience, the fuel industry has adopted certain strategies to mitigate EAC
defects. Some illustrative examples are given below.
 CANLUB: CANLUB significantly increases the defect threshold. It is believed that the ef-
fectiveness of CANLUB increases with thickness. However, it is also believed that after a
certain thickness, additional increases in CANLUB thickness likely yield diminishing re-
turns. At the same time, there is concern that very thick layers of CANLUB may poten-
tially release excessive hydrogen into the sheath. Therefore, a CANLUB layer of a se-
lected optimal thickness is applied to the inside of the sheath [Wood and Hardy, 1979].
 Fuel Management: The key operating parameters for EAC—size of the power ramp, ini-
tial and final powers, and burnup—are all controllable to some degree through refuelling
strategy. Therefore, optimal refuelling strategies have been devised to avoid EAC failures
in CANDU reactors. A compelling example comes from the Pickering reactors, where
early EAC failures were eliminated in part by improved refuelling schemes.
 Sheath/Endcap Junction: A significant contribution to EAC defects at sheath/endcap
junctions arises from the stress concentration at re-entrant corners of these junctions.
Therefore, the detailed design of the sheath/endcap junction must be carefully con-
trolled, including the radius of the re-entrant corner, the pellet profile at the end of the
fuel stack, and the axial distance between the pellet stack and the re-entrant corner.
 Element Rating: Smaller ramps and lower heat generation rates reduce the threat from
EAC, as shown in Figure 24. These can be achieved by, for example, spreading a bundle’s
power among more fuel elements, e.g., by using more but smaller-diameter fuel ele-
ments within a fuel bundle. CANDU fuel has indeed evolved in this direction, starting
with a fuel bundle that used 19 fuel elements, then moving to one that used 28, to 37
elements, and to the CANFLEX design that uses 43 fuel elements [Hastings et al., 1989].
 Detailed Internal Design: Sheath strain and fission gas release can also be influenced by
the detailed internal design of the fuel element, such as pellet density and clearances
between the sheath and the pellets.
Fuel 39
9 Vibration and Fatigue

Fuel elements produce large amounts of heat and therefore must be cooled aggressively.
Coolant velocity is usually high (about 10 m/s), and the flow is at least partly turbulent. More-
over, the coolant may sometimes contain standing waves of pressure pulses. The pressure
pulses are created when a pump’s vanes pass over the pump outlet and may be subsequently
amplified by resonance in pipes that carry the coolant. These combinations of conditions can
lead to vibrations and rocking in fuel elements and bundles and even in fuel strings. This in turn
creates the potential for fatigue and fretting.
In a significant defect excursion, these conditions did indeed occur in 1991 in the Darlington
reactors in Canada. This led to excessive axial vibrations in the fuel strings, which cracked
endplates in fatigue and also caused significant fretting of spacer pads [Lau et al., 1992]. This in
turn led to shutdown of the reactors for a few months and significant loss of revenue.
To avoid damage due to fatigue and fretting, one needs to understand the science behind fuel
vibrations and the resulting alternating stresses and incorporate appropriate mitigation strate-
gies in the design and operation of the fuel and the plant. The next section describes some
aspects of the science of fuel vibration and fatigue.
9.1 Alternating Stresses

Axial flow and turbulence induce lateral vibrations in fuel elements, whereas pressure pulses
induce axial vibrations. Both types of vibrations generate alternating stresses in the fuel bundle
endplate and in the assembly weld (that is, the weld between the endplate and the endcap).
Quantification of alternating stresses and the resulting fatigue due to lateral vibrations involves
a number of calculations, such as the bending moment in the assembly weld and in the end-
plate due to bending of a fuel element, the resulting alternating stresses, and a comparison of
these stresses to the fatigue strengths. Mathematical models to do this are available [Tayal et
al., 1984] and are summarized below.
9.1.1 Moment Due to Fuel Element Deflection

The overall model is shown in Figure 26. If a fuel element of length ℓ experiences a uniformly
distributed load p that deflects it in a direction radial to the fuel bundle, its lateral deflection v is
resisted by its own moment of inertia as well as by the stiffness of the two endplates. The
endplates are twisted by the fuel element and therefore act as torsional springs in resisting the
twisting moment W applied by the fuel element.

Figure 26 Lateral vibration of a fuel element
If the fuel element’s flexural rigidity is J, the classical equation for beam deflection v can be
written as [Gere et al., 1997]:
d 2v
J  W  px(  x) 2 , (6)
dx 2
where x represents the distance along the length of the fuel element.
For boundary conditions of:
v  0 at x  0 and at x   , (7)
Equation (6) integrates to:
Jv  Wx(x  ) 2  px(x3  2x2  3 ) 24 . (8)
To derive an expression for the moment transmitted through the assembly weld, consider the
relationship between the moment and rotation of the spring:
dv
   W S at x  0
dx , (9)
where ω represents the rotation of the fuel element.

By differentiating Equation (8) to obtain the fuel element’s slope at the end and inserting this
into Equation (9), we obtain a relationship for the moment W transmitted by the fuel element to
the endplate through the assembly weld [Tayal et al., 1984]:
 
p 2  1 
W  . (10)
12 1  2 J 
 S 
Equation (10) expresses the moment carried by the endplate and the assembly weld in terms of
Fuel 41
the uniformly distributed load p. However, p is usually not known. Instead, the mid-span
deflection δ can be estimated more easily through other means. Therefore, p can be replaced in
Equation (10) by the mid-span deflection δ using the following additional information:
  v at x   2 . (11)
This results in the following equation for the moment carried by the assembly weld and the
endplate as a function of the mid-length deflection of the fuel element:
W  32 J  (2  10 J  S )
. (12)
To calculate the moment through the above equation, one needs to know the spring constant S
of the endplate, which is discussed in the next subsection.
9.1.2 Spring Constant of the Endplate

The moment W as calculated above twists the endplate by a force which decays along the
circumference of the endplate. As the endplate twists, it forces its own local twist on the fuel
elements that are welded to it, which the latter resist. Therefore, the applied twisting moment
on the endplate is resisted not only by the torsional rigidity of the endplate, but also by the
torsional rigidities of the other fuel elements welded to the endplate. For this reason, the
endplate behaves like a beam resting on an elastic foundation. Figure 27 shows such a model.

Figure 27 Endplate as a beam on elastic foundations
The endplate rings are circular. Sensitivity studies show that under on-power conditions of
CANDU fuel elements, the effect of circularity is small. Therefore, to a good first approximation,
the mathematics of the endplate can be simplified considerably by treating it as a straight
(rather than circular) beam on elastic foundations. A mathematical model for this can be
developed based on the science given in classical textbooks on mechanics of materials [for
example, see Timoshenko, 1961].
Let β be the local twist of the beam. A foundation of modulus κ will exert a twisting moment of
- κ β per unit length of the beam. If the beam’s cross section is assumed to carry only a twisting
moment Mz, then the equilibrium of an infinitesimal length yields:

dM z  d 2 
 GC  2    , (13)
dz  dz 
where z is the distance along the length of the beam, G is the shear modulus, and C is the
twisting moment of inertia.
This second-order differential equation can be integrated to yield the rotation and the twisting
moment along the beam’s length. The two constants of integration can be obtained by the
requirement that the rotation be zero at infinite distance from the origin and that the internal
twisting moment at the origin be half the applied twist. This process yields the following
equations:
T   z
 e , (14)
2
where T is the applied twisting moment and
   (GC ) 2 .
1
(15)
If the beam’s response to W is represented as a torsional spring, its spring constant S is:
T
 2  GC  2 .
1
S (16)
 z 0
The above spring constant can be used in Equation (12) to calculate the moment carried by the
assembly weld.
9.1.3 Foundation Modulus

Equation (16) requires knowledge of the foundation modulus κ for the endplate.
Let two endplate rings of radius R connected by N fuel elements be twisted by an angle β. The
ends of each sheath, which are held perpendicular to the endplate surface by the assembly
weld, will also rotate by the same angle β. If W represents the resistive moment applied by
each sheath to the endplate, then m, the resistive moment from the sheath per unit endplate
length, can be evaluated by spreading the total of such moments over the circumference of the
endplate:
m  NW (2 R) . (17)
W is also equal to the bending moment applied by the endplate to the fuel element to rotate its
end by β. Therefore, classical beam theory can be used to relate W and β as follows:
  W (2J ) . (18)
Therefore, by combining Equations (17) and (18), the foundation modulus κ can be calculated
as:
  m   JN ( R) . (19)
To account for the tight contact with the neighbouring fuel bundle under high hydraulic load, it
is customary to assume twice the actual number of fuel elements in Equation 19 above.
Fuel 43
9.1.4 Cross-Sectional Properties of the Endplate

The equations presented above also require knowledge of the endplate’s twisting moment of
inertia (C) and the fuel element’s flexural rigidity (J).
The twisting moment of inertia C of the endplate can be obtained from the classical equation
for a rectangular beam. For the usual shapes of CANDU endplates [Tayal et al., 1984]:
C  0.26ct ,
3
(20)
where c is the width of the beam and t its thickness.
9.1.5 Flexural Rigidity of the Fuel Element

At high power, the pellet expands and grips the sheath tightly, often stretching the latter well
into the plastic range. Therefore, the flexural rigidity of a fuel element usually contains contri-
butions from the sheath and from the pellets. The concept can be illustrated more simply by
examining the experimental results shown in Figure 28.
[Source: Pettigrew, 1993]

Figure 28 Stiffening of a fuel element at operating power
When the instrumented fuel element represented in Figure 28 was initially brought to high
power, its vibration frequency increased significantly [Pettigrew, 1993]. Frequency of vibration
can be related to flexural rigidity through the classical theory of vibration. This suggests that at
initial high power, the flexural rigidity of the fuel element is significantly higher than that of an
empty sheath. The difference reflects the contribution of the pellet when it is in tight contact
with the sheath.
However, the pellet is not always in tight contact with the sheath. With continued irradiation,
the interfacial pressure between the pellet and the sheath relaxes due to creep. This in turn
lowers the net flexural rigidity of the fuel element and hence decreases its vibration frequency.
Lower power also reduces the pellet/sheath interfacial pressure. These dynamics are also
reflected in Figure 28. For this reason, we express the flexural rigidity of the fuel element as
that of the sheath plus that of the pellet multiplied by a “pellet interaction factor” G which
reflects the varying levels of pellet/sheath interaction.
Another complication is that UO2 is brittle and therefore cannot support significant tensile
loads. Therefore, the resulting equation for pellet flexural rigidity is fairly complex, as described
for example in [Tayal et al., 1984]. For simplicity, here we treat the impact of pellet cracking by
means of a pellet cracking factor ∝ which reduces the net contribution of the pellet to the
overall flexural rigidity of the fuel element. As an illustrative number, the examples below use a
pellet cracking factor ∝ of 0.1.
Therefore, the net flexural rigidity J of the fuel element can be described as:
J  J s  GJ p
   (21)
 E s   r14  r24    E p rp4 G
4  4
where J is the flexural rigidity of the fuel element, Js the flexural rigidity of the sheath, Jp the
flexural rigidity of the pellet, Es the Young’s modulus of the sheath, r1 the outer sheath radius, r2
the inner sheath radius, Ep the Young’s modulus of the pellet, rp the pellet radius, ∝ the pellet
cracking factor, and G the pellet interaction factor.
The pellet interaction factor G varies between 0 and 1. Under the conditions of the experiment
shown in Figure 28, the pellet interaction factor G is zero when the fuel element is at zero
power. When the fuel element is first brought to full power, G increases to 1 because the
interfacial pressure between the pellet and the sheath is high. During continued operation at
full power, G gradually decreases to 0 as the interfacial pressure between the pellet and the
sheath relaxes. Therefore, the pellet/sheath interfacial pressure has a major impact on the
pellet interaction factor.
For a given lateral deflection of the fuel element, stresses in the endplate and in the assembly
weld are higher for higher levels of pellet/sheath interaction. Therefore, the upper end of the
pellet interaction factor range gives conservative results for such stresses.
9.1.6 Stresses
The above Equations (12), (16), (19)–(21) provide all the information needed to calculate the
moment carried by the assembly weld and the endplate. This moment can be converted into
nominal stresses using the classical equations given below [Timoshenko et al., 1970].
Nominal shear stress τ in the endplate:
τ = 1.87 W / (ct2). (22)
Nominal normal stress σ in the assembly weld:
σ = 32 W / (πd3), (23)
where d is the weld diameter.
To the nominal stresses given by the above equations, one would need to add the effects of
stress concentrations due to the rapid change in geometry near the end of the fuel element.
This topic is covered as part of fatigue strength in Section 9.2.
Fuel 45
9.1.7 Illustrative Example

As an illustrative example, let us calculate the nominal stresses in the endplate and in the
assembly weld when the outer fuel element of a CANDU-6 bundle undergoes lateral vibrations
25 micrometers in amplitude.
Assume that the fuel dimensions and properties are as given in Tables 1 and 2, except that the
on-power diametric clearance between the pellet and the sheath is zero. For the remaining
inputs, assume that the mean radius R of the endplate’s outer ring is 43 mm, the weld diameter
d is 4 mm, the pellet cracking factor ∝ is 0.1, and the rigidity enhancement factor G is 1.
Equations (20), (21), (19), (16), and (12) yield the following intermediate results: the twisting
moment of inertia C of the endplate is 5.2 mm4, the flexural rigidity J of the fuel element is 46
Nm2, the foundation modulus κ of the endplate is 25 kNm/(m•rad), the spring constant S is 118
Nm/rad, and the bending moment W is 0.017 Nm.
Equation (22) gives the nominal shear stress in the endplate as 2.5 MPa. Equation (23) gives the
nominal normal stress in the assembly weld as 2.7 MPa.
Because these stresses are due to vibration, they are alternating stresses. Their impacts on
fatigue are examined in the next section.
9.2 Fatigue of the Endplate and the Assembly Weld

O’Donnell and Langer recommended 55 MPa as the endurance limit for cold-worked and/or
welded Zircaloy-4 at 300°C [see O’Donnell et al., 1964]. This assumes that residual stresses are
equal to cyclic yield strength, accounts for damage due to irradiation, and includes a safety
factor of two to cover miscellaneous effects such as specimen size, environment, surface finish,
and data scatter.
The diameter of the fuel element endcap decreases sharply near the weld. Hence, stress
concentrations are to be expected in this region. Therefore, the endurance limit specified
above should be reduced by the effects of stress concentration. Peterson's ∂ concept can be
used for this purpose [see Peterson, 1943]. It states that in a region of sharp stress concentra-
tions, fatigue is best evaluated through stress at a distance ∂ below the surface. ∂ is a material
property.
Based on experimental data, O’Donnell and Langer estimated ∂=50 micrometers for Zircaloy-4.
Fatigue strength reduction for this delta can be determined using the results of O’Donnell and
Purdy [see O’Donnell et al., 1963]. A review of O’Donnell and Purdy’s results shows that for a
fairly wide range of notch radii (30 to 200 micrometers), the fatigue strength reduction factor K
is between 2.0 and 2.5 when the section width is 5 mm (close to the CANDU assembly weld).
For purposes of illustration, let us conservatively assume a fatigue strength reduction factor K of
2.5 for the assembly weld (i.e., the weld between the endcap and the endplate).
There are no major geometric stress raisers in an endplate that would bear significantly on
fatigue due to alternating shear stress near the assembly weld. Nevertheless, a small unde-
tected fabrication flaw (e.g., a crack) will act as a notch and reduce local fatigue strength. Such
a small crack may exist due to a flaw in the sheet metal used for endplate fabrication or may be
introduced during bundle assembly and subsequent handling. Using the curves of O’Donnell
and Purdy, the fatigue strength reduction factor for an endplate in the presence of such a small
crack can be conservatively estimated as 2. Therefore, the fatigue strength of the endplate is 28
MPa.
Exercise:
Consider the potential of fatigue from lateral vibrations of the fuel element assessed in Section
9.1.6.
Assembly weld: The alternating stress is 2 MPa. The fatigue strength is 22 MPa.
Endplate: For comparison to fatigue strength, the classical von Mises formulation for effective
stress tells us that shear stress needs to be multiplied by the square root of 3 to be comparable
to normal stress. Therefore, the effective alternating stress in the endplate is (2.5 * √3) = 4.4
MPa. In comparison, the fatigue strength is 28 MPa.
For both the assembly weld and the endplate, the alternating stresses are much lower than the
corresponding fatigue strengths. Therefore, the weld and the endplate are not at risk of fatigue
failures from this level of vibration.
10 Fuel Design Verification

During design, fabrication, and operation of nuclear fuel, a key objective is to preclude, as a
minimum, significant, chronic, systematic fuel failures. This needs to be confirmed up front for
credible combinations of the most demanding design, fabrication, and operating conditions that
can be expected in the reactor.
During design, likelihood of acceptable fuel performance can be assessed through a mixture of
three types of tools: engineering judgment, tests, and analyzes. Each tool has its pros and cons,
as explained later in this section, and therefore judicious combinations of them tailored to each
specific situation are often used.
Tests can be done on prototype fuels either inside or outside a reactor. In-reactor tests can be
done either in a test reactor such as the NRU and NRX reactors in Chalk River, Ontario, or in an
operating commercial power reactor. In-reactor tests automatically include the effect of neu-
trons, which out-of-reactor tests miss. However, out-of-reactor tests can include certain influ-
ences that are not available in current in-reactor experimental facilities, such as the effect of
gravity (current test loops are vertical) and the effect of the precise support provided to the fuel
string by shield plugs, sidestops, or latches (in current test loops, fuel is supported by a central
structural tube). In addition, out-of-reactor tests often tend to be significantly less expensive
and quicker than in-reactor tests. Judicious choices must therefore be made in each situation.
Tests have a significant advantage over analyzes in that they automatically include all processes
encompassed by the test conditions. Therefore, within the parameters of the experiment, there
is virtually no risk that an important process has been missed. Table 3 is an illustrative list of
tests done on some CANDU prototype fuels.
Fuel 47
Table 3 Illustrative list of tests on prototype fuel bundles
Test Purpose and Configuration
Bent-tube passage Bench test done at room temperature and atmospheric pressure to verify
that a prototype bundle, when not damaged, can pass through the “bent-
tube” gauge, which simulates the most restrictive passage in a fuel chan-
nel.
Bundle strength – Single-bundle test done in stagnant water at reactor pressure and tem-
shield plug sup- perature to verify that a prototype bundle, supported by a simulated shield
port plug, can withstand the maximum reactor load in the fuel channel under
normal conditions without failing. Test bundle failure is indicated when:
(1) the bundle fails the bent-tube gauge test and/or (2) shows permanent
deformation in any of its components that exceeds the as-built dimen-
sional tolerances for that component and/or (3) develops cracks in any of
its structural components or joints between components.
Single-bundle test done in stagnant water at reactor pressure and tem-

Bundle strength – perature to verify that a prototype bundle, supported (normally) by two
double side stop side stops, can withstand the maximum reactor load in the fuel channel
support during refuelling without failing. Test bundle failure is indicated when: (1)
the bundle fails the bent-tube gauge test and/or (2) shows permanent
deformation in any of its components that exceeds the as-built dimen-
sional tolerances for that component and/or (3) develops cracks in any of
its structural components or joints between components. Applies to C6-
type reactors.
Bundle strength – Single-bundle test done in stagnant water at reactor pressure and tem-
single side stop perature to verify that a prototype bundle, supported (abnormally) by a
support single side stop, can withstand the maximum reactor load in the fuel
channel during refuelling without failing. Test bundle failure is indicated
when the bundle fails the bent-tube gauge test. Applies to C6-type reac-
tors.
Refuelling Impact Out-of-reactor test done in a representative full-length fuel channel using
circulating light water at reactor operating temperature and pressure, to
verify that prototype bundles can withstand the maximum impact loads
expected in the fuel channel during refuelling without failing. Test bundle
failure is indicated: (1) when the bundle fails the bent-tube gauge test
and/or (2) shows permanent deformation in any of its components that
exceeds the as-built dimensional tolerances for that component and/or (3)
develops cracks in any of its structural components or joints between
components. Applies to C6-type reactors.
Endurance Out-of-reactor test, done inside a representative full-length fuel channel

containing a full complement of prototype fuel bundles, using circulating
light water at reactor operating temperature and pressure, in increments
of time until the spacer pads of the test bundles cease to exhibit additional
wear with increases in test time, to verify: (1) that spacer pad wear on test
bundles does not exceed its wear allowance; (2) that bearing pad wear on
test bundles, projected over the maximum in-reactor bundle residence
time, does not exceed its wear allowance; (3) that test bundles do not
show permanent deformation in any of their components that exceeds the
as-built dimensional tolerances for those components; (4) that test bun-
dles do not develop cracks in any of their structural components or joints
between components; and (5) test bundles do not cause channel damage,
projected over the lifetime of the channel, which exceeds the channel’s
lifetime material loss allowance, also taking into account possible channel
damage by other means.
Sliding wear Single prototype bundle test done in stagnant light water at reactor
operating temperature and pressure to verify that: (1) bearing pad wear in
the test bundle does not exceed the bearing pad wear allowance and (2)
projected lifetime pressure-tube damage, caused by sliding bearing pads of
the test bundle, does not exceed the channel’s lifetime material loss
allowance, also taking into account possible channel damage by other
means.
Pressure drop Out-of-reactor test done in a representative full-length fuel channel,

light water at reactor operating temperature and pressure, to verify that
the hydraulic pressure drop generated in the portion of the fuel channel
that contains fuel bundles does not exceed the pressure drop allowance
for the same portion of the channel.
Bundle rotation Out-of-reactor test done in a representative full-length fuel channel and/or
pressure drop a partial-length channel, using circulating light water at reactor operating
temperature and pressure and/or at room temperature and low pressure,
to determine the variation in hydraulic pressure drop generated across the
endplate-to-endplate junction between two abutting prototype bundles as
a function of their relative radial (mis)alignment.
Fuelling machine Out-of-reactor test done in a representative full-length fuel channel, using
compatibility circulating light water at reactor operating temperature and pressure, in
conjunction with a production fuelling machine, to verify that the fuelling
machine can charge and discharge prototype fuel bundles into the chan-
nel.
Wash-out Out-of-reactor test done in a representative full-length (or partial-length)

Fuel 49

fuel channel, using circulating light water at reactor operating temperature
and pressure, in conjunction with apparatus that simulates the relevant
functions of the fuelling machine, to determine what malfunctions in use
of carrier tubes might cause “wash-out” of the type observed in some
power reactors. Applies to CANDU rectors that use carrier tubes during
refuelling.
Frequency sweep Out-of-reactor test done in a representative full-length fuel channel

light water at reactor operating temperature and pressure and a pressure
pulse generator, to: (1) determine the frequency of pressure pulsations at
which the column of test bundles, supported normally inside the channel,
may resonate, and (2) verify that the resonant frequency of the column of
test bundles is not the same (or nearly the same) as the coolant pump
vane passing frequency.
Critical heat flux Out-of-reactor test done in a representative full-length fuel channel using
circulating light water at reactor operating temperature and pressure and
electrical heaters to simulate the geometry and power produced by fuel
bundles, to determine the spectrum of channel thermal-hydraulic and
power conditions in the range of interest to reactor operation, which lead
to a step change in heater surface temperature, referred to as “dryout”.
High-power Irradiation test done on prototype bundles cooled by circulating water at

irradiation reactor operating temperature and pressure to verify that a bundle irradi-
ated at power/burnup conditions that encompass the power/burnup
conditions of all bundles in the power reactor will be defect-free at the end
of the irradiation period. Such a test can be done, for example, in a
vertical channel of the NRU test reactor, called a “hot loop”, in which test
bundles are cooled by light water; enriched uranium would need to be
used to generate the high element ratings expected in a commercial
CANDU power reactor.
Power ramp Irradiation test done on prototype bundles cooled by circulating water at
irradiation reactor operating temperature and pressure, to verify that a test bundle
can survive, defect-free, the most severe power ramps expected in a
reactor. Typically, this would involve: (1) irradiation of a test bundle at
relatively low power for some fraction of its residence in the reactor, and
then (2) an increase in power for the remainder of its residence in the
reactor, such that the burnup accumulated at the initial power and the
increased power surpasses the equivalent conditions of all bundles in the
power reactor. Using C6 as an illustrative example, the low-power irradia-
tion would reflect power and burnup in positions 1 to 4 in a high-power
channel, and the size of the ramp would reflect, in an appropriately
conservative manner, the power ramps in bundles 1 to 4 during or after

their shift to positions 9 to 12. Such tests can be performed in, for exam-
ple, the vertical hot loops at NRU in CRL in which test bundles are cooled
by light water; enriched uranium pellets would be required to achieve the
high element powers expected in a commercial CANDU reactor.
Irradiation of Irradiation test done on prototype elements which contain the novel
special materials material (e.g., an inert matrix for burning of actinides) cooled by circulating
water at reactor operating temperature and pressure, to verify that the
elements containing the special material will be defect-free at the end of
their irradiation period. Such a test can be done, for example, in “de-
mountable” bundles in a vertical channel of the NRU test reactor called a
“hot loop”, in which test specimens are cooled by light water.
At the same time, fully representative tests are not always practical in some situations, for
example because of equipment, cost, schedule, or combinations thereof. As specific illustrative
examples, let us consider the following:
 Comprehensiveness: As one illustrative example, a CANDU fuel designer needs to con-
firm, among other aspects, that a fuel bundle has adequate mechanical strength to resist
the mechanical loads imposed on it during discharge from the fuel channel. The test fa-
cility, to verify this experimentally, would need to provide for all the important effects in
this scenario, meaning that (a) the test facility should be shielded so that it can accom-
modate irradiated fuel so that reduction in ductility due to prior irradiation can be ac-
counted for; (b) it should be long enough to accommodate a string of approximately 12
fuel bundles; (c) it should produce power to simulate the effect of on-power interaction
of the pellets and the sheath on load shedding; (d) it should be connected to a pump of
sufficiently large capacity to provide the necessary flow for the hydraulic drag load; and
(e) it should contain a representative fuel support configuration to simulate the appro-
priate load concentrations. We currently do not have a shielded facility that meets all
these requirements, and building a new one would be very expensive and time-
consuming. For this reason, such tests have been done in the past in unshielded out-of-
reactor facilities using non-irradiated fuel bundles. Although such tests are useful, they
do miss the order-of-magnitude drop in ductility of the fuel bundle’s structural materials
that occurs during irradiation, and hence they may not always determine the actual
adequacy of an irradiated fuel bundle’s mechanical strength. Another similar situation
that is also not covered by existing test facilities is the effect of gravity on long-term irra-
diation-enhanced creep sag of fuel in a horizontal orientation. These situations can be
rectified through analyzes.
 Scatter: As another illustrative example, consider the need to verify whether or not a
new fuel design meets the expected power ramp challenges in a new operating enve-
lope for which the EAC defect threshold is not well known ahead of time (a priori). For a
new operating envelope and fuel design, one would not know a priori which specific
ramp is the most demanding. Therefore, one would need to test many ramps, and to
cover the expected scatter in the results, one would need to test a statistically significant
number for each condition. This would require testing many bundles for a large number
Fuel 51
of ramps, and for each ramp, significant in-reactor irradiation would be required to ac-
cumulate the required burnup. Because only limited facilities for in-reactor irradiation
are available, the full test matrix could require large amounts of time and expense.
Again, analyzes can help reduce the cost as well as the timeline.
 Margins: Knowledge of available margins is needed for a variety of purposes, e.g., to es-
tablish a safe operating envelope or to deal quickly with abnormal situations. To deter-
mine the margin available in a design solely from tests, one would need to perform a
large number of experiments. Again, time and expense for these can be reduced
through analyzes.
For these conflicting reasons, some situations are better addressed through tests, whereas
others require a judiciously balanced combination of tests and analyzes. Appropriate combina-
tions of tests and analyzes must be tailored to a variety of factors such as (a) features of the
specific change in design, (b) completeness and level of knowledge of relevant damage mecha-
nisms, and (c) specific features, capabilities, and capacities of available test facilities and analyti-
cal tools. For example, different combinations of tests and analyzes have been used to verify
the CANFLEX fuel design [Hastings et al., 1989], the ACR fuel design [Reid et al., 2008], and the
37M fuel design [Daniels et al., 2008]. In the following sections, we describe a fuel design
verification process that combines tests and analysis.
To craft an optimal program of comprehensive design analyzes, a sound knowledge is needed of
credible damage mechanisms during the conditions for which fuel integrity needs to be ascer-
tained. A few selected damage mechanisms have been described in detail in earlier sections of
this chapter. An overview compilation of all currently known credible and significant damage
mechanisms is given in the next subsection.
10.1 Damage Mechanisms

Most discussions about fuel damage revolve around two types of damage: primary and secon-
dary. A primary damage mechanism is one that creates the initial hole or break in Zircaloy or is
an important precursor to it. Secondary damage—usually from zirconium hydrides—is a
consequence of primary damage. This section focusses on mechanisms for primary damage.
Secondary damage is discussed in Section 12.1.
Based on first principles and confirmed by a comprehensive search of the available literature
[Sun et al., 2010], 18 primary damage mechanisms have been identified for CANDU fuel [Tayal
et al., 2008-1], of which only three are described in detail in the preceding sections for reasons
of space. All eighteen are, however, listed in the following sections. For simplicity in the follow-
ing discussion, the primary damage mechanisms are classified into three broad groups: thermal
damage, structural damage, and compatibility considerations [Tayal et al., 2008-1] and are
described below in turn.
10.1.1 Thermal Damage (“T” Series)

Four thermally driven mechanisms can potentially damage the fuel, as described below.
Pellet melting
If the pellet melts, the resulting volumetric expansion of the pellet may potentially push the
sheath past its breaking point. Alternatively, molten UO2 may potentially flow into contact with
Zircaloy and melt it by its contained heat.
Zircaloy-UO2 is another system with a low-melting-point alloy. Hence, strictly speaking, UO2
does not have to heat the Zircaloy all the way to its melting point to form a liquid alloy.
Melting of Zircaloy or its alloys
Insufficient cooling may potentially melt the Zircaloy, impeding its ability to contain fission
products.
The fuel sheath also contains Zircaloy/beryllium eutectic and other alloys in that neighbour-
hood. Their melting points are lower than that of the parent Zircaloy, and therefore insufficient
cooling may potentially melt these alloys. This could detach the pads from the sheaths, destroy-
ing their functionality.
Crevice corrosion
Crevice corrosion can potentially occur in crevices such as between bearing pads and the
pressure tube or between pads and the sheath. The crevices restrict the flow of coolant.
Coolant can boil off in the near-stagnant conditions in the crevice, which in turn can increase
the concentration of chemicals such as lithium in the crevice. Lithium, in the form of lithium
hydroxide, is added to the coolant to control its pH. The resulting elevated concentration of
lithium can potentially accelerate Zircaloy corrosion.
Overheating by contact
If a heated surface such as the sheath contacts a neighbouring surface such as another sheath
or the pressure tube, there is potential for reduced local cooling and hence overheating of the
fuel, the pressure tube, or both.
10.1.2 Structural Damage (“S” Series)

Structural damage can potentially be caused through nine mechanisms, as described below.
Internal overpressure
In conjunction with a corrosive internal environment and local hydrides, excessive internal
pressure in a fuel element can in principle cause cracks at two locations of stress concentration
for primary stresses: at the sheath/endcap junction, and at the junction of the sheath and the
pad. In practice, to date such cracks have been observed primarily at the sheath/endcap
junction.
Environmentally assisted cracking during power ramps
During power ramps, irradiation-embrittled Zircaloy can experience high stresses and strains in
the presence of a corrosive internal environment and local hydrides. This combination can
potentially crack the sheath at locations of concentration for secondary stresses and strains—at
circumferential ridges and at sheath/endcap junctions—through environmentally assisted
cracking.
Static mechanical overstrain
During a variety of situations such as refuelling, structural components of the fuel bundle can
potentially be exposed to relatively high loads, relatively sparse support, or both, leading to a
potential for static mechanical overstrain.
Fuel 53
Uncontrolled loss of geometry

This term refers to situations that can potentially lead to buckling, which in turn can lead to
uncontrolled loss of geometry.
Fatigue
Alternating stresses caused by vibration (e.g., induced by flow, turbulence, or both), power
manoeuvring, and load following can expose the fuel to potential failure through fatigue.
Fuel mechanical rupture due to impact
Situations such as refuelling or starting and restarting of coolant pumps can potentially impose
significant impact loads. This is more damaging if it occurs after some amount of irradiation
embrittlement of Zircaloy.
Primary hydride failure
A fuel element can pick up hydrogen or deuterium from a variety of sources. The hydrogen and
deuterium tend to concentrate at relatively cooler locations and/or at locations of relatively
higher stress. Excessive local hydrides and deuterides can reduce the local ductility of Zircaloy,
rendering it less capable of carrying its load.
Oxide spalling and hydride lens formation
Some amount of corrosion is unavoidable in Zircaloy at the high temperatures and chemistry of
the coolant. If the oxide surface is too thick, it can spall away. This can create local temperature
gradients, which can form local hydride/deuteride deposits called “hydride lenses”. Loss of crud
from a uniform layer of crud would have a similar effect. A hydride lens embrittles Zircaloy
locally, reducing its capacity to carry load.
Insufficient ductility during post-irradiation handling
To enable post-irradiation handling of fuel, it is prudent to limit the amount of hydrides and
deuterides in irradiated Zircaloy so that it has reasonable residual ductility to resist its antici-
pated loads.
10.1.3 Compatibility Considerations (“C” Series)

Five mechanisms can cause potential incompatibilities, as described below.
Interaction loads on the fuel string
This mode of damage refers to the potential expansion of the fuel string to become longer than
the available length of the cavity in the fuel channel.
Dimensional incompatibility of a fuel bundle
This mode of damage refers to the possibility of the in-service on-power dimensions of the fuel
bundle becoming larger than the available space through processes such as creep and thermal
expansion. This can potentially generate unanticipated large stresses that may damage the fuel.
Spacer pad fretting
Excessive fretting of spacer pads due to vibration can potentially rub the corner of a spacer pad
into the adjacent sheath, potentially creating a hole in the sheath, or can reduce the gap be-
tween elements below the design minimum required for heat removal by the coolant.
Jamming
This mode refers to the possibility that in-service on-power dimensions of the fuel bundle may
become such that the fuel bundle cannot pass—easily or at all—through the fuel channel. For
example, this can potentially occur due to interlocking of multiple spacers or due to excessive
distortion.
Damage to pressure tubes
Bearing pads can cause fretting, sliding wear, and crevice corrosion in pressure tubes.
10.2 Acceptance Criteria and Integration

Acceptance criteria are needed to determine whether or not damage from the above mecha-
nisms is within acceptable limits. Table 4 lists a set of acceptance criteria [Tayal et al., 2008-1].
Table 4 Fuel design acceptance criteria for normal operating conditions

[Source: Tayal et al., 2008-1]
Damage Consideration Acceptance Criterion
“T Series”: Thermal Integrity
Fuel element failure due to T1: Local temperature in all parts of the pellet shall stay below the
fuel melting melting point of the pellet, with a minimum acceptable margin.
Fuel element failure due to T2: Local temperature in all parts of the sheath and the endcap
sheath melting shall stay below the local melting point of the material, with a
minimum acceptable margin.
Fuel element failure due to T3: Underneath a bearing pad or spacer pad, the temperature at
crevice corrosion the sheath outer surface shall be less than that required to cause
crevice corrosion of the sheath, with a minimum acceptable
margin.
Fuel or pressure tube T4: Fuel bundle dimensional changes (e.g., due to irradiation,
failure due to overheating loads, creep, bowing, etc.) shall maintain a minimum acceptable
by contact clearance between neighbouring sheaths or endcaps and also
between the pressure tube and its sheath/endcap.
“S Series”: Structural Integrity
Fuel sheath failure due to S1: The excess of internal pressure over coolant pressure shall be
less than the pressure that causes cracking in the fuel sheath or in
Fuel 55

overpressure the sheath/endcap junction, with a minimum acceptable margin.
Fuel sheath failure due to S2: Stresses and strains (or related powers and ramps) during
environmentally assisted power increases in fuel elements at circumferential ridges and at
cracking due to power sheath/endcap junctions shall be less than the appropriate defect
ramps thresholds (with a minimum acceptable margin), including the
effects of pellet chips, if any.
Fuel failure due to static S3: Local principal strain (elastic plus plastic) shall be less than the
mechanical overstrain available ductility minus the minimum acceptable margin, and
local creep strain shall be less than the creep rupture strain, with
a minimum acceptable margin.
Fuel failure due to uncon- S4: Axial and related loads on the fuel bundle shall be less than
trolled loss of geometry the bundle buckling strength, with a minimum acceptable margin.
Fuel failure due to fatigue S5: Cumulative fatigue damage from repeated cycles of alternat-
ing stresses and strains shall be less than the allowable fatigue
life, with minimum acceptable factors of safety on the magnitude
of cyclic strain and on the number of cycles.
Fuel mechanical rupture S6: Strain energy density during impact shall be less than that
due to impact of loads such required to crack or break any metallic component of the fuel
as refuelling and/or bundle, with a minimum acceptable margin.
start/restart
Primary hydride failures S7: Equivalent concentration of internal hydrogen of an as-

fabricated fuel element, from all sources excluding the sheath,
shall not exceed the minimum acceptable limit.
Formation of a local hy- S8: The combined thickness of oxide and crud on the fuel sheath
dride lens due to oxide and outer surface shall be less than the amount required for spalling
crud from the surface, with a minimum acceptable margin.
Failure due to insufficient S9: The volume-average concentration of hydrogen (in the form of
ductility during post- soluble atomic hydrogen and equivalent hydrides and deuterides,
irradiation handling including their orientation) over the cross section of load-bearing
components shall be less than the amount required to retain
sufficient ductility, with a minimum acceptable margin.
“C” Series: Compatibility
Failure due to excessive C1: Maximum length of the fuel string (e.g., in the fuel channel)
interaction loads along the shall be less than the minimum available cavity (e.g., between the

fuel string shield plug or latches), with a minimum acceptable margin.
Failure from damage by C2: Net dimensions, including dimensional changes, throughout
interfacing equipment fuel bundle residence in the reactor shall be within specified
limits for interfacing equipment.
Fuel sheath failure due to C3: At spacer pads, total wear from all sources such as lateral
fretting of pads vibrations, axial vibrations, fretting, sliding, and erosion shall be
less than that which brings any part of a spacer in contact with a
neighbouring sheath, with a minimum acceptable margin.
Fuel bundle jamming C4: To enable passage of fuel through the reactor in all fuel
handling operations, the axial force required to move the bundle
shall be within design allowance, including all pertinent consid-
erations such as on-power deformations, in-service contacts with
neighbouring components, and changes in material properties.
Protection of pressure tube C5: Depth of crevice corrosion, sliding wear, and fretting wear in
from bearing pads the pressure tube from fuel bearing pads shall be within specified
allowances.
Many complex processes occur in a fuel element. Therefore, in real life, one seldom has com-
plete knowledge of all credible combinations of important variables involving design, manufac-
turing, and operating conditions that might fail the fuel, nor the luxury of performing an essen-
tially infinite number of experiments that would cover them all. Therefore, in real life, we must
deal satisfactorily, not only with what is known, but also with what is not.
In some situations, we know what we don’t know—the known unknowns. In other situations,
we don’t even know what we don’t know—the unknown unknowns. “Known unknowns” are
usually included in assessments through quantified uncertainties. However, “unknown un-
knowns” are usually addressed through margins contained within acceptance criteria.
Figure 29 illustrates the concept of margin within an acceptance criterion; the figure labels it as
"Minimum Acceptable Margin". Such a margin reflects expert judgment about the “unknown
unknowns” in a given technology area. Therefore it can change with time as additional knowl-
edge becomes available and/or as improved technologies for design, production, and/or as-
sessments are developed and incorporated into the product.
Fuel 57
[After Sun et al., 2009]

Figure 29 Margins in criteria and analyzes
As noted earlier, a variety of assessment techniques are used to evaluate a product, including
engineering judgment, tests, and analyzes. Their combinations are used through a large set of
assessments that can potentially be performed as needed to confirm that acceptance criteria
have been met. Some illustrative tests for prototype CANDU fuel, shown in Table 3, have
already been mentioned.
Most fuel analyzes tend to be highly non-linear and multi-disciplinary. In addition, several
analyzes require strong considerations of significant feedbacks among many complex processes.
For this reason, the nuclear fuel industry frequently uses a number of computer codes to
perform analyzes. An illustrative list is given in Table 5.
Table 5 Illustrative list of computer codes used at AECL for fuel analyzes
Program Description
ABAQUS This finite-element code is used for structural analyzes of fuel strings, espe-
cially in the presence of non-linearities of intermediate level. Typical examples
include: (a) determination of stresses in the fuel string resting on side stops
and experiencing hydraulic drag load, and (b) impact stresses during loading of
fresh fuel bundles. For an overview and further details of this code, contact
Dassault Systèmes, France.
ASSERT This sub-channel code is used to assess thermal-hydraulic conditions in the fuel
channel, including the effects of heat transfer between the fuel and the cool-
ant. Key results include distributions of flow and voids, pressure drop, and
critical heat flux. For an overview and further details of this code, contact
Atomic Energy of Canada Limited, Canada.
BEAM This code provides a fast and simple tool to assess several aspects of the
mechanical behaviour of a fuel element. Key results include axial and lateral
stiffnesses of a fuel element, frequencies of lateral vibrations and resulting
elastic stresses, collapse pressure of a sheath, and buckling strength of a fuel
element. An overview is available from M. Tayal et al., “Assessing the Me-
chanical Performance of a Fuel Bundle: BEAM Code Description”, Proceedings,
Third International Conference on CANDU Fuel, Canadian Nuclear Society,
Pembroke, ON, 1992, October 4-8.
BOW This finite-element code is used to assess the lateral deformations of fuel
elements and bundles, e.g., bow, sag, and droop. An overview is available from
M. Tayal, “Modelling the Bending/Bowing of Composite Beams such as Nuclear
Fuel: the BOW Code”, Nuclear Engineering and Design, vol. 116, pp. 149-159
(1989).
ELESTRES This finite-element code is used to calculate several thermal, mechanical, and
microstructural responses of a fuel element. Key results include sheath and
pellet temperatures; fission gas release and internal pressure; and sheath
strain as a result of expansion, contraction, and hourglassing of pellets and the
sheath. An overview is available from M. Tayal, “Modelling CANDU Fuel under
Normal Operating Conditions: ELESTRES Code Description”, Report CANDEV-86-
110, published by CANDU Owners’ Group; also Report AECL-9331, published by
Atomic Energy of Canada Limited, February 1987.
FEAST This finite-element code is used to model detailed local stresses, strains, and
deformations such as at circumferential ridges, at sheath/endcap junctions,
and in a sheath above an axial gap. An overview is available from M. Tayal,
“FEAST: A Two-Dimensional Non-Linear Finite-Element Code for Calculating
Stresses”, Proceedings, Seventh Annual Conference of the Canadian Nuclear
Society, Toronto, Ontario, Canada, June 8–11, 1986.
Fuel 59
Program Description
FEAT This finite-element code is used to model local temperatures such as the
effects of end-flux peaking and local temperatures pertinent to crevice corro-
sion and braze voids. An overview is available from M. Tayal, “The FEAT Finite-
Element Code to Calculate Temperatures in Solids of Arbitrary Shapes”, Nuclear
Engineering and Design, Vol. 114, pp. 99-114 (1989).
FEED This finite-element code is used to model diffusion of hydrogen/deuterium in

Zircaloy. Key results include concentrations of hydrogen at various locations.
An overview is available from L. Lai et al., “A Method to Model Hydrogen
Precipitation”, Proceedings, 11th International Conference on CANDU Fuel,
Canadian Nuclear Society, Niagara Falls, 2010.
INTEGRITY This model uses a semi-mechanistic approach to improve calculations of

environmentally assisted cracking due to power ramps by accounting for a
variety of effects that are not included in the current database of available
empirical correlations, such as varying pellet density, sheath diameter, coolant
pressure, and coolant temperature. An overview is available from M. Tayal et
al., “INTEGRITY: A Semi-Mechanistic Model for Stress Corrosion Cracking of
Fuel”, IAEA Technical Committee Meeting on Water Reactor Fuel Element
Modelling at High Burnup and its Experimental Support, Windermere, United
Kingdom, AECL-10792, September 19–23, 1994.
LONGER This code calculates conditions related to sheath collapse. Key results include
pressure for elastic instability, critical pressure for excessive longitudinal ridges,
and pressure for axial collapse. An overview is available from U. K. Paul et al.,
“LONGER: A Computer Program for Longitudinal Ridging and Axial Collapse
Assessment of CANDU Fuel”, Proceedings, 11th International Conference on
CANDU Fuel, Canadian Nuclear Society, Niagara Falls, Ontario, 2010.
LS DYNA This finite-element code overlaps the functionality of ABAQUS code and can
also be used for analyzes of strength, impact, and vibration of fuel strings. For
an overview and further details of this code, contact Livermore Software
Technology Corporation, U.S.A.
NUCIRC For purposes of fuel analyzes, this thermal-hydraulic code is used to calculate
critical channel power. For an overview and further details of this code,
contact Atomic Energy of Canada Limited, Canada.
11 Operating Constraints and Inputs to the Safe Operating Envelope

For a specific project application, one would perform assessments (tests and analyzes) as
mentioned above to ascertain that the chosen fuel design meets the specified design require-
ments without failing (i.e., by showing that the fuel acceptance criteria are met). The assess-
ments would at the same time establish the limits, or constraints, on certain plant parameters
such that the specified fuel acceptance criteria are met. These plant operating constraints, as
derived from fuel assessments, and similar constraints derived from assessments of all other
systems are then used in safety analysis to establish the plant’s safe operating envelope.
12 Removal of Fuel Bundles Containing Defects

CANDU reactors have an enviable record of very low fuel defect rates, as noted in Section 2.3.
Nevertheless, defects in fuel bundles do occasionally occur. CANDU reactors also have a unique
ability to detect, locate, and discharge fuel bundles which contain defective elements without
shutting the reactor down. This capability is described in this section. However, we first give a
brief summary of how a small primary defect in a fuel element can grow with time into a bigger
hole that releases progressively larger amount of radioactivity to the coolant.
Fundamental aspects of such deterioration have been discussed by Lewis et al. [1988]. Section
12.1 below is based largely on a document published by the CANDU Owners’ Group (COG).
Sections 12.2 and 12.3 below are based on an IAEA document [IAEA NF-T-2.1, 2010].
12.1 Deterioration of Primary Defects in Fuel Elements

As noted in Section 10.1 of this Chapter, there are eighteen mechanisms that can cause a failure
(primary defect) in CANDU fuel, some in fuel bundles and some in fuel elements. In this sec-
tion, we consider only one form of defect, albeit the most important, that of a crack or a hole
developing in a fuel element, to illustrate how the defective fuel detection and removal system
functions in a CANDU reactor. Detection and removal of fuel bundles that have experienced
other types of defects would follow a similar procedure, although some fuel bundle defects
(such as a bundle becoming stuck in a fuel channel) may require different or additional steps.
Please see Section 10.1 for a description of primary defects and how they are caused.
If a crack or hole in the sheath of a fuel element (a primary defect) permits coolant water to
penetrate the fuel element, the hole can grow progressively bigger with time through a process
called “post-defect deterioration”, resulting in release of progressively larger amounts of radio-
active fission products into the coolant. This can eventually pose a health hazard to station
staff. Therefore, fuel bundles that contain defective fuel elements must be discharged from the
reactor before the crack or hole in the fuel element becomes too big.
As explained below, post-defect deterioration of this type of primary defect in fuel elements
occurs due to a combination of three main factors:
 secondary hydriding of the sheath (or “deuteriding” when heavy water is used as a cool-
ant, as it is in CANDU; “hydriding” will be used hereafter in this section for simplicity);
 stresses in the sheath; and
 oxidation and swelling of UO2.
Secondary Hydriding: Secondary hydriding refers to Zircaloy-hydrogen compounds that are
formed in the sheath as a consequence of a primary defect. Water enters the fuel element
through the primary crack or hole. At the higher temperature inside the fuel element, the
water turns into steam and can better react chemically with Zircaloy and with the UO2 pellet.
The chemical reaction oxidizes some of the Zircaloy and turns some of the UO2 into higher
oxide(s). This chemical reaction also releases hydrogen radicals from steam. Radiolytic decom-
position also releases hydrogen radicals from water. The free hydrogen enters the metal matrix
of the Zircaloy sheath and migrates by diffusion to areas of higher stress and lower tempera-
Fuel 61
ture. Therefore, the hydrogen often accumulates in localized areas such as at the relatively
cooler outer surface of the sheath. When the local concentration of hydrogen exceeds the local
solubility limit, the excess hydrogen precipitates as hydrides, which are called secondary hy-
drides. Once hydrides have formed, they can grow by preferentially absorbing additional
hydrogen from within the pellet/sheath gap. High levels of hydrides embrittle Zircaloy, leaving it
less able to resist stresses and strains.
A number of pertinent factors may vary along the length of the fuel element, for example the
local temperature, the local rate of decomposition, and the local oxidation kinetics of UO2 and
of Zircaloy. For this reason, secondary hydriding generally occurs at some distance away from
the primary hole.
Sheath Stresses: The volume of compound hydride is about 16% larger than the volume of the
parent Zircaloy. This creates geometric incompatibility among neighbouring materials, which in
turn leads to large stresses in the hydride and in the Zircaloy surrounding the hydride. When
these stresses become high enough, they can cause cracks in the layer of zirconium oxide which
exists on the inside surface of the sheath. This in turn enables additional hydrogen to enter the
sheath. Moreover, the hydrides can produce cracks that pass—either fully or partly—through
the sheath wall. All cracks enable further ingress of hydrogen; through-wall cracks also enable
more water to enter the fuel element.
Oxidation and Swelling of UO2: Reaction of water and UO2 results in higher concentrations of
uranium oxides, commonly labelled UO(2+x). The higher oxides have lower density, and there-
fore lower thermal conductivity and higher nominal volume, than UO2. The lower thermal
conductivity increases the temperature and hence the thermal expansion of the pellet, which
combines with the higher nominal volume to swell the pellet. The expanding pellet then in-
creases the width of the crack in the Zircaloy sheath.
These processes can lead to a hole, often caused by a loss of material from the outside surface
of the sheath, called a “blister”. Figure 30(a) provides an illustrative example. Holes due to
secondary damage are often much bigger than holes from primary damage and hence can
release significant radioactivity, and even some UO2, into the coolant. For this reason, it is
prudent to discharge a fuel bundle that contains a failed fuel element before significant secon-
dary damage to the sheath occurs. Fortunately, the process of forming secondary defects often
takes some time, and CANDU reactors are unique in being able to: (1) detect fuel element
defects that release radioactive fission products into the coolant, (2) identify the fuel channels
that contain bundles with defective fuel elements, (3) discharge these bundles before the
defects in the fuel-element sheath become too big, and (4) do all this without shutting the
reactor down. The next section summarizes the techniques to do this.
[Courtesy: COG] [Source: IAEA, 2010]

Figure 30 Secondary hydriding
12.2 Defect Removal Systems

Three systems have been developed to detect and locate defective fuel elements that release
radioactive fission products into the coolant: (1) gaseous fission product (GFP), (2) delayed
neutron (DN), and (3) feeder scanning (FS) systems.
The GFP system monitors continuously when the reactor is at power and detects a general
increase of radioactivity in the coolant when a defect occurs in a fuel element anywhere in the
core. This system alerts the operator that a defect has occurred and identifies the coolant loop
in which the defect is located, thereby eliminating 50% of fuel channels from the “suspects” list.
The DN system is also used when the reactor is at power, but does not monitor continuously.
The activity-sensing portion of the DN system is put into operation after a defect signal is
indicated by the GFP system. The DN system is used to identify the fuel channel that contains
the defective elements by detecting activity in the coolant at the downstream end of a fuel
channel [Manzer et al., 1984]. In a closed-loop configuration, sampling lines attached to outlet
feeders continuously carry small streams of coolant to a common room within the reactor
building (see Figure 31) and return the streams back into the coolant circuit downstream of the
outlet header. In the sampling room, BF3 sensors detect delayed neutrons emitted by short-
lived fission-products (137I and 87Br). The lengths of the sampling lines (custom-sized for each
channel) are such that the delivery of the sample to the sensor takes advantage of the decay
half-life of the delayed-neutron precursors. If neutron activity in a coolant sampling line is
higher than normal, the corresponding fuel channel is suspected of containing a fuel bundle
with a defective element. The DN system has sufficient sensitivity to locate fuel element
defects with very small holes. Bruce and CANDU-6 reactors use the DN system.
Fuel 63
[Source: IAEA, 2010]

Figure 31 Schematic diagram of a delayed neutron monitoring system in a CANDU-6 reactor
The FS system is used when a reactor is shut down. It locates the fuel channel that contains the
defective element by scanning for activity on the inside surfaces of the outlet feeders connect-
ing the fuel channels to a common outlet header [Lipsett et al., 1976]. The presence of gamma-
emitting fission products inside the channel is detected by Geiger-Muller detectors moving
within guide tubes that traverse the outlet feeders. If gamma activity in a specific feeder is
higher than normal, then the corresponding fuel channel is suspected of containing a fuel
bundle that has a (severely) defective element. The FS system locates only defects that have
deteriorated to the point of releasing uranium and fission products which deposit immediately
downstream of the defect. The Darlington reactors use the FS system.
12.3 Confirmation of Defect Removal

After the channel which has been identified as containing the defect has been refueled in a
CANDU reactor, the defective fuel element is confirmed to have been discharged by various
methods, again depending on the station:
 Inspecting the discharged bundles in the bay;
 Monitoring gamma activity near the spent fuel handling system when bundles are en
route from the reactor to the fuel bay (wet or dry sipping) and/or when they are residing
in the fuel bay (wet sipping);
 Monitoring delayed neutron activity of the coolant at the outlet end of the channel dur-
ing refuelling.
The first method provides direct confirmation that a bundle that contains a defective fuel
element has been removed. Portable underwater TV cameras are used at multi-unit stations
operated by Ontario Power Generation, whereas periscopes are used at single-unit CANDU-6
stations [Lipsett, 1976]. Photographs of the defective elements confirm that a defective bundle
has been removed.
The second and third methods provide indirect confirmation that a bundle containing a defec-
tive element has been removed by monitoring gamma activity near the spent fuel handling
system or in fuel bays. At Bruce, dry sipping techniques [MacDonald et al., 1990] are used to
monitor airborne gamma activity in the air chamber that separates the heavy water side of the
spent fuel transfer mechanisms from the light water environment of the fuel bay. A higher-
than-normal signal that lingers after bundles have been transferred usually indicates the pres-
ence of a defective element.
Two other techniques have also been developed at CANDU-6 stations to confirm that fuel
bundles containing defective elements have been discharged from the core. One technique
depends on the radiation levels of fission products in the heavy water inside the fuelling ma-
chine. Before discharge of irradiated fuel bundles to a bay, heavy water from the fuelling
machine is transferred to a nearby drain tank. The presence of a defect is indicated when
gamma fields near the tank trigger an alarm that monitors gamma in the area. Another tech-
nique developed in the inspection bay at Point Lepreau is based on “wet sipping”, or measuring
the gamma activity of water samples near recently discharged bundles. Again, a defect is
confirmed if gamma activity is unusually high.
Monitoring delayed neutron activity at the outlet end of a fuel channel during refuelling also
provides some confirmation that a fuel bundle containing a defective element is being dis-
charged from the core [Manzer, 1985]. At CANDU-6 sites, special refuelling procedures are
sometimes used which involve allowing the fuel bundles inside the channel to move slowly with
the flow while monitoring the DN signal of the channel. When the signal drops to below “pre-
defect” levels, this indicates that the fuel bundle containing the defect has been pushed outside
the core boundary.
13 Closure
To restate a point made in Section 2.3, the defect rate is remarkably low in CANDU fuel—almost
in the range of impurities found in most substances. An IAEA survey has reported the following
world-wide fuel defect rates between 1994 and 2006 [IAEA, 2010]:
 94 failed rods per million discharged (ppm) in WWERs;
 87 ppm in PWRs;
 65 ppm in BWRs; and
 3.5 ppm in Canada.
These figures affirm the basic soundness of practices built into all major aspects of CANDU
fuel—from research to development, design, fabrication, and operation and the feedbacks
among them.
Fuel 65
14 Problems
Section 2
Q2.1 – Why is a CANDU fuel “element” able to produce 30 times more heat than an electric
element of a similar size?
Q2.2 – List at least six conditions that occur during at-power operation of CANDU fuel that
challenge the integrity of the fuel.
Q2.3 – Name the components of a typical CANDU fuel bundle.
Q2.4 – Why is the uranium that is used in CANDU fuel bundles in the form of uranium dioxide,
rather than uranium metal, for instance?
Q2.5 – Why is helium used to purge the air out of fuel elements during fabrication?
Q2.6 – For a fixed bundle power, the only means that the fuel designer has at his disposal to
change the individual element rating is to use more or fewer elements. Using more elements
(subdivision) permits the coolant to be nearer the source of the heat at more locations, leading
to lower UO2 temperature and reduced threats from many thermally driven damage mecha-
nisms. What are some of the negative effects of subdivision?
Q2.7 – Experience and fundamental engineering considerations have shown that CANDU fuel
bundles can experience 18 different modes of failure in a reactor. Name the three general
groups into which the modes are grouped and provide the principal characteristic of each
group.
Section 3
Q3.1 – What does “burnup” measure, and what units is it expressed in?
Q3.2 – A CANDU fuel element contains UO2 pellets 20 mm in diameter and having a density of
10.5 g/cm3. At 60 kW/m, how long does it need to accumulate a burnup of 40 MWh/kgU? Why
does this need so much longer than the fuel element of Section 3?
Section 4
Q4.1 – A fuel element has the following characteristics: diameter = 15 mm; sheath thickness =
0.5 mm; Young’s modulus = 80 GPa; Poisson’s ratio = 0.37. Is it elastically stable at 6 MPa? At 7
MPa? At 8 MPa?
Section 6
Q6.1 – What is fission gas?
Q6.2 – Why is it important to maintain fission gas pressure as low as possible?
Q6.3 – Describe the processes involved in the generation, accumulation and release of fission
gas into “open” spaces inside the fuel sheath.
Q6.4 – What design features does the fuel designer have at his disposal to control or minimize
fuel element internal gas pressure?
Q6.5 – Using the equations presented in Section 6.5, determine the effect of storing the fission
gas at pellet centreline temperature compared to that of the pellet interface with the sheath.
What do we learn?
Q6.6 – A fuel element’s free space is 3 ml, and it contains 15 ml of gas at STP. Assuming that all
the gas is stored at a uniform temperature of 1400 K, what is the internal pressure of the fuel
element? [Hint: Use Equation 5].
Section 7
Q7.1 – Describe three defect types in CANDU fuel that may be caused by excessive stress.
Q7.2 – Describe in qualitative terms the reasons that cracks develop in uranium dioxide pellets
at operating conditions.
Q7.3 – A temperature profile is parabolic in a pellet of radius 6 mm. Is the hoop stress tensile or
compressive at: 1 cm? 2 cm? 3 cm? 4 cm? 5 cm?
Section 8
Q8.1 – Describe qualitatively the conditions that might lead to environmentally assisted crack-
ing.
Q8.2 – Provide examples of EAC defects that have been observed in CANDU fuel elements.
Q8.3 – Recognizing the three different effects that can cause EAC defects in CANDU fuel ele-
ments, provide a qualitative description of reactor operation that could lead to EAC defects.
Q8.4 – What parameters from operating reactors have been correlated to provide an indication
of potential fuel failure due to EAC? Draw a figure to show the relationship.
Q8.5 – Describe the mitigating parameters that have been adopted by CANDU designers and
operators to minimize EAC defects.
Q8.6 – A non-CANLUB fuel bundle experiences a power ramp from 25 kW/m to 50 kW/m at 140
MWh/kgU. What is its probability of having a defect? What would the defect probability be if it
were CANLUB fuel?
Section 9
Q9.1 – Describe the coolant conditions that may lead to fuel vibration.
Q9.2 – Describe the type of damage that fuel can experience as a result of vibration.
Q9.3 – A C6 fuel element has the same dimensions as in the illustrative example of Section
9.1.6, with the exception that its assembly weld has inadvertently been manufactured with a
diameter of 3 mm. Coincidentally, it is also destined to be loaded into a channel where the
amplitude of lateral vibration is expected to be 30 micrometers. What nominal stresses are
expected in the endplate and in the assembly weld? Are fatigue failures likely by this mechan-
ism?
Section 10
Q10.1 – How many damage mechanisms have been shown to affect CANDU fuel?
Q10.2 – How many thermal damage mechanisms (“T” series) are there? Provide a brief descrip-
tion of each.
Q10.3 – How many structural damage mechanisms (“S” series) are there? Provide a brief
description of each.
Fuel 67
Q10.4 – How many compatibility damage mechanisms (“C” series) are there? Provide a brief
description of each.
Q10.5 – List the design acceptance criteria for CANDU fuel.
Q10.6 – With the aid of a diagram, show the relationship of acceptance criteria to other rele-
vant parameters, e.g., the material damage limit.
Q10.7 – What is the purpose of assigning margins in relation to acceptance criteria?
Q10.8 – List the qualification tests and assessments used to qualify CANDU fuel bundles.
Section 11
Q11.1 – Name at least one plant operating constraint that is related to ensuring fuel integrity.
Are there others?
Q11.2 – How are plant operating constraints defined by system designers such as fuel designers
used to ensure safe plant operation?
Section 12
Q12.1 – What is the most common defect type that occurs in CANDU fuel during plant opera-
tion?
Q12.2 – What are the principal characteristics of the most common defect type that makes its
detection possible while the plant is in operation?
Q12.3 – What is secondary hydriding? Explain the mechanism and how it affects the integrity of
the fuel element.
Q12.4 – Explain why it is important to remove a defective fuel element from an operating
reactor as soon as the defect is detected.
Q12.5 – Name the three systems used in CANDU reactors to assist in the detection and removal
of defective fuel while the reactor is at power and explain briefly how each is used.
Q12.6 – Name three techniques used to confirm that a bundle containing a defective fuel
element has been discharged from the channel and explain briefly how each works.
15 References
A. H. Booth, “A Method of Calculating Fission Gas Diffusion from UO2 Fuel and its Application to
the X-2-f Loop Test”, Atomic Energy of Canada Limited, Report AECL-496 (1957).
G. H. Bryan, “Application of the Energy Test to the Collapse of a Long Thin Pipe under External
Pressure”, Proceedings of Cambridge Phil. Society, Vol. 6, p. 287 (1888).
R. M. Carroll and O. Sisman, “In-Pile Fission Gas Release from Single-Crystal UO2”, Nuclear
Science and Engineering, vol. 21, pp. 147–158 (1965).
T. Daniels, H. Hughes, P. Ancker, M. O’Neil, W. Liau, and Y. Parlatan, “Testing and Implementation
Program for the Modified Darlington 37-Element Fuel Bundle”, Proceedings, 10th Interna-
tional Conference on CANDU Fuel, Canadian Nuclear Society (2008).
A. Fick, “Über Diffusion”, Phil. Mag., vol. 10, pp. 59–86 (1855).
M. R. Floyd, D. A. Leach, R. E. Moeller, R. R. Elder, R. J. Chenier, and D. O’Brien, “Behaviour of
Bruce NGS-A Fuel Irradiated to a Burnup of ~500 MWh/kgU”, Proceedings, Third Interna-
tional Conference on CANDU Fuel, Chalk River, Canadian Nuclear Society, pp. 2-44 to 2-60
(1992).
M. Gacesa, V. C. Orpen, and I. E. Oldaker, “CANDU Fuel Design: Current Concepts”, IAEA/CNEA
International Seminar on Heavy Water Fuel Technology, San Carlos de Bariloche, Argentina,
1983. Also Report AECL-MISC 250-1 (Rev. 1), published by Atomic Energy of Canada Lim-
ited, 1983.
J. Gere and S. P. Timoshenko, Mechanics of Materials, PWS, Fourth Edition (1997).
I. J. Hastings, “Structures in Irradiated UO2 Fuel from Canadian Reactors”, Atomic Energy of
Canada Limited, Report AECL-Misc-249 (1982).
I. J. Hastings, A. D. Lane, and P. G. Boczar, “CANFLEX: An Advanced Fuel Bundle for CANDU”,
International Conference on Availability Improvements in Nuclear Power Plants, Madrid,
Spain. Also Report AECL-9929, published by Atomic Energy of Canada Limited (1989).
IAEA, “Review of Fuel Failures in Water-Cooled Reactors”, International Atomic Energy Agency,
Vienna, Report NF-T-2.1 (2010).
J. H. K. Lau, M. Tayal, E. Nadeau, M. J. Pettigrew, I. E. Oldaker, W. Teper, B. Wong, and F. Iglesias,
“Darlington N12 Investigation: Modelling of Fuel Bundle Movement in Channel under Pres-
sure Pulsing Conditions”, Proceedings, 13th Annual Conference, Canadian Nuclear Society
(1992).
B. J. Lewis, “Fundamental Aspects of Defective Nuclear Fuel Behaviour and Fission Product
Release”, Journal of Nuclear Materials, Vol. 160, pp. 201–217 (1988).
J. J. Lipsett and W. B. Stewart, “Failed Fuel Location in CANDU-PHW Reactors using a Feeder
Scanning Technique”, IEEE Transactions on Nuclear Science, Vol. NS-23, No. 1, pp. 321-324
(1976).
R. D. MacDonald, M. R. Floyd, B. J. Lewis, A. M. Manzer, and P. T. Truant, “Detecting, Locating,
and Identifying Failed Fuel in Canadian Power Reactors”, Report AECL-9714, published by
Atomic Energy of Canada Limited, Canada (1990).
Fuel 69
A. M. Manzer and R. W. Sancton, “Detection of Defective Fuel in an Operating CANDU-600

MW(e) Reactor”, Proceedings, Conference on Fission Product Behaviour and Source Term
Research, Snowbird, American Nuclear Society, LaGrange Park, Illinois (1984).
M. J. F. Notley and I. J. Hastings, “A Microstructure-Dependent Model for Fission Product Gas
Release and Swelling in UO2 Fuel”, IAEA Specialists’ Meeting on Fuel Element Performance
Computer Modelling, Blackpool, UK, 1978. Also Report AECL-5838, published by Atomic
Energy of Canada Limited, 1979.
W. J. O’Donnell and C. M. Purdy, “The Fatigue Strength of Members Containing Cracks”, Pro-
ceedings of ASME Petroleum Division, ASME Paper 63-PET-1 (1963).
W. J. O’Donnell and B. F. Langer, “Fatigue Design Basis for Zircaloy Components”, Nuclear
Science and Engineering, Vol. 20, pp. 1–12 (1964).
D. E. Olander, “Fundamental Aspects of Nuclear Reactor Fuel Elements”, Energy Research and
Development Administration, U.S.A., Report TID-26711-P1 (1976).
R. D. Page, “Canadian Power Reactor Fuel”, Atomic Energy of Canada Limited, Report AECL-5609
(1976).
W. J. Penn, R. K. Lo, and J. C. Wood, “CANDU Fuel—Power Ramp Performance Criteria”, Nuclear
Technology, Vol. 34, pp. 249–268 (1977).
R. E. Peterson, “Application of Stress Concentration Factors in Design”, Proceedings of the
Society for Experimental Stress Analysis, Vol. 1, pp. 118–127 (1943).
M. J. Pettigrew, “The Vibration Behaviour of Nuclear Fuel under Reactor Conditions”, Nuclear
Science and Engineering, Vol. 114, pp. 179–189 (1993).
P. Reid, M. Gacesa, and M. Tayal, “The ACR-1000 Fuel Bundle Design”, Proceedings, 10th Interna-
tional Conference on CANDU Fuel, Canadian Nuclear Society (2008).
A. D. Smith, I. J. Hastings, P. J. Fehrenbach, P. A. Morel, and R. D. Sage, “Dimensional Changes in
Operating UO2 Fuel Elements: Effect of Pellet Density, Burnup, and Ramp Rate”, Atomic En-
ergy of Canada Limited, Report AECL-8605 (1985).
A. Sun and M. Tayal, “Technical Basis for ACR-1000 Fuel Acceptance Criteria”, Proceedings, 11th
International Conference on CANDU Fuel, Canadian Nuclear Society (2010).
M. Tayal and C. K. Choo, “Fatigue Analysis of CANDU Nuclear Fuel Subjected to Flow-Induced
Vibrations”, Paper AIAA-84-0959-CP, 25th Structures, Structural Dynamics and Materials
Conference, American Society of Mechanical Engineers, Palm Springs, California, 1984. Also
Report AECL-8331, published by Atomic Energy of Canada Limited (1984).
M. Tayal, “Modelling CANDU Fuel under Normal Operating Conditions: ELESTRES Code Descrip-
tion”, Atomic Energy of Canada Limited, Report AECL-9331 (1987).
M. Tayal, L. M. MacDonald, E. Kohn, and W. P. Dovigo, “A Model for the Transient Release of
Fission Products from UO2 Fuel: GASOUT Code Description”, Nuclear Technology, Vol. 85,
pp. 300–313. Also Report AECL-9794, published by Atomic Energy of Canada Limited
(1989).
M. Tayal and G. Chassie, “Extrapolating Power-Ramp Performance Criteria for Current and
Advanced CANDU Fuels”, Proceedings, 21st Annual Conference, Canadian Nuclear Society
(2000).
M. Tayal, P. Reid, M. Gacesa, A. Sun, E. Suk, and D. Gossain, “ACR-1000 Fuel Acceptance Criteria
for Normal Operation and for Anticipated Operational Occurrences”, Proceedings, Tenth In-
ternational Conference on CANDU Fuel, Canadian Nuclear Society, Ottawa, Canada (2008-
1).
M. Tayal, A. Sun, M. Gacesa, P. Reid, P. Fehrenbach, B. Surette, and E. Suk, “Mechanism for
Power Ramp Failures in CANDU Fuel”, Proceedings, 10th International Conference on
CANDU Fuel, Canadian Nuclear Society (2008-2).
S. P. Timoshenko, Strength of Materials, Part 2, McGraw-Hill, Second Edition (1961).
S. P. Timoshenko and J. N. Goodier, Theory of Elasticity, McGraw Hill, Third Edition (1970).
J. C. Wood, D. G. Hardy, and A. S. Bain, “Power Ramping Fuel Performance and Development”,
Atomic Energy of Canada Limited, Report AECL-6676 (1979).
16 Further Reading
Design Verification
P. G. Boczar, "CANDU Fuel Design Process", Ninth International Conference on CANDU Fuel,
Belleville, Canada, Canadian Nuclear Society, 2005.
J. Chang, S. Abbas, A. Banwatt, M. DiCiano, M. Gacesa, A. Gill, D. Gossain, J. Hood, F. Iglesias, L.
Lai, T. Laurie, Y. Ornatsky, U. K. Paul, A. Sun, M. Tayal, S. G. Xu, J. Xu, and P. Reid, "Proof of
Design Adequacy for the ACR-1000 Fuel Bundle", 10th International Conference on CANDU
Fuel, Ottawa, Canada, October 5-8, 2008.
History (Early)
J. A. L. Robertson, “Fuel for Thought”, Nuclear Journal of Canada, Vol. 1, Issue 4; pp 332-341.
Also published in Proceedings of the Engineering Centennial Conference, Montreal, 18-22
May, 1987.
Irradiation Tests and Performance
F. R. Campbell, L. R. Borque, R. Deshaises, H. E. Sills, and M. J. F. Notley, “In-Reactor Measure-
ments of Fuel-to-Sheath Heat Transfer Coefficients Between UO2 and Stainless Steel”,
Atomic Energy of Canada Limited, Report AECL-5400, 1977.
T. J. Carter, "Experimental Investigation of Various Pellet Geometries to Reduce Strains in
Zirconium Alloy Cladding", Nuclear Technology, Vol. 45, No. 2, September 1979, pp. 166-
176.
M. R. Floyd, J. Novak, and P. T. Truant, "Fission-Gas Release in Fuel Performing to Extended
Burnups in Ontario Hydro Nuclear Generating Stations", IAEA Technical Committee Meeting
on Fission Gas Release and Fuel Rod Chemistry Related to Extended Burnup, Pembroke,
Canada. Also Atomic Energy of Canada Limited, Report AECL-10636 (1992).
M. R. Floyd, Z. He, E. Kohn, and J. Montin, "Performance of Two CANDU-6 Fuel Bundles Contain-
ing Elements with Pellet-Density and Clearance Variables", 6th International Conference on
CANDU Fuel, Niagara Falls, Canada. Also Atomic Energy of Canada Limited, Report AECL-
12033, 1999.
Fuel 71
M. R. Floyd, "Extended-Burnup CANDU Fuel Performance", International Conference on CANDU

Fuel, CNS, Kingston. Also Atomic Energy of Canada Limited, Report AECL-CONF-01135,
2001.
D. G. Hardy, A. S. Bain, and R. R. Meadowcroft, "Performance of CANDU Development Fuel in
the NRU Reactor Loops", ANS Topical Meeting on Water Reactor Fuel Performance, St.
Charles, Illinois, pp. 198-206, 1977.
M. J .F. Notley, A. S. Bain, and J. A. L. Robertson, “The Longitudinal and Diametral Expansion of
UO2 Fuel Elements”, AECL-2143, 1964.
Manufacturing and Quality Assurance
C. Ganguly and R. N. Jayaraj (editors), Characterization and Quality Control of Nuclear Fuels,
Allied Publishers, 2004.
M. Gacesa, G. R. Quarrington, W. R. Tarasuk, I. R. Carrick, J. Pawliw, G. McGregor, H. R. Debnam,
and L. Proos, "CANDU Fuel Quality and How it is Achieved", Atomic Energy of Canada Lim-
ited, Report AECL-7061, 1980.
R. Sejnoha, "Quality Assurance in CANDU Fuel Design, Development, and Manufacturing",
IAEA/INSTN, Saclay (France), October 1992.
Material Science
R. Adamson, B. Cox, J. Davies, F. Garzarolli, P. Rudling, and S. Vaidyanathan, "Pellet-Cladding
Interaction", IZNA-6 Special Topical Report, ANT International, 2006.
D. L. Hagrman, G. A. Reyman, and R. E. Mason (editors), "MATPRO-Version 11 (Revision 2): A
Handbook of Material Properties for Use in the Analysis of Light Water Reactor Fuel Rod
Behaviour", EG&G Idaho, Inc., Report NUREG/CR-0479, TREE-1280, Rev.2, 1981.
A. Sawatzky and C. E. Ells, "Understanding Hydrogen in Zirconium", Zirconium in the Nuclear
Industry: Twelfth International Symposium, ASTM STP 1354, 2000, pp 32-48.
Modelling and Computer Codes
J. A. L. Robertson, A. M. Ross, M. J. F. Notley, and J. R. MacEwan, "Temperature Distribution In
UO2 Fuel Elements”, Journal of Nuclear Materials, pp. 225-262, December 1961. Also AECL-
1679.
M. Tayal, D. Lim, “Recent Uses of the Finite Element Method in Design/Analysis of CANDU Fuel”,
Sixth Annual Conference of the Canadian Nuclear Society, Ottawa. Also Atomic Energy of
Canada Limited, Report AECL-8754, 1985.
17 Relationships with Other Chapters

Chapter 8 provides an overview of fuel bundle configuration. Chapters 3 to 5 explain neutron
physics that generates heat in the fuel. Chapters 6 and 7 explain how that heat is removed from
the fuel bundle and illustrate the internal temperature distribution within a fuel rod. Chapters
14 and 15 explain chemical and metallurgical aspects that relate to corrosion of the fuel sheath.
Chapter 13 explains the performance of fuel during postulated accidents. Chapter 18 discusses
advanced fuels and fuel cycles; it also summarizes fuel manufacturing. Finally, Chapter 19
describes interim storage and disposal of used fuel.
18 Acknowledgements
ments during the development of this chapter; their feedback has much improved it.
Peter Boczar
Lawrence Dickson
Rosaura Ham-Su
Paul Chan
Kwok Tsang
Of course the responsibility for any errors or omissions lies entirely with the authors.
Thanks are also extended to Media Production Services, McMaster University, for producing the
figures, and to Diana Bouchard for expertly editing and assembling the final copy.
1
CHAPTER 18
Fuel Cycles
prepared by
Summary
Most commercial reactors currently use the once-through fuel cycle. However, because there
is still a significant amount of useful material (and available energy) in the fuel discharged
from these reactors, many fuel cycles are possible in which some fuel components are recy-
cled for further reactor use. In addition, several other fuel cycles are also possible using tho-
rium, which is much more abundant on Earth than the uranium that is the primary source of
commercial nuclear power today. The CANDU reactor is designed to use natural uranium
fuel, which is less expensive and more efficient in the use of uranium than any known alterna-
tive. Some of the features that enable a CANDU reactor to operate on natural uranium also
make it eminently capable of using alternative fuels. In spite of increased fuel cost, the use of
alternative fuel cycles in a CANDU reactor can lead to benefits such as significant extensions
of humanity’s energy resources, reduced capital costs of nuclear power plants, reduced vol-
ume and duration for long-term storage and disposal of nuclear waste, longer life of some
components of nuclear power plants, increased efficiency of the thermal cycle, and reduced
severity of some postulated accidents. Several alternative fuel and fuel cycles are described
in this chapter, including the enriched uranium fuel cycle, the recovered-uranium cycle, the
MOX cycle, the thorium cycle, the DUPIC cycle, the tandem cycle, low void reactivity fuel, and
actinide burning fuel. In addition to increasing fuel cost, alternative cycles impose operating
requirements that are more challenging than those experienced when using natural-uranium,
low-burnup fuel. Incremental technical challenges in the following subject areas are de-
scribed here: internal gas pressure, power ramps, corrosion, deuterides and hydrides, depos-
its, end-temperature peaking, bowing, and high burnup structure in the pellet.
©UNENE, all rights reserved. For educational use only, no assumed liability. Fuel Cycles – March 2015
Table of Contents
1 Introduction ............................................................................................................................ 4
1.1 Overview ......................................................................................................................... 4
1.2 Learning outcomes.......................................................................................................... 4
2 Neutronics of Fuel Cycles........................................................................................................ 5
2.1 Definitions....................................................................................................................... 5
2.2 Dynamics of Depletion.................................................................................................... 6
3 Overview of Some Possible Fuels and Fuel Cycles.................................................................. 8
3.1 Reprocessing ................................................................................................................... 9
3.2 Natural Uranium Fuel Cycle .......................................................................................... 10
3.3 Enriched Uranium Fuel Cycle ........................................................................................ 10
3.4 Recovered-Uranium Cycle............................................................................................. 11
3.5 MOX Cycle ..................................................................................................................... 11
3.6 Thorium Cycle ............................................................................................................... 11
3.7 DUPIC Cycle................................................................................................................... 12
3.8 Tandem Cycle ................................................................................................................ 12
3.9 Low Void Reactivity Fuel ............................................................................................... 12
3.10 Actinide Burning Fuel.................................................................................................... 13
3.11 Breeding Cycles ............................................................................................................. 13
4 Key Drivers for Advanced Fuel Cycles ................................................................................... 13
4.1 Increase the Sustainability of the Nuclear Fuel Cycle................................................... 13
4.2 Improve Fuel Cycle Economics...................................................................................... 14
4.3 Improve the Health and Environmental Impacts of the Fuel Cycle .............................. 15
4.4 Improve Safety Characteristics...................................................................................... 15
4.5 Improve the Proliferation Resistance of Fuel Cycles..................................................... 15
5 Resource Extension: A Major Incentive ................................................................................ 15
6 Other Considerations ............................................................................................................ 18
7 Pertinent Features of the CANDU Reactor............................................................................ 19
8 Natural Uranium Fuel Cycle .................................................................................................. 20
8.1 Production of Ceramic UO2 Powder.............................................................................. 20
8.2 Fuel Bundle Manufacturing .......................................................................................... 21
9 Details of Selected Alternative Fuels and Fuel Cycles........................................................... 25
9.1 Natural Uranium Equivalent (NUE) Fuel ....................................................................... 25
9.2 Extended Burnup Fuel................................................................................................... 26
9.3 Thorium Fuel ................................................................................................................. 42
9.4 Mixed Oxide (MOX) Fuel ............................................................................................... 46
9.5 Breeder Fuel.................................................................................................................. 48
10 Synergy between CANDU Reactors and LWRs ...................................................................... 48
11 Closure .................................................................................................................................. 49
12 Problems ............................................................................................................................... 50
13 References............................................................................................................................. 53
Fuel Cycles 3
List of Tables
Table 1 Energy output per megagram of mined uranium............................................................. 16
Table 2 Years of uranium availability............................................................................................. 17
List of Figures
Figure 1 Transformation of specific isotopes .................................................................................. 7
Figure 2 Generalized isotope transformation processes ................................................................ 8
Figure 3 CANDU fuel manufacture: assembly flow chart ............................................................. 22
Figure 4 Illustrative notch at the sheath/endcap junction ........................................................... 27
Figure 5 Illustrative effect of irradiation on ductility of Zircaloy................................................... 27
Figure 6 Illustrative effect of burnup on power ramp defect threshold....................................... 28
Figure 7 Waterside corrosion in the sheath.................................................................................. 30
Figure 8 Illustrative distribution of hydrides in the sheath at about 25 MWd/kg HE................... 31
Figure 9 Concepts for end-flux peaking ........................................................................................ 33
Figure 10 Illustrative peaking of flux at the end of a 28-element fuel bundle ............................. 34
Figure 11 EFP: neutron flux and heat transfer .............................................................................. 34
Figure 12 End-temperature peaking: illustrative isotherm........................................................... 35
Figure 13 Model for element bowing ........................................................................................... 37
Figure 14 Illustrative circumferential variation in flux of thermal neutrons................................. 38
Figure 15 Off-centre grain growth ................................................................................................ 38
Figure 16 Increase of bowing during continued irradiation ......................................................... 39
Figure 17 High burnup structure near the rim of a pellet ............................................................ 40
Figure 18 Illustrative heat generation rate across the radius of the pellet .................................. 41
Figure 19 Thermal conductivities of unirradiated pellets............................................................. 43
Figure 20 Illustrative structures in irradiated pellets.................................................................... 44
Figure 21 High fission area (HFA) in an irradiated MOX fuel pellet .............................................. 47
1 Introduction
The term “fuel cycle” commonly refers to all the steps through which the nuclear fuel goes
from mining, to use in a reactor, to discharge, possible recycle, and eventual disposal. There-
fore, a fuel cycle consists of certain steps in the front end, which involves preparation of the
fuel; other steps in the service period, in which the fuel is used during reactor operation; and
still other steps in the back end which are necessary to manage, contain, and either reproc-
ess or dispose of spent nuclear fuel safely.
1.1 Overview
This chapter focusses primarily on selected aspects of the “front ends” and the “service peri-
ods” of illustrative fuel cycles that use natural uranium, enriched uranium, plutonium, tho-
rium, and their combinations. The discussion covers simplified neutronics, fuel manufactur-
ing, and selected considerations of associated in-reactor fuel performance. The remaining
aspects of fuel cycles are covered elsewhere in this book.
The simplest fuel cycle involves a single pass of the fuel through the reactor; this is called the
once-through cycle. Current CANDU reactors run on the once-through natural uranium fuel
cycle. A significant amount of useful material (and available energy) still exists in the fuel
which is discharged from LWR reactors. Therefore, many fuel cycles are possible in which
some fuel components are recycled for further reactor use, particularly in a CANDU reactor.
In addition, several other fuel cycles are also possible by using thorium, which is much more
abundant on Earth than the uranium that is the primary source of commercial nuclear power
today.
We can derive a number of benefits by these means, including significant extensions of hu-
manity’s energy resources, reduced capital costs of nuclear power plants, reduced volume
and duration for long-term storage and disposal of nuclear waste, longer life of some com-
ponents of nuclear power plants, increased efficiency of the thermal cycle, and reduced se-
verity of some postulated accidents. However, several other important factors such as cost
also need to be considered, as outlined later in Section 6.
1.2 Learning outcomes

The goal of this chapter is for the student to acquire a broad initial overview of some of the
more interesting variants of fissile and fertile materials and the associated fuels and fuel cy-
cles. These include fuel cycles based on the use of natural uranium, enriched uranium, plu-
tonium, thorium, and their combinations. The student will also learn how to evaluate some
aspects of fuel cycles and the performance of alternative fuels.
Fuel Cycles 5
2 Neutronics of Fuel Cycles

2.1 Definitions
Fissile Nuclides: Nuclides that have a significant fission cross section at zero energy.
Fissile nuclides can fission by reacting with a neutron of any kinetic energy, including zero or
near-zero kinetic energy (i.e., a slow/thermal neutron).
Example of fissile nuclides: 235U, 239Pu, 233U, 241Pu.
235
U is the only naturally-occurring fissile nuclide. The interaction of 235U with a thermal neu-
tron gives 236U, which may fission in about 10-14 seconds, releasing on average about 2.5 neu-
trons. Therefore, we can write:
n + 235U → Fission Products + 2.5n + γ.
Fissionable Nuclei: Nuclides that can undergo fission (either by a slow or by a fast neutron)
are called fissionable. Fissile nuclides are therefore a sub-category of fissionable nuclides.
Nuclides which can fission only by reacting with a neutron of high kinetic energy (i.e., a fast
neutron) are fissionable, but not fissile.
Examples of fissionable nuclides that are not fissile: 232Th, 238U.
In the case of some non-fissile but fissionable nuclides, the energy-dependent fission cross
section displays a threshold with incident neutron energy. The cross section is negligible up
to the high energy threshold and has a step increase at the threshold.
Fertile Nuclides: Fertile nuclides are nuclides that can capture a neutron and transmute into
a fissile nucleus following one or more nuclear decays.
Examples of fertile nuclides: 232Th, 238U (these happen to be fissionable as well).
Conversion Ratio C
To yield fissile nuclides, fertile nuclides must be irradiated with neutrons, which are created
as a result of fission. Fission, in turn, consumes fissile nuclei, as does capture. We can look
at the process of converting fertile nuclides into fissile ones as consuming some fissile nu-
clides [Lamarsh, 1966].
The conversion ratio C can be defined as:
rate at which fissile nuclides are produced in the entire reactor by conversion of fertile nuclides
C
rate at which fissile nuclides are consumed (by fission and radiative capture ) ;
i.e., C is the average number of fissile nuclides produced by conversion per fissile nuclide
consumed (by fission and radiative capture).
The total number of fissile nuclides produced starting from N initial fissile nuclides can be
calculated as:
NC
NC  NC 2  NC 3   
1 C ,
where C < 1.
Including the initial fissile nuclides, the total number of fissile nuclides available for fission is
therefore:
NC N
N 
1 C 1 C .
It can be seen that the total number of nuclides available for fission increases with the con-
version ratio. Because C is a function of the aggregate nuclear properties of fuel which can
change with time, the value of C can also change with time.
If C > 1, more fissile nuclides are created than destroyed. C is then called the breeding ratio.
Breeding can therefore be thought of as conversion with C > 1. When breeding occurs, the
breeding gain can be defined as:
G  C 1 .
Fuel Burnup
The specific burnup (sometimes called just burnup, ω) equals the fission energy produced per
unit mass of heavy elements in the fuel:
Fission Energy Generated

Total Initial Mass of Heavy Nuclides in Fuel .
The usual units for fission energy are either MWh or MWd. Because 235U is by far the most
frequently used fissile material, initial mass is usually expressed either as kg or T (tonnes) of
U. However, if other fissile materials are used, e.g., plutonium or a mixture of uranium and
plutonium, then the initial mass is usually expressed as kg or T of HE (heavy elements).
2.2 Dynamics of Depletion

The reactions by which fertile nuclides produce fissile nuclides are, for example:
n  232Th    233 Th  233
Pa  233
U (1)
 decay  decay
n  238U    239 U  239

Np  239
Pu (2)
  decay  decay
and subsequently:
n  239Pu    240Pu (3)
n  240Pu    241Pu . (4)

Figure 1 illustrates a more generalized conversion of fertile nuclides to fissile nuclides.
Fuel Cycles 7
(Courtesy Bill Garland)

Figure 1 Transformation of specific isotopes
Recall from neutron physics that moving to the right in Figure 1 involves adding a neutron
(i.e., +n), whereas moving left and up involves beta decay (i.e., β-).
). For clarity, Figure 1 shows
arrows and appropriate symbols (i.e.(i.e., +n, β-,, etc.) mainly for the key reactions discussed
above.
The general nuclide history can be followed by modelling all processes involved [Lamarsh,
[
1966; Garland, 2005]:
1. loss by decay of the nuclide
nuclide;
2. loss by neutron capture by the nuclide, which transmutes it into another isotope of
the same element;
3. production by decay from a parent nuclide
nuclide;
4. production by capture transmutation
transmutation;
5. imposed fuel loading changes.
The first four processes in this list are illustrated in Figure 2.
(Courtesy Bill Garland)

Figure 2 Generalized isotope transformation processes
The governing rate equation is:
dN A
  A N A  aA N A   B N B   C  N C  F(t)
 . (5)
dt 
   
 (3)parent  
(1)decay decay fuel loading
(2)neutron capture (4)capture transmutation
where NA, NB, NC are the number of nuclei (#/cm3) for isotopess A, B and C respectively,
respectively λ is
the associated decay constant (sec-1),  is the neutron flux (#/cm2-sec) and σ a ,   are the
microscopic cross sections (cm2) for neutron absorption and capture respectively.
respectively We can
safely assume that the flux is constant over the integration time step and can easily solve this
equation numerically. In general, there are many nuclides to track, so they are all solved for
simultaneously, thereby track
tracking the depletion and growth of the various fuel nuclides. Note
that the two production processes (3 and 4 above
above)) mean that there is potential for fuel crea-
cre
tion, called conversion or breeding as explained earlier.
The above process yields information on whether or not a given combination and arrange-
arrang
ment of materials will achieve criticality and sustain the chain re
reaction
action and how much energy
can be so produced. The detailed “physics” evaluations of specific fuel cycles use the same
general principles that are outlined in earlier chapters on physics and are usually accom-
plished mainly by changing the inputs of appropriate “generic” physics codes.
3 Overview of Some Possible Fuels and Fuel Cycles

This section outlines some possible fuel
fuels and fuel cycles such as the natural uranium cycle,
the enriched uranium cycle, the thorium cycle, the breeding cycle, actinide burning fuel, and
low void reactivity fuel.
In addition, this section also outlines some interesting and clever fuel cycles that involve ex-
e
tracting significant additional energy from Light Water Reactor (LWR) fuel that would nor- no
mally be considered as “spent
spent”—that is, fuel that has been discharged from the reactor after
having delivered all the energy that is reasonably expected from it. Some cycles that extract
additional energy in this manner are: the recovered uranium cycle, the he MOX cycle, the
Fuel Cycles 9
DUPIC cycle, and the Tandem cycle. They are also outlined in this section.
3.1 Reprocessing
Before we delve into various fuel cycles, we first outline reprocessing of LWR fuel as it per-
tains to this chapter.
Considerable quantities of fission products gradually accumulate inside the fuel element dur-
ing irradiation. Some fission products are parasitic in that they absorb neutrons without
creating proportionate heat. This leaves fewer neutrons to continue the chain reaction, and
at some stage, further irradiation of that material risks shutting down the chain reaction.
Therefore that fuel is removed from the reactor and is frequently labelled as “spent” or
“waste” or “discharged”.
Uranium fuel used in LWRs has an initial 235U concentration (enrichment) of between 3% and
5%. After many months (sometimes even a few years) of generating energy in the reactor,
most but not all of the original 235U in typical enriched fuels in LWRs is largely used up (e.g.,
by fission and by radiative capture). Likewise, some fertile nuclides in the fuel, as well as
some fissionable nuclides produced by fertile nuclides, are also used up. However, fission-
able nuclides are not at zero concentration when the fuel is discharged from the reactor.
Spent fuel retains some of the original enriched uranium, and it also contains some fissile
plutonium that is produced by the chain reaction (see equations and discussion in Section 2
and in earlier chapters on physics).
The residual fissile content in spent LWR fuel depends on the starting enrichment and the
discharge burnup; nevertheless, for illustration purposes, let us say that it is typically in the
range of about 1.5% (235U + 239Pu + 241Pu). From the perspective of a CANDU reactor, this is a
much higher fissile concentration than even the usual fresh (non-enriched) CANDU fuel
(0.7%). Even though LWRs cannot continue to “burn” that fuel, CANDU reactors can, due to
the latter’s superior neutron efficiency. This creates an opportunity to generate additional
energy from spent LWR fuel in a variety of ways, as described later in this section.
However, to make effective use of available thermal neutrons, we first need to remove suffi-
cient parasitic neutron absorbers—such as fission products—from spent LWR fuel. This is
done through “reprocessing”.
Very simplistically, let us think of fuel as having the following initial and final states before
and after its irradiation in an LWR:
235
U (3-5%) + 238U → {235U + Pu} (about 1.5%) + 238U + fission products (FPs)
(fresh LWR fuel) (spent LWR fuel)
Let us next focus only on the right-hand side of this expression. Discussions later in this sec-
tion use the terms explained below:
 For purposes of this chapter, “conventional reprocessing” or “full reprocessing” means
extracting, separating, and purifying uranium (mostly 235U and 238U) and plutonium
(mostly 239Pu and 241Pu) in spent fuel. This, however, leads to some concerns that col-
lected plutonium can potentially be diverted towards nuclear proliferation. Full repro-
cessing generally consists of the following main steps: (Step #1) Irradiated fuel is dis-
solved; (Step #2) Plutonium and uranium are separated from fission product and then
(Step #3) from each other; and (Step #4) purified for enrichment (in the case of REU) or
(Step #5) recycling as MOX.
 In one variant involving partial reprocessing, all fission products are removed from spent
fuel, but plutonium is neither separated from uranium nor extracted separately, and
therefore concerns about proliferation are reduced compared to “full reprocessing”.
This product is used in the “Tandem cycle” (discussed later).
 In another variant involving partial reprocessing, no dissolution of spent fuel takes place.
The gaseous fission products are removed by crushing and heating the spent fuel; the
solid fission products are retained in the fuel. The high residual radio-toxicity of this
product significantly reduces its attractiveness for potential diversion, which further di-
minishes concerns about nuclear proliferation. This product is used in the “DUPIC cycle”,
as described later in this chapter.
 Plutonium recovered from full reprocessing is used to generate electricity in the “MOX
cycle”.
 Uranium that is collected from full reprocessing is called “recovered uranium” (RU). Its
235
U fraction in spent LWR fuel is higher than in natural (as-mined) uranium. RU is used
to generate electricity in the “recovered uranium” cycle.
In summary, a variety of fissile and fertile materials are available from different sources in
multiple forms such as natural uranium, enriched uranium, thorium, various products of full
or partial reprocessing, and others (e.g., dismantled nuclear weapons). This presents oppor-
tunities to design a variety of fuels and fuel cycles in neutron-efficient CANDU reactors, some
of which are outlined in the following sections.
3.2 Natural Uranium Fuel Cycle

In this cycle, the initial fissile nuclide is 235U as found in mined (natural) uranium. This yields
an inexpensive fuel which is the staple of the current generation of CANDU nuclear power
plants.
Fission of initial 235U is subsequently augmented by plutonium, produced by neutrons cap-
tured in 238U. By the time the fuel is discharged from the reactor, almost half the fissions
come from plutonium.
3.3 Enriched Uranium Fuel Cycle

This cycle uses 235U as the initial fissile nuclide that has been enriched, typically to less than
5%.
Enriched uranium fuels are the staples of a variety of reactor types, such as pressurized wa-
ter reactors (PWRs), boiling water reactors (BWRs), and Russia’s VVER/RBMK reactors. In
these reactors, enrichment is used to provide an abundance of fissile atoms, which (1) in-
creases the probability of neutron fission and decreases the importance of moderator effi-
ciency, and (2) prolongs the period before fissile atoms must be replenished in the reactor.
During the late 1900s and early 2000s, CANDU reactor designers seriously considered using
enriched fuels in designs such as the Advanced CANDU Reactor (ACR) [Reid et al., 2008] and
in low void reactivity fuels [Boczar et al., 2004]. In CANDU reactors, the additional neutrons
Fuel Cycles 11
generated by enrichment can be used for a variety of purposes such as: (1) using less expen-
sive coolant (i.e., light water) to reduce capital cost; (2) enabling use of thicker pressure
tubes to reduce their creep rate and thus to extend their service life; (3) enabling use of
burnable neutron absorbers which can reduce the magnitude of a power pulse during a pos-
tulated accident; and (4) reducing the volume of spent fuel through extended burnup.
Enriched uranium fuel could also be used in existing CANDU reactors to achieve other bene-
fits such as higher reactor power (through a flatter radial channel power profile).
3.4 Recovered-Uranium Cycle

This is a variant of the enriched uranium fuel cycle. In this cycle, LWR fuel is reprocessed to
recover unused uranium. The recovered uranium is reused to make new uranium-based fuel
(if necessary, with the addition of fresh uranium).
Thanks to superior neutron efficiency, the CANDU reactor can use RU fuel from LWRs as-is,
without additional enrichment, to achieve the benefits mentioned above for enriched ura-
nium fuel.
A number of publications have described the salient attributes of RU fuel, for example Boczar
et al. [1993 and 2010]. Briefly, the 235U content of RU depends mainly on initial enrichment
and discharge burnup of LWR fuel. Conventional reprocessing yields very pure RU with very
few contaminants from fission products.
In principle, residual fissile material is also available in used CANDU fuel. However, used
natural-uranium (NU) fuel’s residual fissile content is much lower than that of used enriched
fuels from LWRs. Hence, recovery of residual fissile material for re-use would be significantly
more expensive from used NU fuel than from used enriched fuel [Boczar et al., 2010].
Significant amounts of high-grade highly enriched uranium (HEU) can also be obtained from
dismantled nuclear warheads for use in power reactors.
3.5 MOX Cycle

In this cycle, plutonium recovered through reprocessing LWR fuel is mixed with either fresh
or depleted uranium (from enrichment tails) to make mixed uranium- and plutonium-oxide
fuel (MOX). The plutonium concentration is adjusted to give the required enrichment and
burnup.
Significant amounts of high-grade plutonium can also be obtained from dismantled nuclear
warheads for use in power reactors.
3.6 Thorium Cycle

232
Th is fertile and can be used to produce 233U, which is fissile and is used in situ as it is pro-
duced. In the CANDU reactor, a once-through thorium cycle is economically viable, which
enables energy to be derived from 233U without reprocessing and recycling. Full benefit from
the thorium cycle would require reprocessing of the used fuel to recover and recycle the 233U
at an appropriate concentration as new reactor fuel.
Thorium cycles are important because thorium is much more abundant on Earth than ura-
nium. Some countries or regions have thorium resources, but lack plentiful uranium. They
can extend the energy from natural uranium (which is the source of the only natural fissile
material) by using the thorium cycle. In addition, thorium cycles provide other benefits, such
as in fuel performance and fuel safety, as explained in a later section devoted specifically to
thorium.
3.7 DUPIC Cycle

This is one of the cycles that can use discharged LWR fuel in CANDU reactors after partial
reprocessing. In the DUPIC cycle, the LWR fuel is reprocessed using dry thermal or mechani-
cal methods to remove gaseous fission products [Boczar et al., 1996]. This yields a fuel that
neutron-efficient CANDU reactors can use.
Uranium and plutonium are not separated during this operation, and therefore this fuel is
attractive from the perspective of nuclear non-proliferation. Irradiation tests have confirmed
satisfactory performance of DUPIC fuel [Karam et al., 2010(a)].
3.8 Tandem Cycle

This is another cycle based on the use of discharged LWR fuel in CANDU reactors after partial
reprocessing. In this cycle, the plutonium and uranium in LWR fuel are separated from all the
fission products, but are not separated from each other, as is done in conventional LWR fuel
reprocessing operations [Hastings et al., 1991]. Therefore, the extraction of fissile materials
is more attractive from the non-proliferation perspective than the full reprocessing opera-
tion, but less attractive than DUPIC, as explained at the beginning of this section.
3.9 Low Void Reactivity Fuel

One important aspect of CANDU reactor safety is its coolant void reactivity (CVR). If a signifi-
cant volume of the coolant were to be lost in an accident, a complex change in the neutron
spectrum would initially rapidly increase the reactivity in the affected reactor core. This
would automatically trigger corrective action(s) by the reactor’s shutdown system(s), which
would quickly decrease the reactivity again. Therefore, the affected core's reactivity would
peak for a very short duration. The initial increase in reactivity is called “positive coolant
void reactivity”. In CANDU reactors, the magnitude of the CVR during postulated loss of
coolant accidents sets several requirements for the shutdown system, such as speed, redun-
dancy, and depth. A smaller value of CVR (including a negative value) could loosen these
performance requirements and/or increase the safety and/or licensability of the reactor.
One option to reduce such reactivity is the appropriate use of a very small amount of burn-
able neutron absorber (BNA) in the fuel bundle. The low void reactivity fuel (LVRF) bundle
uses a burnable neutron absorber (BNA) such as dysprosium and locates it strategically in the
centre element. Enriched uranium is used in the outer elements to compensate for neutrons
absorbed by the BNA. The amount of neutron absorber and the level of enrichment can be
varied to give the desired values of coolant void reactivity and burnup, although the de-
crease in void reactivity is at the expense of uranium utilization [Boczar et al., 2004]. Such a
fuel design has been developed and successfully tested in the reactor [Boczar et al., 2004].
Fuel Cycles 13
3.10 Actinide Burning Fuel

The back end of the fuel cycle continues to attract considerable public scrutiny. A potential
amelioration involves reducing the duration for which used fuel remains a radioactive hazard.
This could potentially be achieved by fissioning some of the long-lived actinides from used
LWR fuels in fast reactors. In a variant of this concept, actinides can first be removed from
used LWR fuel (through an enhanced reprocessing route) and blended with a fuel matrix
such as uranium, thorium, or a neutronically inert material such as zirconia and used as fuel
in a CANDU reactor. The resulting waste would have significantly lower decay heat and ra-
diotoxicity. These actinide-burning CANDU reactors could be used on their own to dispose of
spent fuel or in conjunction with fast burner reactors. Use of the CANDU burner reactor
would significantly reduce the number of fast reactors that would need to be built for ulti-
mate destruction of LWR actinide waste [Hyland et al., 2010; Bhatti et al., 2013].
As a by-product, this process would also create many rare earths, which are economically
very valuable.
3.11 Breeding Cycles

In fast reactors, breeding cycles are possible: more fissile fuel is produced from fertile nu-
clides than the amount of fissile fuel destroyed. This cycle in fast breeder reactors can then
be used multiple times to increase the use of nuclear-fuel resources. For example, 238U is
one of the possible fertile fuels for breeder cycles. This nuclide makes up most of mined
uranium—over 99%—compared to 0.7% of 235U. Although current thermal reactors do also
convert a tiny fraction of the original 238U into energy, designs of thermal reactors are tai-
lored primarily towards extracting energy from the relatively miniscule fraction of 235U that
exists in mined uranium. Through breeder cycles, we can produce energy from a far greater
fraction of the much more abundant 238U, thus increasing humankind’s nuclear-fuel re-
sources by factors of 50–100 compared with their use in thermal reactors.
4 Key Drivers for Advanced Fuel Cycles

This section provides a broad overview of some of the more significant advantages of alter-
native fuel cycles.
4.1 Increase the Sustainability of the Nuclear Fuel Cycle

Energy resources can be expanded, sometimes dramatically, through the following methods
individually or in combination:
 By using thorium, which is much more abundant than uranium. Thorium fuel has other
advantages as well in safety, in-reactor performance, and long-term waste reduction, as
discussed in Section 9.3;
 By using plutonium, which can be obtained from a variety of sources such as dismantled
warheads or by reprocessing used fuel, as discussed in Section 9.4;
 By using enriched uranium, which can be obtained from a variety of sources such as
dismantled warheads or by reprocessing used LWR fuel, as discussed in Section 9.2;
 By using used LWR fuel, which has been reconfigured (but not reprocessed), in a CANDU
reactor;
238
 By using a breeder reactor to extract energy from U, which constitutes over 99% of
natural uranium, as discussed in Section 9.5.
Section 5 gives a more detailed discussion of resource extension.
4.2 Improve Fuel Cycle Economics

The natural uranium (NU) fuel cycle, which is currently used in CANDU power reactors, pro-
vides low-cost fuel and the highest uranium utilization (energy derived from mined uranium)
of any current commercial reactor. The fuel bundles for this cycle are also relatively easy to
manufacture. Nevertheless, the economics of the CANDU reactor can be further improved
by using enriched uranium. Enriched fuel can be used to:
 Reduce the initial capital cost of a CANDU reactor by enabling the use of lower-cost light
water coolant.
 Safely uprate reactor power: Enriched uranium fuel can also be used to uprate reactor
power (if allowed by the rest of the plant) by flattening the radial channel power distri-
bution across the core. This means that the channels produce close to the same amount
of power and that about 15% more power can be extracted from the same number of
fuel channels in a given size of reactor.
 Safely increase bundle power: Different amounts of power are produced in different
rings of fuel elements. By adding different degrees of enrichment to the fuel elements
that produce low power in different rings, their heat output can be raised without ex-
ceeding the safe limits of local power (including considerations of heat removal).
 Increase the lifetime of plant components by measures such as thickening the pressure
tubes, thus reducing their creep rate and increasing their service life. The enriched fuel
compensates for the increase in parasitic material in the core. Longer service life means
that mid-term replacement of pressure tubes can be delayed, thus extending the total
life of the plant.
 Increase the efficiency of the steam cycle. Thicker pressure tubes also permit operation
at higher coolant pressure and hence at higher outlet temperature. At the higher tem-
perature, steam converts a higher fraction of its thermal energy into electricity. This in
turn increases the overall efficiency of the steam cycle.
 Reduce the frequency of on-power fuelling. Higher burnup means a longer residence
period of the fuel in the channel. This, in some cases, translates into less frequent chan-
nel fuelling and hence a decreased number of duty cycles on the fuelling machines. This
in turn may potentially also increase the time available for maintenance of these ma-
chines.
 Reduce the volume of spent fuel by extending the burnup. The cost of storage, however,
may not be reduced proportionally due to other considerations.
 Reduce the cost of the long-term disposal of spent fuel through the transmutation of
hazardous fuel components such as long-lived actinides.
Fuel Cycles 15
4.3 Improve the Health and Environmental Impacts of the Fuel Cycle
The health and environmental impacts of nuclear power can be improved by using enriched
uranium in CANDU reactors and by transmuting hazardous components of the fuel, such as
long-lived actinides, in a reactor. This reduces the mass of these materials that requires long-
term disposal. This in turn reduces the size of the repositories that will be required and re-
duces potential impacts on human health and the environment.
4.4 Improve Safety Characteristics

Additional neutrons produced by enriched fuel can be used to offset the reactivity reduction
when burnable neutron absorbers are used to reduce coolant void reactivity, as discussed in
the section on low void reactivity fuel.
4.5 Improve the Proliferation Resistance of Fuel Cycles

The proliferation resistance of used nuclear fuel can be improved by irradiating derivatives in
CANDU reactors, as follows:
 Burn both weapons plutonium and plutonium from used reactor fuel, thereby reducing
the inventory of material for potential use in a nuclear explosive device.
 Use new fuel types (such as thorium denatured with 238U) which create spent fuel that is
unattractive for use in nuclear explosives.
5 Resource Extension: A Major Incentive

This section expands on one benefit of alternative fuel cycles noted in the section above:
extension of world-wide resources for fission energy. Although the subject has been consid-
ered by many analysts, the discussion below is based largely on Meneley’s analysis [2006] of
the world energy situation over the next 100 years, focussing largely on oil and nuclear.
As the world’s population and living standards continue to rise, demand for energy is also
likely to increase, despite conservation. Oil supply analysts agree that world oil production
must decline at some time during the 21st Century; some say that we can expect to see the
peak of world oil production very soon. Therefore, there could be a substantial long-term
gap (demand minus supply) in availability of primary energy. If nuclear energy is used to fill a
significant fraction of the expected energy gap, would it also soon run into resource limita-
tions?
Table 1 summarizes the energy requirements of various fuel cycles in thermal reactors. For a
given amount of mined uranium, the fuel cycle that is neutronically the most efficient—1.2%
enriched uranium in a CANDU reactor—produces about 90% more energy than the cycle that
is neutronically the least efficient.
Table 1 Energy output per megagram of mined uranium
MWy(e)/Mg
Enriched U in PWR, BWR 4.61
Pu Recycle in PWR, BWR 5.41
DUPIC (PWR → CANDU reactor) 6.37
Natural U in CANDU reactor 6.37
1.2% Enriched U in CANDU reactor 8.77
[Source: D. Meneley, 2006]

Based largely on the projections of the International Energy Agency (IEA) supplemented as
required by assumptions, Meneley deduced that to fill the projected world-wide gap in pri-
mary energy, we would need some 6,000 nuclear reactors, each with a capacity of 1 GWe.
This is significantly more than the currently installed nuclear capacity world-wide.
Table 2 summarizes Meneley’s results for years of uranium availability for the above scenario,
i.e., a nuclear capacity of 6 TWe. It suggests that if enriched 235U were used to fuel pressur-
ized water reactors (PWRs) or boiling water reactors (BWRs), the planet’s identified uranium
resources would last about 30 years. If the same 235U were instead used in natural-uranium
CANDU reactors, the uranium resources could be extended to last about 42 years. If 1.2%
enriched uranium were used in CANDU reactors, these uranium reserves would last about 58
years. The extension from about 30 years in LWRs to about 58 years in CANDU reactors is
largely due to the superior neutron efficiency of the latter.
If plutonium is recovered from used fuels and reused in LWRs, reactors can be run for 36
years. Alternatively, if used LWR fuel is reused in CANDU reactors in a DUPIC cycle along with
top-up enriched uranium, reactors can be run for 42 years—even without the expensive in-
termediate step of extracting plutonium.
Note, however, that when uranium eventually becomes scarce, prices will likely increase,
exploration will be spurred, and additional reserves can be expected to be discovered.
Fuel Cycles 17
Table 2 Years of uranium availability
Megawatt-Years Years of Uranium

per Tonne Availability for 6000
1000-MWe Units
Enriched U in PWR, BWR 4.61 30
Pu Recycle in PWR, BWR 5.41 36
DUPIC (PWR → CANDU Reactor) 6.37 42
Natural U in CANDU Reactor 6.37 42
1.2% Enriched U in CANDU Reactor 8.77 58
Fast Reactor 1,000 6700
Fast Reactor using Seawater* 1,000 >100,000
* Seawater contains about 0.003 ppm of uranium, for a total of over 4 billion tons of ura-
nium.
[Courtesy D. Meneley; Extracted from a presentation to Canadian Nuclear Society by D. Me-
neley, 2006]
A more dramatic way to extend our energy resources is to breed fissile materials from fertile
materials. This can usually be done in two ways: one, fertile 232Th can be converted into fis-
sile 233U; and two, fertile 238U can be converted into fissile 239Pu, as summarized below.
“The absorption cross-section for thermal neutrons of 232Th (7.4 barns) is nearly three times
that of 238U (2.7 barns). Hence, a higher conversion (to 233U) is possible with 232Th than with
238
U (to 239Pu). Therefore, 232Th is a better ‘fertile’ material than 238U in thermal reactors, but
232
Th is inferior to 238U as a ‘fertile’ material in fast reactors. For the ‘fissile’ 233U nuclide, the
number of neutrons liberated per neutron absorbed is greater than 2.0 over a wide range of
the thermal neutron spectrum, unlike 235U and 239Pu. Therefore, contrary to the 238U-to-239Pu
cycle in which breeding can be obtained only with fast neutron spectra, the 232Th-to-233U fuel
cycle can operate with fast, epithermal, or thermal spectra” [IAEA, 2005].
Thorium resources have not yet been explored as extensively as uranium; nevertheless, ex-
perts estimate that thorium in the Earth’s crust is some three to four times as abundant as
uranium [IAEA, 2005]. Boczar et al. [2010] have estimated that “if all the plutonium from
used natural uranium fuel from existing CANDU reactors operating over their lifetime were to
be used in a mixed thorium/plutonium cycle, two to three times the existing nuclear capacity
in Canada (currently about 15 GWe) could be sustained indefinitely” in a self-sufficient tho-
rium cycle in a CANDU reactor, e.g., a near-breeder. With a breeder, an expanding nuclear
system could be fueled essentially indefinitely (depending on the breeding ratio).
Other experts have estimated that 238U in “waste” CANDU fuel alone that has already been
accumulated in Canada can, after reprocessing and recycling in fast breeder reactors, pro-
duce some 4,000 years of electricity at today’s consumption rates [Ottensmeyer, 2012].
Moreover, additional 238U is available underground. Therefore, in principle, fast breeder re-
actors can provide energy for millennia to come, as shown in Table 2.
At this time, these are only theoretical possibilities. Many technical, economic, and practical
concerns remain about fast breeder reactors, such as the very large fissile inventory needed
to start a fast reactor and the very daunting concerns about proliferation.
6 Other Considerations
In addition to the drivers described above, various other technical and non-technical factors
should be considered in deciding on the most appropriate fuel cycle for a specific power
plant, utility, or country. Although a detailed discussion of this subject is beyond the scope of
this book, five major aspects are summarized below: energy independence, resource utiliza-
tion, resistance to nuclear proliferation, fuel performance, and cost.
Energy Independence
Plants that enrich large quantities of uranium are very expensive and are operational in only
a few nations. Fuel cycles based on natural uranium do not require a nation to build a plant
to enrich uranium, nor do they require a nation to rely on a few other nations for continual
supply of enriched uranium over a very long period. Therefore, natural uranium cycles pro-
mote national energy independence.
Resource Utilization
To produce a given amount of energy, some fuel cycles require less mined uranium than oth-
ers. For example, the combination of heavy water and natural uranium used in current
CANDU reactors requires some 40% less mined uranium per unit energy than fuel cycles that
use light water and enriched uranium. This can be a consideration in countries that do not
have much uranium and would like to stretch their domestic uranium to the extent practical.
Resistance to Nuclear Proliferation
There is concern in some quarters that during reprocessing, potent fissile material can poten-
tially become accessible and hence available for diversion to uses other than electricity gen-
eration. Some fuel cycles such as DUPIC are comparatively more resistant to nuclear prolif-
eration.
Fuel Performance
Overall, current fuels perform very well in current fuel cycles, as noted in Chapter 17. In fuel
cycles that differ significantly from current cycles, satisfactory fuel performance must be de-
signed and demonstrated. Depending on the specific cycle, new R&D may potentially also be
required.
Cost
Generally, enriching and reprocessing fuel is expensive, and the radioactivity and radiotoxic-
ity of spent enriched fuel tends to be higher than that of natural uranium. At the same time,
enriched fuel generally yields higher burnup, which tends to reduce fuelling cost (within a
range) and also reduces the volume of spent fuel to be stored. Therefore, this is a complex
topic, and its impact also changes with local conditions and with time.
Fuel Cycles 19
In summary, the choice of a specific fuel cycle must consider not only the drivers discussed
earlier, but also other techno-commercial and geo-political aspects, some of which are noted
above.
7 Pertinent Features of the CANDU Reactor

A number of features of a CANDU reactor enable it to use alternative fuel cycles more effec-
tively than other reactor types, mainly through higher neutron efficiency and also through
easier processing and fabrication of recycled fuel. These features are explained below.
Higher Neutron Efficiency
Natural uranium has low fissile content. To enable fission in a sustained chain reaction,
CANDU reactors are, of necessity, designed with high neutron efficiency. This in turn makes
CANDU reactors attractive for using other low-fissile materials in alternative fuel cycles.
The following main features contribute to the high neutron efficiency of CANDU reactors
[Boczar, 2012]:
 CANDU reactors use heavy water (D2O) as both moderator and coolant. This reduces
parasitic absorption of neutrons compared to light water.
 The moderator, which is separate from the coolant, is at low temperature. This pro-
motes neutron efficiency.
 In CANDU reactors, a small cluster of fuel elements (i.e., a fuel bundle) is surrounded by
a large amount of low-temperature moderator. This thermalizes a higher fraction of
neutrons.
 The principal structural components inside the core, e.g., the calandria tubes, pressure
tubes, and fuel bundle, are made from zirconium alloys. This reduces parasitic absorp-
tion of neutrons, for example in comparison to components made of stainless steel.
 CANDU reactors use relatively thinner structural materials. For example, as a proportion
of fuel element diameter, sheaths are significantly thinner in CANDU reactors than in
LWRs, reducing parasitic absorption of neutrons in the former.
 CANDU reactors are refueled on-power, as necessitated by the use of NU. This has a
consequential benefit in that the reactor can be operated effectively with low excess re-
activity in the core, obviating the need for burnable neutron absorbers and minimizing
neutron absorption in control materials.
 CANDU reactors contain equipment for on-power refuelling, which provides flexibility in
accommodating alternative fuels.
The resulting high neutron economy of a CANDU reactor enables high fuel utilization regard-
less of the fissile material (233U, 235U, or plutonium). About 90% more energy can be ex-
tracted from a particular fissile material than in an LWR reactor [Boczar et al., 1996], as
shown in Table 1.
Easier Fabrication and Reprocessing of Fuel
The simple, small fuel bundle design of a CANDU reactor (see Chapters 8 and 17) facilitates
remote processing and fabrication for highly radioactive recycled fuels.
A key “enabler” of these advantages is the separation of moderator and coolant in a CANDU
reactor through pressure/calandria tubes. This separation permits the reactor designer to
adjust the moderator and the coolant independently to exploit neutronic characteristics
more effectively. For example, the moderator can be kept at low temperature to promote
neutron efficiency, while the coolant can simultaneously be kept at high temperature to
promote the thermal efficiency of the steam cycle. In another example, the choice of cool-
ant fluid and the moderator-to-coolant-volume ratio can potentially be fine-tuned to reduce
capital cost without adversely affecting fuel cooling, as is done with the Advanced CANDU
Reactor. CANDU reactors are unique among successful contemporary reactor systems in of-
fering this important flexibility to designers.
8 Natural Uranium Fuel Cycle

The front end of the NU cycle starts with mining, refining, and conversion of uranium into
ceramic UO2 powder. The next segment of the front end is fuel manufacturing, including the
procurement and processing of Zircaloy components. These front-end segments are de-
scribed below.
8.1 Production of Ceramic UO2 Powder

The steps involved in production of reactor-grade ceramic UO2 powder include mining, refin-
ing, and conversion, as summarized below.
There are four methods of obtaining uranium from ores:
 Open-pit mining;
 Underground mining (the main method until the recent past);
 In-situ leaching (currently the leading method);
 Recovery from mining of other minerals such as copper, phosphate, or gold.
Regardless of how the uranium is obtained, it cannot be fed directly into a nuclear power
station because it contains mineral and chemical impurities, including radioactive daughter
products of uranium. Processing is required before uranium can be used efficiently to gener-
ate electricity.
If uranium is obtained by open-pit or underground mining, the first process after mining is
crushing (milling) of the ore, followed by dissolution in an acid or alkaline medium to sepa-
rate the uranium metal from unwanted rock and purification with chemicals which selec-
tively leach out the uranium. The uranium-rich solution is then chemically separated from
the remaining solids and precipitated out of solution. Finally, the uranium is dried. The re-
sulting powder is uranium oxide concentrate, U3O8, commonly referred to as “yellowcake”
due to the bright yellow colour that it frequently, but not always, exhibits. The unwanted
solids, called “tailings”, are stored at the mine site for safe disposal.
If the in-situ leaching method is used, an acidic or alkaline mining solution is passed directly
through the underground ore body by means of a series of bores or wells, and uranium is
brought to the surface in a dissolved state for purification. The uranium-rich solution is
treated as described above, the principal difference being that no tailings are produced by
this method.
Uranium meeting nuclear-grade specifications is usually obtained from yellowcake through a
Fuel Cycles 21
tributyl phosphate solvent-extraction process. First, the yellowcake is dissolved in nitric acid
to prepare a feed solution. Uranium is then selectively extracted from this acid feed by tri-
butyl phosphate diluted with kerosene or some other suitable hydrocarbon mixture. Finally,
uranium is stripped from the tributyl phosphate extract into acidified water to yield highly
purified uranyl nitrate, UO2(NO3)2. Uranyl nitrate is the starting material for conversion to
uranium dioxide (UO2) powder or to uranium hexafluoride (UF6). Both these conversion
routes conventionally begin with calcining the nitrate to UO3.
A large percentage of the UO3 produced in Canada is converted to uranium hexafluoride
(UF6) destined for enrichment plants for use in LWRs, and the remainder is converted to
natural uranium dioxide (UO2) ceramic-grade powder for use in CANDU reactors. The proc-
esses involved in producing natural uranium (NU) UO2 powder for CANDU fuel pellets are
described in more detail below. The UO3 is dissolved in nitric acid, and the resulting solution
is diluted and reacted with ammonium to precipitate ammonium diuranate (ADU). The liq-
uid is decanted, and the ammonium diuranate slurry is dried to remove the water. The re-
sulting ammonium diuranate powder is dried in a rotating kiln in hydrogen gas to produce
UO2 powder. The powder may then be subjected to further physical “conditioning” to im-
prove its mechanical “reactivity” during pellet sintering. Small batches of each powder lot
may be subjected to sintering trials as part of the final QA acceptance test before it is sent to
manufacturers of CANDU fuel bundles.
At each production stage mentioned above, the uranium compound is subjected to stringent
quality control surveillance to ensure that the final product meets industry technical specifi-
cations with respect to the content of specified impurities as well as certain physical charac-
teristics related to subsequent manufacturability. In-reactor behaviour of CANDU fuel bun-
dles has demonstrated that some aspects of fuel performance (e.g., fission product
behaviour) are related not only to powder characteristics, but also to the process parameters
used in producing the powder.
8.2 Fuel Bundle Manufacturing

A simplified CANDU fuel bundle manufacturing flow sheet is shown in Figure 3.
Manufacturing of bundles starts in two separate production streams: production of UO2 pel-
lets (left-hand side (LHS) of Figure 3), and production of sheath sub-assemblies and other
Zircaloy components (right-hand side (RHS) of Figure 3). These two production streams re-
quire such diverse processes and expertise that they are not only carried out in separate
production lines, but also may be carried out at different geographical locations.
Figure 3 CANDU fuel manufacture: assembly flow chart

The UO2 pellet production stream starts with receipt of UO2 powder (Step 1, LHS of Figure 3).
As received, the powder consists of very fine particles that tend to agglomerate and impede
rapid and consistent filling of the pellet pressing dies. To improve the “flow”
low” of the powder,
pow
it is subjected to two physical processes: compaction and granulation (Step 2). During this
step, a lubricant is also added to aid pellet pressing (Step 3), which is followed by pellet sin-
si
tering (Step 4). Sintering is one of the most critical steps in pellet production and requires
precise control of temperature, time, and the atmosphere used during the sintering cycle.
Pellet density and density distribution as well as physical integrity (no cracks or “chips”) are
critically dependentt on proper control of this process step.
As sintered, the pellets are hourglass
hourglass-shaped
shaped (see Chapter 17 for additional discussion of this
phenomenon). The sintered pellets are ground (Step 5) in a “cent “centreless”
less” grinder to (1) re-
r
move the hourglass shape an and
d produce a straight cylindrical pellet shape, (2) size the pellet
Fuel Cycles 23
to the desired diameter, and (3) impart a fine finish to the cylindrical pellet surface.
The ground pellets are sorted by diameter and placed into stacks (Step 6) ready for insertion
into sheath sub-assemblies, which come from the second stream. Pellet stack length is con-
trolled to tight tolerances because it is the principal means for ensuring adherence to the
required axial clearance within the fuel element. Pellet stacks and sheath sub-assemblies are
sorted by diameter and matched to satisfy the diametric clearance requirements. In addi-
tion, to prevent interference with the sheath, the pellets at both ends of the stack may be
tapered, or their diameter may be reduced. Additional discussion on this topic is provided
below.
Some of the greatest challenges in CANDU fuel production occur in the UO2 pellet production
stream. These challenges derive from the pellet technical specifications, some of which are
unique to CANDU fuel. A few of the major challenges are high pellet density; precise dimen-
sional control, especially of pellet diameter; small size of permissible visible surface flaws,
referred to as “chips”; and high quality of surface finish, especially the pellet cylindrical sur-
face (see Chapter 17 for additional discussion on pellet/sheath interaction). These chal-
lenges have been successfully met by manufacturers of CANDU fuel. This is substantiated by
the almost complete absence of defects ascribed to pellet manufacturing deficiencies, the
only notable exception being failures observed in “overstuffed” fuel elements, which were
intentionally manufactured to challenge the extreme limits of the relevant parameters:
maximum pellet density, minimum diameter clearance, and minimum axial clearance. Once
defects were witnessed in “overstuffed” elements, the practice was subsequently discontin-
ued.
The second manufacturing stream, the production of sheath sub-assemblies, starts with re-
ceipt of tubes finished to the required diameter and cut to length (Step 1, RHS of Figure 3).
Some manufacturers (including both Canadian manufacturers) have “vertically integrated”
Zircaloy tubing production into this stream, so that the stream actually starts with receipt of
zirconium tube hollow (TREX), which undergoes several stages of reduction and associated
processing (not shown in Figure 3), culminating in cut-to-length tubes.
In parallel with receipt of tubes, Zircaloy sheet/strip, bar/rod, wire, graphite slurry, and beryl-
lium are also received. The sheet (or wire) is used to manufacture spacers and bearing pads
(appendages). When manufactured from sheet, the appendages are punched and “coined”
into their final shape from the sheet after it is coated with beryllium. When manufactured
from wire, the appendages are cut to length, coined, and subsequently coated with beryl-
lium. The appendages are “tacked” to the tubes with spot welds (Step 2) and heated until
the beryllium coating and the Zircaloy metal in the appendage and tube form an alloy, i.e.,
“braze” (Step 3). Appendage dimensions, the shape of the coined surfaces, the thickness of
the brazing metal, and the location and tacking of the appendages on the tube must be con-
trolled to tight tolerances to enable the manufacturer to achieve the required dimensional
tolerances for the assembled bundle. The manufacturers have successfully met the chal-
lenges of this part of the manufacturing process, as evidenced by the absence of any fuel
failures caused by detached appendages or appendages causing bundle incompatibility
within the fuel channel.
Graphite slurry is diluted with water or an industrial alcohol, applied to the inside of the tube
sub-assemblies, and dried to form a coating of precisely controlled thickness and hydrogen
content referred to as “CANLUB” (Step 4). The effect of CANLUB on in-reactor performance
of fuel elements is discussed in Chapter 17. The ends of the sheath sub-assemblies are ma-
chined (Step 5) to (1) achieve the required final length, and (2) produce the profile required
for end-cap welding. In addition, CANLUB and any debris are removed from the end of the
sub-assembly lest they interfere with the subsequent welding process.
The two process streams (LHS and RHS in Figure 3) come together at the element assembly
step, i.e., the insertion of pellet stacks into the sheath sub-assemblies and the welding of
element end closures (endcaps). Endcaps are machined from bar or rod to match the sheath
sub-assembly weld preparation and to include other features (internal and external) which
are crucial to the final configuration of the element assembly. They are then welded to the
sheath using resistance welding, which is a thermo-electric process.
During resistance welding, the sheath is held by a collet. Heating during resistance welding
softens the sheath locally. The combination of the heating and cooling cycle that occurs dur-
ing the welding cycle and the restraint provided by the collet causes the sheath diameter to
shrink immediately adjacent to the weld. To maintain the minimum required diametric
clearance in this region of the element, the pellets at both ends of the stack have a reduced
diameter (usually tapered to mimic the sheath diameter profile).
After insertion of the pellet stack into the sheath sub-assembly, and before welding of the
end caps, the air trapped inside the sheath sub-assembly is purged with helium (or a mixture
of helium and argon). The end cap is then welded while the weld region is blanketed by inert
gas.
Similarly to the pellet production stream, the Zircaloy component stream has its own unique
CANDU-related challenges, the most unique and critical being the dimensional and process
parameters related to welding of endcaps to sheath sub-assemblies. To appreciate the im-
portance of this operation more fully, it is useful to remember that the largest CANDU reac-
tors have close to one-half million of these welds in the active part of the core and that one
defective weld is cause for concern, requiring urgent action to remove it from the core. In
comparison, LWR reactors have no discontinuities or joints of any type in fuel cladding within
the active part of the core. To make sure that the contribution of endcap welding to defect
statistics in CANDU reactors remains insignificant, defects due to endcap welding must be
maintained at or below the part-per-million (ppm) level, which is extremely challenging.
Careful examination of the fuel manufacturing flow chart, including the relevant component
design and process parameters, reveals more than 15 design and process parameters which
must be maintained within control limits and must come together at the precise moment the
weld is made. It is not surprising that of the very few defects in operating CANDU fuel which
have been ascribed to manufacturing causes, more than 90% are related to endcap welding.
Fuel integrity statistics for CANDU reactors and LWRs are discussed in Chapter 17; notwith-
standing the contribution from endcap welding to defects in CANDU fuel, defects in CANDU
fuel from all causes are significantly fewer than defects in LWR fuel (see Chapter 17).
Following endcap welding, the fuel elements are inserted into an assembly jig and the ele-
ment ends resistance-welded to endplates to form the final fuel assemblies, i.e., the fuel
bundles. End plates are stamped out of an incoming strip. Although this final process in it-
self is not as challenging as some of the processes described above, it is important to note
Fuel Cycles 25
that some of the dimensional requirements on the bundle assembly are equivalent in preci-
sion to those which are normally associated with machined components. To achieve this,
manufacturers must have stringent control, not only of the final assembly process, but also
of the processes used to produce the component parts, particularly element sub-assemblies
and endplates.
Manufacturing is the final step in the realization of a product, converting the intent specified
in the design drawings and technical specifications into functioning hardware. For one unit
of energy produced, current CANDU reactors use five to six times the mass of fuel used in
LWRs. Despite this relative disadvantage, fuelling cost per unit energy has always been much
lower for CANDU than for LWR. This can be attributed mainly to three factors: (a) CANDU
reactors avoid the high cost of enrichment, (b) CANDU reactors use 40% less mined uranium
per unit of energy produced than LWRs, and (c) CANDU fuel manufacturers have managed to
keep their fabrication costs very low while keeping fuel reliability very high. In addition, Ca-
nadian fuel manufacturers have also been very successful in transferring access to their
technology to all non-Canadian manufacturers of CANDU fuel.
9 Details of Selected Alternative Fuels and Fuel Cycles

From the ten possible alternative fuel cycles summarized in Section 3, selected five alterna-
tive cycles are described in greater detail in the sections below.
9.1 Natural Uranium Equivalent (NUE) Fuel

In all respects other than the sourcing of the fissile material (235U) and the starting amount of
fissile material (%235U in total U), this “cycle” is essentially the same as the natural uranium
(NU) cycle. The starting feedstock (UO2 ceramic-grade powder) for this cycle is a blend of
recycled enriched uranium (REU) and depleted uranium (DU) from enrichment plant tails.
The proportion of each component is determined by the respective content of 235U in each.
For instance, if the 235U content of REU is 0.9% 235U in total U and that of DU is 0.2% 235U in
total U, then the proportion of the two components per unit weight of the blend, with an
enrichment of 0.71% 235U in total U, can be calculated as follows:
If x is the fraction of REU in the blend, then the fraction of DU is (1 – x). If the enrichment of
the final mixture is 0.71%, then:
x*0.9 + (1 - x)*0.2 = 0.71. (7)
By solving Equation (7) for x, one can conclude that the blend needs to consist of 73% REU
and 27% DU. Similar calculations can also be performed for other combinations of starting
235
U contents of the constituent components. It is likely that the blended 235U fissile content
for this cycle will have to be very slightly higher than for NU to compensate for impurities in
the REU and DU, which are expected to be higher than the respective impurities in NU. The
exact 235U content in the blend would be determined for each batch of fuel procured, based
on the measured amount of impurities in each component.
Once the NUE powder has been obtained, the fuel bundle is manufactured using the same
processes that are used to produce NU fuel bundles, as described earlier.
Satisfactory performance of NUE fuel has been confirmed in “demonstration” irradiations in
a commercial power plant in China [Jioa et al., 2009].
9.2 Extended Burnup Fuel

As noted in Section 3, enriched uranium can be used for several purposes in a CANDU reac-
tor. One of them is to increase significantly the discharge burnup of CANDU fuel. This means
that a given amount of energy can be produced by fewer fuel bundles, reducing the amount
of spent fuel (i.e., high-level waste) that needs to be stored and disposed of.
To balance the above, one must consider (a) the increased cost of fuel, because enrichment
is expensive, (b) technical considerations in several key areas of fuel performance, such as
internal gas pressure, power ramps, corrosion, hydriding, deposits, end-temperature peak-
ing, bowing, high burnup structure, and (c) other effects related to plant operation, such as
the effect of enriched uranium on coolant void reactivity (CVR) and on refuelling operations.
Explanations of how fuel integrity is affected by some of these mechanisms have already
been discussed in Chapter 17 from the perspective of normal CANDU burnup. High burnup
exacerbates some of these, especially those discussed below. The sub-sections that follow
provide additional science that can aid in developing successful fuel designs for extended
burnup, which is a common feature of many of the alternative fuel cycles described in Sec-
tion 3.
From the perspective of maximum utilization of mined uranium, the optimal enrichment
(and hence burnup) for a CANDU reactor is around 1.2%, which is much lower than the typi-
cal enrichments in LWRs. Therefore, many of the performance challenges discussed below
for extended burnup fuel are significantly lower in CANDU fuel than in high-burnup LWR fu-
els.
However, in CANDU fuel, the high burnup occurs in conjunction with high element power,
which frequently exacerbates the challenge to fuel integrity. On the other hand, in CANDU
reactors, on-line fuelling provides a measure of flexibility in shaping the axial and radial
power profiles throughout the core. This can be used to help keep peak element ratings
within limits and reduce the size of power ramps resulting from refuelling.
9.2.1 Internal Gas Pressure
Extended burnup has the potential to increase, among other things, the internal gas pressure
within a fuel element (see Chapter 17). If excessive, the higher gas pressure can potentially
overstress and crack the fuel sheath. Some CANDU fuel elements indeed failed in the Bruce
reactor when they stayed in the reactor for unusually long periods and their internal pressure
inadvertently exceeded failure limits. Likewise, excessive internal pressure has also failed a
few fuel elements in experimental fuels as well [Floyd, 2001].
Three techniques are available to avoid such failures. First, fission gas release can be re-
duced, e.g., by reducing element ratings. Second, the void space in the element can be in-
creased, e.g., by using bigger chamfers, gaps, or plena [Floyd, 2001]. Third, stress concentra-
tions at critical locations can be reduced by using a larger notch radius at the sheath/endcap
junction [Tayal et al., 1993], see Figure 4.
Fuel Cycles 27
[Source: Page, 1976]

Figure 4 Illustrative notch at the sheath/endcap junction
9.2.2 Zircaloy ductility and power ramp
As noted in Chapter 17, a power ramp causes thermal expansion of a pellet,pellet which in turn
pushes the sheath. The sheath can crack if its strain exceeds its ductility in the corrosive
corros en-
vironment of fission products.
Irradiation reduces the ductility of Zircaloy significantly
significantly, as illustrated in Figure 5 [Tayal et al.,
1995]. For example, Figure 5 suggests that irradiation to 200 MWh/kg reduces the ductility
of Zircaloy by a factor of three to four compared to its initial as-fabricated
fabricated value.

Figure 5 Illustrative effect of irradiation on ductility of Zircaloy
Decreased ductility leads to lower tolerance to power ramps at extended burnups. In Chap- Cha
ter 17 (Section 8.2), the power ramp defect threshold of the sheath was considered in a nar-
row burnup range of 140 ± 20 MWh/kgU. Fuel can be operated to much higher burnups,
especially LWR fuel and extended burnup CANDU fuel. TTo o design safe operating conditions
for such fuel, we need to know the power ramp defect thresholds at the higher burnups. To
obtain an illustrative idea, let us use Figure 5 of this chapter to extend the defect threshold
curve of Figure 24 in Chapter 17 tto

o other burnups. The result, Figure 6, illustrates how higher
burnups can decrease resistance to power ramps in non non-CANLUB fuel. The decrease is ini- in
tially rapid, but slows down after burnup of about 100 MWh/kgU. Similar trends are ex- e
pected in CANLUB fuell as well.
Figure 6 Illustrative effect of burnup on power ramp defect threshold

Exercise: As an illustrative example, assume that a power ramp of 10 kW/m is being consid-
consi
ered in non-CANLUB fuel at a burnup of 300 MWh/kgU. Before the ramp, the fuel is operat-
opera
ing at 20 kW/m. (a) Will the fuel survive this ramp? (b) Will the fuel survive the same ramp
if its pre-ramp
ramp power is 30 kW/m?
Figure 6 shows that the first ramp is below the defect threshold, and therefore the fuel will
wil
survive it. The second ramp, however, is above the defect threshold
threshold; therefore,
therefore the fuel will
have a non-zero
zero probability of developing defects.
If the fuel’ss expected ramps are significantly higher than the defect threshold, appropriate
steps must be taken to increase the sheath
sheath’ss tolerance to power ramps. These could include,
for example, optimizing the shape of the pellet to decrease its thermal expansion; a thicker
coating of CANLUB to reduce the concentration of aggressive species of fission products
prod at
the sheath surface; or even an entirely new CANLUB formulation.
The decreased ductility of Zircaloy at extended burnup also affects several other areas of fuel
performance, for example the mechanical strength of the endplate during discharge. TheseThe
factors must all be considered carefully during the design process.
9.2.3 Corrosion
Higher burnup is usually associated with longer residence period
periods.. The latter can potentially
increase sheath corrosion. This tends to be much more pronounced at the outer surface of
the sheath, which is in contact with water.
Fuel Cycles 29
When part of the sheath corrodes, its ability to carry mechanical load decreases. Strength
assessments must account for this loss, and therefore we need to quantify sheath corrosion.
Furthermore, very thick layers of oxide tend to flake, causing a potential for debris and de-
posits in the heat transport system. In addition, flaking often leaves behind non-uniform
patches of oxides. These cause significant axial and circumferential variations in sheath tem-
perature, which in turn cause hydrogen and deuterium to diffuse and concentrate at local
peaks. The shape of the resulting hydrogen- (and deuterium)-rich areas often resembles a
lens, and therefore the product is often called a “hydrogen lens”. If local peaks of hydrogen,
hydrides, deuterium, and deuterides become pronounced, they can embrittle the sheath,
which in turn can significantly challenge its mechanical integrity. For all the above reasons,
the extent of corrosion must be kept within acceptable limits.
In CANDU reactors, lithium hydroxide (LiOH) is added to maintain coolant pH in the 10.5
range, which is intended to minimize deposition of iron-based system corrosion products
(“crud”) on both fuel surfaces and steam generator tubes [Barber et al., 1982]. From out-of-
reactor autoclave tests, several investigators have determined that oxidation of zirconium
alloys is not affected by the LiOH concentrations that are required to maintain coolant pH
near 10.5. Similarly, in-reactor, no acceleration of zirconium-alloy oxidation has been ob-
served in the pre-transition oxidation period, with additions of lithium hydroxide to control
pH in the 10.5 range. However, under very long exposures and high burnups which develop
relatively thick post-transition oxide films, some acceleration of oxidation has been observed
on fuel elements in light water reactors [Garzarolli et al., 1982; Cox, 1985]. The latter has
been attributed to possible concentration of the lithium hydroxide within the thick oxide
films due to temperature rise during heat transfer [Courtesy COG].
Waterside corrosion has indeed been identified in LWRs as a process that warrants the use of
more corrosion-resistant materials in the sheaths of very high burnup fuels. In CANDU reac-
tors, however, the burnup, and therefore the residence period, is far less, and therefore cur-
rent CANDU reactors are still quite far from requiring such an expensive change. Even in en-
riched CANDU fuel with burnup of about 30 MWd/kg, corrosion is not expected to be a
problem.
In summary, there are many detailed nuances in sheath corrosion. Nevertheless, for pur-
poses of this initial introduction, Figure 7 shows an illustrative progression of sheath corro-
sion with burnup. Note the large scatter in the data, which is typical. Also note that the
overall trends are similar in CANDU and PWR sheaths.

Figure 7 Waterside corrosion in the sheath
Example: How
ow much corrosion can be typically expected in a CANDU sheath (a) at 240
MWh/kgU; (b) at 480 MWh/kgU?
From Figure 7(b), we would expect an oxide layer of about 5 µm at 240 MWh/kgU and one of
about 15 µm at 400 MWh/kgU. These are both very thin layers and pose no risk of flaking.
Note the disproportionate increase in oxidation with burnup.
9.2.4 Deuterides and hydrides
Waterside corrosion liberates deuterium or hydrogen:
In heavy water: Zr + 2D2O = ZrO2 + 2D2 (8)
In light water: Zr + 2H2O = ZrO2 + 2H2. (9)
In this section, for simplicity, we will use the terms “deuterium/deuteride” to include generi-
cally “hydrogen/hydride” depending on context.
Some fraction of the liberated deuterium is dissolved in the sheath
sheath,, a process sometimes
also called “uptake”.. Over time, the deuterium diffuses due to gradients of concentration,
temperature, and stress and can accumulate preferentially at a few spots. When the local
concentration exceeds the solubility limit, deuterium precipitates in the form of deuteride
platelets,
atelets, referred to as “deuterides”. The solubility limit decreases with lower temperature,
and therefore deuterides tend to precipitate preferentially in relatively colder regions.
Depending on their orientation, deuterides can embrittle Zircaloy
Zircaloy, thereby
eby reducing its ability
to carry tensile loads. For this reason, deuteride levels must be maintained within accept-
able limits.
If the coolant is light water rather than heavy water, the same qualitative assessment applies,
Fuel Cycles 31
except that instead of deuteride platelets, hydride platelets (referred to as “hydrides”) are
formed. Figure 8 shows an illustrative example of hydrides in a sheath that was irradiated to
a burnup of about 600 MWh/kg HE in light water coolant. Note the relatively higher concen-
trations of hydrides near the outer surface because that surface is cooler than the inner one.
[Source: Karam et al., 2010(b)]

Figure 8 Illustrative distribution of hydrides in the sheath at about 25 MWd/kg HE
There is no evidence that excess hydrogen in the coolant causes primary defects in fuel-
bundle metallic components, or that it participates in the formation of secondary damage in
fuel metallic components. This may be because there is no thermodynamic mechanism for
excess H+ (or H2) to enter the matrix of metallic components. The mechanism for introducing
deuterium (not hydrogen) into the matrix of CANDU-bundle metallic components is depend-
ent on D2O coming into contact with an unoxidized metallic component, at which point oxy-
gen, having a greater affinity for the metal than deuterium, combines with the metal, causing
the D2O molecule to dissociate and leaving the liberated D+ radical (or D2) inside the matrix
of the metallic component. D+ radical (or D2) then dissolves within the metal matrix. Excess
D+ (or D2) within the metal then precipitates to form deuterides in the coldest part of the
metallic component, as explained earlier.
Typical values of hydrogen/deuterium (H/D) pickup in extended-burnup CANDU sheaths are
available from a compilation presented by Floyd [2001]. For comparison purposes, sheath D
concentrations are expressed in terms of equivalent H (=D/2). With the exception of silox-
ane-coated fuels, equivalent H concentrations up to 200 µg/g are widely observed, with most
of the data bounded by 150 µg/g. The data contain large variability with respect to burnup.
This suggests that besides burnup, other parameters such as coolant chemistry, neutron flu-
ence, coolant/sheath temperatures, and sheath properties exert significant influence on H/D
pickup.
9.2.5 Deposits
[This section was contributed primarily by John G. Roberts, JGRChem Inc. and Dr. Craig R.
Stuart, Canadian Nuclear Laboratories].
Deposits are sometimes observed on the external surfaces of fuel bundles and are some-
times also called “crud”. Most deposits on CANDU fuel tend to be magnetite-based com-
pounds. In BWRs, “scale-type” crud contains high concentrations of copper, up to and some-
times even in excess of 50%. In PWRs, deposits tend to be compounds based mainly on
nickel ferrite. Deposits on LWR fuels can also incorporate lithium and boron.
Deposits on fuel are undesirable because they can potentially impede heat transfer. In
BWRs, they have sometimes even caused through-wall holes in the sheath [IAEA, 2010].
Deposits on fuel are very rare in CANDU reactors, but relatively more frequent and also
thicker in BWRs [IAEA, 2010]. The causes of these deposits are rooted mainly in materials
and coolant chemistry, which differ in different types of reactors. This is a very complex sub-
ject, and its detailed treatment is beyond the scope of this chapter. However, pertinent ex-
perts have identified the following four significant differences between the various reactor
types that affect in-core deposits:
a) different materials among reactor types, e.g., carbon steel in the CANDU heat transport
system and stainless steels in LWRs;
b) presence of boron in PWR coolant and its absence in CANDU coolant;
c) higher degree of success in CANDUs in maintaining consistent chemistry over time in the
primary heat transport system (PHTS); and
d) much longer fuel residence period in an LWR (usually a few years) than in a CANDU reac-
tor (usually less than two years).
As for boron in the coolant, recall that CANDU reactors separate the coolant from the mod-
erator using pressure tubes and calandria tubes. This enables CANDU designers to maintain
separate chemistries in the coolant and in the moderator. Each system can then be focussed
on optimally addressing its own respective needs. For instance, in a CANDU reactor, the ad-
ditives required for neutron management are injected into the moderator, not the coolant,
and therefore the chemistry of the PHTS is not affected. In contrast, when boron is added in
an LWR’s moderator-cum-coolant to manage the neutronics, it also has a chemical impact on
the entire primary heat transport system (PHTS), including the steam generators and all the
pipes. If this promotes crud in the PHTS of an LWR, it is then carried by the coolant into the
core, where it can deposit on fuel surfaces. In this way, separation of moderator and coolant
in separate systems is one aspect that contributes to fewer crud deposits on CANDU fuel.
In summary, the lesser quantities of deposits on CANDU fuel than on PWR fuel are due prin-
cipally to different materials in the coolant circuit, lack of boron in the coolant, higher consis-
tency over time in coolant chemistry, and shorter fuel residence period.
9.2.6 End-temperature peaking
Consider two neighbouring fuel bundles in a CANDU channel. Their fuel stacks do not abut
each other; they are separated by the two endplates, endcaps, and axial gaps, as shown in
Figure 9. Because of the absence of UO2 in this region of bundle-to-bundle contact, neutron
Fuel Cycles 33
absorption cross sections are smaller here than in the bulk of the fuel bundle. Some fraction
of the unabsorbed neutrons from the bundle bundle-to-bundle junction—the “excess
excess” neutrons—
becomes available for captureture by neighbouring UO2. This results in an increase in neutron
flux in this region; the flux is higher a few centimet
centimetress into the fuel pellets on either side of
this junction than in the axial centre of the bundle. This is called end-flux peaking (EFP) and
is illustrated schematically in Figure 9.
Figure 9 Concepts for end-flux peaking

End-flux
flux peaking is even more severe at the very end of the fuel string because of the larger
step change in absorption cross section at that location, the “bundle-to-coolant
coolant contact”.
contact
The end-flux peaking factor is defined as the ratio of the local thermal flux at the end of the
fuel rod to the thermal flux at the centre of the rod. Figur
Figure
e 10 illustrates typical magnitudes
of end-flux
flux peaking. For bundle
bundle-to-bundle
bundle contact, thermal flux at the end of the pellet stack
can be about 20% higher than that at the centre of the rod. For bundle-to-coolant contact,
the local flux at the end can be considerably higher—some
some 80% higher than the flux at the
centre [Roshd et al., 1978].
Flux peaking in UO2 fuel leads directly to higher local heat generation and potentially also to
higher local temperature, depending on heat transfer in the end region of the fuel element.
This is called end-temperature
temperature peaking
peaking.. The higher temperature in turn is associated with
other detrimental effects such as higher thermal expansion and increased surface heat flux.
[After Roshd et al., 1978]

Figure 10 Illustrative peaking of flux at the end of a 28
28-element
element fuel bundle
From the perspective of local temperature, increased local heat generation can be partially,
or even wholly, offset by better heat transfer in the end region. The area near the end of the
pellet stack is not only cooled by radial transfer of heat, as in other locations, but also by ax-
a
ial transfer of heat through the eendcap, as shown in Figure 11.
Figure 11 EFP: neutron flux and heat transfer

During bundle-to-bundle
bundle contact
contact, the strong axial heat transfer usually depresses the local
temperature at the very end of the pellet stack
stack, despite EFP.
P. For this reason,
reason the peak local
temperature occurs not at the very end of the fuel stack
stack, but a short axial distance inward of
it, as shown in Figure 11. Therefore, although at high power levels, the local temperature in
Fuel Cycles 35
the end region does tend to inc

increase due to end-flux
flux peaking, it increases less than would be
expected from the amount of end
end-flux peaking alone.
Figure 12 illustrates a typical temperature distribution at high power in the end region during
bundle-to-bundle
bundle contact [Tayal, 1989(a)]. Isotherms in the end region are shaped like tear-
tea
drops, a shape determined by the net balance between local generation and local transfer of
heat. Teardrop-shaped
shaped voids are indeed observed in UO2 pellets that are irradiated at very
high power levels. In Figure 12 12, the local thermal flux in the end region is about 10% higher
than at the element’s mid-point
point. Due to axial heat transfer, the increase in peak local tem-
te
perature is limited to about 150
150°C in this case.
[Source: Tayal, 1989(a)]

Figure 12 End
End-temperature peaking: illustrative isotherm
In contrast, at low power level
levels and for small values of end-flux
flux peaking, e.g., during bundle-
bundle
to-bundle contact, axial transfer of heat is stronger than the local increase in heat flux. Un-
der such conditions, the local pellet temperature does not peak at all in the end region; in-i
stead, it decreases only near the very end of the pellet stack [Tayal, 1989(a)].
Enrichment increases the degree of end
end-flux
flux peaking and hence also the amount
am of end-
temperature peaking. This tends to decrease the margin to melting in enriched fuel [Girgis et
al., 1990].
9.2.7 Bowing
Bowing means lateral deformation. It is of two kinds
kinds: element bowing and bundle bowing.
Element bowing refers to lateral bending of individual fuel elements. Bundle bowing refers
to lateral bending of the fuel bundle as a whole or of a number of elements acting in concert.
In an experiment in Ontario, inte
intentional
ntional local dryout in a fuel element led to excessive local
bowing, which led to contact between the fuel sheath and a neighbouring component. The
latter in turn led to local overheating and therefore significant local oxidation in the region of
contact.. Failures of fuel sheaths due to bowing have also been reported in the WR1 experi-exper
mental reactor in Manitoba [Veeder et al., 1974].
Excessive outward bowing in the outer ring of elements has also led to “sticking
sticking” of the fuel
bundle in the channel of Mani
Manitoba’ss WR1 reactor, creating difficulties in removing the fuel
bundle [Veeder et al.,, 1974]. The WR1 reactor was cooled by organic oil which dissociated
into components, some of which were solids at operating temperature. Some of the solids
deposited on the bundle surfaces
surfaces, including the surfaces of matching element spacers. This
increased the overall diameter of the fuel bundle and caused it to stick in the WR1 reactor
channel.
In another experiment in a reactor in Ontario, excessive element bowing caused many cen-
tral bearing pads to press tightly against the pressure tube for a significant duration. This
trapped some coolant water between the bearing pad and the pressure tube. With time, the
trapped, stagnant water boiled off. This increased the local concentration of LiOH in the
coolant (LiOH is used to control coolant pH). At high concentration, an LiOH-water solution is
corrosive. This led to local corrosion in the bearing pads and also likely in the pressure tube.
This process is called crevice corrosion. There is always a potential for crevice corrosion at
the bottom bearing pads; excessive bowing can lead to crevice corrosion at other locations
around the bundle circumference if the bearing pads come into contact with the pressure
tube. Although a small amount of crevice corrosion is acceptable, its extent needs to be lim-
ited to reasonable levels lest it threaten the integrity of the pressure tube.
Excessive bowing can also alter the thermal hydraulics in the channel, which in turn can
complicate the evaluations and confirmations of heat transfer coefficients between the
sheath and the coolant. In light water reactors, excessive bowing of a fuel assembly has led
to concerns about control rod movements and dryout. Therefore, bowing and dryout also
have a mutual feedback cycle. For these reasons, it is important to limit bowing to accept-
able levels. Bowing can increase with time, and therefore it needs to be addressed in high-
burnup fuels.
When CANDU sheaths do bow, spacer pads prevent them from touching each other at their
mid-planes. Likewise, bearing pads at three axial locations reduce the likelihood of a sheath
touching the pressure tube in the absence of a severe local dry patch.
Major Driving Forces
Fabricated fuel elements are not necessarily perfectly straight, i.e., they may be bowed. In
addition, during normal operations, fuel elements initially bow due to non-uniform tempera-
ture around the circumference of a fuel element and also because of mechanical loads (a list
of causes of bowing is given later in this section).
Circumferential variations in temperature lead to differences in thermal expansion around
the sheath circumference. The hotter side of the fuel element becomes relatively longer
than the colder side. The fuel element then deforms laterally (bows) towards the longer side
to accommodate the extra length. Bowing due to thermal effects is illustrated schematically
in Figure 13(a).
Fuel Cycles 37
[Source: Tayal, 1989(b)]

Figure 13 Model for element bowing
Gravitational pull also acts to bend a horizontal fuel element. Once a fuel element has
bowed, a moment-arm arm is available to the axial forces generated by hydraulic drag; this mag-
ma
nifies the amount of bow. ThTherefore, the initial bow is the net result of the
he following major
influences:
 initial as-fabricated bow;
 thermal effects, e.g.,
 non-uniform
uniform heat transfer between the sheath and the coolant due to variations
in sub-channel
channel geometry and local flow conditions;
 non-uniform
uniform coolant temperature due to imperf
imperfect
ect mixing among sub-channels;
sub
local hotspots, if any (such as a dry patch);
 asymmetric heat production due to neutron flux gradients.
 mechanical effects:
 gravitational pull;
 axial load, e.g., hydraulic drag.
Across the diameter of a CANDU fuel bundle, the flux profile is approximately parabolic,
parabolic as
shown in Figure 14. That means that in any given fuel element, heat production may not
always be centred at the geometric centre of the fuel element. Figure 15 provides an illus-
trative example of this: the centre of grain growth has shifted in the direction of maximum
neutron flux. Asymmetric heat production will necessarily set up circumferential tempera-
ture variations in the pellet and in the sheath, contributing to their bowing.
[After Veeder et al., 1974]

Figure 14 Illustrative circumferential variation in flux of thermal neutrons
[Source: Veeder et al., 1974]

Figure 15 Off-centre grain growth
The loads mentioned above are resisted mainly by the flexural rigidities of the sheath and of
the pellets, by endplates, and by lateral restraints (if any) from the pressure tube and from
neighbouring fuel elements. Most of thethese loads as well as the major restraints are shown
Fuel Cycles 39
schematically in Figure 13(b).

After initial bowing has occurred, creep
creep, additional deposits on spacer pads,
pads or both can in-
crease the bow further over time. For example
example, Figure 16 shows post-irradiation
irradiation bow in a
fuel element measured after 100, 240
240, and 304 days in the reactor under similar irradiation
conditions. Note the significant increases in bow with continued irradiation. Therefore, the
design process needs to consider the potential for additional bowing in extended-burnup
extended
fuels.
[Source: Veeder et al., 1974]

Figure 16 Increase of bowing during continued irradiation
Simplified Model for Initial Thermal Bowing
Veeder et al. [1974] and Tayal [1989(b)] have outlined models for bowing calculations. As a
simplified illustrative example, let us consid
consider the empty fuel sheath shown in Figure 13(a)
that experiences a linear variation in temperature around its circumference and uniformly
along its length. In the absence of hydraulic drag load and gravity, the equation for thermal
bending of this beam can be written as [Gere et al., 1997]:
d2y/dx2 = α ΔT/d, (10)
where y is the transverse deflection, x is the axial distance, α is the coefficient of linear ther-
the
mal expansion of the sheath in the axial direction, ΔT is the difference in temperature across
a
the diameter, and d is the diameter.
Furthermore,, let us assume for simplicity that neighbouring fuel elements or pressure tubes
tube
do not impede lateral deflection of the fuel elements and that the endplates provide simple
supports to the fuel element.. Under these assumptions, the boundary and symmetry condi-
cond
tions simplify to:
y = 0 aat x = l/2, and dy/dx = 0 at x=0. (11)
Equation (10) can be integrated under boundary and symmetry conditions (11) to yield the
following equation [Gere et al

al., 1997] for unconstrained thermal bowing at the mid-plane,
mid δ:
δ = α ΔT l2/(8d). (12)
As a specific example, let us consider an empty sheath with a diameter of 13 mm, a length of
500 mm, and a coefficient of linear thermal expansion of 6.72 µm/(mm•°C).C). It experiences a
thermal variation of 5°CC around its circumference. Equation 12 predicts that in the absence
of mechanical loads, its initial mid
mid-span
span thermal bow would be about 0.08 mm.
The pellet operates at a much hotter temperature than the sheath, and its coefficient of
thermal expansion is about twice that of Zircaloy. Hence
Hence, a pellet can have a significant im-
i
pact on the thermal bow of a fuel element, especially if it is in tight diametric
diametr contact with
the sheath. Mechanical loads can also have a si significant impact on element bow. Their
quantitative treatments are complex and beyond the scope of this chapter; interested read-
rea
ers can explore several detailed papers that have been published on this subject, e.g.,
e.g. Veeder
et al. [1974] and Tayal [1989(b)
[1989(b)].
9.2.8 High burnup structure
Very high local burnup can cause the formation of “high burnup structure” (HBS) at the pe-
riphery of the pellets [Baron et al., 2009, 2012], also called the “rim effect” and “cauliflower
structure”. To date, this has been observed mainly in LWR fuels because their burnups tend
to be significantly higher than those of current CANDU fuels. The discussion below is based
largely on Baron et al.. [2012].
Figure 17 shows a typical HBS. It is believed to be formed by subdivision of pellet grains
when the local burnup typically exceeds 55
55–60 MW•d/kg (or 1300–1400
1400 MW•h/kg).
[Source: Baron et al., 2009, 2012]

Figure 17 High burnup structure near the rim of a pellet
In UO2 fuels, the local burnup has a sharp peak at the radial periphery of the pellet due to
“neutronic self-shielding” (also called the “skin effect”), as shown in Figure 18. Therefore,
Therefore in
UO2 fuels, HBS tends to be confined to a very narrow shell at the pellet
pellet’s radial
ial periphery.
Fuel Cycles 41
[Source: Tayal, 1987]

Figure 18 Illustrative heat generation rate across the radius of the pellet
Generation of HBS is a complex process and is described in detail in many publications such
as Baron et al. [2012]. At a very simplified level
level, irradiation
rradiation produces, among other things,
point defects and dislocation loops in the pellet. They do recover thermally, but the recovery
is relatively slow in the comparatively colder regions near the periphery of the pellet. There-
fore, the dislocations gradually become denser (more numerous) and more tangled. This
eventually causes the crystalline structure of the pellet to become unstable, initiating a re-
structuring driven by the energy stored in the material. This in turn results is subdivision of
grains, often into range of 0.2 to 0.5 µm diameter.
In comparison, typical grain diameter is about 10 µm in as
as-fabricated
fabricated CANDU pellets. By con-
co
sidering the ratio of volumes of as-fabricated vs. subdivided grains, i.e., (about 10 µm / (0.2
to 0.5 µm))3, one can deduce that HBS subdivides a typical grain into a few thousand smaller
grains.
Compared to the parent material, UO2 restructured by HBF is typically less dense and more
viscous.
9.3 Thorium Fuel

[Main source: World Nuclear Organization, WNA, 2014]
Thorium does not have a fissile isotope, which means that it does not undergo fission by
thermal neutrons. It is, nevertheless, a fertile material. This means that thorium absorbs
thermal neutrons to create, or “breed”, 233U, which in turn undergoes fission by thermal neu-
trons.
Section 2.2 summarizes the neutronics of how fertile 232Th converts to fissile 233U. Fission of
a 233U nucleus releases about the same amount of energy (200 MeV) as that of 235U.
During the early years of nuclear power development, interest in thorium was prompted by a
concern that economically available uranium resources would not be sufficient to support
the ambitious nuclear program then envisioned. Subsequently, new uranium resources have
been discovered that are thought to be sufficient to support the currently envisioned nuclear
program for at least a few decades. Development aimed at exploiting the natural advantages
of thorium and demonstrating the feasibility of its introduction into all known types of ther-
mal reactors (and breeder reactors) has continued, with considerable success. Some coun-
tries are planning to introduce commercial thorium-fueled reactors, mainly for strategic rea-
sons of energy independence. A comprehensive review of the work done internationally is
provided in [IAEA, 2005], which is recommended for additional reading.
Experts estimate that thorium is about three to four times as abundant as uranium. There-
fore, thorium can significantly moderate the cost of nuclear fuel in the future if it can be used
to generate electricity. In addition, thorium and uranium are distributed differently among
the various geographic locations on Earth. For example, countries such as India, China, and
Turkey have comparatively little uranium, but substantial amounts of thorium. Such coun-
tries may potentially find thorium especially attractive for power generation.
The preferred method of using thorium is in the form of thorium dioxide, also called thoria,
which is a ceramic. The reasons are analogous to those for using uranium in the form of UO2
(see Chapter 17). The following properties of thorium dioxide pellets contribute to improv-
ing various aspects of their in-reactor performance (compared to UO2 pellets) [Boczar, 2012]:
 Figure 19 shows that for a wide range of temperatures, the thermal conductivity of unir-
radiated thoria is about 25–35% higher than that of unirradiated UO2 [Pillai et al., 2000;
MacDonald et al., 1976]. Therefore, for a given power level, thorium pellets operate at
lower temperatures;
 The lower temperature of ThO2 reduces fission gas release, which reduces internal gas
pressure, which in turn helps achieve higher burnup;
 The melting temperature of unirradiated ThO2 is about 3350°C [IAEA, 2005], which is
among the highest of any oxide. This compares with a melting point of about 2840°C in
non-irradiated UO2. This increases the margin to melting in ThO2. It also introduces
challenges in pellet production, requiring a higher sintering temperature;
 Thorium has only one oxidation state; therefore, it has a higher resistance to oxidation;
 Thorium is chemically inert. Therefore, it can be expected to yield a more stable form of
waste that is expected to be more resistant to leaching.
Fuel Cycles 43
Sources: [UO2: MacDonald et al., 1976] [ThO2: Pillai et al.,

., 2000]
Figure 19 Thermal conductivities of unirradiated pellets
Test irradiations to date have confirmed the beneficial impacts of the these properties.
properties Com-
pared to UO2 fuel irradiated under similar conditions, (Th, Pu)O2 fuel has exhibited minimal
grain growth (as shown in Figure 20 20), benign fission gas release, and low sheath strain
[Karam et al.,., 2010(b); Livingston
Livingstone et al., 2013; Kupferschmidt et al.,
., 2013]. All these obser-
vations suggest that thorium pellets indeed operate at considerably lower temperature.
Extensive Columnar Grains Equiaxed Grains

(a) UO2 (65 kW/m) (b) (Th,Pu)O2 (67 kW/m)
[Source: Kupferschmidt, 2013]
Figure 20 Illustrative structures in irradiated pellets
9.3.1 Energy from thorium
The paragraphs below are based largely on and/or reproduced from Boczar et al. [2012],
IAEA [2005], and WNA [2014].
As illustrated in Section 2.2 , 232Th must first be converted into 233U for use in power genera-
tion. “Therefore, thorium fuel concepts require that 232Th be first irradiated in a reactor be-
fore fissile material can be made available. The 233U that is produced can either be chemically
separated from the parent thorium fuel and recycled into new fuel, or the 233U may be us-
able in-situ in the same fuel form”.
“Thorium fuels therefore need a fissile material as a ‘driver’ so that a chain reaction can be
initiated and maintained. The only fissile driver options are 233U, 235U, or 239Pu, none of which
is easy to supply”.
A number of configurations are possible for arranging thorium and a fissile driver in a reac-
tor; some have been used successfully to demonstrate the feasibility of thorium implementa-
tion. In a CANDU reactor, three illustrative configurations are the following [Boczar, 2012]:
 Homogeneous fuel. In this option, each pellet contains a mixture of thorium and the
fissile material.
 Heterogeneous “mixed bundles”. Such bundles contain thorium and the fissile material
in separate fuel elements.
 Heterogeneous “mixed channels”. In this option, each bundle contains either only tho-
rium or only the driver fuel. Each channel contains only one type of bundle.
“In the heterogeneous fuel arrangements, the high fissile (and therefore higher-power) fuel
Fuel Cycles 45
zone is called the seed region, and it is physically separated from the fertile (low-power) tho-
rium part of the fuel, called the blanket”.
“It is possible—but quite difficult—to design thorium fuels that produce more 233U in ther-
mal reactors than the fissile material they consume. Thermal breeding with thorium is only
really possible if neutron economy in the reactor is very good (i.e., low neutron loss through
escape or parasitic absorption). The possibility to breed significant amounts of fissile mate-
rial in slow (thermal) neutron systems is unique to thorium-based fuels and is not possible
with uranium fuels. Irrespective of whether breeding is achieved, more practical thorium
fuel cycles in a CANDU reactor offer a significant reduction in uranium requirements”.
“Another distinct option for using thorium is as a ‘fertile matrix’ for fuels containing pluto-
nium (and even other transuranic elements like americium). No new plutonium is produced
from the thorium component, unlike for uranium fuels, and so the level of net consumption
of this metal is rather high”.
“In fresh thorium fuel, all the fissions (thus power and neutrons) derive from the driver com-
ponent. As the fuel operates, the 233U content gradually increases, and it contributes more
and more to the power output of the fuel. The ultimate energy output from 233U (and hence
indirectly from thorium) depends on numerous fuel design parameters, including fuel bur-
nup attained, fuel arrangement, neutron energy spectrum, and neutron flux (affecting the
intermediate product protactinium-233, which is a neutron absorber)”.
Specific thorium cycles can be designed to suit a country’s specific circumstances.
9.3.2 Margin to melting
As an illustrative example, consider two neighbouring fuel elements, one fueled with UO2,
the other with ThO2. Both are operating at 60 kW/m; the coolant temperature is 300°C. If
end-temperature peaking is ignored, what are the margins to melting in the two fuel ele-
ments?
To work this out, we can use the theoretical background given in Chapters 6 and 7 about the
temperature distribution within a fuel element, along with some assumptions to simplify this
illustrative calculation. Recall that if other parameters are the same, to a first approximation,
the temperature difference between the centre and the surface of the pellet is inversely pro-
portional to the thermal conductivity of the pellet.
Guided by the previous illustrative examples in this book, assume that the UO2 fuel element
has a central temperature of about 1900°C and that the melting point of unirradiated UO2 is
about 2840°C. Therefore, in the UO2 fuel element, the thermal margin is about 2840 – 1900
= 940°C.
The peak operating temperature in the ThO2 pellet can be approximated as follows:
 Guided by the previous illustrative examples, assume that the temperature increase be-
tween the coolant and the pellet surface is about 100°C.
 Therefore, the temperature at the surface of the pellet is: 300 + 100 = 400°C.
 In the UO2 pellet, the temperature increase from the surface to the centre of the pellet
is 1900 – 400 = 1500°C.
 Guided by Figure 19, assume that to a first approximation, the thermal conductivity of
ThO2 is about 30% higher than that of UO2. Therefore in a thoria pellet operating at the
same power, the temperature increase from the surface to the centre of the pellet is
about 1154°C (=1500°C / (1 + 0.3)).
 Therefore, the temperature at the centre of the thoria pellet is 400 + 1154 = 1554°C.
 The melting point of ThO2 is about 3350°C [IAEA, 2005], and therefore the margin to
melting in the ThO2 pellet is about 3350 – 1554 = 1796°C.
Therefore, for the same operating power of about 60 kW/m, the margin to melting in ThO2
fuel is 1796°C, which is significantly higher than in UO2 fuel.
9.3.3 Internal gas pressure
Pellet temperature is a main driver of fission gas release. Above a threshold temperature,
fission gas release increases rapidly with local temperature.
In the preceding illustrative example, the UO2 fuel operates a few hundred degrees above
the threshold temperature for fission gas release. This results in significant fission gas re-
lease in UO2 fuel.
Assuming that the threshold temperature is similar in ThO2 and UO2, the ThO2 fuel described
above operates much closer to the threshold temperature for fission gas release. This means
that the ThO2 pellet will release much less fission gas.
The above illustrative examples point to some of the large benefits in operating performance
that are possible with thorium. Along with thorium’s impressive benefits, we do also need to
consider, as noted earlier, the higher costs of some thorium cycles.
9.4 Mixed Oxide (MOX) Fuel

MOX stands for Mixed Oxide, which means that the fuel contains a mixture of uranium and
plutonium (Pu).
Section 2.2 summarizes the neutronics of Pu generation and fission. A slow neutron can fis-
sion 239Pu into mainly barium and strontium. This releases 207 MeV of energy, which is very
similar to the energy released by fission of 235U (202 MeV).
In a real-world application of “swords-to-plowshares”, plutonium extracted from dismantled
nuclear warheads can be used to generate electricity as MOX fuel. Plutonium can also be
obtained by reprocessing used LWR fuel. Although feasible, reprocessing of used CANDU
fuel is seen as impractical because of the relatively small quantity of plutonium (and fissile
uranium) contained in used CANDU fuel. LWRs are already being used to dispose of military
plutonium, and MOX is a commercial reality in European LWRs. More detailed overviews of
the MOX cycle are described in many publications such as [Boczar, 2012; Dyck et al., 2005].
One interesting aspect of the common variant of MOX fuel is its tendency to form “agglom-
erates” arising out of non-homogeneities in the mixture of UO2 and PuO2. The amount of
agglomerates depends on the specific manufacturing process, as explained below.
Reasonably homogeneous mixtures of UO2 and PuO2 can be produced by vibratory milling of
UO2 and PuO2 powders in the proportions desired in the final fuel. The resulting homogene-
ity is expected to yield the best performance in MOX fuel. However, this process is quite
time- and resource- consuming. Alternative manufacturing processes which produce less
Fuel Cycles 47
homogenous mixtures require less effort, but they also contain agglomerates, meaning that
trade-offs
offs must be considered [Harrison et al., 2010].
For purposes of illustration, two such alternative processes are
are: [Harrison et al.,
al 2010]:
Alternative Method # 1: A master mix with high Pu concentration is prepared by vibratory
milling of the UO2 and PuO2 powders together. The required amount of this “master mix” is
added to additional UO2 powder to achieve the desired final Pu concentration. This ap-
proach results in discrete “islands”
“islands”—also called agglomerates—of
of master mix contained in
the UO2 matrix.
Alternative Method # 2: Same as Alternative Method # 1 above except that only PuO2 is
milled (rather than a mixture of PuO2 plus UO2). The resulting powder is then added to UO2,
resulting in nearly pure PuO2 “islands” distributed within the UO2 matrix.
The two alternative fabrication processes above create islands of relatively higher plutonium
concentration. The islands undergo relatively higher amounts of fission and produce
produc corre-
spondingly higher quantities of fission gases. Figure 21 shows a typical island in a fuel ele-
el
ment that was fabricated using Alternative Method # 1. The arrow points to the area of high
fission. Similar areas were also observed in pellets produced by the second alternative
method.
[Source: Harrison et al., 2010]

Figure 21 High fission area (HFA) in an irradiated MOX fuel pellet
Within an island, the pellet behaves as if it were operating at significantly higher local power.
This results in larger local fission gas release. This is partly offset in the remaining parts of
the pellet, which behave as if they are operating at a lower local power. Because fission gas
release increases non-linearly
linearly with local temperature, the overall fission gas release from
non-homogeneous
homogeneous pellets is greater than from homogeneous pellets. However, there is sig- si
nificant scatter in both the UO2 and the MOX data, and the limited MOX data to date in Can- Ca
ada are in the general range of scatter of the UO2 data.
Non-homogeneous
homogeneous MOX fuel also shows larger post
post-irradiation
irradiation sheath strain than homoge-
homog
neous fuel.
Note that despite the higher gas release and strain noted above, the overall performance of
MOX fuel is quite acceptable in many LWRs worldwide [IAEA, 2003].
9.5 Breeder Fuel

[Main source: Wikipedia, 2015]
In a typical thermal reactor fuel cycle, energy is extracted from 235U as well as from 239Pu, the
latter being bred from 238U. Because some fraction of the starting 235U remains unused when
the fuel is discharged, less than 1% of the available energy is currently extracted from mined
uranium.
The dominant isotope—over 99%—in natural uranium is 238U, which is not fissile. However, it
is fertile, which means that it can be converted into a fissile isotope (239Pu) through fast neu-
trons, and energy can then be extracted from it as well.
238
U can be sourced from virgin uranium, from the recycled uranium from reprocessing irra-
diated fuel, or from depleted uranium from enrichment plant tails.
As noted earlier, some amount of fertile 238U is routinely converted to fissile 239Pu in CANDU
reactors. One measure of a reactor’s performance is the conversion ratio—the ratio of the
rate at which fissile nuclides are produced in the entire reactor by conversion, to the rate at
which fissile nuclides are consumed through radiative capture and fission. All practical nu-
clear reactor fuels experience some degree of conversion.
As long as any amount of a fertile material exists within the neutron flux of the reactor, some
new fissile material is always being created. For example, commonly used light water reac-
tors have a conversion ratio of approximately 0.6. Pressurized heavy water reactors (PHWR)
running on natural uranium have a conversion ratio of 0.8. In a breeder reactor, the conver-
sion ratio is greater than one. “Breakeven” occurs when the conversion ratio is exactly one,
and the reactor produces exactly as much fissile material as it uses. (This is the basis of the
so-called self-sufficient equilibrium thorium CANDU reactor, which at equilibrium contains as
much 233U in the used fuel as is in the fresh fuel, accounting for reprocessing losses.)
Historically, breeder reactor development has focussed on reactors with low breeding ratios,
from 1.01 for the Shippingport reactor running on thorium fuel and cooled by conventional
light water to over 1.2 for the Russian BN-350 liquid-metal-cooled reactor. Theoretical mod-
els of breeders with liquid sodium coolant flowing through tubes inside fuel elements (“tube-
in-shell” construction) show that breeding ratios of at least 1.8 are possible.
Some experts have estimated that thousands of years’ worth of 238U is available world-wide
for use in fast breeder reactors.
10 Synergy between CANDU Reactors and LWRs

As mentioned in previous sections, the superior neutronic characteristics of CANDU reactors
can be used for a variety of synergies with LWRs, e.g.:
 extracting additional energy from used LWR fuel, e.g., through RU/NUE, through ex-
Fuel Cycles 49
tended burnup, through DUPIC, or through plutonium recycling (e.g., MOX), and/or
 destroying some of the long-lived actinides removed from spent LWR fuels through re-
processing.
Based on the fissile content of used LWR fuel, one can calculate the number of LWRs from
which the used fuel can be used as fuel by a single CANDU reactor [Boczar et al., 1996].
The quantities used in the example below are not necessarily actual values for a specific
CANDU or LWR reactor, although they are reasonably representative of the two reactor types
and are used to illustrate how the actual quantities would be used to determine the synergy
between an actual LWR and an actual CANDU reactor by substituting real, specific values.
LWR reactor:
 capacity: 1000 MW(e)
 total uranium in core: 80,000 kgU
 amount of U refueled annually: 20,000 kg (one-quarter of the core replaced annually)
235
 U contained in spent fuel (annually): 180 kg (0.9% 235U in total U).
CANDU reactor:
 capacity: 1000 MW(e)
 total uranium in core: 120,000 kg (6000 bundles containing 20 kg U each)
 annual requirements:
o annual refuelling: 140,000 kg of U (7000 bundles);
o annual 235U requirements: approximately 1,000 kg 235U (0.71% 235U in total of
140,000 kg U).
Synergy:
 annual 235U arising from spent LWR fuel: 180 kg
 annual 235U requirements for CANDU reactor of equal electrical capacity: 1,000 kg
 proportion of 235U used in CANDU fuel that is derived from LWR arising: 0.73% (see Sec-
tion 9.1 above); the remainder of 235U is derived from depleted U from tails of enrich-
ment plants.
 CANDU reactor’s annual 235U requirements supplied by LWR: 730 kg (73% of 1,000).
 Number of LWRs required to supply 235U for CANDU reactors of equal electrical capacity
(rounded to the nearest whole number): 4 (= 730/180).
The example above assumes that only 235U arisings would be used. If plutonium were also
used, the proportion of LWR per CANDU reactor would be accordingly reduced.
11 Closure
In summary, some illustrative fuels and fuel cycles are the natural uranium cycle, natural-
uranium equivalent fuel, extended burnup fuel, thorium, mixed oxide fuel, DUPIC fuel, low
void reactivity fuel, actinide burning fuel, and breeder fuel.
The natural-uranium fuel cycle that is currently used in CANDU power reactors is simple, reli-
able, and low-cost. It is also very amenable to localization, which promotes national energy
independence.
Alternative fuel cycles offer several potential strategic advantages, such as significantly ex-
tending humanity’s energy resources, reducing the initial capital cost of a power plant, in-
creasing the lifespan of a CANDU power plant, increasing the thermal efficiency of the steam
cycle, reducing the coolant void reactivity, reducing the volume of spent fuel, and increasing
the power output from a fuel bundle and the reactor.
Many alternative fuel cycles result in more expensive fuel, additional challenges in the front
end, enhanced challenges during the service period, and reduced challenges in the back end.
Therefore, a careful assessment of net benefit is required before implementing any particular
fuel cycle.
12 Problems
Section 1 – Introduction
Q1.1 – Why would one consider fuel cycles other than the once-through cycle?
Section 2 – Neutronics of Fuel Cycles
Q2.1 – Define fissile nuclides and give examples. Which is the only naturally occurring fissile
nuclide?
Q2.2 – Define fissionable nuclide and provide examples.
Q2.3 – Define fertile nuclides and provide examples.
Q2.4 – What is the conversion ratio as related to nuclides?
Q2.5 – In the context of conversion ratio, what is breeding and breeding gain?
Q2.6 – What is burnup as applied to nuclear fuel?
Section 3 – Overview of Some Possible Fuels and Fuel Cycles
Q3.1 – List the fuel cycles which are used, or can be considered for use, in CANDU rectors
and provide a brief rationale for using each.
Section 4 – Key Drivers for Advanced Fuel Cycles
Q4.1 – List the principal advantages of the once-through natural uranium (NU) fuel cycle
used in CANDU compared to fuel cycles using enriched uranium.
Q4.2 – With reference to the three parts of the fuel cycle, provide a general assessment of
the advantages or disadvantages of alternative cycles compared to the once-through
CANDU-NU fuel cycle.
Q4.3 – Compared to the NU fuel cycle used in CANDU, many alternative fuel cycles tend to
increase the cost of fuel and also pose various degrees of additional technical challenge.
Nevertheless, they are considered for use in CANDU because they offer other advantages.
List these advantages and provide a brief explanation of each.
Fuel Cycles 51
Section 5 – Resource Extension: A Major Incentive

Q5.1 – Compared to LWRs, CANDU reactors can better use alternative fuel cycles, principally
because of their superior neutron efficiency. How much more efficient is CANDU than an
LWR in the use of uranium?
Q5.2 – In current CANDU reactors, peak neutron efficiency occurs at about 1.2% enrichment.
If 1.2% enriched uranium is used in CANDU reactors, how long will the planet’s identified
uranium resources (mentioned above) last?
Section 6 – Other Considerations
Q6.1 – List the considerations that have to be taken into account when deciding whether or
not to “recycle” spent fuel.
Section 7 – Pertinent Features of the CANDU Reactor
Q7.1 – CANDU reactors can use alternative fuel cycles principally because of their superior
neutron efficiency. List the seven features of the CANDU reactor that enable it to have supe-
rior neutron efficiency.
Q7.2 – Explain how the design of CANDU fuel facilitates the adoption of alternative fuel
cycles.
Section 8 - Natural Uranium Fuel Cycle
Q8.1 – List the four methods by which natural uranium can be obtained.
Q8.2 – Unlike other types of fuels (e.g., coal, gas, or oil), as-mined uranium ore cannot be fed
directly into a nuclear power station. Explain why.
Q8.3 – Produce a simplified flow sheet of the processes involved in the manufacture of
CANDU fuel bundles.
Q8.4 – Why is as-received uranium powder subjected to two physical processes, compaction
and granulation?
Q8.5 – What are the three process parameters that the operator controls during pellet sin-
tering?
Q8.6 – Why are pellets ground after sintering?
Q8.7 – Pellet stack length is the principal means by which axial clearance is maintained inside
the finished fuel element. What are the other means?
Q8.8 – Why do the pellets at the ends of the pellet stack have a reduced diameter?
Q8.9 – List the four features of uranium dioxide pellets that render their production challeng-
ing.
Q8.10 – Similarly to the pellet production stream, the Zircaloy component stream has its own
unique CANDU-related challenges, the most unique and critical being welding of endcaps to
sheath sub-assemblies. Explain why endcap welding is a critical and challenging process.
Q8.11 (a) – Identify five design and/or process parameters that must be controlled to achieve
a good endcap weld; Q8.11 (b) for bonus marks, identify an additional five parameters;
Q8.11 (c) for additional bonus marks, identify five more parameters.
Q8.12 – For one unit of energy produced, current CANDU reactors use five to six times the
mass of fuel as is used in LWRs. Despite that relative disadvantage, fuelling cost per unit
energy has always been much lower for CANDUs than for LWRs. List the three reasons for
this.
Section 9 – Details of Selected Alternative Fuels and Fuel Cycles
Q9.1 – A fuel manufacturer has purchased REU fuel containing 1% 235U and DU fuel contain-
ing 0.4% 235U. He wishes to blend them to obtain NUE fuel. What should be the fraction of
REU fuel in the blend?
Q9.2 – List the benefits which can be derived from using enriched fuel in a CANDU.
Q9.3 – List the commercial and technical aspects that must be considered in increasing ele-
ment rating, burnup, and bundle power by using enrichment.
Q9.4 – Why is it important to control fuel-element internal gas pressure, and what can fuel
designers do to achieve this?
Q9.5 – A power ramp of 15 kW/m is contemplated in CANLUB fuel at a burnup of 400
MWh/kgU. The pre-ramp power is 20 kW/m. Is this fuel element expected to survive the
power ramp?
Q9.6 – List at least two reasons for keeping sheath corrosion below the allowable limit.
Q9.7 – List the reasons for the reduced amount of corrosion on CANDU fuel compared to
PWR fuel
Q9.8 – In the example of Section 9.2.3, what is the expected thickness of waterside corrosion
in a CANDU sheath at 800 MWh/kgU?
Q9.9 – What causes end-flux (and end-temperature) peaking?
Q9.10 – Pellet/sheath interaction causes an unrestrained thermal bow of 0.6 mm in a fuel
element that has a diameter of 13 mm and a length of 500 mm. What is the equivalent tem-
perature variation around its circumference?
Q9.11 – What are the two main reasons for considering the use of thorium in place of ura-
nium?
Q9.12 – Which is the fissile isotope in thorium?
Q9.13 – The preferred method of using thorium is in the form of thorium dioxide, also called
thoria, which is a ceramic. The reasons are analogous to those for using uranium in the form
of UO2 (see Chapter 17). List the properties of thorium dioxide pellets that contribute to
improving their in-reactor performance compared to UO2 pellets.
Q9.14 – Describe three ways that thorium and the fissile driver can be arranged to power a
CANDU reactor.
Q9.15 – (a) What are the two possible sources of plutonium that can be used in CANDU? (b)
Is reprocessing of spent CANDU fuel a likely source of plutonium?
Q9.16– What is a common challenge in the production of MOX?
Fuel Cycles 53
Q9.17 – Why are “agglomerates” undesirable in MOX pellets?

Section 10 – Synergy between CANDU Reactors and LWRs
Q10.1 - Assume that the waste of LWRs contains 0.9% 235U and 0.6% Pu. Further assume
that DU can be procured with 235U content at any value in the range from 0.2%235U to
0.4%235U. We wish to use the 235U and Pu from LWR to feed a CANDU of equal capacity, after
blending with DU. If each CANDU and each LWR is 1000 MW(e), determine the required
content of 235U in DU, such that a single LWR will feed one CANDU of equal capacity. Solving
this problem highlights and reinforces six (6) aspects covered in this chapter. What are they?
13 References
D. Barber, D. H. Lister, 1982. “Chemistry of Water Circuits in CANDU Reactors”, Proceedings,
Symposium on Water Chemistry, Report IAEA-SM-264/15 and AECL-8217, 1982.
D. Baron, M. Kinoshita, P. Thevenin, R. Largenton, 2009. “Discussion about HBS Transforma-
tion in High-Burnup Fuels”, Nuclear Engineering and Technology, Vol. 41, No. 2, pp. 199-214,
March 2009.
D. Baron, L. Hallstadius, 2012. “Fuel Performance of Light Water Reactors (Uranium Oxide
and MOX)”; Chapter 40 in “Comprehensive Nuclear Materials”, book edited by R. Konings,
Elsevier, ISBN-9780080560274, 2012.
Z. Bhatti, B. Hyland, G. Edwards, 2013. “Minor Actinide Transmutation in Thorium and Ura-
nium Matrices in Heavy Water Moderated Reactors”, Proceedings, Global 2013, Salt Lake
City, Utah, U.S.A., September 29–October 3, 2013.
P. G. Boczar, J. D. Sullivan, H. Hamilton, Y. O. Lee, C. J. Jeong, H. C. Suk, C. Mugnier, 1993. “Re-
covered Uranium in CANDU—A Strategic Opportunity”, Proceedings of International Nuclear
Conference and Exhibition (INC 93), Toronto, Ontario, Canada. October 3-6, 1993.
P. G. Boczar, P. J. Fehrenbach, D. A. Meneley, 1996. “CANDU Fuel Cycle Options in Korea”,
Proceedings, KAIF/KNS Annual Conference, Seoul, Korea, April 1996; also Atomic Energy of
Canada Limited, Report AECL-11586, 1996.
P. G. Boczar, J. D. Sullivan, 2004. “Low Void Reactivity Fuel”, Proceedings, 25th Annual CNS
Conference, Toronto, 2004.
P. G. Boczar, J. T. Rogers, D. H. Lister, 2010. “Considerations in Recycling Used Natural Ura-
nium Fuel from CANDU Reactors in Canada”, Proceedings, 31st Annual Conference of the Ca-
nadian Nuclear Society, Montreal, Quebec, Canada, May 24–27, 2010.
P. Boczar, 2012. “CANDU Nuclear Reactor Design, Operation, and Fuel Cycles”, Chapter 11,
“Nuclear Fuel Cycle Science and Engineering”, Woodhead, 2012.
B. Cox, 1985. “Assessment of PWR Waterside Corrosion Models and Data”, EPRI Report NP-
4287, 1985.
G. R. Dyck, T. Mochida, T. Fukasawa, 2005. “Application of Fluoride Volatility to the Recycling
of LWR Spent Fuel into CANDU”, Proceedings, Global 2005, Paper number 493, Tsukuba, Ja-
pan, October 9–13, 2005.
M. R. Floyd, 2001. “Extended-Burnup CANDU Fuel Performance”, Proceedings, International
Conference on CANDU Fuel, CNS, Kingston. Also Atomic Energy of Canada Limited, Report
AECL-CONF-01135, 2001.
W. J. Garland, 2005. EP4D03 course notes, Chapter: Core Composition Changes, McMaster
University, http://www.nuceng.ca/ep4d3/ep4d3home.htm, 2005.
F. Garzarolli, W. Jung, H. Schoenfeld, A. M. Garde, G. W. Parry, P. G. Smerd, 1982. “Waterside
Corrosion of Zircaloy Fuel Rods”, Electric Power Research Institute, U.S.A., Report EPRI-NP-
2789, 1982.
J. M. Gere, S. P. Timoshenko, 1997. “Mechanics of Material”, PWS, 4th edition, 1997.
S. Girgis, N. Singhal, M. Tayal, 1990. “Influences of Load Following Partial Dryout and End-
Flux Peaking on Temperature Distributions in CANDU Fuel”, Proceedings, 11th Annual Confer-
ence, Canadian Nuclear Society, June 3–6, 1990.
N. F. Harrison, T. Janathasing, F. C. Dimayuga, 2010. “Post-Irradiation Examination of MOX
Fuel with Varying Plutonium Homogeneity”, Proceedings, 11th International Conference on
CANDU Fuel, Sheraton Fallsview Hotel and Conference Centre, Niagara Falls, Ontario, Canada,
CW-124920-CONF-004, October 17–20, 2010.
I. Hastings, P. G. Boczar, C. J. Allan, M. Gacesa, 1991. “Synergistic CANDU-LWR Fuel Cycles”,
Korea Nuclear Society, Annual Conference, 1991. Also Atomic Energy of Canada Limited,
Report AECL-10390, 1991.
B. Hyland, B. Gihm, 2010. “Scenarios for the Transmutation of Actinides in CANDU Reactors”,
Proceedings, 18th International Conference on Nuclear Engineering (ICONE18), Paper
ICONE18-30123, Xi’an, China, May 17–21, 2010.
IAEA, 2003. “Status and Advances in MOX Fuel Technology”, International Atomic Energy
Agency, Vienna, Austria, Technical Report Series, Report TRS 415, 2003.
IAEA, 2005. “Thorium Fuel Cycle”, International Atomic Energy Agency, Report IAEA-TECDOC-
1450, 2005.
IAEA, 2010. “Review of Fuel Failures in Water-Cooled Reactors”, International Atomic Energy
Agency, Vienna, IAEA Nuclear Energy Series, report number NF-T-2.1, ISBN 978-92-0-102610-
1. 2010.
Y. Jioa, D. Li, M. Chen, Z. Meng, J. Wang, C. Cottrell, S. Kuran, 2009. “Test Irradiation of Recy-
cled Uranium in Chinese CANDU Reactors”, Proceedings, 30th Annual Conference of the Ca-
nadian Nuclear Society, Calgary, Alberta, Canada, May 31–June 3, 2009.
M. Karam, N. Harrison, J. Montin, 2010(a). “Post-Irradiation Examination of DUPIC Fuel”, Pro-
ceedings, 11th International Conference on CANDU Fuel, Sheraton Fallsview Hotel and Con-
ference Centre, Niagara Falls, Ontario, Canada, October 17–20, 2010.
M. Karam, F. C. Dimayuga, J. Montin, 2010(b). “Post-Irradiation Examination of CANDU Fuel
Bundles Fueled with (Th, Pu)O2”, Proceedings, 11th International Conference on CANDU Fuel,
Sheraton Fallsview Hotel and Conference Centre, Niagara Falls, Ontario, Canada, October 17–
20, 2010.
W. C. H. Kupferschmidt, 2013. “Advanced Fuel Development at AECL: What does the Future
Fuel Cycles 55
Hold for CANDU Fuels/Fuel Cycles?”, Proceedings, 12th International Conference on CANDU
Fuel, Canadian Nuclear Society, Kingston, Ontario, Canada, September 15–18, 2013.
J. R. Lamarsh, 1996. “Introduction to Nuclear Reactor Theory”, Addison-Wesley, 1966.
S. J. Livingstone, M. R. Floyd, 2013. “Thoria Irradiation and Post-Irradiation Examination Ex-
perience at AECL”, Proceedings, 12th International CANDU Fuel Conference, CNS, Kingston,
2013.
P. E. MacDonald, L. B. Thompson (eds.), 1976. “A Handbook of Materials Properties for Use in
the Analysis of Light Water Reactor Fuel Rod Behaviour”, MATPRO-Version 09, TREE-NUREG-
1005, 1976.
D. A. Meneley, 2006. “Transition to Large-Scale Nuclear Energy Supply”, Proceedings, 27th
Annual CNS Conference, Toronto, 2006.
P. Ottensmeyer, 2012. “CANDU Fuel Waste Re-Used, Recycled, Eliminated: $48 Trillion Car-
bon-Free Electricity via Fast-Neutron Reactors”, Engineering Dimensions, Professional Engi-
neers of Ontario, Volume 33, No. 4, July/August 2012.
R. D. Page, 1976. “Canadian Power Reactor Fuel”, Atomic Energy of Canada Limited, Report
AECL-5609 (1976).
C. G. S. Pillai, P. Raj, 2000. “Thermal Conductivity of ThO2 and Th0.98U0.02O2”, Journal of Nu-
clear Materials, Vol. 277, pp. 116-119, 2000.
P. Reid, M. Gacesa, M. Tayal, 2008. “The ACR-1000 Fuel Bundle Design”, Proceedings, 10th
International Conference on CANDU Fuel, Canadian Nuclear Society, 2008.
M. H. M. Roshd, H. C. Chow, 1978. “The Analysis of Flux Peaking at Nuclear Fuel Bundle Ends
using PEAKAN”, Atomic Energy of Canada Limited, Report AECL-6174, 1978.
M. Tayal, 1987. “Modelling CANDU Fuel under Normal Operating Conditions: ELESTRES Code
Description”, Atomic Energy of Canada Limited, Report AECL-9331, February 1987.
M. Tayal, 1989(a). “The FEAT Finite Element Code to Calculate Temperatures in Solids of Arbi-
trary Shapes”, Nuclear Engineering and Design, Vol. 114, pp. 99–114, 1989.
M. Tayal, 1989(b). “Modelling the Bending/Bowing of Composite Beams such as Nuclear
Fuel: The BOW Code”, AECL-9792, Nuclear Engineering and Design, Vol. 116, pp. 149–159,
North-Holland, Amsterdam, 1989.
M. Tayal, K. D. Hallgrimson, R. Sejnoha, P. N. Singh, R. DaSilva, 1993. “Elastic-Plastic Stress
Distributions near the Endcap of a Fuel Element”, Proceedings, 4th International Conference
on Simulation Methods in Nuclear Engineering, Canadian Nuclear Society, Montreal, Canada.
Also Atomic Energy of Canada Limited, Report AECL-10646, June 1993.
M. Tayal, B. Wong, Y. Shudoh, 1995. “Effect of Radial Power Profile on Endplate Integrity”,
Proceedings, 4th International Conference on CANDU Fuel, Canadian Nuclear Society, Pem-
broke, Canada. Also Atomic Energy of Canada Limited, Report AECL-Conf-00452, 1995.
J. Veeder, M. H. Schankula, 1974. “Bowing of Pelletized Fuel Elements: Theory and In-Reactor
Experiments”, Nuclear Engineering and Design, Vol. 29, pp. 167–179, 1974.
Wikipedia, 2015. “Breeder Reactor”, http://en.wikipedia.org/wiki/Breeder_reactor, January
2015.
WNA, 2014. “Thorium”, World Nuclear Association, http://www.world-nuclear.org/info/

inf62.html, September 2014.

Chapter 8 provides an overview of fuel bundle configuration. Chapters 3 to 5 explain neutron
physics that generates heat in the fuel. Chapters 6 and 7 explain how that heat is removed
from the fuel bundle and illustrate the internal temperature distribution within a fuel rod.
Chapters 14 and 15 explain chemical and metallurgical aspects that relate to corrosion of the
fuel sheath. Chapter 13 explains the performance of fuel during postulated accidents. Chap-
ter 17 describes the design and performance of fuel, focussing on the current natural-
uranium cycle (i.e., low burnup). The current chapter (#18) describes a few selected fuel
cycles, summarizes key aspects of fuel manufacturing related primarily to the natural-
uranium fuel cycle, and explains a few selected aspects of fuel performance that become
relatively more important for advanced fuel cycles. Finally, Chapter 19 describes interim
storage and disposal of used fuel.
15 Acknowledgements
The authors thank Dr. Peter Boczar for permission to borrow heavily from his extensive publi-
cations on fuel cycles. Sections 2.1 and 2.2 of this chapter were provided by Dr. Dorin Ni-
chita, Dr. Ben Rouben, and Dr. William Garland. Dr. William Garland also contributed Figures
1 and 2. Dr. D. Baron (EdF, retired) graciously provided material on high burnup structure.
John Roberts and Dr. Craig Stuart contributed key material on fuel deposits. The above con-
tributions are gratefully acknowledged with thanks.
In addition, the following reviewers are gratefully acknowledged for their hard work and ex-
cellent comments during the development of this chapter; their feedback has much im-
proved it:
Peter Boczar
Bronwyn Hyland
Geoff Edwards
William Garland
Of course the responsibility for any errors or omissions lies entirely with the authors. Thanks
are also extended to Simon Oakley and Media Production Services, McMaster University, for
nicely drawing the figures, and to Diana Bouchard for expertly editing and assembling the
final copy.
1
CHAPTER 19
Storage and Disposal of Irradiated Fuel
prepared by
Summary
Nuclear wastes, in particular irradiated nuclear fuel, must be handled, stored, and placed into
permanent disposal facilities safely to prevent harm to people and the environment.
Radioactive wastes can be grouped into four classes: (i) low-level radioactive waste, (ii)
intermediate-level radioactive waste, (iii) high-level radioactive waste such as irradiated nuclear
fuel, and (iv) uranium mine and mill waste. Storage and disposal of irradiated fuel in Canada
follows a three-step process: storage in water-cooled pools, storage in air-cooled storage
cylinders, and final disposal. The two stages of storage (the first two steps above) are fully
proven and have been in practical operation for some time. The associated technical challenges
that must be addressed and the engineered solutions are discussed in this chapter. Several
configurations for final disposal have been shown to be technically feasible. Public acceptance
and implementation remain to be achieved; ongoing work aimed at achieving these goals is
discussed. Nature’s “reactors” that existed billions of years ago are examined for useful
analogies that can be applied to engineered disposal facilities.
©UNENE, all rights reserved. For educational use only, no assumed liability. Storage and Disposal of Irradiated Fuel – December 2015
Table of Contents
1 Introduction ............................................................................................................................ 4
2 Types of Waste ........................................................................................................................ 4
3 Strategies to Manage Radioactive Waste ............................................................................... 5
4 Irradiated Fuel Storage and Disposal: Main Considerations................................................... 6
5 Fuel Integrity ........................................................................................................................... 8
5.1 Requirements.................................................................................................................. 8
5.2 Key Drivers and Precursors ............................................................................................. 9
5.3 Fuel Damage Mechanisms ............................................................................................ 12
6 Wet Storage at Reactor Sites................................................................................................. 14
6.1 Flow of Fuel Bundles Through a CANDU....................................................................... 14
6.2 Irradiated Fuel Bundles Temporarily in Air ................................................................... 16
6.3 Throughput of Irradiated Fuel Bundles......................................................................... 16
6.4 Irradiated Fuel Bay: Typical Dimensions ....................................................................... 17
6.5 Irradiated Fuel Bay: Heat Removal................................................................................ 18
6.6 Corrosion of Irradiated Fuel in the Pool........................................................................ 19
6.7 Shielding of Irradiated Fuel in the Pool......................................................................... 20
6.8 Criticality Not an Issue for CANDU Irradiated Fuel ....................................................... 20
7 Dry Storage at Reactor Sites.................................................................................................. 20
7.1 Dry Storage Containers ................................................................................................. 20
7.2 Concrete Canisters ........................................................................................................ 20
7.3 Storage Baskets ............................................................................................................. 21
7.4 Transfer of Bundles from the Pool into Baskets............................................................ 22
7.5 Heat Transfer................................................................................................................. 22
7.6 Fuel Integrity ................................................................................................................. 22
7.7 Shielding........................................................................................................................ 23
8 Final Disposal or Isolation ..................................................................................................... 23
8.1 Strategy ......................................................................................................................... 23
8.2 Phases of Disposal......................................................................................................... 24
8.3 Repository ..................................................................................................................... 25
8.4 Used Fuel Containers .................................................................................................... 27
8.5 Buffer Material.............................................................................................................. 27
8.6 Rock Temperature ......................................................................................................... 28
8.7 Health Protection.......................................................................................................... 29
8.8 Breach of Containment ................................................................................................. 32
8.9 Nature’s Analogues for a Waste Repository ................................................................. 33
9 Problems ............................................................................................................................... 35
10 References............................................................................................................................. 37
11 Further Reading Material...................................................................................................... 39
Storage and Disposal of Irradiated Fuel 3
List of Figures
Figure 1 Decay power of an irradiated fuel bundle ...................................................................... 10

Figure 2 Illustrative temperatures in an LWR fuel and canister.................................................... 11
Figure 3 Gas volume inside a fuel element................................................................................... 12
Figure 4 Flow of fuel through a CANDU reactor ........................................................................... 15
Figure 5 A typical irradiated fuel bay ............................................................................................ 19
Figure 6 Overview of MACSTOR.................................................................................................... 21
Figure 7 Basket for dry storage of fuel ......................................................................................... 22
Figure 8 Concept of a deep geological repository ........................................................................ 26
Figure 9 Containers for Permanent Disposal of Irradiated Fuel ................................................... 28
Figure 10 Decay of radioactivity in CANDU fuel............................................................................ 30
Figure 11 Toxicity in ground water................................................................................................ 31
Figure 12 Dose rate at the surface of a repository ....................................................................... 33
1 Introduction
This chapter focuses mainly on the challenges that must be met during storage and disposal of
high-level nuclear waste, specifically storage and disposal of irradiated fuel (also called “spent”
fuel or “used” fuel). Some illustrative potential solutions are also summarized.
To this end, Sections 2 to 5 describe the types of wastes, the overall strategy to manage them,
the associated design considerations, and the potential mechanisms for damage to fuel
integrity. Then each major phase of managing irradiated fuel is described: initial wet storage at
reactor sites (Section 6), followed by dry storage at reactor sites (Section 7) and final disposal
(Section 8). IAEA [2003] provides an extensive glossary of relevant terms.
This chapter draws heavily from previous publications. Significant passages that have been
copied verbatim from other sources are enclosed within quotes “...”.

The goal of this chapter is for the student to acquire a broad initial understanding of the main
challenges that must be considered in designing facilities to store irradiated fuel (also called
"spent" fuel), the evaluations performed, and the safety and other criteria used for selecting
the principal features of the facilities for initial wet storage of irradiated fuel at reactor sites,
subsequent dry storage of irradiated fuel at reactor sites, and the current Canadian concepts for
eventual disposal of irradiated fuel (e.g., geological isolation).
2 Types of Waste
All biological life forms produce waste as part of their normal cycle of existence. Within the
biosphere, dissimilar life forms (e.g., plants and animals) have evolved to exist in harmony and
dependence on exchange of wastes. All human activity, starting with the most primitive
prehistoric activity associated with subsistence, up to and including the most sophisticated
industrial activity of today, produces waste. Regulations for managing all aspects of human-
generated wastes have evolved from “non-existent” for most of human existence to the
comprehensive and sophisticated regulations of today. Human industrial activity has not been
in existence sufficiently long to test the ability of ecological systems to create synergies,
interdependency, or both. Indeed, some industrial wastes are known to be detrimental to
people and the environment. Such wastes must not be left untreated or allowed to enter the
biosphere. Some nuclear wastes also fall into this category.
Under the terms of its statute, the International Atomic Energy Agency (IAEA) has a mandate to
establish or adopt safety standards to protect people and the environment from the harmful
effects of ionizing radiation. IAEA identifies six classes of nuclear waste ranked in order of their
harmful radiological effects (from the lowest to the highest) and the escalating requirements for
their safe disposal, as follows [IAEA, 2009]:
 Exempt waste
 Very short-lived waste
 Very low-level waste

 Low-level waste
 Intermediate-level waste
 High-level waste.
The crafters of the IAEA standard provided this list of waste classes as an aid to individual users,
recognizing that it may not be necessary or suitable for all users to adopt the list explicitly.
Users have the option of adopting the list literally or adapting it to their own needs, particularly
users who have had prior experience with managing radioactive waste.
The Canadian Standards Association (CSA), in collaboration with industry, government, and the
Canadian Nuclear Safety Commission (CNSC), has developed the “Radioactive Waste” standard
[CNSC, 2015a] that recognizes four main classes of radioactive waste:
 Low-level radioactive waste

 Intermediate-level radioactive waste
 High-level radioactive waste
 Uranium mine and mill waste.
Although these two lists of radioactive waste classes are not identical, it can be readily
discerned that the Canadian standard is not inconsistent with that of the IAEA. The Canadian
standard omits “exempt” waste and combines the three classes of “low-level” waste into a
single class. The two standards are explicitly consistent with respect to “intermediate” and
“high-level” waste. The Canadian standard adds a separate class for “uranium mine and mill”
waste, which reflects the importance of uranium mining in Canada. Every existing source of
radioactive waste in Canada and its class is listed, and the entity responsible for its management
is identified. Some entities may be responsible for managing more than one class of waste.
3 Strategies to Manage Radioactive Waste

Low-Level and Intermediate-Level Wastes
Radioactive waste has been produced in Canada since the early 1930s; see, for example,
“Historic Nuclear Waste” [CNSC 2015a], which describes the period when radium and uranium
were mined in the Northwest Territories and transported to Port Hope, Ontario, for refining.
Subsequently, uranium in much greater quantities has been mined and milled, most notably in
northern Ontario and Saskatchewan, and transported to Port Hope for refining and subsequent
use in nuclear fuel.
The following steps are taken to manage low-level and intermediate-level radioactive wastes
[CNSC, 2015b]:
 Decontamination and clean-up;
 Long-term storage and management by the user;
 Return to the manufacturer for long-term storage and management;
 Storage and management at large central facilities such as those operated or proposed
by Ontario Power Generation and Canadian Nuclear Laboratories.
The remainder of this chapter deals with the management of irradiated CANDU fuel.
Irradiated Fuel
Irradiated fuel continues to produce power and emit radiation after it is removed from the
reactor. The power and radiation decay with time, and therefore the need to cool the fuel and
to monitor its activity also decrease with time. For this reason, irradiated nuclear fuel passes
through two phases of storage before final disposal in the third phase, as follows:
 Immediately after its discharge from reactors, the fuel is stored for several years in deep
cooling pools adjacent to the reactor.
 After several years of forced-circulation cooling in water, the fuel is transferred to
concrete containers which are air-cooled by natural convection. The fuel can reside in
these storage cylinders for up to 100 years before its ultimate disposal.
 In the third phase, irradiated fuel is intended to be permanently sequestered in specially
designed facilities. As an illustrative example, one concept is to put the fuel in a sealed
container which is surrounded by clay and placed in a room that is built deep
underground in impervious rock.
Examples of these facilities, particularly as they apply to CANDU irradiated fuel, are described
later in this chapter.
Irradiated CANDU fuel may well be commonly considered “waste” to be (eventually)
permanently sequestered in specially designed facilities. At the same time, it also has the
potential to be recycled to generate significant energy through advanced fuel cycles in the
future (see Chapter 18), and therefore it can also be rightly considered as an asset worth a few
trillion dollars. Therefore it is prudent to adopt a flexible approach. On the one hand, the
disposal facility should safely isolate the irradiated fuel from the human population and the
environment with minimal ongoing expense, and on the other hand, the disposal facility should
permit its retrieval at a later date if appropriate and if so chosen.
IAEA [2012] provides guidance and recommendations for design, safe operation, and safety
assessments of fuel storage facilities, both wet and dry.
4 Irradiated Fuel Storage and Disposal: Main Considerations

The main features that irradiated fuel storage and disposal facilities must incorporate are listed
below. Other more detailed design and jurisdictional requirements exist, but are not covered in
this book. The relative importance of a specific feature may differ from one facility to the next;
therefore, an individual feature listed below may or may not be absolutely essential in a given
facility.
(i) Avoid overheating of irradiated fuel
Even after discharge from the reactor, fuel continues to produce power, as explained in more
detail in other chapters. Both the fuel and the materials used in storage and disposal structures
and facilities have their respective temperature limits for safe operation. Therefore, the
potential temperature increase caused by decay power must be controlled effectively to
prevent overheating.
(ii) Limit chemical and metallurgical damage to irradiated fuel bundles
A variety of factors can damage stored and disposed fuel, e.g., chemical, metallurgical, and
mechanical. Damaged fuel can release highly radioactive substances into the irradiated fuel bay
and also potentially into the air above. Such radiological contamination must be controlled to
acceptably low levels to protect the workers and the general public. Therefore, degradation and
damage to bundles must also be limited to acceptably low levels.
(iii) Avoid mechanical damage to irradiated fuel bundles
Irradiated fuel bundles are handled remotely by fuelling machines (pre-programmed robots)
during their removal from the reactor into the irradiated fuel pool, and by operators using
remotely operated tools for subsequent transfers. These tools must be designed and operated
so that the irradiated bundles are not damaged.
(iv) Avoid irradiated fuel configurations that could achieve criticality
CANDU fuel is made of natural uranium oxide and Zircaloy. Therefore, whether it is new or
irradiated, it cannot be put into a configuration that will achieve criticality in ordinary water
[Tsang, 1996]. This is not true, however, of LWR irradiated fuel, which contains higher levels of
fissile isotopes than CANDU irradiated fuel.
(v) Protect workers and public from radiological exposure
Adequate shielding must be provided to protect the workers and public from the radiotoxicity of
nuclear fuel. Whereas the cooling challenge reduces very quickly, shielding requirements last
longer because some long-lived isotopes (particularly the transuranics) take thousands of years
to decay. Nevertheless, the radioactivity of irradiated fuel does decrease with time, and
therefore so does the radiological threat. Hence, the most demanding conditions exist when
the fuel is handled during the initial wet and dry storage phases; special shielding must be
designed to protect workers from radiation exposure.
(vi) Adequately safeguard irradiated fuel against proliferation
IAEA takes proactive measures to observe and record movements of nuclear fuel at reactor sites
wherever practical. CANDU reactors are refuelled essentially continuously. This presents
unique safeguards (tracking) challenges, particularly during discharge and initial storage of
irradiated fuel in water-cooled bays. Continuous refuelling makes in-person monitoring
impractical; hence, IAEA-approved monitoring by remote means is used. Several monitoring
devices positioned at a number of locations in the reactor are used [Feiveson et al., 2011]. For
example, neutron and gamma radiation detectors are used in the reactor vault and in the
transfer port between the vault and the irradiated fuel bay (see Section 6.2); closed-circuit
video cameras are used in the irradiated fuel bays; and tamper-indicating enclosures that have
been inspected and sealed by IAEA are installed within the irradiated fuel bay.
(vii) Provide appropriate retrievability of irradiated fuel
Irradiated fuel may need to be retrieved from storage, especially after the fuel completes its
designated storage periods in phases 1 and 2.
However, during phase 3—disposal/isolation—a nation may or may not choose to incorporate
fuel retrievability into the design of its disposal facility, as explained in Section 3. Therefore, a
reasonable balance must be struck regarding ease of access to irradiated fuel: access must be
made quite difficult for a member of the general public, but not prohibitively so for an
approved, legitimate, large, organized agency. The desired duration for such retrievability must
also be chosen. These conscious choices influence the final configuration of the disposal facility.
Canada’s choice is described in a later section.
(viii) Avoid damage to facilities that interface with irradiated fuel
Similarly to the requirement for protection against damage to irradiated fuel, interfacing
equipment must also be adequately protected, e.g., from corrosion.
5 Fuel Integrity
The above considerations promote safety during storage. Even though this section lists a
number of items that focus on fuel integrity, several other aspects of storage also contribute
significantly to overall safety; some of these are discussed in later sections.
As with safety of the reactor core, safety during storage and disposal must be confirmed during
normal situations, during anticipated operational occurrences, and during postulated accidents.
For brevity, this chapter focusses largely on normal situations. Broader discussions of safety
during accidents are given elsewhere; see, for example, OPG [2009].
Recognizing that there will invariably be significant uncertainties in projections involving such
long time periods, it is also common practice to vary a number of important parameters and
assumptions and to perform bounding assessments. Furthermore, a number of hypothetical
“what if” scenarios are usually also developed to explore the influence of parameters and
scenario uncertainties in assessing long-term safety [NWMO, 2012a].
The remainder of this section provides some details on one aspect of safety during storage—
fuel integrity.
5.1 Requirements
IAEA [1994] outlines, among other aspects, requirements for fuel integrity during storage. Lian
[2010] cites the following two illustrative requirements from this IAEA document:
“The spent fuel cladding shall be protected during storage against degradation that leads
to gross ruptures, or the fuel shall otherwise be contained in such a manner that
degradation of the fuel during storage will not pose operational problems” (Article 223),
and:
“The heat removal capability shall be such that the temperature of all fuel (and fuel
cladding) in a storage facility does not exceed the maximum temperature recommended
or approved by the national nuclear regulatory body for the type and condition of fuel to
be stored” (Article 225).
Furthermore, from IAEA [2002]: “The prediction of the integrity and retrievability of spent fuel
constitute the main discussion topics for spent fuel behaviour regardless of the storage system
and time period envisaged.” Therefore, fuel integrity is an important consideration in the
overall design of storage facilities.
As well, some fuel bundles do occasionally develop defects during operation in the reactor, i.e.,
before they are brought to storage facilities (see Chapters 17 and 18). Their pellets may well
undergo faster degradation than those of intact fuel due to, for example, exposure of pellets to
air, which would increase the rate of oxidation of UO2. Therefore, the impact of fuel with
defects on safety during storage must also be considered.
To quantify up-front the expected degradation of fuel in any given design or scenario, one must
first identify credible damage mechanisms. Towards this end, the authors have culled from the
literature some illustrative fuel damage mechanisms pertinent to storage and disposal and their
key precursors and drivers. These are discussed in the next two sections.
5.2 Key Drivers and Precursors

Decay Power
As noted earlier, even after discharge from the reactor, nuclear fuel continues to produce
power.
Standard methods are available to quantify decay power; for example, see Garland [1999], ANS
[2014], and Glasstone [2014] for illustrative examples. Figure 1 shows how power decays with
time in typical irradiated CANDU fuel. The figure shows power decay in a fuel bundle that
produced 493 kW in the reactor just before its discharge. Immediately upon removal from the
reactor core, an irradiated CANDU fuel bundle generates less than 10% of the power that it
produced in the core. This figure drops to less than 1% in only a day after removal and to about
0.013% after a year has passed. The average power generated in a bundle at this point (one
year) is less than 100 W, which is comparable to a household light bulb. This keeps dropping
substantially with time, e.g., to about 5 W after 10 years and to about 1 W after 100 years, as
shown in Figure 1. Hence, the decay power is small compared to the power produced by the
fuel in the reactor; nevertheless, the fuel’s temperature increase must be controlled effectively
to avoid overheating the fuel and the storage structures.
Fuel Temperature
As an illustrative example of how fuel temperature depends on decay power, Figure 2 shows
temperature decay in LWR1 fuel and in its storage canister after emplacement in a long-term
Note 1: LWR means Light Water Reactor
repository for permanent disposal [Rothman, 1984]. In the repository, LWR fuel temperature
peaks at about 330°C and then declines. For much of the latter half of the first millennium after
emplacement, maximum LWR fuel temperature ranges from 90°C to 150°C. CANDU fuel
temperature is expected to be comparatively lower due to its lower burn-up and therefore
lower decay power.
Changes in UO2 composition and microstructure
In the dry sealed environment inside containers, there are few processes that would
significantly alter the local composition and microstructure of UO2. Over long periods, however,
some changes are likely to develop by processes such as ongoing decay, diffusion of
radionuclides, and damage from alpha radiation.
[Illustrative Example; Courtesy Kwok Tsang]

Figure 1 Decay power of an irradiated fuel bundle
Build-Up of Helium Gas and Internal Gas Pressure
As explained in Chapter 17, fuel elements contain: (a) initial filling gas that is added inside a fuel
element during fabrication and (b) fission gas that is generated during irradiation. In addition,
alpha decay during storage generates helium (He) gas. All three components together can
create internal pressure that can frequently be higher than atmospheric pressure. The excess
pressure (=internal minus external pressure) can create tensile stresses in the fuel element’s
Zircaloy, which can persist over significant durations. If stress is excessive, even nominally slow
mechanisms can potentially threaten fuel element integrity over the long periods pertinent to
storage and disposal.
The following description has been reproduced largely from NWMO [2012a]. Helium is stable
(i.e., not radioactive) and does not react chemically with other elements. Therefore, the total
amount of helium gas in the fuel elements would be expected to increase with time during
storage. Figure 3 gives an illustrative example, assuming that all fission-generated gases escape
the UO2 matrix into the open space inside the fuel element (see Chapter 17 for explanations of
these terms). Fission gases are formed in the reactor, and therefore, except for radioactive
decay, their amount would not change significantly during storage and disposal. In contrast,
after about 30,000 years, the amount of helium would be equal to that of fission gases, so that
the total amount of gas present would be double that under initial conditions. After about one
million years, the rate at which helium is produced would slow down due to changes in the
composition of the decay chain, but the total amount of helium within the fuel element would
continue to increase.
[Source: Rothman, 1984]

Figure 2 Illustrative temperatures in an LWR fuel and canister
Figure 3 illustrates the amount of helium produced within grains of UO2. Before it can
contribute to internal gas pressure, this helium would need to reach the open void inside the
fuel element. Pertinent processes that enable this are described in Chapter 17 and consist
mainly of diffusion to grain boundaries, storage within grain boundary bubbles, inter-linkage of
bubbles, and micro-cracking in grains. See Chapter 17 for more detailed descriptions of these
phenomena.
[At normal temperature and pressure. Source: NWMO, 2012]

Figure 3 Gas volume inside a fuel element
5.3 Fuel Damage Mechanisms

Over extended periods of time, fuel temperatures generated as described above can lead to
mechanisms for fuel damage during storage and disposal, as documented in a number of
publications such as Rothman [1984], IAEA [2002], McMurry et al. [2003], Lian [2010], and
NWMO [2012a]. Some illustrative damage mechanisms are summarized in the following
paragraphs, mainly from NWMO [2012a].
Zircaloy Corrosion
Obviating excessive corrosion is a generic design requirement for fuel. This topic is addressed in
Chapter 17 in the context of fuel residence in the reactor. In addition, during storage, sheath
corrosion must be limited to acceptable levels.
Zircaloy is relatively resistant to pitting corrosion in the pure chloride-free water encountered in
the fuel bay, and pitting does not occur in the atmosphere provided during dry storage.
Hydrogen Absorption and Zircaloy Embrittlement
As noted in Chapter 18, fuel sheaths can contain hydrogen and deuterium from several sources,
mainly residual hydrogen produced during fabrication, hydrogen or deuterium picked up from
the coolant during irradiation, and hydrogen or deuterium generated by in-reactor corrosion of
the sheath. Still more hydrogen can enter the Zircaloy as a result of radiolysis in the storage
pool; this can be limited by chemistry control and by purification of pool water. If quantities are
excessive, some hydrogen and deuterium may precipitate in the Zircaloy as zirconium hydride or
deuteride, especially at the relatively cooler temperatures of storage after removal from the
reactor. These hydrides and deuterides result in more brittle (less ductile) Zircaloy that is more
susceptible to fracturing from the mechanical loads imposed on it, for example during fuel
transfers and/or during fuel extraction from or loading into a disposal container.
Delayed Hydride Cracking (DHC)
DHC is similar to the process mentioned above (embrittlement of Zircaloy caused by hydrogen),
but occurs after a time delay, as explained below.
In some situations, this hydrogen or deuterium may be distributed fairly uniformly. In such
situations, their local concentrations could be acceptably low. However, if the material is under
significant temperature and/or stress gradients for a significant duration, hydrogen and
deuterium can migrate and concentrate preferentially at locations of relatively lower
temperature and higher stress. Over time, this can potentially lead to excessive local hydride
and deuteride concentrations and therefore to higher susceptibility to brittle fracture.
Sheath Creep and Rupture
Internal gas pressure can lead to long-term creep of Zircaloy. If excessive, this can potentially
rupture the Zircaloy.
Stress Corrosion Cracking (SCC)
As noted in Chapter 17, SCC (also called environmentally-assisted cracking, or EAC, in Chapter
17) occurs when irradiation-embrittled Zircaloy experiences sufficiently high tensile stresses in
the presence of a corrosive environment for a sufficiently long duration. Depending on the
magnitude of stress and of intensity of the corrosive environment, through-wall cracks due to
SCC can occur quickly, or slowly, or not at all. During storage, long-term stresses are provided
by internal gas pressure. A corrosive internal environment is provided by fission gases that were
released during irradiation in the reactor.
Mechanical Overstress
From NWMO [2012a]: “As long as the fuel bundles are supported by baskets in intact
containers, they are not subjected to significant load-bearing stresses. If tremors associated
with earthquakes caused the fuel bundles to vibrate sufficiently, presumably some of the fuel
pellets or the cladding could be damaged. The damaged material would remain in an intact
container, and the overall evolution of the used fuel bundles would not be significantly
changed.”
Fatigue
This would be a consideration mainly during transportation of irradiated fuel bundles, e.g., to
sites for dry storage and/or long-term disposal or isolation.
Defective Fuel
Sometimes fuel can develop defects in the reactor, that is, the Zircaloy develops a hole or a
crack. Detailed discussions of this phenomenon are provided in Chapters 17 and 18.
A hole or crack in Zircaloy will allow oxygen-containing air to come in contact with UO2. This,
over time, can potentially oxidize the pellet to higher states of oxygen, such as U3O8. The latter
is less dense than the former, and therefore the pellet can swell. Excessive swelling of the pellet
can potentially split the sheath. Under these conditions, in addition to the original hole or
crack, the Zircaloy can develop further splits during storage.
Other Effects and Remarks
The literature also mentions other mechanisms pertinent to potential fuel damage, for example,
hydraulic processes and biological processes. They are not covered in this chapter for the sake
of brevity. Combinations of these various mechanisms are also possible.
These fuel failure mechanisms during storage have been postulated partly from first principles
and partly from exploratory experiments. Because several of these mechanisms are thermally
activated and increase non-linearly with temperature, it would be good design practice to keep
storage temperatures reasonably low. Design studies as well as experience to date indicate that
under realistic as-designed operating conditions, the actual risk of fuel failure from any of these
postulated mechanisms is negligibly small. Some of these aspects are illustrated in the
following sections.
6 Wet Storage at Reactor Sites

Storage of irradiated fuel entails engineered facilities that can and must be monitored for the
entire time that the fuel is stored and that make retrieval and movement of irradiated fuel to
other facilities readily feasible.
6.1 Flow of Fuel Bundles Through a CANDU

Because CANDUs use natural uranium fuel, they need to be refuelled essentially continuously to
maintain the required level of reactivity within the core. From practical considerations of work
scheduling and equipment maintenance, refuelling is not done continuously, but rather during
parts of each day. The flow of fuel through a CANDU is shown in Figure 4.
[Courtesy Ralph Granz]

Figure 4 Flow of fuel through a CANDU reactor
Inside the core, fuel bundles reside in horizontal fuel channels and are cooled by high-pressure
heavy water. During refuelling, new fuel bundles are inserted into a fuel channel (usually in the
same direction as the coolant flow), and irradiated bundles are removed from the same fuel
channel by two fuelling machines operating in tandem. One machine, labelled as the “charge”
machine in the figure, inserts new fuel bundles at one end of the channel, and the other,
labelled the “accept” machine, removes irradiated bundles from the other end. The two
machines reverse charge/accept roles depending on the channel being fuelled (i.e., the flow
direction in the channel being fuelled). Fuelling machines, when connected to the fuel channel,
act as extensions of the fuel channel pressure boundary and provide the required cooling for
irradiated fuel bundles while they are inside the fuelling machine. The “accept” machine
delivers the irradiated bundles to the irradiated fuel (spent fuel) port. From there, the bundles
are moved by other remotely operated means to the irradiated fuel discharge room, which
contains ordinary water at atmospheric pressure.
During their residence in the reactor, and during all movements within the reactor and the
irradiated fuel storage bay, the bundles are oriented in the horizontal position. This is the
orientation for which the bundles are designed when they are subjected to the flow and power
conditions inside the reactor fuel channel (Chapter 17), and therefore this would be the
preferred orientation for the bundles throughout their storage and disposal life. However, a
single tier of vertical bundles supported on their end plates is also an acceptable configuration,
provided that the bundles are not required to support more than their own weight.
6.2 Irradiated Fuel Bundles Temporarily in Air

At the spent fuel port (Figure 4), the irradiated fuel bundles transit from the fuelling machine,
which contains heavy water, through an air-cooled interlock and elevator at atmospheric
pressure, and are then deposited in the spent fuel discharge room, which contains ordinary
(light) water. CANDU designs that use enriched fuel and ordinary water for coolant, such as the
ACR [AECL, 2007], obviate the passage of irradiated fuel through air, and the transfer of the
bundles can be performed underwater from the fuelling machine to the spent fuel discharge
room (or possibly directly to the reception bay). In existing CANDUs, movement of bundles
through the air interlock and elevator is timed so that irradiated bundles do not overheat while
in air.
As an illustrative exercise, you are encouraged to determine the maximum length of time that
the bundles can safely stay in air. Start with the information on decay heat given in Section 5.
Assume power equivalent to zero time after discharge and no heat removal by air. Use the
standard equation for adiabatic heat-up and apply it to a fuel element. How long does it take
for the sheath temperature to reach the melting point of Zircaloy?
CANDU plants are equipped with water spray nozzles to cool the bundles if their residence in air
exceeds the maximum permissible time.
6.3 Throughput of Irradiated Fuel Bundles

From the spent fuel discharge room (Figure 4), irradiated bundles are moved to the spent fuel
reception bay through the water-filled transfer canal. From the reception bay, the bundles are
moved to the spent fuel storage bay, where they are loaded underwater by a remotely
controlled bundle loading apparatus into storage trays that rest on storage racks. More recently,
consideration has been given to replacing the storage trays or racks with storage baskets which
are directly transferrable to dry storage containers. Both the trays or racks and the baskets are
designed to ensure:
 adequate cooling of the bundles in the storage bay,
 avoidance of damage to bundles during their residence in the pool,
 avoidance of damage to pool components,
 compliance with safeguards monitoring requirements (e.g., bundle serial numbers are
readily visible),
 direct transfer of bundles (baskets only) into dry storage containers, and
 use of the most compact arrangements of bundles in the pool.
As an illustrative exercise, one can determine the number of bundles discharged from a CANDU
6 reactor in one year, assuming the following:
 The plant’s electrical output is 600 MW(e).
 The plant overall efficiency of conversion of thermal power to electricity is ~30%.
 Capacity factor (fraction of time the plant is running at full power) is 90%.
 Each bundle contains approximately 19 kg of U and produces thermal power (average
discharge burn-up) of 6,500 MWd(th)/teU.
Using the above parameters, one can determine:
 Total thermal energy produced in one year: 657,000 MWd(th) (=600x0.9x365/0.3).
 Average thermal energy produced by an individual fuel bundle: 123.5 MWd (th)
(=6500x19/1,000).
 Number of bundles required to produce the reactor’s annual power: 5,320 bundles
(=657,000/123.5).
The storage pool is designed to accommodate the spent fuel generated from 10 years of
operation (~53,200 bundles in this example) plus one full core-load of bundles (4560 bundles in
this example; 380 channels containing 12 bundles each). The additional capacity of one core-
load provides for the possibility of having to unload the entire core to accommodate
refurbishment. In practice, bundles could begin to be removed from the pool as early as seven
years after their placement into the pool, reducing the risk that pool capacity would be reached
anytime during reactor life.
6.4 Irradiated Fuel Bay: Typical Dimensions

In Canada, there are several locations where irradiated fuel is either currently stored in fuel bays
or has been stored in the past. Some of these locations are: Chalk River Laboratories (CRL) at
Chalk River, ON; Whiteshell Laboratories (WL) at Pinawa MN; McMaster Nuclear Research
Reactor, Hamilton, ON; Point Lepreau Nuclear Power Plant at Point Lepreau, NB; Gentilly
Nuclear Power Plant at Trois-Rivières, QC; Darlington Nuclear Power Plant in Darlington ON;
Pickering Power Plant at Pickering ON; and Bruce Power Plant at Tiverton ON.
Designs, layouts, and dimensions of irradiated fuel bays (IFBs) differ from one station to
another. As an illustrative example, the following discussion concerns an IFB with three
sections: a reception bay; a long-term (or main) storage bay; and a transfer bay.
Irradiated fuel is received in the reception bay and stored there for one to two weeks (Figure 4).
Then it is transferred to the long-term storage bay and stored there for a few years. The long-
term storage bay usually also has a fuel inspection station where fuel can be inspected
underwater, for example to monitor and assess fuel performance. Finally, in the transfer bay
(not shown in Figure 4), fuel is loaded (underwater) into storage baskets en route to placement
in dry storage
Figure 5 illustrates a typical irradiated fuel storage bay. The sizes of the bays differ in different
stations, e.g., as small as 20 m x 12 m or as large as 34 m x 17 m [OECD/NEA, 2015]. These have
a surface area similar in size to an Olympic-size swimming pool, but the IFBs are constructed of
double-walled reinforced concrete and are much deeper. To provide sufficient radiological
shielding to workers above, water in IFBs is usually 6–8 m deep. Typically, the volumes of water
in C6 IFBs are: reception bay, 700 m3, and main storage bay, 2,000 m3.
6.5 Irradiated Fuel Bay: Heat Removal

Heat must be removed from the fuel bay to maintain water at a predetermined temperature
and to continue to provide cooling to the fuel. Generally, an irradiated fuel bay with 10-year
storage is maintained at <38°C; with 10-year storage plus one full core load, the pool is
maintained at <49°C. Major elements of the cooling system are a pump, a heat exchanger, and
a resin bed.
As industrial installations go, incorporating the required heat removal capability into the
irradiated fuel pool cooling system is not a significant challenge for the designer. However, in
case of impaired heat removal capability, it is important for safety reasons to determine how
much time may be allowed to elapse before auxiliary cooling must be provided. In the example
above, the reader is encouraged to determine how much time is required for the cooling water
to evaporate and the fuel bundles to be exposed to air, and how that time changes with the
number of bundles in the bay.
In the post-Fukushima era, factors such as available response time based on pool size and the
availability of off-site emergency cooling have achieved heightened prominence.
[Source: Villagran, 2014]

Figure 5 A typical irradiated fuel bay
6.6 Corrosion of Irradiated Fuel in the Pool

Corrosion of CANDU fuel during its residence in the reactor is covered in Chapter 18. To in-
reactor corrosion, any additional corrosion occurring during the residence of the fuel in storage
pools must be added. To help reduce the latter, demineralized water is used in the pool.
Because CANDU fuel’s residence time in the reactor is relatively short (especially compared to
LWR fuel), the resulting in-reactor corrosion is very low (see Chapter 18). Experience to date
has been that additional corrosion during storage is trivial compared to the corrosion picked up
during residence in the reactor. Hence, the total corrosion in-reactor and during pool storage is
very small and does not significantly reduce the fuel bundle sheath thickness.
6.7 Shielding of Irradiated Fuel in the Pool

For shielding to protect workers and the public, irradiated fuel storage pools typically contain
about 7 m of water above the last rack of fuel.
6.8 Criticality Not an Issue for CANDU Irradiated Fuel

The IAEA (and CNSC) design guidelines for irradiated fuel storage and disposal include a need to
assess conditions that could lead to criticality. There are no criticality-based restrictions on
proximity of irradiated CANDU bundles in the pool because the remaining fissile content in
irradiated bundles is low. Bundles can be placed into the most compact configuration (based on
heat transfer considerations only) and do not need to be “re-racked” during their residence in
the pool.
7 Dry Storage at Reactor Sites

As noted earlier, after a few years in water pools, fuel bundles are transferred to dry storage
because it is a comparatively less expensive method of storing large quantities of fuel and
because after a few years out of the reactor, fuel requires much less cooling.
A variety of options are available for dry storage. As one illustrative example, dry storage
containers are used at the Pickering and Bruce plants [OPG, 2015]. As another illustrative
example, Atomic Energy of Canada Limited (AECL) has developed concrete canisters that are
used at some C6 plants and decommissioned prototype plants. These are described in the
following sections.
7.1 Dry Storage Containers

Each dry storage container (DSC) can store a few hundred fuel bundles [OPG, 2015]. It is made
of reinforced high-density concrete approximately 51 cm thick and is lined inside and outside
with 12.7 mm-thick steel plate. The thick concrete provides an effective barrier against
radiation. Irradiated fuel bundles are loaded into a DSC within the IFB. The DSC is then sealed,
dried, and moved indoor into a secure storage building.
7.2 Concrete Canisters

AECL has developed modular “concrete canisters” such as CANSTOR and MACSTOR [CANDU
Energy, 2014]. The two are similar in concept, but MACSTOR is significantly bigger. A brief
overview of the latter is provided here; Figure 6 illustrates its principal features.
In the MACSTOR design, many fuel bundles are stored in a “storage basket”; many baskets are in
turn placed in a steel “storage cylinder”. A large number of storage cylinders are inserted into a
concrete labyrinth called MACSTOR. Shielding is provided by the concrete of the module, and
heat removal is achieved by air circulation due to natural convection. In Figure 6, note the
staggered openings at the bottom and top of the module; they allow air circulation, but do not
provide a direct line of sight through the concrete to the storage cylinder. Three barriers
prevent escape of radionuclides from the fuel: the fuel sheath, the sealed basket, and the
storage cylinders. Irradiated bundles can be retrieved using appropriate shielded equipment.
The cross section of each module is about 8 m wide and about 7 m high. Each MACSTOR unit
can store about 12,000 CANDU fuel bundles.
[Courtesy Jim Lian]

Figure 6 Overview of MACSTOR
7.3 Storage Baskets

Figure 7 [OPG, 2009] illustrates the main features of a typical storage basket used in MACSTOR.
Fuel bundles are loaded into baskets underwater in the fuel transfer bay, dried and weld sealed
inside the baskets, then moved in shielded flasks and loaded into storage cylinders. Bundle
orientation is changed from the normal horizontal orientation in the reactor and the irradiated
fuel storage bay to the vertical orientation during the basket loading process. In this design, the
basket is typically designed to hold 60 CANDU fuel bundles oriented vertically. Ten such baskets
are loaded into a metal storage cylinder.
[Source: OPG, 2009]

Figure 7 Basket for dry storage of fuel
7.4 Transfer of Bundles from the Pool into Baskets

Shielding requirements dictate that bundles must be loaded into dry storage baskets
underwater using remotely operated equipment. Design and process details differ from one
station to another.
7.5 Heat Transfer

MACSTOR facilities enable dry storage of fuel bundles with up to about 6 W of power per
bundle. This is typical of C6 fuel that has been cooled for six years.
Considering the size and good heat transfer configuration of the fuel elements, it can be easily
recognized that effective dissipation of this amount of heat from CANDU fuel can be achieved
by heat transfer through convection, conduction, and/or radiation without any parts of the
element or the bundle becoming overheated. Ultimately, the heat is passively transferred from
the storage cylinder to the atmosphere through natural convection.
7.6 Fuel Integrity

For dry storage, the target criterion is to limit sheath failures to within 1% of fuel rods during
100 years of storage [Lian, 2010].
Lian [2010] provides an overview of the integrity of irradiated CANDU fuel during interim dry
storage in MACSTOR and the associated research and development. Lian concludes the
following:
 Highest fuel temperature is less than 150°C; and
 Maximum sheath hoop stress is less than 4 MPa.
The above conditions are less demanding than those experienced by LWR fuel during dry
storage. Therefore, based on detailed assessments of LWR fuels done by the Electric Power
Research Institute in the United States, Lian [2010] concluded the following for CANDU fuel
stored in MACSTOR:
 Creep-rupture and external oxidation should not cause failure at storage below 300°C;
 For storage temperatures below 300°C, stress-corrosion cracking (SCC) is the most likely
mode of fuel failure, but even it is unlikely to occur to any significant extent during dry
storage;
 Fatigue is not a limiting failure mechanism for stored fuel sheaths; and
 Sheath splitting by UO2 oxidation is a limiting mode only for fuel rods that have already
developed defects.
Based on the above discussion, Lian [2010] projects that the failure rate in CANDU fuel during
dry storage in MACSTOR would likely be in the neighbourhood of 0.001% of fuel rods in 100
years. This is 1,000 times lower than the criterion noted earlier.
7.7 Shielding
As also noted in Section 5, radiation (i.e., the shielding requirement) does not diminish
proportionately as much as heat generation and requires continuing attention in handling of
irradiated fuel. As an illustrative example, nominal wall thickness in MACSTOR is approximately
one metre of ordinary concrete to control direct radiation exposure at the outside wall. This is
an indication that significant radiation can be expected from stacks of irradiated fuel bundles,
even six years after discharge.
8 Final Disposal or Isolation

As of 2014, Canada had about 2.5 million irradiated fuel bundles in storage [Garamszeghy,
2014]. Following the two interim phases of irradiated fuel storage described above, the
irradiated fuel may be either reprocessed or placed into permanent disposal or isolation
facilities. This section touches upon the following aspects of Canada’s current concept for
permanent disposal or isolation of irradiated fuel: approach and strategy; repository design;
container design; and health protection.
The remainder of this section has borrowed heavily from lectures delivered by J.E. Villagran at a
course sponsored by the Canadian Nuclear Society [Villagran, 2013, 2014].
8.1 Strategy
Several options are available for permanent disposal or isolation of irradiated nuclear fuel,
including isolation in stable geological formations or emplacement in abandoned salt mines.
After many decades of research [Boulton, 1978], detailed evaluations, and public consultations
[Seaborn, 1998], Canada has opted for centralized containment and isolation of irradiated
nuclear fuel in a deep geological repository (DGR) [NWMO, 2015a]. Several other nuclear
countries are also inclined towards geological repositories, e.g., Finland, Sweden, France,
Germany, Japan, Switzerland, and the United Kingdom.
“The goal of (Canada's) plan is to place the (irradiated) fuel deep underground in rock where
it will be constantly watched to make sure it is secure and not affecting anything around it.
Then at some point in the future, people can decide if they want to close the facility and
return the ground to its natural state” [NWMO, 2015b]. This approach includes an optional
step of interim shallow underground storage.
A key aspect of the current concept is to surround the irradiated fuel by multiple barriers, as
explained later in this section. This minimizes the potential for release of radionuclides should
any individual barrier fail.
The Canadian strategy for long-term waste fuel management is called “adaptive phased
management” (APM) [NWMO, 2015a] and is being administered by the Nuclear Waste
Management Organization (NWMO). APM covers all phases of high-level waste management,
including dry storage of fuel at the reactor site; the end result is final disposal (e.g., isolation in a
DGR).
8.2 Phases of Disposal

One can define two distinct phases in the life of a disposal or isolation facility:
 The initial, operational period during which the facility is constructed, filled with fuel,
and closed; and
 The long period of isolation of irradiated fuel after closure of the facility.
Initial operational period: Canada’s strategy and design preserves the retrievability option and
provides for continuous monitoring. Thus, APM engenders flexibility in design and promotes
ongoing technical and sociological research, enabling continuous learning and adaptation. The
process is collaborative, open, inclusive, and transparent, and decision-making is phased. The
current expectation is that this phase will last a few decades.
Isolation period after closing the facility: The facility’s closure is a few decades away, and
society’s needs and preferences that far away cannot be reasonably anticipated at this time
with sufficient certainty. Therefore, it is best to be flexible for now about the facility’s
configuration after closure. NWMO’s intention is to “remove underground equipment and
backfill and seal the access tunnels and shafts. Surface facilities will also be dismantled at a
pace and in a manner determined collaboratively with the community, regulators, and other
interested individuals” [NWMO, 2012b]. In this sense, there is an option, even an intention,
eventually to “entomb” the fuel.
From “first principles” many factors can be postulated as important for detailed quantitative
assessment of repository and container design. The following sub-sections highlight a few such
factors. Experience and actual analyzes have shown that some are relatively more important
than others.
8.3 Repository
Figure 8 shows a conceptual design of an illustrative repository. The overall concept starts with
placing many fuel bundles inside a sealed, corrosion-resistant container called a “used fuel
container” (UFC). This step isolates the fuel bundles from groundwater and delays release of
radionuclides into the environment.
The containers in turn are surrounded by a buffer material, bentonite clay, which possesses two
physical characteristics that enhance its function of isolating the fuel. First, this natural clay can
absorb up to 10 times its weight in water. If moisture were to approach the containers filled
with used fuel, the bentonite clay would absorb the water and swell up, forming a seal around
the container. This would delay the groundwater from reaching the container. Second, this clay
has low permeability. Therefore, should the container fail, the clay would slow the diffusion of
radionuclides out of the repository.
Many such containers would be placed in a horizontal position inside placement rooms in the
repository, at a depth of some 500 m. Other concepts are also being considered, such as
vertical placement of containers in boreholes.
Thus, multiple barriers isolate the fission products from the biosphere: the UO2 matrix; the
Zircaloy sheath; the sealed copper containers; bentonite clay; and rock.
Current projections are that by their end of life, CANDU power reactors will irradiate some 4.6
million fuel bundles. To dispose of these, a typical repository would require an area of about 2
km x 3 km, or about 600 Ha or 1,480 acres.
[Illustrative concept; Source: Villagran, 2014]
Figure 8 Concept of a deep geological repository
8.4 Used Fuel Containers

This section illustrates the main features of two designs of used fuel containers (UFCs) that use
a copper shell.
8.4.1 Conceptual designs

Figure 9 shows two slightly different container concepts; each container can hold many CANDU
fuel bundles. In the first concept, an important barrier to corrosion is provided by a shell made
from oxygen-free phosphorus-doped (OFP) copper 25 mm thick. Structural support is provided
by an inner liner made of carbon steel 96 mm thick. In the second concept, the copper shell is
replaced by a copper coating about 3 mm thick [Hatton, 2015].
The number of UFCs in a typical repository depends on the detailed design of the UFC. For
example, about 100,000 UFCs of Concept # 2 (as shown in Figure 9) would be needed to
accommodate some 4.6 million fuel bundles.
8.4.2 Container lifetime

The major damage mechanisms are expected to be uniform corrosion and pitting of the copper
shell or coating. Corrosion in turn can be driven by two major environmental factors [Villagran,
2014]:
 By initially trapped oxygen: the resulting corrosion is equivalent to about 0.080 mm
of wall penetration;
 By chloride and/or sulphide from groundwater, from pyrite impurities in the clay, or
by microbial activities. These cumulative corrosive interactions can cause additional
wall thickness loss of about 0.27 mm over a period of one million years.
After increasing nominal estimates by applying conservative factors, the total potential loss of
container wall thickness is estimated to be about 1.7 mm over one million years. Therefore,
depending on the thickness of copper, a typical UFC should stay intact for over one million
years.
8.5 Buffer Material

Main design requirements for the buffer material (bentonite clay in the above example) include
preventing groundwater migration to the disposal container, slowing migration of radionuclides
away from the container, and maintaining the long-term integrity of the chemical and
mechanical properties of clay.
[Source: Villagran, 2013, 2014]

Figure 9 Containers for Permanent Disposal of Irradiated Fuel
8.6 Rock Temperature

The power generated by the fuel mass will increase its temperature, which will cause heat to be
transferred from the fuel to the surrounding rock mass. Therefore, the temperature of the
disposal vault is a design criterion for a repository. For hard-rock (granite, sedimentary rock,
clay) geological disposal facilities, the temperature must remain below 90°C–100°C, but for salt
repositories (such as the Asse facility in Germany or the WIPP facility in New Mexico), a higher
temperature of up to 250°C may be acceptable. The higher-temperature facilities are more
appropriate for disposal of reprocessed waste, which has a higher short-term heat load.
Consequently, the thermal calculations revolve around geometry: how much fuel can be placed
into each container and how close to each other the containers can be placed to maintain a
temperature below the design requirement. The limiting parameters for heat transfer to the
rock vault are usually the surface temperatures of the container and the vault, rather than the
maximum fuel temperature in the container.
Initial calculations indicate that it takes between 25 and 40 years for the temperature of a fully
filled vault to go from ambient to 90°C, which would be a rather high temperature for workers.
Hence, one might need to consider potential arrangements for cooling the vault during the
waste emplacement period.
The rock temperature will be higher near the fuel and lower further away. This will lead to non-
uniform local expansion of the rock, which in turn will lead to thermal stresses. From “first
principles”, one can postulate that if the temperature gradients and the resulting thermal
stresses are large, the rock can potentially crack locally. However, are the thermal gradients
expected to be large enough to crack the rock?
As an illustrative example, let us assume that the rock temperature inside the repository is
100°C. The minimum temperature in the Earth’s crust at the depth of the repository must be
greater than the freezing temperature of water (because otherwise leaching and migration
would be impossible). Therefore, the maximum temperature gradient in the rock is ~100°C.
Considering that it takes at least 25 years for the vault temperature to increase by 70°C (see the
discussion earlier) provides a strong indication that the heat flux into the rock is extremely small
and therefore that the local stress anywhere in the rock due to the temperature gradient is also
very small.
8.7 Health Protection

8.7.1 Radioactivity and relative toxicity
Figure 10 [Boulton, 1978] illustrates the typical decay of radioactivity of irradiated nuclear fuel
for up to ten million years after discharge. Shortly after 100,000 years, the radioactivity of
irradiated fuel is very similar to that of natural uranium plus its decay chain.
The only credible escape route of radioactivity from disposed fuel to the biosphere is by
transport through groundwater. Therefore, an illustrative perspective on the potential hazard of
nuclear waste can be obtained by comparing its toxicity in water with that of naturally occurring
materials.
The International Commission on Radiological Protection (ICRP) has defined a “relative toxicity
index” as the ratio of the volume of water needed to dilute the radionuclides in waste material
to drinking water standards to the volume of water needed to dilute an equal weight of 0.2
percent uranium ore to the same standards [Boulton, 1978]. Figure 11 shows some illustrative
results of these relative toxicity calculations [Boulton, 1978]. The calculations necessarily
contain some assumptions, but nevertheless, for purposes of an initial illustration, the figure
suggests that after about a century following removal from the reactor, the relative toxicity of
used nuclear fuel (in groundwater) is similar to that of a naturally occurring ore of mercury
which contains 2.6% mercury. Within another century, the radiotoxicity of irradiated nuclear
fuel becomes similar to that of naturally occurring ores of lead (of 5.8% concentration) and of
uranium (3% concentration).
[Illustrative example; Fuel burnup: 220 MWh/kgU]

[Source: Villagran, 2014]
Figure 10 Decay of radioactivity in CANDU fuel
In a repository, access of groundwater to radionuclides is delayed by the engineered barriers
described earlier. It must, however, be recognized that some engineered barriers may
eventually fail, and therefore it is of key importance to ensure that significant time has elapsed
by the time groundwater gains access to radionuclides and carries them to the biosphere. If so,
the radionuclides would have decayed to harmless levels by the time they reach the biosphere
(as illustrated in Figures 10 and 11). Therefore two important aspects of a safe disposal facility
are orderly failure of barriers and the specific sequence of events. Hence, performance
modelling is a critical component of the safety case and is explicitly dealt with in the CNSC
Regulatory Guide G320 [CNSC, 2006].
Discussions on safety can be conveniently classified into two broad categories: (i) safety during
the phase of waste emplacement in the facility, and (ii) safety during the post-closure and
entombment phase.
[Illustrative example; Source: Boulton, 1978]

Figure 11 Toxicity in ground water
8.7.2 Operational Phase

The relatively more important aspects of worker safety during this period are:
 radiation dose,
 rock falls,
 container stability during transport,
 hazards and difficulties in retrieval, and
 accident analysis.
8.7.3 Post-closure phase

In any disposal system, potential release of pathogens to the public is the main health concern.
Two important health protection objectives are that the committed dose be negligible and that
the risk of increased cancer be negligible. Therefore, an important element of public health
protection is provided by making the fuel inaccessible. This is enhanced by entombment of the
fuel.
In general, the relatively more important aspects of a disposal facility are container lifetime,
radionuclide diffusion, groundwater chemistry, vault sealing, and performance modelling.
It must also be ascertained that the Earth provides sufficient radiological shielding to humans
on the surface. Figure 12 shows the results of illustrative dose rate calculations at the surface of
a conceptual repository at a hypothetical site. The predicted dose rates provide good margins
of safety compared to the natural background dose rate and also compared to the acceptance
criteria recommended by the International Commission on Radiological Protection (ICRP).
8.8 Breach of Containment

Write-ups in this and the following sections are based heavily on a variety of previous
publications, mainly Whitlock [2015].
“In the design of Canadian technology, one aspect that was seriously considered was: ‘what if
both the copper container and the Zircaloy sheaths are breached?’ Would the radioisotopes in
used fuel reach the biosphere and pose a significant hazard to human health?”
If both fuel containers were breached, migration through groundwater is the one credible
mechanism through which harmful radionuclides from the used fuel could be transported to the
food chain and/or to water reservoirs accessible to humans.
Extensive experiments and detailed computer simulations have established that in the current
design, transport times to the surface are in the hundreds of thousands of years. In that time,
the radionuclides would decay to harmless levels (see Figure 10), and therefore the effects of
used fuel on the biosphere would be negligible. In fact, after a few hundred years of decay, the
actinide content in the used fuel repository is comparable to that of naturally occurring high-
grade uranium deposits found in Canada, as well as those of other toxic ores such as lead and
mercury. This period—a few hundred years—is well within human experience for safe design of
engineered structures.
Although this science is based on extensive experiments and detailed computer modelling, its
complete verification over the timescales of interest is obviously not possible through
laboratory experiments alone. Therefore, this technological solution has also been verified
using “natural analogues”. Natural analogues are systems that occur naturally and possess key
attributes that are very similar to components of the used fuel disposal system. They are
described next.
[Illustrative calculations; Source: Villagran, 2014]

Figure 12 Dose rate at the surface of a repository
8.9 Nature’s Analogues for a Waste Repository

From the discussion above, it should be clear that the two aspects of the Canadian disposal
strategy that are most critical to its success and that would benefit from having natural
analogues to demonstrate their viability are a demonstration that radionuclide migration
through groundwater is equal to or less than that used in the assessment models, and a
demonstration that the effectiveness of the buffer material (bentonite clay) is as good as or
better than that assumed in the assessment models.
 Migration of plutonium and uranium
In 1972, a French uranium mining company deduced that a natural “chain reaction” existed on
earth some 2 billion years ago, long before the first chain reaction was demonstrated by human
beings. These “natural reactors” existed at Oklo in Gabon, Africa. The key ingredient that made
these natural reactors possible at that time is that natural uranium contained a significantly
higher proportion of 235U than it does today. At this higher enrichment level, natural uranium,
with access to ordinary water, could be configured to constitute a critical mass. In fact, about
16 such “reactor” locales were identified at Oklo. It is postulated that these natural reactors
“operated” intermittently for nearly a million years, with their power being regulated by access
to water. When the reactor power increased, groundwater heated up and evaporated; this
caused the reactor to shut down. When the water cooled down and flooded the uranium
deposit, the power started up again, and so on. These reactors shut down permanently when
the concentration of 235U became depleted to a value that could no longer sustain a chain
reaction. However, the groundwater that had made the chain reaction possible continued to
flow over the ore body until it was discovered 2 billion years later.
Unfortunately, by the time the phenomenon of “natural reactors” was understood, most of the
locales where these natural reactors operated had been mined and the evidence of radionuclide
migration 2 billion years earlier destroyed. Fortunately, some sites were undisturbed, which
provided evidence that neither the plutonium (now totally decayed) nor the uranium had
migrated from the location in the grain where they existed when fission occurred. This is
powerful evidence that the model used for radionuclide migration in the Canadian waste
disposal assessments (which assume that the radionuclides percolate to the surface of the
repository together with groundwater) is extremely conservative.
 Effectiveness of the buffer material
The most direct natural analogue that provides evidence of the effectiveness of the buffer
material is the uranium deposit in the Athabasca Basin of northern Saskatchewan, known as
Cigar Lake. This deposit was discovered in 1981 and, after a lengthy period fraught with
production difficulties unrelated to its geology, came into production in 2014. It is one of the
richest (average ore grade of 20%) and largest deposits of uranium in the world. It occurs at a
depth of about 500 metres as a contiguous deposit about 2000 m long, 100 m wide, and 20 m
thick. A well-defined ridge of basement rocks about 2 billion years old underlies the mineralized
zone over its entire length.
The most remarkable feature that has relevance to the effectiveness of the buffer material is
the clay deposit that completely covers the ore deposit, on top of which there is an overburden
of sandstone. The fact that the ore deposit has survived undisturbed for about 2 billion years
speaks to the effectiveness of the clay barrier in preventing water from reaching the ore deposit
and leaching out uranium atoms. The effectiveness of this entombment is further verified by
the fact that the ore deposit provides no chemical or radioactive signal of its existence that can
be detected at the surface. The age of the ore deposit and the traumatic geological events that
have occurred during that time, and which events the clay barrier has survived intact, provide
indisputable evidence for the effectiveness of the clay barrier. Finally, it must be remembered
that the buffer material designated for use in the Canadian disposal facility will be superior to
the Cigar Lake clay, and that the host rock in the disposal facility will be far less permeable than
the sandstone of the Cigar Lake deposit. Additional thoughts on natural analogues can be found
in Whitlock [2015].
9 Problems
Section 2
Q2.1 – Why is it necessary to regulate and control storage and disposal of nuclear waste?
Q2.2 – Why do regulatory bodies such as the IAEA classify nuclear wastes?
Q2.3 – Which Canadian agency is responsible for regulating the manner in which nuclear wastes
are to be handled in Canada, and how many classes of nuclear waste are designated?
Section 3
Q3.1 – What are the two stages of nuclear fuel storage that are generally accepted and used in
Canada?
Q3.2 – Why is nuclear fuel stored in water-cooled pools immediately after discharge from the
reactor and for several years thereafter?
Q3.3 – After being stored in a water-cooled pool for several years, why is it acceptable and
desirable to remove irradiated CANDU fuel bundles from the pool and store them in air-cooled
storage cylinders?
Section 4
Q4.1 - Designers of storage facilities must be mindful of a number of design requirements that
must be met to ensure the integrity and security of irradiated fuel while it is being stored. List
them.
Section 5
Q5.1 – What are the two overriding requirements specified by the IAEA that must be met to
ensure the integrity of spent fuel while it is being stored?
Q5.2 - Whether stored in water-cooled pools or air-cooled storage cylinders, irradiated fuel and
its surroundings are potentially susceptible to a number of damage mechanisms and other
challenges. List them.
Section 6
Q6.1 – What is the preferred orientation for CANDU fuel bundles in-reactor and in storage
facilities?
Q6.2 – Why are irradiated CANDU fuel bundles required to reside temporarily in air as they pass
from the fuelling machine to the irradiated fuel discharge room?
Q6.3 – How are the irradiated bundles, when residing in air, prevented from overheating?
Q6.4 – How much water is contained in a typical irradiated fuel storage pool?
Q6.5 – What can the operators of irradiated storage pools do to minimize oxidation of fuel
Zircaloy components?
Q6.6 – How are the workers inside irradiated fuel storage facilities protected from radiation
produced by irradiated bundles?
Q6.7 – What is the optimum stacking arrangement of irradiated CANDU bundles to prevent
criticality in the irradiated fuel bay?
Section 7
Q7.1 – Describe the main features of a storage cylinder.
Q7.2 – Although the thermal power of the irradiated bundles decreases dramatically while the
bundles are stored in water-cooled pools, the bundles’ emission of ionizing radiation remains
high (due to decay of long-lived isotopes and elements). How are people and the environment
protected from radiation when bundles are stored in concrete casks?
Section 8
Q8.1 – What types of geological formations are thought to be suitable for final entombment of
irradiated fuel (or fuel waste)?
Q8.2 – What is the proposed method in Canada for permanent disposal of high-level radioactive
waste?
Q8.3 – What is “adaptive phased management” and how does it influence Canada’s high-level
waste disposal program?
Q8.4 – What are the two distinct phases in the “life” of a disposal facility?
Q8.5 – Describe the main components of a disposal facility.
Q8.6 – What are the two physical properties of bentonite clay that make it eminently suitable
for sealing storage containers inside rock caverns?
Q8.7 – Once the irradiated bundles have been placed into containers and the containers sealed
inside the vault, how many barriers are at work to prevent the radioactivity from affecting
people and the environment? Name the barriers.
Q8.8 – What are the life-limiting factors for a storage container, and what is the container’s
expected lifetime?
Q8.9 – What is the principal characteristic of the Deep Geological Repository that provides
shielding protection to people and the environment from the radiological effects of irradiated
fuel?
Q8.10 – The initial three barriers to release of radiologically harmful effects of irradiated fuel to
people and the environment are UO2 pellets, Zircaloy sheaths, and steel/copper containers,
which can delay the release of radioactivity for several million years and are assumed
(conservatively) to do so for 105 years. If these three barriers were accidentally breached
immediately after the fuel was entombed, how would the effects of radiation on people and
the environment be affected?
Q8.11 – Empirical verification of the efficacy of the proposed disposed concept would require
several hundred years, if not hundreds of thousands of years, which is not feasible. Therefore,
verification to date has been based on short-term experiments, computer modelling, and
natural analogues. Name and describe the natural analogues that have produced information
that provides direct support (verification) of the design of the disposal facility proposed in
Canada.
10 References
AECL, 2007. “ACR-1000 Technical Summary”, Atomic Energy of Canada Limited.
http://www.uea.ac.uk/~e680/energy/energy_links/nuclear/ACR1000-Tech-Summary.pdf
ANS, 2014. “Decay Heat Power in Light Water Reactors”, American Nuclear Society, Publication
ANSI/ANS-5.1-2014.
J. Boulton, 1978. “Management of Radioactive Wastes: The Canadian Disposal Program”,
Atomic Energy of Canada Limited, Report AECL-6314.
CANDU Energy, 2014. “MACSTOR: Modular Air-Cooled Storage”, published by CANDU Energy,
1016/2014_05. http://www.candu.com/site/media/Parent/CANDU%20brochure-
MACSTOR-0618B.pdf.
CNSC, 2006. “Assessing the Long-Term Safety of Radioactive Waste Management”, Canadian
Nuclear Safety Commission, Regulatory Guide G-320.
CNSC, 2015a. “Historic Nuclear Waste”, Canadian Nuclear Safety Commission, Web site,
http://nuclearsafety.gc.ca/eng/waste/historic-nuclear-waste/index.cfm.
CNSC, 2015b. “Radioactive Waste”, Canadian Nuclear Safety Commission,
http://nuclearsafety.gc.ca/eng/waste/index.cfm.
H. Feiveson, Z. Mian, M.V. Ramana, and F. von Hippel (editors), 2011. “Managing Spent Fuel
from Nuclear Power Reactors Experience and Lessons from Around the World”,
International Panel on Fissile Materials, ISBN 978-0-9819275-9-6,
http://fissilematerials.org/library/rr10.pdf.
M. Garamszeghy, 2014. “Nuclear Fuel Waste Projections in Canada – 2014 Update”. Nuclear
Waste Management Organization, Report NWMO TR-2014-16.
http://www.nwmo.ca/uploads_managed/MediaFiles/2471_nwmo_tr-2014-
16_nuclear_fuel_waste_projections_201.pdf.
W.J. Garland, 1999. “Decay Heat Estimates for the McMaster Nuclear Reactor”, McMaster
University Nuclear Reactor, Technical Report MNR-TR 1998-03, 11 pages.
S. Glasstone, A. Sesonske, 2014. “Nuclear Reactor Engineering: Reactor Design Basics / Reactor
Systems Engineering”, Springer, ISBN-13: 978-1-4615-7527-6 / 978146157276, ISBN: 0-
442-02725-7, Fourth Edition.
C. Hatton, 2015. “Development of the Canadian Used Fuel Repository Engineered Barrier
System”, 35th Annual Conference of the Canadian Nuclear Society, St. John, NB, Canada.
IAEA, 1994. “Design of Spent Fuel Storage Facilities”, International Atomic Energy Agency, IAEA
Safety Series No. 116.
IAEA, 2002. “Long-Term Storage of Spent Nuclear Fuel: Survey and Recommendations”,
International Atomic Energy Agency, Report IAEA-TECDOC-1293.
IAEA, 2003. “Radioactive Waste Management Glossary”, International Atomic Energy
Commission, STI/PUB/1155 ISBN 92–0–105303–7.
IAEA, 2009. “Classification of Radioactive Waste”, General Safety Guide GSG-1, International
Atomic Energy Agency, Vienna.
IAEA, 2012. “Storage of Spent Nuclear Fuel”, International Atomic Energy Agency, Specific
Safety Guide, Number SSG-15.
J. Lian, 2010. “Integrity Assessment of CANDU Spent Fuel During Interim Dry Storage in
MACSTOR”, IAEA International Conference on Management of Spent Fuel from Power
Reactors, Vienna, May 31–June 4, 2010.
J. McMurry, D.A. Dixon, J.D. Garroni, B.M. Ikeda, S. Stroes-Gascoyne, P. Baumgartner, T.W.
Melnyk, 2003. “Evolution of a Canadian Deep Geological Repository: Base Scenario”,
Ontario Power Generation, Nuclear Waste Management Division, Report 06819-REP-
01200-10092-R00.
NWMO, 2012a. “Used Fuel Repository Conceptual Design and Post-Closure Safety Assessment
in Crystalline Rock”, Nuclear Waste Management Organization, Toronto, Canada. Report
NWMO TR-2012-16.
NWMO, 2012b. “Description of Canada’s Repository for Used Nuclear Fuel and Centre of
Expertise”, Nuclear Waste Management Organization, Toronto, Canada.
http://www.nwmo.ca/uploads_managed/MediaFiles/2011_projectdescriptionbrochure-
english.pdf.
NWMO, 2015a. “Implementing Adaptive Phased Management, 2015 to 2019”, Nuclear Waste
Management Organization, Canada.
NWMO, 2015b. “Managing Canada’s Used Nuclear Fuel”, Nuclear Waste Management
Organization, Toronto, Canada. http://www.nwmo.ca/aboriginalbrochure.
OECD/NEA, 2015. “Status Report on Spent Fuel Pools under Loss-of-Cooling and Loss-of-
Coolant Accident Conditions”, Nuclear Energy Agency, Report NEA/CSNI/R(2015)2.
OPG, 2009. “Darlington Environmental Assessment”, Ontario Power Generation, Nuclear Waste
Management, Technical Support Document, New Nuclear, Report NK054-REP-07730-
00027 Rev 000.
http://www.ceaa.gc.ca/050/documents_staticpost/cearref_29525/0104/nwm.pdf.
OPG, 2015. “Pickering Waste Management Facility”, Ontario Power Generation, Brochure.
http://www.opg.com/generating-power/nuclear/nuclear-waste-
management/Documents/PWMFbrochure.pdf.
A.J. Rothman, 1984. “Potential Corrosion and Degradation Mechanisms of Zircaloy Cladding on
Spent Nuclear Fuel in a Tuff Repository”, Lawrence Livermore National Laboratory,
Report UCID-20172.
B. Seaborn (Chair), 1998. “Environmental Assessment Panel: Review of Nuclear Fuel Waste
Management and Disposal Concept”, Report to Canada’s Ministers of Environment and
of Natural Resources, published by Minister of Public Works and Government Services
Canada. Catalogue No. EN-106-30/1-1998E, ISBN: 0-662-26470-3.
http://www.ceaa.gc.ca/default.asp?lang=En&n=0B83BD&printfullpage=true.
K. Tsang, 1996. “Storage of Natural Uranium Fuel Bundles in Light Water: Reactivity Estimates”,
Proceedings of the CNS Simulation Symposium, Montreal, Quebec.
J.E. Villagran, 2013 and 2014. “Long-Term Management of Canada’s Used Nuclear Fuel”, CANDU
Fuel Technology Course, Canadian Nuclear Society.
J. Whitlock, 2015. “The Canadian Nuclear FAQ, Section E: Waste Management”,
http://www.nuclearfaq.ca/cnf_sectionE.htm#v2.
11 Further Reading Material

F.K. Hare (Chair), A.M. Akin, and J.M. Harrison, 1977. “The Management of Canada's Nuclear
Wastes”, Energy, Mines, and Resources Canada, Report EP77-6.
A. Porter (Chair), 1978. “A Race Against Time”, Interim Report on Nuclear Power in Ontario
(Ontario), Royal Commission on Electric Power Planning, Queen’s Printer for Ontario.
A. Porter (Chair), 1980. “The Report of the [Ontario] Royal Commission on Electric Power
Planning: Vol. 1, Concepts, Conclusions, and Recommendations”, Queen’s Printer for
Ontario.
P.K. Kuroda, 1982. “The Origin of the Chemical Elements and the Oklo Phenomenon”,
Springer, ISBN 978-3-642-68667-2.

Chapter 8 provides an overview of fuel bundle configuration. Chapters 3 to 5 explain the
neutron physics that generates heat in the fuel. Chapters 6 and 7 explain how that heat is
removed from the fuel bundle and illustrate the internal temperature distribution within a fuel
rod. Chapter 13 explains fuel performance during postulated accidents. Chapters 14 and 15
explain chemical and metallurgical aspects that relate to fuel sheath corrosion. Chapter 17
describes fuel design and performance, focussing on the current natural-uranium cycle (i.e., low
burnup). Chapter 18 describes a few selected fuel cycles, summarizes key aspects of fuel
manufacturing related primarily to the natural-uranium fuel cycle, and explains a few selected
aspects of fuel performance that become relatively more important for advanced fuel cycles.
Finally, the current chapter (#19) describes interim storage and disposal of irradiated fuel.
13 Acknowledgements
The following reviewers are gratefully acknowledged for their contributions, hard work, and
excellent comments during the development of this chapter; their feedback has much improved
it.
 Jim Lian
 Brian Ikeda
 Kwok Tsang
 Jennifer Noronha
Of course the responsibility for any errors or omissions lies entirely with the authors.
Thanks are also extended to J. Villagran for providing many figures; to Jeremy Whitlock for
providing initial write-ups on natural analogues and breach of containment; to Simon Oakley
and Media Production Services, McMaster University, for nicely drawing the figures; and to
Diana Bouchard for expertly editing and assembling the final copy.
1
CHAPTER 20
Fuel Handling and Storage
prepared by
Diane Damario P. Eng,
Senior Engineer – Mechanical, Fuel Handling, Nuclear Power,
SNC-Lavalin Nuclear Inc., Candu Energy Inc.
Summary:
CANDU® reactors are fuelled on a somewhat continuous basis while on-power. This fuelling
capability is mainly performed by two remotely operated fuelling machines. CANDU®1 6 fuelling
machines are part of the fuel handling and storage system, which is divided into three main sub-
systems. New fuel transfer and storage involves the receipt, handling and storage of palletized
crates of unirradiated, or new, fuel, the inspection and handling of individual new fuel bundles
and the transfer of this fuel into a fuelling machine via a new fuel port. Fuel changing involves a
pair of fuelling machines transporting new fuel from a new fuel port to the reactor face, loading
and unloading fuel from a reactor fuel channel and transporting discharged irradiated fuel from
the reactor face to a spent fuel port. Spent fuel transfer and storage involves the discharge of
irradiated fuel from a fuelling machine and the transfer and storage of this irradiated fuel in
water-filled bays until the fuel is transferred to dry storage.
Table of Contents
1 Introduction ............................................................................................................................ 3
1.1 Overview ......................................................................................................................... 3
2 Fuel Handling and Storage System Overview ......................................................................... 3
3 New Fuel Transfer and Storage ............................................................................................... 7
3.1 System Description ......................................................................................................... 7
3.2 Component Description.................................................................................................. 7
3.3 System Operation.......................................................................................................... 12
4 Fuel Changing........................................................................................................................ 14
4.1 System Description ....................................................................................................... 14
4.2 Component Description................................................................................................ 14
5 Spent Fuel Transfer and Storage ........................................................................................... 29
5.1 System Description ....................................................................................................... 29
1 CANDU® (CANada Deuterium Uranium) is a registered trademark of Atomic Energy of Canada Limited (AECL)
used under exclusive licence by Candu Energy Inc.
©UNENE, all rights reserved. For educational use only, no assumed liability. Fuel Handling and Storage – January 2017
5.2 Component Description................................................................................................ 30

6 Problems ............................................................................................................................... 40
7 Further Reading .................................................................................................................... 40
List of Figures
Figure 1 Fuel handling and storage system..................................................................................... 5

Figure 2 Fuel handling and storage sequence ................................................................................ 6
Figure 3 New fuel transfer equipment.......................................................................................... 10
Figure 4 New fuel transfer mechanism......................................................................................... 11
Figure 5 Fuel changing equipment................................................................................................ 21
Figure 6 Fuelling machine head .................................................................................................... 22
Figure 7 Front end of fuelling machine......................................................................................... 23
Figure 8 Ram assembly – exploded view ...................................................................................... 24
Figure 9 Ram adapter and operation............................................................................................ 25
Figure 10 Fuelling machine snout plug ......................................................................................... 26
Figure 11 Carriage assembly ......................................................................................................... 27
Figure 12 Fuel changing sequence................................................................................................ 28
Figure 13 Spent fuel discharge equipment ................................................................................... 37
Figure 14 Spent fuel transfer equipment...................................................................................... 38
Figure 15 Failed fuel canning equipment...................................................................................... 39
Fuel Handling and Storage 3
1 Introduction
1.1 Overview
This chapter provides an overview of a typical CANDU 6 fuel handling and storage system
handling natural uranium fuel. It does not address the potential use and handling of alternative
fuels. Although the fuel handling and storage systems of different CANDU 6 stations are almost
identical to each other, there are some differences resulting from evolution over time. As a
result, some of the equipment described herein, may not be applicable to all CANDU 6 stations.
Although this chapter does not address the fuel handling and storage system of CANDU stations
other than CANDU 6, the equipment and processes described herein share many similarities
with their equipment and processes.
This chapter explains how the fuel handling and storage system is divided into three main sub-
systems: new fuel transfer and storage, fuel changing and spent fuel transfer and storage. It
discusses the structures, equipment and purpose for each sub-system, and provides descrip-
tions of major components and system processes.
Section 3 covers the new fuel transfer and storage system, Section 0 covers the fuel changing
system and Section 5 covers the spent fuel transfer and storage system.

The principal learning objectives of this chapter are to gain a general understanding of the
following aspects of a typical CANDU 6 fuel handling and storage system:
 The purpose of this system and that of its three main sub-systems: new fuel transfer and
storage, fuel changing and spent fuel transfer and storage;
 The main structures and equipment of these sub-systems;
 The operation of these sub-systems.
2 Fuel Handling and Storage System Overview

The purpose of the fuel handling and storage system, depicted in Figure 1, is to provide
on-power fuelling capability at a rate sufficient to maintain continuous reactor operation at full
power. While fulfilling its purpose, the fuel handling and storage system is also required to
maintain all fuel sub-critical outside the reactor core and in conditions preventing fuel damage
and contamination by foreign materials. Refer to Chapter 8 for information on the main
components of a CANDU nuclear power plant and Chapter 9 for information on the operating
concepts of a CANDU nuclear power plant.
The main equipment necessary to achieve on-power refuelling capability, together with associ-
ated services and controls, includes:
 New fuel transfer and storage;
 Fuel changing; and
 Spent fuel transfer and storage.
On-power refuelling, which involves replacing discrete amounts of fuel in the core as required
while the reactor is operating, is a major contributor to the high capacity factor of existing
CANDU reactors. Other advantages of on-power refuelling include the following:
 Optimized burnup of natural uranium fuel and, therefore, lower fuelling costs.
 Low excess reactivity available, which reduces the requirement for reactivity control during
operation.
 Flexibility to plan scheduled shutdown activities that do not have to include fuelling opera-
tions.
 Low levels of fission products in the heat transport system, by early detection and removal
of failed fuel bundles.
The sequence of movements for typical fuel handling and storage operations is illustrated in
Figure 2. The average equilibrium fuelling rate in the CANDU 6 reactor at full power is ap-
proximately 112 fuel bundles per week.
New fuel transfer and storage, described in Section 3, involves new, or unirradiated, fuel being
received and stored in the new fuel storage room in the service building. New fuel is trans-
ferred to the new fuel loading room in the reactor building as required. There, fuel is loaded
into one of two new fuel transfer mechanisms for transfer into one of the fuelling machines via
the new fuel ports.
Fuel changing, described in Section 0, involves new fuel being received by a fuelling machine via
one of the new fuel ports, fuelling machines travelling to the reactor face, new fuel being
inserted into and irradiated fuel being removed from one of the 380 fuel channels by the
fuelling machines, fuelling machines travelling from the reactor face, and irradiated fuel being
discharged via one of the spent fuel ports.
Spent fuel transfer and storage, described in Section 5, involves irradiated fuel being received
from fuelling machines via spent fuel ports and placed in water-filled fuel storage bays.
The fuelling machines, the new fuel transfer equipment and the spent fuel discharge equipment
are normally operated remotely and automatically from the fuel handling control console in the
control room. The automatic control system for the fuel handling and storage system uses one
of the station computer-controllers and includes the fuel handling control console in the Main
Control Room and local control panels specific to the fuel handling system.
Personnel are normally only required to enter the reactor building to load new fuel and to
maintain the fuel handling system components. . Regular and frequent maintenance of fuel
handling equipment, particularly the fuelling machines, is required to prevent or minimize
unexpected breakdown of fuel handling equipment and to maintain or achieve station target
capacity factors. Breakdown of fuel handling equipment at the reactor face and/or while
manipulating irradiated fuel is particularly undesirable and problematic due to very high radia-
tion fields.
The term ‘irradiated fuel’, which is common in the industry and is more precise, is used when
referring to the fuel itself; however, the term 'spent fuel' is used when referring to applicable
fuel handling subsystems and components per traditional fuel handling naming conventions for
these subsystems. Refer to Chapter 17, for information about the fuel.
Note that the figures provided are for illustrative purposes only. Actual fuel handling and
storage equipment may vary in appearance.
Figure 1 Fuel handling and storage system
Figure 2 Fuel handling and storage sequence
7
3 New Fuel Transfer and Storage

3.1 System Description
The new fuel transfer and storage system covers the structures and equipment involved in
receiving, handling and storing palletized crates of new fuel, inspecting and handling individual
new fuel bundles and transferring these new fuel bundles into the fuelling machine.
New fuel transfer and storage equipment, shown on Figure 1, includes racks in the new fuel
storage room, a fork lift truck, crane lifting attachments, a pallet cart and new fuel transfer
equipment located in the new fuel loading room in the reactor building, just adjacent to the
reactor vault and in between the two fuelling machine maintenance locks. The new fuel
transfer equipment, shown on Figure 3, and located in the new fuel loading room, includes an
air balance hoist, jib crane, fuel bundle lifting tool and fuel spacer interlocking gauge for han-
dling and inspecting fuel bundles, and two new fuel transfer mechanisms for transferring fuel
bundles into a fuelling machine magazine. Each new fuel transfer mechanism, shown on Figure
4, includes a new fuel port, a loading trough and loading ram, an air lock valve, a magazine, a
shield plug, a transfer ram and drive and a magazine/port adapter. New fuel transfer auxiliaries
use instrument air to power the new fuel transfer equipment. They also connect the new fuel
transfer mechanism to the vapour recovery system and provide drainage for it. A local control
panel located on each new fuel transfer mechanism allows semi-automatic control of the
process for loading new fuel into the new fuel transfer mechanism magazine.
Up to 144 pallets each containing 36 fuel bundles can be stored on racks in the new fuel storage
room as part of the normal storage capacity, which accommodates more than eight months of
operation at equilibrium. Additional pallets of fuel can also be stored on the floor as part of the
abnormal storage capacity.
3.2 Component Description

Major new fuel transfer and storage components are as follows.
New Fuel Storage Room Racks
Racks are available in the new fuel storage room in the service building for storing palletized
crates of new fuel. Some of the racks are fixed while some of them are movable.
New Fuel Storage Room Lift Truck
A fork lift truck is available in the new fuel storage room to move the pallets of new fuel. It is
an electrically operated, stand-up rider type, and has a capacity of one fuel pallet. This fork lift
truck is also used to transport the pallets to the pick-up point for the service building crane
during transfer to the new fuel loading room in the reactor building.
Crane Lifting Attachment
Two fork lifting crane attachments are available; one located in the service building for use with
the service building crane, and one in the reactor building for use with the reactor building
crane. Two pallets can be lifted at one time with these attachments.
Cart
A powered, pallet cart is available at the equipment airlock level to transfer up to two pallets at
a time between the service building and the reactor building through the equipment airlock.
Air Balance Hoist and Jib Crane
The air balance hoist is mounted on a jib crane. This crane is pivoted from the building wall
above the two new fuel transfer mechanisms and covers the area of the new fuel loading room
through which new fuel bundles have to be moved. The air balancing hoist is free to travel
along the length of the jib. It is air-powered and can be adjusted to balance the load from a
control on the operator’s handgrip.
Fuel Bundle Lifting Tool
The fuel bundle lifting tool is used with the air balance hoist to manually load fuel bundles into
the new fuel transfer mechanism loading trough. The bundle lifting tool is a manually operated
device which clamps around the end plates of the fuel bundles, through lifting adaptors. When
the bundle is suspended, it is clamped securely by the toggle action of the tool linkage. The
lifting adaptors are mounted on spherical bearings which allow the adaptors to align with the
bundle end plates and permit the bundle to be rotated for inspection. The air balancing hoist
used with the bundle lifting tool allows the operator to transfer the fuel bundles without
damage and with a minimum of effort.
Fuel Spacer Interlocking Gauge
The fuel spacer interlocking gauge is used for checking the diameter of each fuel bundle over
the central spacers before the bundles are loaded into the new fuel transfer mechanism maga-
zine. This gauge is used while the fuel bundle is supported on the fuel bundle lifting tool. This
gauge consists of two pivoted segments and a dial gauge having a shaded area which indicates
the go/no-go range. The dial gauge is used to verify that the bundle has not changed shape
during transit and that the pencil spacers have not interlocked. This is to verify that the bundle
diameter has not expanded and that it will pass through the fuel handling tubes and the reactor
channel without interference.
Loading Trough and Loading Ram
The loading trough and loading ram are provided for moving the new fuel into the magazine of
the new fuel transfer mechanism. Two fuel bundles are normally loaded into the trough, and
after the bundles are loaded, the lid is closed and the bundles pushed into the magazine by the
loading ram. The trough is provided with a hinged lid which is interlocked to prevent operation
of the ram unless the lid is closed. Limit switches are mounted in the trough and provide an
indication to the control system when a fuel bundle is placed in the trough. The loading ram is
in line with the top magazine station of the new fuel transfer mechanism. The ram is an oil/air
operated, double acting cylinder, supplied from the new fuel transfer auxiliaries. Oil pressure is
used to extend the ram to provide smooth ram operation and adequate lubrication.
Air Lock Valve
The air lock valve is pneumatically operated and is installed between the loading trough and the
new fuel transfer mechanism magazine, to seal off the magazine whenever fuel is not being
loaded into the magazine. This valve prevents the spread of contamination from the fuelling
machine head, the maintenance lock or the fuelling machine vault, into the new fuel loading
room. Limit switches indicate when the valve is in the open and closed positions.
New Fuel Transfer Mechanism Magazine
The new fuel transfer mechanism magazine assembly consists of a leak-tight housing, a rotor,
and a drive unit. The magazine housing is a drum-like enclosure with a normally closed drain
connection to the active drainage system and a vent to the reactor building vapour recovery
system to remove any contamination through purging. The magazine rotor contains seven
tubular channels, six for accommodating the fuel bundle pairs and one for the new fuel transfer
mechanism shield plug.
New Fuel Transfer Mechanism Shield Plug
The shield plug, which is normally installed in the new fuel port except when new fuel is being
transferred to the fuelling machine, reduces radiation streaming into the new fuel loading room
when a fuelling machine containing irradiated fuel passes the end of the new fuel port in the
fuelling machine maintenance lock.
Transfer Ram and Drive
The function of the transfer ram is to remove and install the shield plug between the new fuel
port and the transfer mechanism magazine and to push the new fuel bundles from the new fuel
transfer mechanism magazine into the fuelling machine. The openings for the transfer ram on
the magazine front and rear covers are located at a bottom position, in line with the respective
magazine station. The ram head incorporates a latch assembly for engagement with the shield
plug. Ram position, for control of ram stopping and speed changing operation, is detected by a
shaft encoder mechanically connected to the chain drive sprocket shaft.
Magazine/Port Adapter
An adapter assembly is bolted to the new fuel transfer mechanism front cover, at a bottom
position in line with the respective magazine station and the fuel transfer ram. The adapter
connects the magazine to the new fuel port and consists of a spool piece with two dou-
ble-acting pneumatic cylinders. Its function is to hold the shield plug in position in the new fuel
port.
New Fuel Ports
The new fuel ports are mounted in embedments in the walls between the new fuel loading
room and the fuelling machine maintenance locks. Each port is a tubular connection with an
end fitting at one end extending into the fuelling machine maintenance lock and the other end
engaging with the fuel transfer mechanism port adaptor. When loading fuel into a fuelling
machine, the new fuel port becomes the passageway for bundle movement from the fuel
transfer mechanism to the fuelling machine.
10
Figure 3 New fuel transfer equipment
Figure 4 New fuel transfer mechanism
12
3.3 System Operation

New fuel is received on pallets and stored on racks in the new fuel storage room in the service
building. When required, one or two pallets of new fuel are transferred to the new fuel loading
room in the reactor building.
1) One new fuel pallet is removed from the new fuel storage room and transferred to the
service building crane using the new fuel storage room lift truck. This is repeated for the
second pallet.
2) Then the two new fuel pallets are lifted from the crane hall floor elevation and set on the
floor of the equipment airlock using the main service building crane.
3) The two new fuel pallets are transported through the equipment airlock into the reactor
building with the new fuel pallet cart.
4) From the air lock floor elevation, the two pallets are then lifted using the reactor building
crane and set in the new fuel loading room.
Keeping the new fuel bundles on the new fuel pallet protects the new fuel from damage and
contamination during transportation from the fuel supplier, during storage at the station and
during transfer from storage to the new fuel loading room. The new fuel pallet has slots the
fork lift has to engage into prior to lifting and thereby eliminates the chance of dropping the
entire pallet.
The new fuel storage room is kept locked and access to fuel is restricted to people who have a
valid reason to access or handle fuel. Security procedures are to be followed and the key to the
new fuel storage room is controlled by the shift supervisor who authorizes all access to it. The
serial number of each fuel bundle is recorded prior to loading into the new fuel transfer
mechanism.
In the new fuel loading room, the new fuel pallet is opened and individual fuel bundles are
manually unwrapped. One fuel bundle at a time is then picked up and moved onto the inspec-
tion table via the bundle lifting tool which clamps around the end plates of the bundle and
which is attached to the air-balance hoist. The bundle size (the bundle diameter over the
centre bearing pads) is carefully inspected with the fuel spacer interlocking gauge to ensure
that it will pass smoothly through the fuel handling system. Each fuel bundle is checked to be
free of abnormalities, damage or foreign matter prior to loading into one of the two new fuel
transfer mechanisms. Experience from operating CANDU plants indicates that very few new
fuel bundles have been rejected because of damage.
The new fuel loading room is heavily shielded and the spread of contamination, from the
fuelling machine and the fuelling machine maintenance locks into the new fuel loading room, is
prevented by the air lock valve, the shield plug, by maintaining a negative pressure at all times
in the magazine of the new fuel transfer mechanism through a duct connected to the vapour
recovery system, and by draining into the active drainage system any liquid which may accumu-
late in the magazine. The shield plug is not removed from the new fuel port unless the air lock
valve is closed, and the air lock valve is not opened unless the shield plug is installed in the new
fuel port.
New unirradiated fuel bundles, which are made up of natural uranium, have a very low level
radioactivity, and can be handled safely by personnel without protective clothing (Note: gloves
are normally worn).
Once the operator has manually placed two fuel bundles in the loading trough and closed the
trough lid, the fuel bundles are pushed into the new fuel transfer mechanism magazine by the
loading ram. The magazine is indexed, using the local control panel mounted on the new fuel
transfer mechanism. This process is repeated until the magazine contains the required number
of fuel bundles.
The transfer of fuel bundles from the new fuel transfer magazine to the fuelling machine
magazine is normally performed under complete computer control, using a series of steps.
Confirmation of the successful completion of each step, which is indicated by either a limit
switch or a shaft encoder, is required prior to starting the following step. Prior to beginning the
transfer sequence, the fuelling machine is clamped onto the new fuel port and its heavy water2
level is lowered to just below the snout.
To ensure that the system is ready to begin the sequence, the computer memory is checked to
confirm that the fuelling machine snout plug is stored in its magazine station and that the guide
sleeve has been installed in the snout. Feedback must indicate that all rams are retracted to
their home positions, that the air lock valve is closed, that there is new fuel in the new fuel
transfer mechanism magazine, and that the new fuel transfer mechanism magazine is not in
use. All conditions being satisfactory, the new fuel magazine is rotated to the shield plug
station, the ram is advanced to remove the shield plug from the port, and then retracted to
deposit the shield plug in the magazine for storage. The commands are then given to rotate the
fuelling machine magazine to an empty station and the new fuel magazine to a full station. The
new fuel transfer ram is then fully advanced to push two new bundles into the fuelling machine
magazine station. The rams are retracted to their respective home positions, the fuelling
machine and new fuel magazines are rotated to the next stations, and two more fuel bundles
are transferred. This process is repeated until the fuelling machine magazine contains the
required number of bundles. After completion of the fuel bundle transfer, the shield plug is
re-installed into the new fuel port and locked in place. The fuelling machine then removes the
guide sleeve from the new fuel port, installs its snout plug, raises its heavy water level and is
unclamped from the new fuel port in order to traverse to the reactor face.
Normal handling of new fuel in air or light water does not pose criticality problems for natural
uranium.
Note: New fuel bundles for the initial core are manually loaded into the fuel channels. Visual
inspections of these new fuel bundles are performed on a loading platform at the reac-
tor face.
2 Heavy water (D2O) has deuterium atoms instead of hydrogen atoms and is approximately 10% heavier than light
or regular water.
4 Fuel Changing
Fuel changing, depicted in Figure 5, involves the structures, equipment and activities required
to transport new fuel from the new fuel port to the reactor, to load and unload fuel from the
fuel channel, and to transport the discharged fuel from the reactor to the spent fuel port. This
is mainly performed by two fuelling machine heads remotely operated from the Main Control
Room.
Each fuelling machine head is suspended on a carriage, which moves horizontally along bridge
rails in its respective fuelling machine vault and on tracks in its respective fuelling machine
maintenance lock. A bridge assembly, adjacent and parallel to each reactor face, provides for
vertical movement of the fuelling machine head and its carriage. Each bridge and carriage
assembly allows a fuelling machine to access all 380 reactor fuel channels at the respective
reactor face. The maintenance lock tracks align with the lowest bridge position to allow trans-
fer of the fuelling machine and carriage between the fuelling machine vault and the mainte-
nance lock.
The fuelling machine D2O system supplies the fuelling machines with heavy water to cool
irradiated fuel temporarily stored in the fuelling machines, operate associated fuelling machine
mechanisms and generate injection flow. The oil hydraulic system supplies the fuelling ma-
chines and carriages with oil to operate and lubricate associated mechanisms. The catenary
system conveys heavy water, hydraulic oil and electrical power and control signals to and from
the mobile fuelling machine heads and carriages.
The fuelling machine heads are normally stored in two maintenance locks, where they can lock
onto the new fuel ports to accept new fuel, onto the service ports for maintenance and re-
hearsal, or onto the spent fuel ports to discharge irradiated fuel.
Powered shielding doors separate the fuelling machine vault and the reactor from the mainte-
nance locks and, when closed, provide shielding that allows personnel access to the fuelling
machines in the maintenance locks while the reactor is operating.
The on-channel fuel changing operations are performed in the direction of coolant flow in the
channel. Each fuel channel contains 12 fuel bundles. Typically, eight new fuel bundles are
inserted at the upstream end of a fuel channel, and eight irradiated fuel bundles are removed
at the downstream end. Since the coolant in adjacent fuel channels flows in opposite direc-
tions, for reasons of reactor symmetry, each fuelling machine is an upstream machine (carrying
new fuel) and a downstream machine (carrying irradiated fuel) at different times.

4.2.1 Fuelling Machine Head
The fuelling machine is a pressure vessel that comprises five major sub-assemblies, the snout,
the magazine, two separators and the ram, which are necessary for it to perform its required
functions. The fuelling machine head, shown on Figure 6, includes the fuelling machine
mounted and supported in a cradle assembly as well as the guide sleeve and tool, and the ram
adapter. The fuelling machine heads engage the end fittings at opposite ends of the same
channel assembly for the purpose of fuel changing. The fuelling machine head contains heavy
water and is designed to form an extension of the heat transport system pressure boundary to
enable refuelling operations to be carried out with the reactor at full power.
Snout Assembly
The snout assembly, located at the front of the fuelling machine and shown in Figure 7, enables
the fuelling machine head to clamp onto any reactor fuel channel and to seal the connection at
high pressure and temperature. To achieve loading of new fuel bundles and the discharging of
irradiated fuel bundles from the fuelling machine, the snout assembly is capable of locking onto
the new fuel port and the spent fuel port. It is also capable of connecting to the service ports as
required.
The snout assembly includes the centre support, the clamping mechanism, the snout emer-
gency lock assembly, the head antenna and the snout probes. The clamping mechanism is
engaged by oil hydraulic pressure. The snout emergency lock, a passive safety lock, prevents
inadvertent removal of the snout assembly from an end fitting when the channel closure is not
installed. When the fuelling machine is not attached to a fuel channel or port, the snout is
sealed with a snout plug.
Magazine Assembly
The magazine assembly, as shown on Figure 7, includes a twelve-chamber rotor inside a pres-
sure-retaining housing assembly for storing fuel bundles, channel closures, shield plugs, a snout
plug, a ram adapter, a guide sleeve and tool, and a flow assist ram extension (FARE) tool.
It also includes a drive system, comprising an indexing mechanism driven by an oil-hydraulic
motor for turning the magazine rotor, and the magazine emergency drive, a manual drive
facility to rotate the magazine rotor if the oil hydraulic motor becomes disabled.
Ram Assembly
The ram assembly, shown in Figure 8, provides the necessary axial movements and forces
required for the transfer and discharge of the fuel bundles, or the installation and withdrawal of
various tools and plugs such as the guide sleeve, the channel closure plug, the shield plug, and
the snout plug. The ram assembly consists of a B ram, a latch ram and a C ram. These rams are
essentially concentric tube assemblies supported within a pressure-retaining ram housing.
Each ram has a different head assembly to perform the necessary plug or fuel changing opera-
tion. Both the B ram and the latch ram are driven by oil hydraulic motors through a gear
system and ball screws. The C ram acts as a multi-stage hydraulic cylinder and is powered by
the fuelling machine D2O system.
Separators
For each fuelling machine head there are two separator assemblies which perform identical
functions and operate in synchrony. They are mounted in penetrations through the magazine
front cover at points just forward of the magazine tubes and each assembly is oriented at 67.5°
from the vertical. The separators sense the gaps between two fuel bundles or between a fuel
bundle and ram adapter or FARE tool, provide a signal to the computer to stop the fuelling
machine ram at the correct position, restrain the axial motion of the fuel column due to hydrau-
lic forces from the fuel channel flow, push the bundles that have been separated from the fuel
column into the magazine to allow clearance for magazine rotation, and verify the presence of
the shield plug and the FARE tool during the refuelling operations. High pressure heavy water
from the fuelling machine D2O system is used to operate the separators.
Ram Adapter
The ram adapter, shown in Figure 9, attaches to the C ram head to provide a flat face for
contacting the end plate of a fuel bundle and to centralize and support the C ram, which
minimizes sagging of the ram. The ram adapter consists of five parts: the body, the sleeve, the
stem, the inner spring and the outer spring. The ram adapter body provides the contact face
with the fuel. The face is machined to simulate the end face of a fuel bundle, so that the gap
between the ram adapter and the fuel is identical to the gap between two fuel bundles. The
body is also machined to allow adequate cooling flow. The stem is screwed into the bore of the
body and is contoured to connect with the C ram. The sleeve is spring-loaded by the concentric
inner and outer springs and locks the C ram to the stem.
Guide Sleeve and Tool
The channel closure seal face in each fuel channel end fitting forms a step in the fuel channel,
the bore of the channel being smaller than the bore of the end fitting. This discontinuity forms
an obstruction to the smooth passage of fuel bundles and shield plugs. The guide sleeve
provides a smooth bore for the passage of fuel bundles and shield plugs between the fuel
channel and the fuelling machine.
The guide sleeve is moved between the fuelling machine magazine and the fuelling machine
snout by the fuelling machine ram using a guide sleeve insertion tool. The guide sleeve and the
insertion tool are locked together except when the guide sleeve is in position in the snout and
end fitting. When not in use, they are stored together in the fuelling machine magazine and are
held in the magazine tube by a locating tube and spacer.
Snout Plug
The snout plug, shown in Figure 10, is used to seal the fuelling machine snout when the fuelling
machine is not attached to a fuel channel or port. It allows the interior of the fuelling machine
to be maintained full of water at a controlled temperature and pressure, to maintain cooling of
the irradiated fuel while in transit from the reactor fuel channel to the spent fuel port. The
fuelling machine snout plug is composed basically of two parts, the latching mechanism and the
seal assembly. The latching mechanism which makes up the rear half of the plug has four jaws
which are extended by a spider mechanism. These four jaws locate the plug in a groove in the
fuelling machine snout centre support. The front end of the snout plug is comprised of the seal
assembly, which is screwed onto the latch assembly. The seal assembly contains a large elas-
tomer O-ring seal and the associated mechanism required to expand and retract it.
4.2.2 Fuelling Machine Bridge and Carriage

Two fuelling machine bridge and carriage assemblies are provided, one at each reactor face. As
shown in Figure 5, each bridge spans the face of the reactor and supports a carriage assembly,
when the carriage is in the fuelling machine vault. The carriage assembly supports a fuelling
machine head. Each bridge moves vertically on two guide columns, to provide vertical motion
(y-motion) of the fuelling machine head. Carriage wheels ride on rails mounted on the bridge
(when in the fuelling machine vault) and on the maintenance lock tracks (when in the mainte-
nance lock) to provide horizontal motion (x-motion) of the fuelling machine. With the bridge in
its lowest position, the rails on the bridge are aligned with matching rails on the maintenance
lock tracks, enabling the carriage, with the fuelling machine head, to transfer between the
fuelling machine vault and the maintenance lock.
Each carriage assembly consists of a drive unit assembly, and upper and lower gimbal assem-
blies, as shown in Figure 11. In addition to horizontal motion along the bridge and maintenance
lock rails, the carriage assembly also provides fine vertical motion and motion toward and away
from the reactor (z-motion) and allows a controlled amount of rotation of the head. It also
provides a termination point for the catenary system.
The carriage also includes clamping mechanisms, which securely anchor the carriage to the
bridge rails when the fuelling machine head is clamped to a reactor end fitting, to prevent
excessive loads from being applied to the end fitting by the fuelling machine head during a
seismic event.
4.2.3 Fuelling Machine D2O System

The D2O system supplies the fuelling machine with heavy water compatible with the heat
transport system, and regulates the pressure and flow of heavy water to and from the various
parts of the fuelling machine. Heavy water is used in the fuelling machine for the following
functions:
 Cooling irradiated fuel stored in the fuelling machine magazine.
 Operating the C ram.
 Operating the separators.
 Preventing high temperature heat transport system water from entering the fuelling ma-
chine when the fuelling machine is coupled to a fuel channel.
The fuelling machine D2O system is divided into the following major parts:
 Two valve stations that contain most of the control components of the stationary portion of
the system.
 The mobile portion of the system located on each fuelling machine head.
 Catenary hoses that connect the stationary portion of the system with the mobile portion
located on the fuelling machine head.
4.2.4 Fuelling Machine D2O Supply System

The fuelling machine D2O supply system can be divided into three subsystems:
1. The supply subsystem, which supplies pressure regulated, filtered heavy water, drawn from
the primary heat transport system storage tank and/or the D2O return subsystem, to the
two D2O system valve stations.
2. The D2O return subsystem, which accepts the D2O discharge from the fuelling machine head
via the D2O System, and cools and purifies it before returning it to the supply subsystem.
3. The D2O leakage collection system, which collects the D2O from all drain, leak and vent
points in the system into two tanks.
4.2.5 Oil Hydraulic System

Each of the two fuelling machines is operated by an individual oil hydraulic system. The oil
hydraulic system provides a continuous oil supply, at controlled conditions of flow, pressure
and temperature, to the mobile fuelling machine head and carriage, via a catenary system. This
oil supply is used to operate associated actuators on the fuelling machine head and carriage.
4.2.6 Fuelling Machine Head Accessories

Fuelling machine head accessories include the FARE tool, the pressure tube seal and latch
assembly and the fuel grapple system.
FARE Tool
The FARE tool provides a restrictive element in low flow fuel channels, which develops the
necessary coolant flow drag force to overcome the frictional resistance of the fuel bundles in
the channel, in order to push the fuel column downstream when the fuelling machine ram
cannot provide this force without entering the core. The FARE tool has an overall length of two
fuel bundles and is handled like a pair of fuel bundles. For simplicity, it can be considered as a
piston in a cylinder with a large amount of leakage.
The FARE tool is carried in the fuelling machine magazine and is pushed into the fuel channel in
the same manner as the fuel bundles after the last new fuel bundle is inserted in the fuel
channel.
After the irradiated fuel bundles are discharged from the fuel channel and the fuel string is
returned to its normal in-core position, the C ram attaches to the FARE tool for its removal and
return to the fuelling machine magazine.
Pressure Tube Seal and Latch Assembly
The pressure tube seal and latch assembly can be inserted into a fuel channel by the fuelling
machine so that it stops the channel flow, thus enabling ice-plugs to be formed in the channel
feeders by means of a freezing operation. This channel flow blockage is sometimes required to
allow inspection of the channel or other possible fuel handling recovery scenarios. It is only
used during reactor shut downs at low pressures. The pressure tube seal and latch assembly
consists of a seal assembly with a mechanism for radially expanding an O-ring to seal the bore
of the pressure tube and a latch assembly which is similar to a fuel channel shield plug.
Fuel Grapple System
The fuel grapple system enables the fuel channel to be defuelled from one end using a single
fuelling machine when the other fuelling machine is disabled or when a channel is to be defu-
elled. It includes a fuel grapple assembly, for latching onto a fuel bundle; grapple extensions,
which attach to the grapple assembly or other extensions for grappling fuel bundles beyond the
reach of the fuelling machine rams; a ram grapple adaptor, which allows the fuelling machine to
operate latches on the extensions; and a grapple reset tool for releasing irradiated fuel bundles
from the fuel grapple.

To begin the fully automated fuel changing process, one of the two fuelling machines clamps
onto the new fuel port, its water level is lowered to just below the fuelling machine snout, its
snout plug is removed, the guide sleeve is installed and the fuelling machine typically accepts
eight new fuel bundles from the new fuel port. The guide sleeve is removed, the snout plug
re-installed and the water level is raised. Fuel bundles are received and stored in pairs.
With the shielding doors open, both fuelling machines travel along the maintenance lock tracks
onto the fuelling machine bridges at each face of the reactor. The bridges are then raised and
the fuelling machines positioned on the bridge so that the fuelling machines are located at
opposite ends of the selected fuel channel. The fuelling machine with the new fuel is at the
upstream end while the other fuelling machine, which contains no fuel, is at the downstream
end of the same fuel channel.
Each fuelling machine then moves forward to home and lock onto the fuel channel end fitting
to form a secure and pressure-tight joint with the channel. Figure 12 depicts a typical
eight-bundle refuelling sequence.
The fuelling machines remove their snout plugs followed by the fuel channel closures. At this
point, the fuelling machines become part of the heat transport system pressure boundary. The
fuelling machine at the upstream end of the fuel channel removes the shield plug and typically
inserts eight new fuel bundles, two at a time, into the channel. The downstream fuelling
machine removes the shield plug and typically receives eight irradiated fuel bundles, two at a
time, into its magazine. The remaining twelve-bundle fuel string is then moved back to the
correct in-reactor position and the shield plugs and channel closures are re-installed in the fuel
channel. The fuelling machines re-install the snout plugs in their snouts, the channel closures
are checked for leaks and the fuelling machines are then retracted from the end fitting.
The fuelling machine containing irradiated fuel returns to the maintenance lock and discharges
the irradiated fuel through the spent fuel port.
An eight-bundle refuelling sequence, from the time that one fuelling machine clamps onto the
new fuel port to accept new fuel to the time that the other fuelling machine unclamps from the
spent fuel port after discharging irradiated fuel, as described above, typically takes approxi-
mately two hours. This includes approximately 30 minutes at the new fuel port, a little over 50
minutes at the reactor face and approximately 35 minutes at the spent fuel port.
Other fuelling schemes are also possible. For example, an alternate scheme is available for the
first refuelling visit to a fuel channel; it involves the fuelling machine at the upstream end of a
fuel channel inserting eight new fuel bundles as usual, and the fuelling machine at the down-
stream end of this fuel channel receiving ten irradiated fuel bundles and returning the first
bundle pair removed to the fuel channel. Although refuelling fewer than eight bundles at a time
is favourable with respect to improving fuel economy; it increases wear and tear on fuel han-
dling equipment, particularly on the fuelling machines. A fuelling machine can accommodate up
to eight fuel bundles when carrying a FARE tool, up to ten fuel bundles without a FARE tool, and
up to twelve fuel bundles when the reactor is shut down. Refuelling uneven numbers of fuel
bundles is expected to be limited to abnormal situations.
Bundle movement is controlled by the downstream fuelling machine, assisted by the coolant
flow inside the channel which moves the string of fuel bundles due to hydraulic drag.
The channel flow, however, varies depending on location. In the outer zone fuel channels
where the flow is lower, a FARE tool is used to provide sufficient hydraulic drag force to move
the fuel string under the reduced channel flow condition. The FARE tool is loaded at the end of
the fuel string after new fuel bundles have been inserted by the upstream fuelling machine.
The FARE tool is removed after the required number of irradiated fuel bundles is received by
the downstream machine and the fuel bundle string is moved back to its correct in-reactor
position. The FARE tool is used rather than the upstream machine ram so that the latter does
not enter the core to become activated and contribute to the dose rate to maintenance per-
sonnel working on the machine. The FARE tool (which also becomes activated when it passes
into the core) is discharged from the machine in the same manner as irradiated fuel bundles
before maintenance work is started. The FARE tool is not required on inner zone channels.
The failed fuel detection and location system, which is part of reactor instrumentation and
control, is used to detect and identify failed fuel in the reactor core. Its gaseous fission moni-
toring system detects the reactor loop in which the fuel failure has occurred. Its failed fuel
location system, also referred to as the delayed neutron monitoring system, is then used to
identify in which channel of the particular coolant loop, the fuel failure occurred and to find, in
this particular channel, which bundle pair contains the failed bundle(s). It does this by monitor-
ing delayed neutrons via a sample station. The delayed neutron count rate decreases sharply
when a failed fuel bundle leaves the channel flow. Failed fuel bundle(s) are then identified via
their fuelling machine magazine position. They are otherwise handled by the fuelling machine
in the same way as undamaged fuel bundles.
The fuel changing system maintains irradiated fuel, temporarily stored in its magazine, at
acceptable temperatures by keeping it submerged and accommodating adequate heavy water
circulation in the fuelling machine magazine, except for the short period of time when fuel
bundles are exposed to air for discharge via the spent fuel port. The temperature of the water
in the fuelling machine magazine is monitored.
In order to prevent damage to fuel:
 Fuel changing mechanisms are normally operated automatically according to pre-defined
sequences.
 Interlocks inhibit mechanism operations unless the configuration is acceptable.
 Fuel changing components that interface with the fuel such as the ram adapter and the
separator assemblies are designed to limit the stresses in the fuel bundle to acceptable le-
vels.
 When handling fuel, the ram forces are limited to prevent unacceptable axial forces on the
fuel bundles.
The geometry of the fuelling machine maintains the fuel subcritical even with heavy water
present.
Fuel changing operations are performed remotely to limit radiation exposure of site personnel.
The fuelling machine shielding doors provide radiation shielding and a ventilation barrier
between the fuelling machine vaults and maintenance locks when the fuelling machine heads
and their carriage and suspension are in their maintenance locks to permit personnel access for
their maintenance. The atmosphere in the maintenance locks can be purged.
Prior to the start of maintenance work on the fuelling machine, plugs and tools, such as the
shield plug, ram adapter and FARE tool, which become activated when they enter the core, are
discharged from the fuelling machine.
21
Figure 5 Fuel changing equipment
Figure 6 Fuelling machine head
Figure 7 Front end of fuelling machine
24
Figure 8 Ram assembly – exploded view
Figure 9 Ram adapter and operation
26
Figure 10 Fuelling machine snout plug
27
Figure 11 Carriage assembly
Figure 12 Fuel changing sequence
5 Spent Fuel Transfer and Storage

Spent fuel transfer and storage, which can be seen on Figure 1, covers the structures and
equipment involved in the discharge of irradiated fuel from the fuelling machine and the
transfer and storage of this irradiated fuel in bays until the fuel is transferred to dry storage. It
consists of spent fuel discharge equipment, shown in Figure 13, spent fuel transfer auxiliaries,
spent fuel transfer equipment, shown in Figure 14, and storage bay equipment. It also includes
failed fuel canning equipment, shown on Figure 15, which is a subset of the spent fuel discharge
equipment, and failed fuel handling equipment, which is a subset of the storage bay equip-
ment. Spent fuel transfer and storage involves three bays filled with light, or regular, water: the
spent fuel discharge bay, spent fuel reception bay and spent fuel storage bay (also referred to
as the main storage bay).
The discharge equipment, which is located in the reactor building, receives irradiated fuel from
the fuelling machines and lowers it into the discharge bay. It includes spent fuel ports that
interface with the fuelling machines, and elevators to lower the fuel bundles into the water. It
also includes the failed fuel canning equipment in the discharge bay, for handling and canning
the small quantity of failed fuel.
The spent fuel transfer auxiliaries provide the fluid flows (air and water) for the spent fuel
discharge mechanisms. They include a standby cooling system, D2O leak collection system, fuel
stop actuating system, port relief system, fuelling machine overflow detection system and port
valve actuating system.
The spent fuel transfer equipment provides for the transportation of irradiated fuel from the
discharge bay in the reactor building to the reception bay and the storage bay in the service
building. It includes the discharge and transfer canal conveyors, the storage tray conveyor, a
conveyor rack and cart, as well as tools and accessories for manipulating the fuel in the recep-
tion bay. The transfer of irradiated fuel between buildings is under water through a contain-
ment gate and a transfer canal, which connects the reception and storage bays under water.
The reception bay is connected under water with the main storage bay by the penetration for
the storage tray conveyor.
The storage bay equipment provides for the storage of irradiated fuel in the reception and
storage bays. It includes the storage bay man-bridge, the storage trays and their supports, a
storage tray lifting tool and the failed fuel handling equipment for handling and storing canned
failed fuel in the reception bay.
An irradiated fuel channel shield plug (with its jaws retracted by a locking mechanism) or FARE
tool can be received from either fuelling machine and accommodated similarly to fuel bundles.
Irradiated elements of the reactivity control mechanisms can be accommodated in the storage
bay for underwater storage.

5.2.1 Spent Fuel Discharge Equipment
Spent fuel discharge equipment, shown in Figure 13, includes the components described as
follows, in addition to the failed fuel canning equipment described in Section 5.2.5.
Spent Fuel Ports
Spent fuel ports are mounted in port sleeves embedded in the walls between the spent fuel
discharge room and the fuelling machine maintenance locks. Two ball valves are mounted in
series on the bay end of each port to seal the port and maintain the containment boundary.
The ports are cantilevered from the discharge room end, via a flange on the port housing, to
provide the degree of flexibility necessary for fuelling machine lock-up. A detachable
stainless-steel liner tube is installed in the bore of the port housing, forming an annulus be-
tween the liner and the housing, for emergency cooling of irradiated fuel, if required. The liner
has a row of 10 holes along the top and 5 holes radially around the fuelling machine end. The
annulus connects to the standby cooling system water supply to provide adequate cooling
water through the liner holes for cooling of any irradiated fuel bundles, which may, inadver-
tently, be held up in the port.
Elevators
Two elevator ladles accept the fuel bundles from the spent fuel ports and are electrically driven
to lower them onto a conveyor rack on a conveyor at the bottom of the discharge bay. The two
elevators consist of ladles running on tubular and angle rails. The 30° inclination of the rails
allows both elevators to terminate at the single conveyor at the bottom of the discharge bay.
The two elevators are interlocked so that only one ladle can be driven down to the unloading
position at any one time. Each ladle accommodates two fuel bundles and is suspended and
driven by a stainless steel cable from a drum mounted above the rails.
5.2.2 Spent Fuel Transfer Auxiliaries

The spent fuel transfer auxiliaries comprise a series of systems, which provide the fluid flows
(air and water) for the spent fuel discharge mechanisms. The systems include a standby cooling
system, a D2O leak collection system, a fuel stop actuating system, a port relief system, a
fuelling machine overflow detection system and a port valve actuating system.
Standby Cooling System
The standby cooling system can be initiated to provide cooling water to the spent fuel port via
its liner tube and/or to the length of travel to the bay water via spray headers mounted parallel
to the elevator ladle rails, to provide cooling to exposed irradiated fuel if the discharge process
is delayed.
D2O Leak Collection System
The D2O leak collection system collects heavy water that leaks past the inner ball valve seals
into the spent fuel port cavity between the inner and outer valves, during fuelling machine
operations at the spent fuel port, when the fuelling machine and spent fuel port are filled with
heavy water and pressurized. The D2O leak collection system includes a pneumatically oper-
ated drain valve, which is connected to the cavity between the port valves and is opened to
drain any heavy water into the D2O system. At the same time, the cavity is supplied with air to
assist in clearing any heavy water from the cavity.
Fuelling Machine Overflow Detection System
The fuelling machine overflow detection system includes a D2O collection container located
below the discharge end of the spent fuel port, and connected to the port. The container is
drained into the spent fuel discharge bay through a removable orifice. If the level of heavy
water in the fuelling machine fails to be lowered, a liquid level detector probe in the container
actuates a transmitter.
Port Valve Actuating System
The port ball valve actuating system is used to operate the spent fuel port ball valves via
pneumatic actuators.
5.2.3 Spent fuel transfer equipment

Spent fuel transfer equipment, shown in Figure 14, includes the following components.
Conveyor Rack
Each conveyor rack (also referred to as a transfer rack) can receive up to 12 fuel bundles from
the spent fuel discharge elevator in two rows of 6; although 8 are normally carried. A conveyor
rack can also accommodate canned failed fuel bundles.
Conveyor Cart
The conveyor cart (also referred to as a transfer cart) supports the conveyor rack and travels
between the discharge bay and the reception bay on conveyors. The cart runs on eight
stainless steel wheels which enable it to traverse the gap between the two conveyors. The cart
engages drive chains on each conveyor.
Discharge Bay and Transfer Canal Conveyors
Two conveyors, the discharge bay and transfer canal conveyors (also referred to as spent fuel
transfer conveyors), allow the conveyor cart to travel between the discharge bay and the
reception bay. The transfer canal conveyor is sometimes also referred to as the reception bay
conveyor.
The conveyors are of similar construction and consist of a welded stainless steel frame with a
flat upper surface on which the cart runs. A rectangular guide rail on the upper surface mates
with three sets of rollers on the cart to provide lateral guidance for the cart. A mechanical stop
is mounted at each end of cart travel. The two conveyors are driven by electrical motors
mounted above water on the discharge bay walkway floors, and vertical shafts used to drive
the underwater gearboxes, conveyor sprockets and chain.
The discharge bay conveyor is mounted on a series of supports secured to embedments in the
bottom of the reception bay. The conveyor is secured to the supports by captive screws and
tapered alignment pins which facilitate conveyor removal and installation. The conveyor is also
provided with lifting lugs.
The transfer canal conveyor is mounted on supports in the bay and also sits directly on the
concrete in the transfer canal. It is located laterally and vertically by a slotted bracket assembly
at the discharge bay end of the canal and is secured by captive screws and tapered alignment
pins only at the reception bay end to facilitate removal and installation.
Fuel Transfer Canal Containment Gate
The fuel transfer canal containment gate is located at the entrance to the fuel transfer canal in
the spent fuel discharge bay. The stainless steel gate is raised and lowered by a pneumatic
actuator and includes a safety latch actuated by a second pneumatic actuator. It is closed and
becomes part of the containment boundary when the spent fuel port ball valves are open
during the discharge of irradiated fuel from the fuelling machine and through the spent fuel
port. Otherwise the gate is normally open to permit water circulation between the bays. The
containment gate opening permits passage of the conveyor cart with a full load of fuel.
Storage Tray Conveyor
The storage tray conveyor transports spent fuel storage trays, in any state of loading, between
the reception bay and the storage bay. The manually driven storage tray conveyor has three
superimposed frames, capable of telescopic extension. The bottom frame is bolted to the floor
of the reception bay, while the other end extends into a rectangular opening in the wall be-
tween the reception bay and storage bay. The storage tray conveyor is manually operated from
a handwheel located on the reception bay walkway and extended or retracted to the limit of its
travel as required.
Tools and Accessories
Spent fuel transfer equipment tools and accessories are located in the reception bay. They
enable fuel bundles to be transferred from the conveyor rack to a storage tray.
5.2.4 Storage Bay Equipment

Storage bay equipment, which can be seen in Figure 1, includes the following components in
addition to the failed fuel handling equipment described in Section 5.2.6.
Storage Bay Manbridge
The spent fuel storage bay manbridge consists of an over-running, electric, travelling crane to
which an under-slung walkway is attached. A hoist with an electrically driven trolley provides
the lifting capability. The walkway spans the width and runs the full length of the storage bay
to provide a working platform for operating personnel engaged in handling irradiated fuel or
other active reactor components. The crane and hoist are controlled from a single pendant
control station suspended from the hoist so that an operator can manipulate the crane from
the walkway or the side of the bay.
Storage Trays
Each storage tray can hold 24 fuel bundles in two rows of 12 each. The trays are of
stainless-steel welded construction, with contoured cradle strips to support and separate the
fuel bundles. The trays are stackable. Each tray is provided with two lifting plates correspond-
ing to the pins on the storage tray lifting tool. Each tray is also provided with two tapered
locating pins and two slots, the pins engaging with the slots of the next-higher tray in the stack.
Storage Tray Supports
Storage tray supports are provided in the reception bay, and in the main storage bay, to sup-
port the stacks of trays. Each support consists of a diagonally braced, stainless-steel frame,
supported on raised pads.
Storage Tray Lifting Tool
This tool is similar to the storage tray lifting tool used in the reception bay, except for the length
of the tool and the orientation of the handle to the head.
5.2.5 Failed Fuel Canning Equipment

The following failed fuel canning tools and equipment, shown on Figure 15, are available in the
spent fuel discharge bay.
Carousel
A carousel provides temporary storage for failed fuel, for a decay period of up to two months
after the bundle has been discharged from a fuelling machine. The carousel base is a weldment
of angles bolted onto the bay floor. The outside walls consist of an octagonal lower section
open at the bottom, with a removable conical hood also of octagonal shape covering the top of
the tray. The 360 tray is divided into 12 sections, with each section capable of holding one
failed fuel bundle. The maximum capacity of the carousel is 11 bundles, to allow an empty
section to line up with the open portion of the carousel canopy. The carousel is manually
operated from a handle located on the discharge bay walkway. A ventilation pipe connects the
hood to the radioactive off-gas waste management system, to collect and purge active gases
released by the failed fuel bundles in the carousel.
Failed Fuel Can
Each stainless steel failed fuel can is able to accommodate one failed fuel bundle after the
necessary decay period in the carousel. The can has one open end for the bundle insertion,
while other end is blanked and punched with a small hole to prevent pressure building up
inside the can, during and after canning operations. The can is sealed by a lid, after the failed
fuel bundle has been pushed into it.
Canning Device
The canning device consists of a can storage rack, a can support, a fuel bundle trough, and a
bundle transfer ram. All these components are submerged and mounted on the canning device
support frame.
Failed Fuel Bundle Lifting Tool
The failed fuel bundle lifting tool enables failed fuel bundles to be lifted from the conveyor rack
into the carousel, and from the carousel to the canning device.
Lid Rack Handling Tool
The lid rack handling tool is used to transfer full lid racks from the ladle onto the canning
device, and move empty lid racks back to the elevating table.
Elevating Table
The elevating table is a carbon-steel rectangular table, which can be hooked onto either of the
elevator ladles. A rack of can lids can then be loaded onto the table and lowered into the bay
by the elevator.
Can Handling Tool
This tool is used to move empty failed fuel cans between the elevator ladle and the can storage
rack, between the rack and the trough, and to move full cans between the trough and conveyor
rack.
Lid Handling Tool
This tool allows the failed fuel can lids to be picked up from the storage rack and placed in
position in the canning equipment ram. The tool consists of a length of pipe, with a hook at the
top and a contoured fork at the bottom.
5.2.6 Failed Fuel Handling Equipment

The following failed fuel handling tools and equipment are available in the reception bay.
Transition Table
The transition table is used to provide a transition point between the reception bay and the
failed fuel storage area, which is lower, and to accommodate the tools of different lengths.
Can Handling Tool
Two can handling tools are provided for handling cans of failed fuel in the reception bay. These
tools are similar to the can handling tool used in the discharge bay. Each tool consists of a
handle connected to a double lifting hook by a length of pipe.
Failed Fuel Handling Storage Tray
These trays are of similar construction to the main storage trays, but can only hold 10 canned
fuel bundles in a single row. These trays are also stackable.

5.3.1 Normal Operation
Irradiated fuel bundles removed from the reactor remain immersed in heavy water within the
fuelling machine until the fuelling machine magazine rotates to align the first pair of bundles
with the spent fuel port. This raises the two bundles above the heavy water level inside the
fuelling machine head and exposes them to air. This rotation starts a timer to measure the
time the fuel bundles are exposed to air. The irradiated bundles will normally be exposed to air
for less than two minutes.
After irradiated fuel is received, two bundles at a time from the fuelling machine via the spent
fuel port, the elevator lowers the bundles into the water of the discharge bay, and deposits
them on the conveyor rack, supported by the conveyor cart and correctly indexed under the
elevator by the discharge bay conveyor drive, in position to receive pairs of fuel bundles.
If, due to a delay, the irradiated fuel remains in air longer than a predetermined time, an alarm
is provided to the operator. In such a case, the operator would then select one of the following
options to supply cooling water to the exposed fuel bundles, to prevent fuel bundle overheat-
ing and possible subsequent bundle failure.
If the fuel bundles are stuck in the port and the outer ball valve can be closed, the level of
heavy water in the fuelling machine can be raised to flood the port with heavy water and cool
the fuel.
The standby cooling system can be initiated to provide cooling water to the spent fuel port via
its liner tube if fuel is stuck in the port such that the outer ball valve cannot be closed, and/or to
the length of the elevator travel to the bay water via spray headers mounted parallel to the
elevator ladle rails, if fuel is stuck on the elevator.
Once all fuel bundles are discharged onto the conveyor rack, the containment is sealed by
closing the spent fuel port globe valves, and the containment gate is then opened. The con-
veyor cart then travels on the discharge bay conveyor up to the containment gate. Once the
cart reaches the space between the two conveyors, one slotted bracket on the cart disengages
the pins of the chain as the pins go around the sprocket of the first conveyor. This automati-
cally aligns the other slotted bracket of the cart to engage the pins of the reception bay con-
veyor chain, which then moves the cart to the reception bay.
The fuel bundles are then transferred individually to a spent fuel storage tray. Full storage trays
are moved onto the extendable storage tray conveyor for transfer to the storage bay via an
opening in the wall between the two bays. Once in the storage bay, individual trays are lifted
by a hoist on the storage bay manbridge and placed in stacks on a storage tray support. The
storage tray conveyor then retracts until it closes the opening in the wall with its flow obstruc-
tion plate. This minimizes the movement of water between the two bays.
An operator on the manbridge manoeuvres all storage trays, and stacks them, using an electri-
cal hoist and storage tray handling tool.
Irradiated fuel bundles are typically stored in the spent fuel bays for a minimum of six years
prior to transfer to dry storage; this allows the heat of a natural uranium fuel bundle with
maximum burnup to decay enough to allow for dry storage. The typical transfer of irradiated
fuel to dry storage involves transferring irradiated fuel from storage trays into a storage basket,
drying and seal welding the basket, transferring the basket, via a transfer flask, to one of the
multiple storage cylinders within a concrete storage module. Fuel transfers are typically carried
in batch processes where the inventory from approximately one year of irradiated fuel produc-
tion is transferred to dry storage. The quantity of fuel baskets transferred to dry storage is
selected to completely fill the storage cylinders involved. Refer to Chapter 19 for additional
information on the dry storage of irradiated fuel.
The spent fuel storage bay can accommodate the storage of irradiated fuel in stacked spent fuel
storage trays, in addition to the space required for equipment to prepare the fuel for dry
storage. The storage capacity of the spent fuel storage bay accommodates the storage of
irradiated fuel for a minimum of six years of wet-storage after removal from the reactor, the
accumulation of inventory from at least one year of irradiated fuel production for transfer to
dry storage, plus a half-core off-load (sometimes referred to as one loop) for emergency
purposes.
Normal handling of irradiated fuel bundles in air or light water does not pose criticality prob-
lems for natural uranium fuel.
Operations are carried under water as much as possible and the amount of time irradiated fuel
spends in air is minimized to maintain acceptable temperatures for the irradiated fuel and
prevent fuel damage.
During handling, the fuel bundle is supported on at least two bands of its wear pads. Minor
axial forces only are directly applied to the fuel sheaths via the endplates and endcaps.
Discharge operations are controlled remotely and the spent fuel discharge room is inaccessible
to personnel while irradiated fuel is being discharged or when there is a high radiation field.
The depth of water in the spent fuel bays has been calculated to limit exposure to any person-
nel in the area to well below allowable levels. Long-handled manual tools have markers to show
safe operating lift heights. Fixed area monitors are permanently installed in the spent fuel
transfer and storage areas and provide safe notification to personnel when radiation levels
exceed allowables. Consequently, operators do not normally require personal protective
equipment for radiation protection when handling irradiated fuel during normal operation.
5.3.2 Failed Fuel Handling

Once a failed fuel bundle has been detected, identified and removed from the reactor, it is
discharged from the fuelling machine head and lowered on the spent fuel elevator in the same
way as a normal fuel bundle. However, in the case of a failed fuel bundle, the conveyor rack is
indexed two positions and the failed fuel bundle is removed from the rack and inserted into the
carousel using the failed fuel bundle lifting tool suspended from the discharge bay monorail
hoist. The carousel is then indexed to position the bundle under the carousel canopy.
The failed fuel bundle is retained in the carousel, where its radioactive gases are captured and
directed to the radioactive offgas waste management system, for a period of up to two months,
during which time radioactive gases will have decayed to safe levels.
Protective equipment is required when handling the small quantity of failed fuel.
If necessary, the failed fuel bundle is canned; otherwise, it is further handled as a normal
bundle. A fuel bundle requiring canning is lifted from the carousel and placed in the trough of
the canning device. A manually driven ram is then actuated to push the bundle into the can. A
lid handling tool is then used to position the lid such that the ram can drive the lid into the
tapered open end of the can.
The canned fuel bundle is then transferred via the discharge bay and transfer canal conveyors
into the reception bay.
Canned failed fuel bundles are handled with a can handling tool and stored on failed fuel
storage trays in a deeper area of the reception bay specially designated for failed fuel storage.
A failed fuel storage tray lifting tool is used to maneuver and stack these trays. The failed fuel
area of the reception bay can accommodate the small quantity of failed fuel.
Figure 13 Spent fuel discharge equipment
38
Figure 14 Spent fuel transfer equipment
39
Figure 15 Failed fuel canning equipment
6 Problems
1. To accommodate eight months of reactor operation at equilibrium,
a. Approximately how many channels are required to be refueled by the fuelling machines?
b. Approximately how many fuel bundles need to be stored in the new fuel storage room?
c. How many pallets of new fuel are required?
2. Estimate the minimum number of fuel trays that the spent fuel storage bay needs to ac-
commodate, assuming that fuel is transferred to dry storage when an accumulation of fuel
from approximately one year of production has decayed adequately for dry storage?
3. Describe the typical direction of on-channel fuel changing operations and the benefit(s) of
operating in this direction.
4. Why does the level of water in the fuelling machine need to be lowered below its snout
level for the transfer of fuel bundles to the spent fuel port and from the new fuel port?
5. What measures can be taken if a fuel bundle(s) is (are) stuck during spent fuel transfer?
6. What is the purpose of the snout emergency lock?
7 Further Reading
[ISA1984] P. Isaac, “Evolution of On-power Fuelling Machines on Canadian Natural Uranium
Power Reactors”, AECL-8337, October 1984
[JAC2004] W.H. Jackson, “Design of On-Power Fuelling Machines”, CNS Bulletin, Vol. 25, No.
3, September 2004
[STE2006] R.G. Steed, Nuclear Power in Canada and Beyond, General Store Publishing
House, ISBN-13: 978-1-897113-51-6, 2006
8 Acknowledgements
George Ilich
Benjamin Rouben
Thanks are also extended to Bill Garland for expertly editing and assembling the final copy.
1
CHAPTER 21
CANDU In-Core Fuel-Management
prepared by
Dr. Benjamin Rouben, 12 & 1 Consulting, Adjunct Professor,
McMaster University & University of Ontario Institute of Technology (UOIT)
Summary:
This chapter describes the physical and mathematical models and the computational methods
used in the management of fuel in CANDU nuclear reactors. The concepts important to the topic
of fuel management are explained: fuel irradiation (fluence), fuel burnup, cross-section averag-
ing. The various levels of physics models used in fuel management are presented in quantitative
detail: these are the continuous refuelling model, the time-average model, patterned-age
snapshot models, and core-follow models. The typical reactivity curve for the CANDU lattice is
presented, and a method to estimate the average attainable fuel burnup from the reactivity
curve is explained. The CANDU initial core and the first few months of operation are discussed.
Finally, the chapter gives a short qualitative survey of advanced fuel cycles which could be
exploited in CANDU reactors.
Table of Contents
1 Introduction ............................................................................................................................ 3
1.1 Overview ......................................................................................................................... 3
2 Definition of Concepts ............................................................................................................ 4
2.1 Fuel Irradiation................................................................................................................ 4
2.2 Fuel Burnup..................................................................................................................... 5
2.3 Relationship Between Fuel Irradiation and Fuel Burnup................................................ 5
3 CANDU Reactor Physics Computational Scheme.................................................................... 6
3.1 Lattice Calculation........................................................................................................... 6
3.2 Reactivity-Device Calculation.......................................................................................... 9
3.3 Full-Core Calculation ....................................................................................................... 9
4 Averaging Nuclear Cross Sections......................................................................................... 11
4.1 Averaging Over Space ................................................................................................... 12
4.2 Averaging Over Time..................................................................................................... 13
5 The Fuel Reactivity Curve...................................................................................................... 13
6 Exit-Burnup Estimation from Lattice Calculations ................................................................ 15
7 Flattening the Power Distribution......................................................................................... 16
8 Fuel-Management Calculations for Core Design .................................................................. 17
8.2 The CANDU Time-Average (Equilibrium-Core) Model .................................................. 19
9 More Design Calculations: Snapshots Based on the Time-Average Model .......................... 28
9.1 Channel-Irradiation-and-Power Cycle........................................................................... 28
©UNENE, all rights reserved. For educational use only, no assumed liability. CANDU In-Core Fuel-Management – December 2016
9.2 Channel Ages................................................................................................................. 29

9.3 Snapshot with Patterned Ages...................................................................................... 29
10 Design of the CANDU Initial Core.......................................................................................... 31
11 Fuel-Management Calculations for Reactor Operation........................................................ 32
11.1 From Initial Core to Onset of Refuelling ....................................................................... 32
11.2 After Onset of Refuelling............................................................................................... 32
11.3 Channel-Power Peaking Factor ..................................................................................... 33
11.4 Criteria for Selecting Channels for Refuelling ............................................................... 34
12 Advanced Fuel Cycles............................................................................................................ 35
13 Summary ............................................................................................................................... 36
14 References............................................................................................................................. 37
List of Figures
Figure 1 Relationship of Fuel Burnup and Irradiation from POWDERPUFS-V Code ....................... 6
Figure 2 Face View of the CANDU Basic Lattice Cell ....................................................................... 7
Figure 3 Supercell for Calculation of Device Incremental Cross Sections....................................... 9
Figure 4 Face View of Full Reactor Model..................................................................................... 10
Figure 5 Top View of Reactor Physics Model ................................................................................ 11
Figure 6 Averaging the properties of a 2-region core ................................................................... 12
Figure 7 The Reactivity of the CANDU Lattice............................................................................... 14
Figure 8 Average Exit Irradiation Attainable with Daily Refuelling ............................................... 16
Figure 9 Example of Possible Time-Average Model, with Eight Irradiation Regions .................... 24
Figure 10 Flow Chart of Time-Average Calculation....................................................................... 25
Figure 11 Dwell Times (in FPD) from a Time-Average Calculation for the CANDU 6.................... 27
Figure 12 Channel-Power Cycle During On-Going Reactor Operation ......................................... 29
Figure 13 Possible Pattern for Order of Refuelling in a 6 x6 Region of Core ................................ 30
Figure 14 Possible Core Refuelling Sequence for the CANDU 6 ................................................... 31
List of Tables
Table 1 Sample Lattice Nuclear Cross Sections versus Fuel Irradiation .......................................... 8
Table 2 Example of Summary Results from a CANDU-6 Time-Average Calculation ..................... 27
CANDU In-Core Fuel-Management 3
1 Introduction
1.1 Overview
CANDU Refuelling is carried out with the reactor at power. This feature makes the in-core fuel
management substantially different from that in reactors that must be refuelled during shut-
downs. The capability for on-power refuelling means that excess reactivity requirements are at
a minimum: only a few milli-k are necessary for continuous and short-term reactivity control.
This leads to excellent neutron economy and low fuelling costs.
The primary objective of CANDU in-core fuel management is to determine fuel-loading and fuel-
replacement strategies to operate the reactor in a safe and reliable fashion while keeping the
total unit energy cost low. Within this context, the specific objectives of CANDU in-core fuel
management are as follows:
 Adjust the refuelling rate to maintain the reactor critical
 Control the core power to satisfy safety and operational limits on fuel and channel
power, thus ensuring that the reactor can be operated at full rating
 Maximize burnup within operational constraints, to minimize fuelling cost
 Avoid fuel defects, to minimize replacement fuel costs and radiological occupational
hazards.
To refuel a channel, a pair of fuelling machines latches onto the ends of the channel. A separate
chapter in this book describes the fuelling machine and the related fuel handling [Damario
2016]. An even number of fresh fuel bundles are inserted into the channel by the machine at
one end, and an equal number of irradiated fuel bundles are discharged into the machine at the
other end of the channel. For symmetry, the refuelling direction is opposite for next-neighbour
channels. In the CANDU-6 reactor, the refuelling direction is the same as that of coolant flow in
the channel.
Several refuelling operations are normally carried out daily, so that refuelling is almost continu-
ous. CANDU reactors offer extreme flexibility in refuelling schemes:
 The refuelling rate (or frequency) can be different in different regions of the core,
and in the limit can in principle vary from channel to channel. By using different re-
fuelling rates in different regions, the long-term radial power distribution can be
shaped and controlled.
 The axial refuelling scheme is not fixed;
 A channel can be refuelled without delay if failed fuel exists or is suspected;
 The fuel can be removed completely from a few channels, in order for example to
limit the channel axial growth.
This chapter will cover the topics relevant to the management of nuclear fuel in CANDU reac-
tors. Fuel management in CANDU has both design and operations aspects; both are covered in
this chapter. The concepts important to in-core fuel management will be explained: fuel irradia-
tion (fluence), fuel burnup, cross-section averaging. The various levels of physics models which
are used to carry out fuel-management calculations will be presented: these are the continu-
ous-refuelling model, the time-average model, patterned-age snapshot models, and core-follow
models. The purpose and use of the various calculations will be explained. Additional informa-
tion can be obtained from [Rouben 2003].

The goal of this chapter is to help the student understand:
 The basic concepts of fuel irradiation (fluence), fuel burnup, and fuel isotopic
changes
 The reactivity curve of the fuel lattice and its importance
 The refuelling process in CANDU
 The significance of the flux and power shapes in reactor
 How to control the flux shape, the means to flatten the flux distribution (adjuster
rods, differential fuelling)
 Time-average, snapshot, and core-follow models for CANDU reactors, and their use.
2 Definition of Concepts
Fuel irradiation (also known as fluence) and fuel burnup are separate but related concepts.
Both are in a way a measure of the “age” of the fuel in the reactor.
2.1 Fuel Irradiation

Let us use the symbol  F  t  for the neutron flux in the fuel in a certain fuel bundle, where t is
the time measured from the moment when the bundle entered the reactor core. The irradia-
tion (or fluence) of the fuel in the bundle, denoted   t  , is defined as the time integral of the
flux to time t:
  t     F  t  dt
t
(1)
0
This definition is not quite complete, since the flux is a function of neutron energy E. We must
remove the vagueness and make the definition complete by specifying which flux is to be used.
The usual choice is to select  F as the thermal flux in the fuel. Excluding for now the time
variable, we can write  F as the integral of the flux as a function of energy:
  F  E  dE
Eth
F  (2)
0
where Eth is the upper bound of the thermal energy interval. Note: Since the thermal energy
interval may be defined differently in different computer codes, the fuel irradiation may vary
from code to code, and caution must therefore be exercised when comparing irradiation values
using different codes.
Having settled on the definition of fuel flux, let us return now to Eq. (1) and note that fuel
irradiation (fluence) is a monotonically increasing function of time, since flux is never negative.
The irradiation starts at 0 when the fuel bundle enters the core, and increases up to the time
when the fuel bundle exits the core. The value of the irradiation when the bundle exits from
the core is then called the exit or discharge irradiation.
The units of irradiation are units of flux multiplied by units of time. This gives for example n.cm-
2 -1
.s *s, i.e., n.cm-2. But a much more appropriate (microscopic) unit for area is the barn (b = 10-
24
cm2) or kilobarn (kb = 10-21 cm2). Thus, the common unit for irradiation is n/kb.
2.2 Fuel Burnup

The fuel burnup of a given fuel bundle, denoted B(t), is defined as the total fission energy
(note: not thermal or electric energy) released in the fuel bundle since the moment when it
entered the reactor core, divided by the initial mass of heavy element in the bundle. (In the
case of the CANDU reactor, where the only initial heavy element in the fuel is uranium, the
initial mass of heavy element is the mass of uranium in the bundle.)
Since the energy release depends on the fission rate in the bundle, we can write
E R  f  t   f  t  dt
t
B t  
 0
(3)
M
where M is the mass of uranium in the bundle, f is the fission cross section in the bundle, and
ER is the fission energy release per fission (~200 MeV).
In contrast to that of fuel irradiation, the definition of fuel burnup does not suffer from vague-
ness, since energy is a measurable, well-defined quantity. For this reason, if comparisons are to
be made between different computer codes, it is much easier, and preferred, to compare on the
basis of fuel burnup rather than of irradiation.
Similarly to fuel irradiation, fuel burnup is a monotonically increasing quantity, since the fuel
continues to fission and release energy with time. The burnup of the fuel in the fuel bundle
starts at 0 when the fuel bundle enters the core, and increases up to the time when the fuel
bundle exits the core. The value of the fuel burnup when the bundle exits from the core is then
called the exit or discharge burnup.
The units of fuel burnup are units of energy divided by units of mass. The most commonly used
burnup units are megawatt-days per megagram of uranium (MW.d/Mg(U)) or megawatt-hours
per kilogram of uranium (MW.h/kg(U)). The typical average exit discharge burnup in existing
CANDU reactors varies depending on the design of the core, and may range between 180 and
225 MW.h/kg(U) (or, equivalently, between 7,500 and 9,375 MW.d/MgU).
2.3 Relationship Between Fuel Irradiation and Fuel Burnup

Both fuel irradiation and fuel burnup increase with time and reflect therefore in a way the “age”
of the fuel. Of course, fuel “ages” faster in a high neutron flux, since the flux appears in the
integrand of both Eqs. (1) and (2).
In a given nuclear lattice, fuel burnup and fuel irradiation have a one-to-one relationship which
depends on the fission cross section f and its variation with time. In the CANDU lattice, there
is a relatively small variation in f (and in ER) over the typical residence time of a fuel bundle in
core, so the relationship between irradiation and burnup is almost linear. For example, the old
lattice code POWDERPUFS-V [Rouben 1995] gives the following relationship for fuel burnup as a
function of irradiation:
14000.0
12000.0
Fuel Burnup (MW.d/Mg(U)) 10000.0
8000.0
6000.0
4000.0
2000.0
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Fuel Irradiation (n/kb)
Figure 1 Relationship of Fuel Burnup and Irradiation from POWDERPUFS-V Code
3 CANDU Reactor Physics Computational Scheme

To properly manage the nuclear fuel in the core, one must be able to calculate the neutron flux
and power distributions in space and in time. The most accurate computational tool for CANDU
neutronics is the Neutron Transport equation (see Section 2.1 of Chapter 4, Reactor Statics).
However, solving this equation over the full core on a routine basis demands too much compu-
tational effort. Therefore, a 3-stage process involving the Neutron Transport equation and the
simpler Neutron Diffusion equation, which involves some approximation (see Section 2.2. of
Chapter 4), has been developed for CANDU neutronics. By experience, this 3-stage computa-
tional system has been found to give generally very good results. It has been described in
Chapter 4 and is reviewed here.
3.1 Lattice Calculation

The first stage involves solving for the flux distribution in the fundamental unit of the CANDU
reactor, i.e., the basic lattice cell, which is defined to consist of the fuel, coolant, pressure and
calandria tubes, and moderator, but no reactivity devices (see face view in Figure 2; in the third
dimension the basic lattice cell is one bundle-length deep).
The lattice calculation is performed using a lattice code. For many years, the lattice calculation
was performed with the code POWDERPUFS-V, which was developed specifically for CANDU
reactors and which makes use of computational “recipes” to incorporate experimental results
obtained with heavy-water lattices. More recently, deterministic multi-group neutron transport
codes with a strong mathematical and theoretical basis, such as WIMS-IST [Irish 2002] or
DRAGON [Marleau 1999], have been adopted in place of POWDERPUFS-V for lattice calcula-
tions. The DRAGON code is currently freely available from École Polytechnique de Montréal.
Figure 2 Face View of the CANDU Basic Lattice Cell

The deterministic transport code uses the geometry and properties of materials in the lattice
cell to do a very accurate calculation of the flux distribution over the lattice cell. It uses a large
or very large number (tens or even hundreds) of neutron energy groups in this calculation, to
take account of the neutron moderation accurately. It then averages the nuclear properties (for
absorption, moderation, fission, etc.) over the cell according to the calculated reaction rates in
each sub-region of the lattice cell to determine the effective neutronic properties of the whole
cell. This averaging (homogenization) of the properties of the basic lattice cell dilutes the strong
absorption of the fuel with the much weaker absorption in the moderator. The code also
“collapses” the cell properties over a very small number of neutron energy groups. Most often
in CANDU neutronics, the collapsing is done over 2 energy groups (a thermal group and a
slowing-down group). This is sufficient since only about 2% of all fissions are induced by non-
thermal neutrons. With these homogenized properties, Fick’s Law (see Section 2.2 of Chapter
4) becomes a good approximation over most of the reactor, and the diffusion equation can be
used to calculate the neutron flux over the full core.
The lattice code must do one more thing, a “depletion” calculation: Here the code updates the
composition of the fuel over a time step, using the various reaction rates that it has determined,
and then recalculate the homogenized cell properties. Repeating this over many time steps, the
code evolve the lattice-cell neutronic properties to reflect the changes in the nuclide composi-
tion of the fuel with time (irradiation/burnup).
The output of the lattice code is then a table of the basic-cell (lattice) nuclear properties as
functions of fuel irradiation or burnup. For example, POWDERPUFS-V gives the properties in
the 2-energy-group formalism, but in the approximation that the fast-fission and upscattering
cross sections are assumed to be 0. A typical fuel table would then appear as in Table 1. Note
that the lattice code must also in addition provide the nuclear properties (cross sections) of the
moderator, to be applied outside the fuelled region of the core.

Table 1 Sample Lattice Nuclear Cross Sections versus Fuel Irradiation
Fast Thermal
Yield
Fast Thermal Absorption Absorption Ratio of
Cross Downscattering
Irradiation Diffusion Diffusion Cross Cross Fuel Flux
Section Cross Section
(n/kb) Coefficient Coefficient Section Section to Cell
2 2 (cm^-1)
D1 (cm) D2 (cm) a1 (cm^- a2 (cm^- Flux
(cm^-1)
1) 1)
0.00 1.269 0.936 7.68E-04 3.79E-03 4.52E-03 7.40E-03 0.383
0.10 1.269 0.936 7.68E-04 3.82E-03 4.56E-03 7.40E-03 0.380
0.20 1.269 0.936 7.68E-04 3.89E-03 4.66E-03 7.40E-03 0.374
0.30 1.269 0.937 7.67E-04 3.95E-03 4.73E-03 7.40E-03 0.369
0.40 1.269 0.937 7.67E-04 3.99E-03 4.78E-03 7.40E-03 0.365
0.50 1.269 0.937 7.67E-04 4.03E-03 4.81E-03 7.40E-03 0.361
0.60 1.269 0.937 7.66E-04 4.07E-03 4.83E-03 7.40E-03 0.359
0.70 1.269 0.937 7.66E-04 4.10E-03 4.83E-03 7.40E-03 0.356
0.80 1.269 0.937 7.66E-04 4.12E-03 4.83E-03 7.40E-03 0.354
0.90 1.269 0.937 7.66E-04 4.14E-03 4.82E-03 7.40E-03 0.352
1.00 1.269 0.937 7.65E-04 4.15E-03 4.81E-03 7.40E-03 0.351
1.10 1.269 0.937 7.65E-04 4.17E-03 4.78E-03 7.40E-03 0.350
1.20 1.269 0.937 7.65E-04 4.18E-03 4.76E-03 7.40E-03 0.349
1.30 1.269 0.937 7.65E-04 4.19E-03 4.74E-03 7.41E-03 0.348
1.40 1.269 0.936 7.64E-04 4.20E-03 4.71E-03 7.41E-03 0.348
1.50 1.269 0.936 7.64E-04 4.20E-03 4.68E-03 7.41E-03 0.347
1.60 1.269 0.936 7.64E-04 4.21E-03 4.65E-03 7.41E-03 0.347
1.70 1.269 0.936 7.63E-04 4.22E-03 4.62E-03 7.41E-03 0.346
1.80 1.269 0.936 7.63E-04 4.22E-03 4.59E-03 7.41E-03 0.346
1.90 1.269 0.936 7.63E-04 4.22E-03 4.56E-03 7.41E-03 0.346
2.00 1.269 0.936 7.63E-04 4.23E-03 4.53E-03 7.41E-03 0.345
2.10 1.269 0.936 7.62E-04 4.23E-03 4.50E-03 7.41E-03 0.345
2.20 1.269 0.936 7.62E-04 4.23E-03 4.47E-03 7.41E-03 0.345
2.30 1.269 0.936 7.62E-04 4.24E-03 4.45E-03 7.41E-03 0.345
2.40 1.269 0.936 7.62E-04 4.24E-03 4.42E-03 7.41E-03 0.344
2.50 1.269 0.936 7.61E-04 4.24E-03 4.39E-03 7.41E-03 0.344
2.60 1.269 0.936 7.61E-04 4.25E-03 4.37E-03 7.41E-03 0.344
2.70 1.269 0.936 7.61E-04 4.25E-03 4.35E-03 7.41E-03 0.344
2.80 1.269 0.936 7.60E-04 4.25E-03 4.33E-03 7.41E-03 0.343
2.90 1.269 0.936 7.60E-04 4.26E-03 4.30E-03 7.41E-03 0.343
3.00 1.269 0.936 7.60E-04 4.26E-03 4.28E-03 7.41E-03 0.343
3.10 1.269 0.936 7.60E-04 4.26E-03 4.27E-03 7.41E-03 0.343
3.20 1.269 0.936 7.59E-04 4.26E-03 4.25E-03 7.41E-03 0.342
3.30 1.269 0.936 7.59E-04 4.27E-03 4.23E-03 7.41E-03 0.342
3.40 1.269 0.936 7.59E-04 4.27E-03 4.21E-03 7.41E-03 0.342
3.50 1.269 0.936 7.59E-04 4.27E-03 4.20E-03 7.41E-03 0.342
3.60 1.269 0.936 7.58E-04 4.28E-03 4.18E-03 7.41E-03 0.341
3.70 1.269 0.936 7.58E-04 4.28E-03 4.17E-03 7.41E-03 0.341
3.80 1.269 0.936 7.58E-04 4.29E-03 4.16E-03 7.41E-03 0.341
3.90 1.269 0.936 7.58E-04 4.29E-03 4.15E-03 7.41E-03 0.340
4.00 1.269 0.936 7.57E-04 4.29E-03 4.13E-03 7.41E-03 0.340
3.2 Reactivity-Device Calculation

The CANDU basic lattice cell contains, by definition, no reactivity devices. Neither does it
contain any structural material other than the fuel channel itself (i.e., the pressure tube and
calandria tube). However, there are reactivity devices in appropriate locations in the core.
Therefore, the effect of the reactivity devices, their guide tubes and other structural material on
the nuclear properties in their vicinity needs to be taken into account. This effect is expressed
in the form of homogenized “incremental” cross sections which are added to the (unperturbed)
homogenized properties of neighbouring lattice cells in a modelled volume around the device,
called a “supercell”; see an example in Figure 3, although the supercell is often defined as
having half the width shown, i.e., with only half the width on either side of the device. The
incremental cross sections are calculated with a transport code, most often DRAGON; they are
identified with the differences in the supercell homogenized properties between the case
where the device is inserted into the supercell and the case where it is withdrawn from it. The
incremental cross sections are added to the lattice-cell cross sections in the modelled volume
around the device’s position in the reactor core, when that particular device is inserted.
Figure 3 Supercell for Calculation of Device Incremental Cross Sections
3.3 Full-Core Calculation

In the final stage, a full-core reactor model is assembled. As explained above, because ho-
mogenized cell and device properties are used, few-group (most often 2-group) diffusion theory
can be applied with this model. The full-core model must incorporate, for each cell in the
reactor, the homogenized unperturbed-lattice-cell cross sections which apply there, together
with the incremental cross sections of nearby reactivity devices in their appropriate locations.
The model therefore consists of homogenized basic-lattice-cell properties in the appropriate
locations, on which are superimposed reactivity-device properties in the devices’ modelled
volumes.
A typical full-core 3-d CANDU reactor model for use with finite-difference diffusion theory is
shown in Figure 4 (face view, showing adjuster modelled volumes) and Figure 5 (top view,
showing adjuster and liquid-zone-controller modelled volumes).
The full-core flux and power distributions are calculated in CANDU with the finite-difference
diffusion-theory code Reactor Fuelling Simulation Program RFSP-IST [Rouben 2002], or for
example with DONJON [Hébert 2012]. The DONJON code is currently freely available from École
Polytechnique de Montréal.
Figure 4 Face View of a Typical CANDU-6 Full Reactor Model
Figure 5 Top View of a Typical CANDU-6 Reactor Physics Model
4 Averaging Nuclear Cross Sections

There are many situations in which one wishes to average nuclear cross sections in space or in
time. For example, we have already seen in Section 2.1 that the lattice code averages (ho-
mogenizes) the nuclear properties within the basic lattice cell. In the finite reactor also, we may
wish to average the properties of a region of the core. How to do this is explained here.
The overarching aim in reactor physics is to calculate reaction rates, because from a knowledge
of reaction rates we can derive the values of the important quantities in the system, for example
the core reactivity or the power distribution in the core, etc.
A reaction rate RR at a point, corresponding to a process with cross section , is calculated as

the product of the cross section with the neutron flux:
RR   (4)
These point reaction rates can then be integrated over space or time. Since the total reaction
rate is the quantity of interest, then we must ensure that this is conserved in any averaging.
4.1 Averaging Over Space

Suppose we consider a non-uniform reactor subdivided into a number of regions with different
values of the cross section, and we would like to know the “average” cross section for the whole
reactor. Let’s take a very simple example of a reactor with 2 regions, with uniform cross section
value 1 and 2 respectively (see Figure 6).
Figure 6 Averaging the properties of a 2-region core
The flux is  1 and  2 in the two regions, and is a function of space in each region. How would
we calculate the overall average cross section  ?
We want to conserve the total reaction rate, which is
RRtotal   11dV    2  2 dV (5)
V1 V2
Therefore  should satisfy
V1
1dV    2 dV  RRtotal
V2
which gives


V1
11dV   2  2 dV
V2
(6)

V1
1dV    2 dV
V2
This shows that the proper definition of the average cross section is obtained by averaging the
cross section with flux and volume weighting.
If we define also the average flux

V1
1dV    2 dV
V2

V1
1dV    2 dV
V2
(7)

V1
dV   dV
V2
V
where V  V1  V2 is the total volume, then we can also write

 V1
11dV    2  2 dV
V2
(8)
V
or
 V   11dV    2  2 dV RRtotal (9)
V1 V2
4.2 Averaging Over Time

Now let’s think about averaging over time at a given point or region (say, for a fuel bundle).
Here the properties will vary with time, i.e., with fuel irradiation or burnup. The reaction rate
which we must conserve is the time-integrated reaction rate    t    t  dt .
The average cross section over a time interval t1 to t2 is then defined by
  t    t  dt
t2

 t1
(10)
   t  dt
t2
t1
Now let’s change variables to , using d  dt , and we get

2 2
    d     d
1 1
 2
 (11)
2  1

1
d
i.e.,  is basically the average of the cross section over irradiation on the irradiation interval.
5 The Fuel Reactivity Curve

The reactivity of the fuel lattice versus irradiation (or burnup) is provided by the lattice code,
together with the lattice properties. For any lattice design, the fuel reactivity curve is a funda-
mental piece of knowledge for fuel management. A typical reactivity curve for the CANDU
lattice is shown below (Figure 7):
Figure 7 The Reactivity of the CANDU Lattice

The reactivity curve has several interesting features:
 The reactivity at zero irradiation (fresh fuel) is quite high, ~76 mk. It is to be noted that
this value is calculated assuming that the saturating fission products (xenon, samarium,
etc.) are already at equilibrium, even though the fuel is fresh. To run the infinite lattice
in steady state, the positive reactivity would have to be cancelled by adding poison to
the moderator, for example ~9.5 ppm of boron in the initial core (assuming a boron
reactivity coefficient of -8 mk/ppm).
 The reactivity reduces slowly at first, then increases to a maximum, called the pluto-
nium peak. This behavior is due to the combined effects of the start of U-235 depletion
and the start of Pu-239 production. The latter is a slightly better fuel than U-235, and
consequently increases reactivity as it is produced, following a short time lag due to the
~2.5-day half-life of Np-239, the immediate parent of Pu-239.
 Following the plutonium peak, the reactivity decreases in a monotonic fashion indefi-
nitely. This is due to the continued net depletion of fissile nuclides and accumulation of
fission products. The plutonium peak, however, does not mark the maximum in pluto-
nium concentration. In fact the latter keeps increasing, however it does so more and
more slowly, since Pu-239 participates in fission at the same time as it is produced.
 Hypothetically, if the infinite lattice could be run in steady state as the fuel gained irrad-
iation, the boron concentration would have to be adjusted continuously to reflect the
positive reactivity.
 In Figure 6, the reactivity reaches a value of zero at an irradiation of ~1.6 n/kb. This
would mark the exit irradiation of the infinite lattice. At this point the boron concentra-
tion required would of course be 0.
6 Exit-Burnup Estimation from Lattice Calculations

The fuel-reactivity curve above gave us the discharge irradiation attainable for the infinite
lattice. Even before doing fuel-management calculations for the full core, we can devise a way
to make a fairly good estimate of the average attainable discharge irradiation for the real
CANDU reactor, if we have an estimate of the leakage and in-reactor-device worth.
For the CANDU-6 reactor, good estimates of the leakage and of the device worth (adjusters and
zone-controller water) are 30 mk and 18 mk respectively. If we subtract a total of 48 mk from
the fuel-reactivity curve, we get a curve (Figure 8) which the full reactor would follow if all parts
of it advanced in irradiation at the same rate. While this is obviously not exactly the case, we
will still get a fairly good approximation of the reactivity curve for the full reactor as per Figure
8, which shows that the CANDU 6 core would initially reach 0 ppm of boron at an irradiation of
about 1.0 n/kb (a burnup of ~4,000 MW.d/Mg(U)). This would be the exit irradiation if the
CANDU were refuelled in batch mode. [Note that in practical applications, the excess reactivity
of the fresh core is reduced with the use of depleted fuel, and the refuelling is started gradually
before the excess reactivity of the core reaches decreases to 0.]
However, CANDU can be - and is - refuelled on line. This means that when the moderator
poison runs out, daily refuelling can be started, replacing just enough fuel each day to cancel
the loss of reactivity due to the increased overall irradiation.
By repeating this day after day, we eventually reach a state where there is approximately an
equal amount of fuel at all irradiations, from 0 to a maximum value which determines the
attainable discharge irradiation with daily refuelling. This value is obtained by doing a “rolling
average” of the reactivity starting from an irradiation of 0, up to the point where the rolling
average is 0. It is clear that the 0 rolling average is obtained when the positive area to the left of
the point where the reactivity is 0 is exactly cancelled by the negative area beyond that point.
By doing the rolling average numerically, we find that the average attainable exit irradiation is
~1.8 n/kb (burnup ~7,500 MW.d/Mg(U)) – see Figure 8. Thus daily refuelling has almost dou-
bled the average exit burnup, compared to batch refuelling!
Figure 8 Average Exit Irradiation Attainable with Daily Refuelling

Note of course that all the above analysis, in the previous and current sections, depend on the
specific values assumed for various lattice parameters, such as heavy-water purity, fuel and
moderator temperatures, etc. A higher value of moderator purity, most importantly, can raise
the reactivity curve and consequently increase the attainable exit burnup: a 0.1% increase in
moderator purity is estimated to increase the exit burnup by ~360 MW.d/Mg(U). Values around
99.9% for the moderator isotopic purity are common.
7 Flattening the Power Distribution

The reactivity curves in the previous sections provide information on the “local” reactivity in
bundles according to their irradiation level. And we have up to now used these figures in the
approximation of uniform irradiation throughout the core.
However, of course, fuel in different locations in core actually picks up irradiation at different
rates, depending on the local flux. Chapter 4 of this book has shown that the flux distribution in
a uniform reactor has a J0 (Bessel function) shape and a cosine shape in the radial and axial
directions respectively. CANDU reactors are not homogeneous, and so the flux distribution will
be different. In addition, we may deliberately want to change the power distribution.
One reason we may want to change the power distribution is that the J0 and cosine shapes are
“peaked” flux shapes. This means that the power decreases relatively quickly from the maxi-
mum value. Now, a peaked flux shape may create a conflict between the reactor power and
peak local powers (channel or bundle power). The plant’s operating license typically puts limits
not only on the total reactor power, but also on the maximum channel and bundle powers.
With these restrictions, it may not be possible with a peaked flux shape to generate the reactor
power desired without exceeding the licensed maximum channel and bundle powers. For this
reason, most power reactors in fact operate with a flattened power distribution, in which the
power is redistributed so that the peak is reduced and more channels (radially) and bundles
(axially) generate more power. In this way, the desired total power can be generated while
complying with the peak local limits.

In CANDU reactors, power flattening is done in two ways:
1) By placing some absorbing rods in the central area of the core. In CANDU reactors,
these are the adjuster rods.
2) By differential fuelling. This means deliberately running the fuel in the central region of
the core to a higher discharge irradiation than fuel in the peripheral region. The higher
irradiation reduces the local reactivity and therefore the local flux.
Differential fuelling of a core region (it could even be a single channel) in CANDU can also be
thought of as being produced by varying the region’s refuelling frequency. This can be seen
from the following relationship between exit burnup and refuelling rate.
Let P be a region’s fission power, B be the average exit burnup in the region, and rm be the
refuelling rate in the region (in mass of uranium per unit time). It is easy then to derive the
relationship
P P
B , also written as rm  (12)
rm B
If for example P is given in MW and rm in MW(U)/d, then the units of B are MW.d/Mg(U). If a
CANDU bundle contains a mass M of uranium, then we can write the above relationship in
terms of the bundle refuelling rate, rb:
P P
B or rb  (13)
Mrb BM
The most convenient units for rb are bundles/d, obtained for example from P in MW, B in
MW.d/Mg(U), and M in Mg(U)/bundle.
From this analysis we can see that by controlling the refuelling rate one can control the average
exit burnup. In fact, in this way we can control the reactivity of the fuel in each region. If we
run the fuel in a certain region to a higher burnup, we decrease the region’s reactivity, and
consequently lower the region’s power, as we will see in greater detail further below.
8 Fuel-Management Calculations for Core Design

CANDU reactors are refuelled on-line. Refuelling operations are typically performed every day,
or a few times per week. A very small fraction of the fuel in core (typically ~8 fuel bundles,
representing something of the order of 0.1-0.2%) is replaced at each refuelling operation. Also,
because refuelling is performed on a quasi-continuous basis, the excess reactivity of the core is
always very low (except at the very start of the core life, or after a long shutdown).
On account of these features, the overall global power distribution in a CANDU core operated in
a well-thought-out manner remains largely constant in time, with relatively small variations
(“power ripples”) superimposed locally as individual bundles and channels go through their
normal burnup cycles from one of their refuellings to the next. Thus, it is meaningful and
useful to want to calculate this long-term average.

In this Section we start to look at fuel-management models which are used at the core-design
stage. They represent the reactor in the average over a long period of time, not on any particu-
lar day. Such models can then be used to perform diffusion calculations (for example, with full-
core codes such as RFSP-IST) to obtain the expected spatial distribution of flux and power in the
long term; these distributions will eventually be used as basis for running the reactor over time.
8.1 The Axially-Uniform Model Based on Continuous Refuelling
The first model which we will develop is the Axially-Uniform Model of nuclear cross sections.
We start by visualizing the reactor as being continuously refuelled, meaning that fuel is pushed
continuously and simultaneously through every fuel channel at a certain speed . If channels
L
have length L, then the time T it takes fuel to traverse the entire channel length is T  .

Let   x  be the axial flux distribution in a fuel channel, from x  0 to x  L (channel inlet to
outlet). Suppose as a first approximation we would like to use the average value  of a basic-
lattice cross section (x) over the channel [this can be the fission cross section, or absorption
cross section, or any other). We can get this average value by applying Eq. (5) to our specific
case:
   x    x  dx
L
 0
(14)
   x  dx
L
Now change variables from x to t by using x   t , dx   dt , which gives
   t    t  dt
T
 0
(15)
   t  dt
T
Now we make a second change of variable, from t to fuel irradiation  :
  t      t  dt , d    dt (16)
And if we define exit as the value of the fuel irradiation at the exit of the fuel channel (but note
that we are not limited to using the same value of exit irradiation in all channels), we get

exit exit
    d     d 
 0
exit
 0
(17)
exit
 0
d
Thus we would use the average values  calculated as per Eq. (16) (for the various cross-section
types) at all bundle positions within each individual fuel channel. Note that actually Eq. (16) is
just another manifestation of Eq. 10). [Once again, note that we can select different values of
exit in different channels. Another way to visualize this is to realize that we have the option of
selecting different speeds  of the fuel in different channels.]
Once all cross sections are defined for each position in core, and the incremental cross sections
of in-core devices are added, then we have a full axially-uniform reactor model for use with the
diffusion equation. The solution of that equation will be the core reactivity and the 3-
dimensional flux distribution in core, from which all powers, etc., can then be calculated.
Of course this is a simplified model of the properties in the reactor, since in each channel the
properties are modelled as uniform and we are losing the axial variation of lattice properties
over the channel. This is actually not as bad an approximation as may be thought, because it
can be shown that the bi-directional refuelling in CANDU reactors leads to approximately
constant lattice properties when these are averaged over 2 neighbouring channels (with the

same exit ). Thus, the axially-uniform model was used extensively in the early stages of CANDU
design development and analysis. However, eventually a model which better models the
variation of lattice properties in each individual channel was developed. This is the time-
average model, described in the next subsection.
8.2 The CANDU Time-Average (Equilibrium-Core) Model

The time-average model [Rouben 2007] is the next, more realistic effort to take into account the
spatial distribution of properties over time, arising from different fuel irradiations in different
fuel bundles. As opposed to the axially-uniform model, the time-average model does not
average lattice properties along a channel. Instead, it takes into account the fact that individual
fuel channels are refuelled once in a while, and that fuel bundles remain in a certain location in
the channel until the channel is refuelled. It models the effect of the axial refuelling scheme
(e.g., 8-bundle-shift (8-bs), 4-bundle-shift (4-bs), etc.) used in each channel on the average
lattice properties at each bundle location in the channel.
8.2.1 Degrees of Freedom in CANDU Operation

Let us look at the breadth of freedom in operating a CANDU reactor. The following are “degrees
of freedom” available to the operator:
 The fuel type used (e.g., the fuel enrichment or the fuel-bundle geometry, that is 28-
element, 37-element, etc.)
 Which channel to refuel on any given day
 Which channel to refuel after the previous one refuelled, and which channel to re-
fuel after that one (i.e., ultimately, the sequence of refuellings)
 The “axial” refuelling scheme to be used (i.e., the number of bundles to change at
each channel refuelling), e.g., 8-bundle shift, 4-bundle shift, 2-bundle shift, or other,
more complicated scheme. In addition, this axial refuelling scheme may vary from
core region to core region, or even in principle from channel to channel. [Note that
refuelling schemes where bundles are reshuffled from one channel to another are
not considered in the current discussion.]
 The refuelling frequency of each channel – the average frequency will depend on the
fuel type(s), on the refuelling scheme, and on the overall core-reactivity decay rate.
 The desired discharge (exit) burnup for the fuel, as an overall average and even by
core region
 The power desired for each core zone or region – yes, this is something that can be
set at design time (of course the sum of the region powers must be equal to the total
core power).
These degrees of freedom are not all independent of one another: refuelling frequencies,
discharge burnup, and region powers are all interrelated. Nonetheless, it is clear that there is a
very broad range of possibilities available to the operator, even if a single fuel type is used.
Therefore, rational and orderly operation of the reactor really requires following guidelines in
terms of the average refuelling frequency of each channel.
8.2.2 Definitions
Let us start with some notation and some definitions:
Let c be a label for channel number, c = 1 to M, where M is the total number of channels in core
 Let b be a label for a bundle-position number within a channel, b = 1 to Nc, where Nc
is the total number of bundles per channel
 Then we can use cb as a label for a specific individual bundle position in core
 Let x be a label for a type of nuclear cross section, e.g., absorption, fission, etc.
 Let cb(t) and cb(t) be respectively the fuel flux and the cell flux at position cb at
some time t. These fluxes are actually multigroup quantities (2-group in the current
RFSP-IST). cb(t) and cb(t) are related by the so-called “F-factor”, ratio of flux in fuel
to average flux in cell; this is calculated by the lattice code.
 Let tav,cb and tav,cb be the corresponding fuel and cell fluxes in the time-average
model. Because the time-average model does not have time as an independent pa-
rameter, these fluxes are not functions of time.
 Let Dc be the average time interval between refuellings of channel c; this is also
known as the dwell time for channel c. [Note: This is not the same as the average
residence time of a bundle in core, since in most refuelling schemes some bundles
remain in core for more than 1 cycle.]
 Let entr,cb be the fuel irradiation at the time of entrance of the fuel at position cb,
and
 Let exit,cb be the fuel irradiation at the time of exit of the fuel from position cb.
 Also, let Ec be the total number of bundles which leave channel c on a refuelling op-
eration; these bundles will be labelled ,  = 1 to Ec. For example, in the 8-bs refuel-
ling scheme, we have Ec = 8, and the bundles  are bundles 5 to 12 from the channel
inlet end.
 Let exit,c be the channel exit irradiation of the fuel from channel c, i.e., it is the av-
erage of the exit burnup over the bundles  = 1 to Ec which leave the core on a refu-
elling.
 Let  x,cb   be the macroscopic cross section for cell cb for reaction type x. This is a
function of the fuel irradiation .
 Then let  x ,cb ,tav denote the time-average cell cross section for cell cb and reaction
type x; this will be calculated in the time-average model.
8.2.3 The Selectable Quantities

We shall take the channel exit-irradiation values exit,c values as the essential degrees of free-
dom, i.e., as the ones most easily selected by the designer. These exit,c values will be depend-
ent on the dwell times Dc, and vice versa. Both of course will depend also on the fuel-flux
distribution tav,cb, which depends on the core flux distribution tav,cb, which will be determined
as the solution of the neutron diffusion equation. The relationship between the fuel and cell
fluxes comes from the lattice (or cell) calculation.
Note that in addition the axial refuelling scheme is also a degree of freedom which can be
selected a priori by the designer, for each fuel channel (or region) in core.
8.2.4 Relationships Between Bundle and Channel Irradiations and Dwell Times
For positions cb in core where a refuelling introduces fresh fuel:
entr ,cb  0 (18)
For positions b to which a bundle moves from a position b’ in the same channel upon refuelling,
entr ,cb  exit ,cb’ (19)
Now, since
  Irradiation  Flux  Time 19  (20)
then
Incremental irradiation accumulated by fuel at cb between refuellings  tav ,cb  Dc (21)
and we can use this to relate the exit irradiation to the entrance irradiation:
exit ,cb  entr ,cb  tav ,cb  Dc (22)
We can now relate the average channel exit irradiation exit,c to the fuel flux at the various
bundle positions in the channel, as follows.
Ec
1
By definition,  exit ,c 
Ec



1
exit , c
(23)
and, using Eq. (22)
Ec
   tav ,c   Dc 
1
exit ,c  entr ,c (24)
Ec  1
The bundle positions  are those from which bundles leave the channel on refuelling. Some of
these positions may at the same time be positions into which bundles enter as fresh bundles,
e.g., bundle positions 5 to 8 for an 8-bundle-shift refuelling scheme. For these bundles, the
term entr,c is zero and does not contribute in Eq. (24).

On the other hand, some  may be positions where bundles come from a previous location in
the channel, e.g., bundles 9 to 12 in the channel for the 8-bundle-shifts refuelling scheme. For
these bundles, the term entr,c is not zero, but can be equated (see Eq. (19)) to exit,c’, where ’
is the position from which the bundle at  came on the previous refuelling. In this way, it can be
seen that the sum of terms in tav in Eq. (24) now extends over bundles  and ’.
In fact, by repeating the exercise for bundles ’ and examining from which position they came,
we see that eventually all the terms entr,cb disappear and are replaced by terms in tav, with the
sum in Eq. (24) covering all bundles b in the channel:
Nc
   Dc 
1
exit ,c  tav ,cb (25)
Ec b 1
Since Dc is a channel quantity and is independent of the bundle position b, we can take Dc out of
the sum. We can also turn the equation around to isolate Dc on one side:
Ec  exit ,c
Dc  Nc
(26)

b 1
tav ,cb
Eq. (26) links the channel dwell time to the channel average exit irradiation, the refuelling
scheme (via Ec), and the total fuel flux in the channel.
8.2.5 Relationships Between Time-Average Cross Sections and Irradiation Values

For any bundle position cb in core and any cross section of type x, we can write the time-
average cross section during the bundle’s residence time at that location using Eq. (11) derived
earlier:
i.e.,
exit ,cb
  x ,cb   d
entr ,cb
 x,cb,tav  (27)
exit ,cb  entr ,cb
The integral in Eq. (27) can easily be computed using the cell cross section obtained as a func-
tion of irradiation with the cell code.
Using the values selected for the entrance and exit irradiations at each position, the  x ,cb ,tav can
be computed for all positions in core. These nuclear lattice properties can then be entered into
the diffusion core model, and the neutron-diffusion equation can be solved by the usual means
to find the time-average flux distribution tav,cb.
8.2.6 Summary of Computational Scheme

The computational scheme for the time-average model is now complete. It consists of the
neutron-diffusion equation plus Eqs. (18), (19), (22), (24)-(27).
Remember that:
 The flux distribution tav,cb depends on the time-average cross sections  x ,cb ,tav
through the diffusion equation;
 The time-average cross sections  x ,cb ,tav depend on the entrance and exit irradiation
values via Eq. (27);
 The entrance and exit irradiation values entr,cb and exit,cb depend on the fuel flux in
the channel (which will be calculated from the time-average flux tav,cb) and the
dwell-time Dc via Eqs. (18), (19) and (22);
 The dwell-time Dc depends on the flux and on the channel-average exit irradiation
exit,c via Eq. (26).
It is clear then that this equation set must be solved as a self-consistency problem. An iterative
solution scheme is applied until satisfactory self-consistency is attained.
The basic independent data which needs to be input to the time-average model consists of:
a) the axial refuelling scheme for each channel (for instance, 8-bs, 4-bs, 10-bs, etc.). Note that
this may vary from channel to channel, i.e., one may choose 8-bs for some channels and
4-bs for others. Typically, however, the axial refuelling scheme must be selected taking into
account also non-physics considerations, e.g., fuelling-machine utilization.
b) the channel-average exit irradiations exit,c. These are the main degrees of freedom in the
problem. While there are in principle as many degrees of freedom as there are channels,
the core is usually divided into a small number of relatively large regions, in each of which
exit is taken as uniform, to reduce the number of degrees of freedom. Figure 9 illustrates a
possible time-average model with 8 irradiation regions, for fine control over the target
power distribution; note in particular the regions at the bottom and far right-hand side,
which represent areas in the calandria with much hardware (to hold device guide tubes in
position) and consequently for which different exit irradiations are needed.
Figure 9 Example of Possible Time-Average Model, with Eight Irradiation Regions

A very important point to note is that there are constraints on the degrees of freedom:
 First, reactor criticality is required. That is, the value obtained for the reactor multiplication
constant keff when solving the diffusion equation must be unity (or close to unity, if we know
that the model does not include all material in core, for instance detector material, in which
case a small bias is required) – otherwise, the results obtained cannot be expected to repre-
sent a realistic time-average picture.
 Secondly, the absolute and relative values of the channel-average exit irradiations
exit,c have a significant impact on the radial power distribution, and in turn on the
refuelling frequencies and the attained discharge burnup. Increasing the exit irradia-
tion in a region tends to decrease the local reactivity of the region, leading to a lower
power. Decreasing the exit irradiation has the opposite effect.
 Thus, the channel exit irradiations need to be selected judiciously to achieve both
criticality and the desired degree of radial flattening of the power distribution and of
consequent burnup and refuelling rates.
The flow chart for the iterative calculation for the time-average flux and power distribution is
Figure 10 Flow Chart of Time-Average Calculation

shown in Figure 10. First, the user selects the axial refuelling scheme(s) and the core regions
with the corresponding values of exit,c. Using an initial guess for the flux distribution, the
program determines the channel dwell times Dc and the entrance and exit values of irradiation
entr,cb and exit,cb. It then determines the time-average cross sections  x ,cb ,tav . The program
then proceeds to solve the neutron diffusion equation. The flux distribution thus obtained is
used to repeat the cycle, until consistency is attained in all the parameters.
At that point, the user must intervene (outside the code) and examine the results obtained for
keff and for the power distribution. If keff is not equal (or at least close) to unity, the exit irradia-
tions selected must be adjusted. A standard way of doing this is to increase or decrease all the
exit,c values uniformly, depending on whether the keff is too high or too low. Also, the radial
power distribution must be examined (this can be done at this time, or after another cycle of
iterations with the new exit,c values).
If the radial power distribution is not satisfactory, e.g., if the degree of radial flattening is not
the desired one, or if peak channel powers are too high (or too low), or if the rate of refuelling
in certain regions is too high, then the regions and/or exit irradiation values must be “manually”
adjusted to move the results in the desired direction. Then another cycle of iterations is
started.
This process is repeated until a critical reactor (or the required bias) and a satisfactory power
distribution are obtained, at which point the time-average model has been established. In
practice, for operating reactors some additional conditions may be imposed when designing the
time-average model, such as matching time-average channel powers determined from relevant
operating histories. This is obtained by varying the individual channel exit irradiations; in such
instances Figure 9 may be replaced by a map with more irradiation regions, or even by a full
map of 380 individual exit irradiations, one for each channel.
8.2.7 Output of Time-Average Calculation

The calculations described above provide or can provide a number of time-average outputs,
which are useful as design outputs and also as targets for the reactor operator. These outputs
include:
 The reactor multiplication constant (keff)
 The 3-dimensional time-average neutron-flux distribution. From this, the time-
average zone fluxes can be obtained, for use as the target distribution by the Reactor
Regulating System.
 The channel and bundle powers. Note that these are the time-average values; they
do not include the instantaneous ripple corresponding to the stage in the refuelling
cycle. The maximum time-average channel power in a CANDU-6 reactor is typically
about 6.6-6.7 MW.
 The entrance and exit irradiation values for each bundle position in core.
 The channel dwell times (intervals between refuellings of the individual channels).
Figure 11 shows dwell times from a CANDU-6 time-average calculation. One-cycle
residence time of bundles at a given position in a CANDU-6 reactor typically range
between 160 and 360 full-power days (FPD).
 The refuelling rates in the various irradiation regions.
 The reactivity decay rate in mk/Full-Power Day. This can be determined in the time-
average model from the change in reactivity (as obtained from the change in keff) for
an overall change in the exit irradiation in all regions and linking that to the change in
residence time corresponding to the change in irradiation. The reactivity-decay rate
per FPD is useful in gauging the expected variation in zone-control-compartment fill
per FPD if there were no refuelling. [Note that the same quantity can also be calcu-
lated by doing 2 core snapshots 1 FPD apart (see next Section) without refuelling any
channels.]
See some of these results from a particular time-average calculation for a CANDU-6 reactor in
Table 2 below.
Table 2 Example of Summary Results from a CANDU-6 Time-Average Calculation

Total Reactor Power (MW) 2061.4
Total Fission Power (MW) 2158.5
Average Channel Power (kW) 5425
Average Bundle Power (kW) 452
Uranium Mass per Bundle (kg) 19.1
Maximum Channel Power (kW) 6604 (N-6)
Maximum Bundle Power (kW) 805 (O-5, Bundle 7)
keff 1.00250
3272 MW.h/Bundle
Average Exit Burnup 171.3 MW.h/kg(U)
7139 MW.d/Mg(U)
Avg. Channel Dwell Time (FPD) 192
1.98 channels/FPD
Feed Rate
15.83 bundles/FPD
Reactivity Decay Rate (mk/FPD) -0.385
Average Zone Fill (%) 50.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
A 310 300 294 294 300 310
B 356 300 253 237 227 226 226 227 237 253 300 356
C 317 270 233 204 193 188 190 190 188 193 204 233 270 317
D 304 258 222 198 177 171 169 173 173 169 171 177 198 222 258 304
E 316 257 220 195 179 163 160 160 164 164 160 160 163 178 195 220 257 316
F 265 222 197 180 169 158 173 174 177 177 174 173 158 169 180 197 222 265
G 293 236 201 186 174 174 170 172 173 173 173 173 172 170 174 174 186 201 235 286
H 258 213 187 178 170 172 170 171 172 172 172 172 171 170 172 169 178 187 212 252
J 319 239 198 178 173 177 173 184 185 187 187 187 186 185 184 173 177 172 177 197 233 311
K 298 224 188 171 169 175 172 185 187 189 192 192 189 187 184 172 175 169 170 187 218 290
L 286 217 183 166 163 170 170 184 187 191 196 196 191 187 184 170 170 162 165 182 210 278
M 286 216 181 164 160 167 168 183 186 191 196 196 190 186 182 168 167 159 163 180 210 278
N 299 223 185 166 160 166 167 181 185 188 192 192 188 185 181 167 165 159 165 184 216 291
O 315 237 195 172 163 167 167 180 183 185 186 186 185 183 179 166 167 163 172 194 230 306
P 259 211 184 173 165 168 167 170 170 170 170 170 169 167 168 165 173 184 211 252
Q 294 237 203 188 175 174 169 170 170 170 170 170 170 169 173 174 187 203 237 287
R 274 232 208 188 173 165 173 172 173 173 172 173 165 172 188 208 231 273
S 339 275 237 207 184 172 167 166 169 169 166 167 172 184 207 237 274 338
T 336 281 238 208 191 182 179 183 183 179 182 191 208 237 281 336
U 339 282 239 213 199 193 195 195 193 199 213 239 281 339
V 350 283 241 221 209 206 206 209 221 241 283 350
W 290 271 260 260 271 290
Boundaries Of Radial Zones
Boundary Of Inner Burnup Region
Figure 11 Dwell Times (in FPD) from a Time-Average Calculation for the CANDU 6
9 More Design Calculations: Snapshots Based on the Time-Average

Model
In this section we look at another type of design calculations, intended to look ahead at power
and overpower results which may be expected during reactor operation.
An important point to remember about the time-average core is that it features no refuelling
ripple: all properties are averaged over residence time, no channel is near the beginning or end
of its refuelling cycle. However, at the core design stage it is important to be able to collect
possible and realistic snapshots of the reactor power distribution during actual reactor opera-
tion, without actually following the history of the reactor. We can obtain snapshots of the
power distribution if we can define instantaneous distributions of bundle irradiation to calculate
instantaneous lattice properties, based on which the diffusion equation can be solved.
On way to obtain instantaneous lattice properties is to create a map of “channel ages”. To
define channel ages let us first consider a typical “channel-irradiation-and-power cycle”.
9.1 Channel-Irradiation-and-Power Cycle

The channel-irradiation-and-power cycle may be described as follows.
 When a channel is refuelled, its local reactivity is high, and its power will be several
percent higher than its time-average power.
 The fresh fuel in the channel then initially goes through its plutonium peak as it picks
up irradiation. The local reactivity increases for ~ 40-50 FPD, and the power of the
channel tends to increase further.
 Following the plutonium peak, the reactivity of the refuelled channel starts to de-
crease as the irradiation increases further, and its power drops slowly.
 The reactivity of the channel and its power then continue to drop. Eventually, the
channel becomes a net “sink” (absorber) of neutrons, the time since the channel’s
last refuelling approaches the channel’s “dwell time” Dc (as defined in the time-
average model), and the fuel irradiation in the bundles which will be discharged at
the channel’s next refuelling increases towards the discharge value, and the time
comes around when the channel must be again refuelled. At this time the power of
the channel may be 10% or more below its time-average power. When the channel
is refuelled, its power may jump by 15 to 20% or even more.
In on-going reactor operation, the power of each channel therefore goes through an oscillation
about the time-average power during every cycle. This cycle repeats - not in exactly the same
quantitative detail each time- at every refuelling of the channel. This is illustrated in Figure 12.
Channel Power Cycles
15
10
5
% CHANGE
IN POWER 0
-5
-10
0 1 2 3
Figure 12 Channel-Power Cycle During On-Going Reactor Operation
9.2 Channel Ages

On any day of operation, we can define the “age” ac of channel c as the instantaneous fraction
that it has achieved in its irradiation cycle since its last refuelling. On any given day different
channels will of course have different ages: channels which were refuelled recently have a low
age (a number close to 0.0), whereas those which were refuelled quite a while ago have a
“high” age (a number close to 1.0). The channel ages ac will be numbers between 0 and 1,
although it is possible for a given channel to have an age somewhat higher than 1, if it has not
been refuelled for a time longer than its dwell time.
From the meaning of the channel’s age, we can define a value for the instantaneous irradiation
of all bundles in the channel. For channel c and bundles b, using the entrance and exit irradia-
tion values for bundle positions (from the time-average model), we can write for bundle’s cb
instantaneous irradiation
inst ,cb  entr ,cb  ac  exit ,cb  entr ,cb  (28)
The lattice properties for position cb can then be determined from the instantaneous irradiation
value.
9.3 Snapshot with Patterned Ages

To model the core on a given day, we can define a core snapshot by assigning to each channel c
in core an age value ac.
We could select random values (between 0 and 1, or slightly higher than 1) for the ac. However,
random selection of ages may easily result in low age values for close-neighbour channels (i.e.,
clusters of low ages). This would not be very realistic, because low ages in close proximity can
lead to “hot spots” in channel power, and a good operator will consciously avoid refuelling
close-neighbour channels very closely in time, so as to avoid low ages close to one another.
It is therefore more desirable to use some “intelligence” in assigning channel ages. One way
which has been found to do this is to start by defining an “intelligent pattern” in the order in
which channels would be refuelled in a relatively small region of the core, e.g., 6-channels-by-6-
channels (see Figure 13). The next step is to subdivide the entire core into 6 x 6 regions and to
define a core “refuelling sequence” by selecting channels intelligently in sequence from one
region to region to another, making sure that whenever revisiting each 6 x 6 region the order
within the region follows the pattern defined for the 6 x 6 square. One obtains (see [Rozon and
Shen 2001] a core refuelling sequence such as shown for example for the CANDU 6 in Figure 14,
with numbers nc from 1 to 380 assigned to the channels. The age assigned to channel c can
then be defined as
380.5  nc
ac  , nc  1,..., 380 (29)
380
Once the snapshot model with lattice properties derived using Eqs. (29) and (28) is solved with
the diffusion code, useful results are obtained for the maximum channel and bundle powers
and, also very importantly, for the Channel Power Peaking Factor (see further below, Section
11.3) and for the overall channel-power, bundle-power, and region-power distributions which
can be expected during on-going reactor operation.
From the one snapshot, many others can be obtained by “massaging” the core refuelling
sequence, for example by reflecting it horizontally or vertically, or alternatively by redefining
each nc as modulo (nc + N), where N is a number which can be selected at will. In this manner
many sets of results and good statistics can be obtained for maximum channel and bundle
powers and the CPPF. This can be very useful in determining rules for selecting channels for
refuelling.
Figure 13 Possible Pattern for Order of Refuelling in a 6 x6 Region of Core
Figure 14 Possible Core Refuelling Sequence for the CANDU 6
10 Design of the CANDU Initial Core

The CANDU initial core is the state of the reactor as it starts life. This is a very specific state, in
which the fuel throughout the core is fresh (unirradiated). At zero irradiation, the CANDU
lattice has high reactivity. Of course this high reactivity must be cancelled in order to allow for
the reactor to be run in steady state.
There is no differential burnup which can assist in flattening the power distribution. Conse-
quently, the power of the central core region would be unacceptably high if no alternate means
of flattening the radial power distribution were provided.
Two means are available in CANDU to cancel the high reactivity of the initial core and to flatten
its power distribution. These are the use of moderator poison (boron, or possibly gadolinium)
and the use of depleted fuel (fuel at an enrichment around 0.48-52% of U-235).
Advance calculations are needed to design how exactly to use these two means to render the
core critical and to provide appropriate power flattening. Design calculations are done with
various concentrations of poison and various numbers and locations of depleted-fuel bundles
until a desirable/attractive result is achieved.
The solution is not unique. There are many ways to design the poison concentration and
depleted-fuel assignment for the initial core. However, note:

 using only depleted fuel and no moderator poison would result in having to initiate
refuelling very early, the consequence being that too much fuel would be removed
with very low burnup – an economic penalty, whereas
 using only boron and no depleted fuel would not achieve sufficient power flattening.
In the first design of the CANDU-6 initial core, the solution which was adopted used two de-
pleted-fuel bundles (of 0.52 atom percent 235U content) in each of the innermost 80 fuel chan-
nels, with the depleted-fuel bundles located in positions 8 and 9, numbered from the channel
refuelling end. In these axial positions, the depleted-fuel bundles are removed from the core in
the first refuelling visit of each of these channels. The boron concentration in this design turns
out to be about 2 ppm at full CANDU-6 power. In recent years some operators have redesigned
the initial core loading to factor in the impact of the required initial poison concentration on the
safety analysis.
11 Fuel-Management Calculations for Reactor Operation

Once the reactor is operating, it is important to follow the reactor history. This is called a core-
follow calculation. It involves calculating updates of the core’s power and burnup distributions
at specific instants in the reactor’s operating history, usually at time intervals of 1 or a few FPD.
The objective is to have good knowledge of the state of the core at all times, to be able to
confirm compliance with the maximum channel and bundle powers allowed by the operating
licence.
Note that the core-follow calculations described here can also be done for a simulated core
history, with simulated channel refuellings and device positions. This is to demonstrate how a
reactor would/could be operated.
11.1 From Initial Core to Onset of Refuelling

At the very start of reactor operation, the entire initial fuel load goes through the plutonium
peak at about the same time (about 40-50 FPD). At this juncture, the core reaches its global
plutonium peak and the core reactivity is the highest it will ever be. Following the plutonium
peak, the plutonium production can no longer compensate for the depletion of 235U and the
build-up of fission products, and the excess core reactivity decreases. Also at this time modera-
tor poison must be removed as the excess reactivity drops gradually to zero, at about FPD 120.
During this entire first period in the reactor life, refuelling is not necessary since there is already
excess reactivity. The core-follow for this period is therefore rather simple. It is a matter only of
simulating the core every few days with instantaneous reactivity devices and a poison concen-
tration which yields a critical reactor.
When the excess core reactivity has fallen to a small value, actually, about 10 or 20 FPD before it
reaches 0 (i.e., typically around FPD 100), refuelling operations start. It is best not to wait until
excess reactivity is exactly 0, because the initial refuelling rate would prove too high.
11.2 After Onset of Refuelling

When on-power refuelling starts, it becomes the primary means of maintaining a CANDU
reactor critical. During the transitional period which follows the onset of refuelling, the reactor
gradually approaches the equilibrium core, with the refuelling rate rapidly tending to the time-
average value (for example, approximately 15-16 bundles per FPD in the CANDU 6).
When refuelling starts, it is channels in the inner core which tend to be refuelled first, because
this region has achieved the highest burnup (on account of its higher flux values). As time goes
on, more channels in the peripheral regions of the core will be refuelled.
Approximately 400 to 500 full-power days (FPD) after initial start-up, a CANDU reactor has
reached an “equilibrium core” state. The overall refuelling rate, the in-core average burnup, the
burnup of the discharged fuel, and the refuelling rates in different core regions have become
essentially steady with time. A number of channels are refuelled every day, on the average.
However, note that refuelling is not necessarily done every calendar day; some stations prefer
to concentrate all refuelling operations to 2 or 3 days within each week.
Doing core-follow calculations at short intervals (1 to a few FPD) then becomes crucial in run-
ning the reactor in a steady and effective manner, to comply with licence limits, to ensure that
there are no “hot spots” or large power tilts.
Each run of the diffusion code in the core-follow requires inputting the instantaneous 3-
dimensional irradiation/burnup distribution. This of course comes from the output of the
preceding run. In addition, each run of the core-follow requires modelling all channel refuel-
lings which have occurred since the previous run (and their timing). This allows modelling
bundle movements and the entry of fresh-fuel bundles in appropriate locations. The instanta-
neous positions of reactivity devices are also input. From the power calculated and the time
step used, the fuel irradiation in each individual bundle are updated, and the lattice properties
updated.
The output of the current run is then obtained and becomes the starting point for the calcula-
tion at the next time step. The current 3-dimensional power and fuel-irradiation distributions
obtained serve then specifically as the basis on which intelligent selections of channels to refuel
are made (see second section below).
In addition, another very significant output of core-follow calculations is the determination of
the Channel Power Peaking Factor, which is used in the calibration of in-core protection-system
(Regional Overpower/Neutron Overpower) detectors. This is defined in the next Section.
11.3 Channel-Power Peaking Factor

At any given time, there are several channels in the core which are at or near the maximum
power in their power-and-irradiation cycle. Therefore, the maximum instantaneous channel
power is always higher than the maximum channel power in the time-average model.
The Channel-Power Peaking Factor (CPPF) quantifies the degree by which the instantaneous
power distribution peaks above the time-average distribution:
 Channel  Power 
ins tan tan eous  c  
CPPF  Max  (30)
c  Channel  Power c  
 c time  average 
where c runs over channels in the core.

The exact value of the CPPF (which varies from day to day) is extremely important because it is
used to calibrate the in-core protection (ROP/NOP) detectors. The protection system(s) are
designed to actuate a shutdown system in a power increase before any fuel channel risks
reaching a “critical channel power” (e.g., a power at which there would be fuel dryout some-
where in the channel). The CPPF has therefore important safety implications.
Hundreds of flux shapes are used in the design of the ROP/NOP detector system to determine
detector positions and setpoints. These flux shapes are calculated in the time-average model,
assuming many different core configurations. But because the real instantaneous channel
powers are higher than the time-average powers, channels would reach their “critical channel
power” earlier than predicted in the time-average model. To take this into consideration and
ensure proper safety coverage in the instantaneous power shape, the in-core detectors are
calibrated each day to the instantaneous value of CPPF.
On the other hand, a high CPPF value could cut into the reactor’s operating margin, i.e., the
CPPF has economic implications. It is therefore important to keep the CPPF as low as possible.
The competing safety and economic considerations mean that a careful selection of channels to
be refuelled needs to be made always. Determining the daily CPPF value, selecting channels to
keep it low, and ensuring detectors are calibrated to the correct value, are on-going duties of
the fuelling engineer or reactor physicist at a CANDU nuclear generating station.
11.4 Criteria for Selecting Channels for Refuelling

One of the main functions of the fuelling engineer or reactor physicist is to establish a list of
channels to be refuelled during the following period (few days) of operation. For instance, for a
5-FPD period, a list of approximately 10 channels must be prepared. To achieve this, the current
status of the reactor core is determined from computer simulations of reactor operation, the
on-line flux mapping system, the ROP and RRS in-core detectors, and zone-control-
compartment water fills. The computer simulations of reactor operation provide the instanta-
neous 3-dimensional flux, power and burnup distributions.
Good combinations of channels selected for refuelling in the few days to follow will typically
contain:
 channels “due to be refuelled”, i.e., channels for which the time interval since the
last refuelling is approximately equal to the channel dwell time (from the time-
average calculation)
 channels with high current value of exit burnup, relative to their time-average exit
burnup
 channels with low power, relative to their time-average power
 channels in (relatively) low-power zones (compared to the time-average zone-power
distribution)
 channels which, taken together, promote axial, radial and azimuthal symmetry and a
power distribution close to the reference power shape
 channels which provide sufficient distance to one another and to recently refuelled
channels (to avoid hot spots)
 channels which will result in acceptable values for the individual zone-controller fills
(20%-70% range), and
 channels which, together, provide the required reactivity to balance the daily reactiv-
ity loss due to burnup (and which will, therefore, tend to leave the zone-controller
fills in the desired operational range: average zone fill between 40 and 60%) .
Note that channel selections for refuelling are seldom unique. Many options are available. A
good way of being confident about a channel selection is to perform a pre-simulation of the
core following the refuellings. This pre-simulation, especially if the code can predict the result-
ing zone-control-compartment fills (i.e., the bulk-control and spatial-control responses of the
Reactor Regulating System), will show whether the various power, burnup, and zone-fill criteria
are likely to be satisfied, or whether a different channel selection should be made. In fact,
CANDU operators employ pre-simulations routinely to ensure optimum fuelling selections,
including achieving low CPPF values, i.e., good operating margins.
12 Advanced Fuel Cycles

On account of the high neutron economy provided by heavy-water moderation, the CANDU
reactor can run on natural-uranium fuel (unlike Light-Water Reactors). But this also means that
the CANDU reactor can run on fuels which have higher effective enrichments than natural
uranium. This allows a number of Advanced Fuel Cycles to be adopted in CANDU reactors in the
future. Many of these cycles illustrate the great synergy possible between CANDU and PWRs, as
CANDU can in some way re-use PWR fuel and extend the achievable burnup and reduce waste
volumes.
Fuel-management simulations for Advanced Fuel Cycles would proceed in essentially the same
manner as for the once-through natural uranium (NU) cycle. Of course the calculation of lattice
properties would have to reflect the presence of different nuclides and/or of different nuclide
concentrations.
In many cases the Advanced Fuel Cycles can be adopted in CANDU reactors without any
changes to the reactor or the control devices. In other cases it may be necessary to adopt
different axial refuelling schemes (replacing fewer bundles per visit to a channel). Or it may be
appropriate to alter or remove some reactivity devices to optimize reactivity and burnup
considerations.
Some examples of Advanced Fuel Cycles in CANDU are:
 The use of Slightly Enriched Uranium (SEU). In this cycle uranium with enrichment in
the range 0.9% to, say, 1.2% is used instead of NU. The discharge burnup can be in-
creased significantly, with a consequent significant decrease in the volume of irradi-
ated fuel replaced per year.
 The use of recycled uranium (RU). This uranium would be recovered from the re-
processing of irradiated fuel from CANDU, or more likely, from PWRs. In this case,
the presence of certain concentrations of other isotopes of uranium in the fuel has
to be recognized in the calculation of lattice properties.
 The MOX (Mixed Oxide) Fuel Cycle, where fuel would be produced from a mixture of
uranium and plutonium separated in the reprocessing of PWR fuel. Here again a
range of fissile fractions can be accommodated in the MOX fuel.

 The DUPIC Cycle (Direct Use of PWR Fuel in CANDU). This cycle also uses fuel made
from reprocessed PWR fuel, however in this case dry reprocessing is used which
does not separate the plutonium from the uranium. A typical fissile fraction in irra-
diated PWR fuel is about 1.5% (0.9% uranium and 0.6% plutonium). Studies of the
DUPIC cycle made from this fuel have shown that the additional burnup obtainable
in CANDU can be about double the burnup achieved in the original Pressurized Wa-
ter Reactor.
 The Thorium Cycle. Although Th-232 is not fissile, it is a fertile nuclide: Upon absorp-
tion of a neutron and two beta decays it becomes (via Pa-233) the fissile fuel U-233.
This cycle draws a lot of interest, as thorium is roughly 3 times as abundant as ura-
nium in the earth’s crust, so its use would significantly increase our nuclear natural
resources. The thorium cycle requires mixing uranium with a “driver” fuel which
provides the neutrons to be absorbed by the thorium. The driver fuel can be ura-
nium or plutonium. In CANDU, the thorium and the driver fuel can be placed in dif-
ferent core regions or different channels, or can even be mixed within the same fuel
bundle. The U-233 can be burned in situ or can be separated by reprocessing the
fuel after it has been sufficiently irradiated. Much study of the thorium cycle is on-
going currently.
13 Summary
This Chapter has summarized the concepts, models and calculations involved in the manage-
ment of nuclear fuel in CANDU reactors.
Fuel management in CANDU has both design and operations aspects.
The design component consists of establishing:
 the desired time-average power distribution for the equilibrium core, which will be
used as the target power shape by the site fuelling engineer, and
 the configuration of depleted fuel in the initial core.
The design of the time-average distribution is facilitated by the flexibility in selecting region-
specific (or, in the limit, channel-specific) target exit-irradiation values and axial refuelling
schemes, allowed by the CANDU on-power-refuelling feature.
The operations component is the responsibility of the site fuelling engineer or reactor physicist.
It involves:
 core-follow calculations, typically performed 2 or 3 times per week to keep close
track of the in-core flux, power, and burnup distributions and of the discharge bur-
nup of individual bundles,
 the selection of channels for refuelling, based on the current core state, power and
burnup distributions and zone-control-compartment water fills, and
 the determination of the CPPF (channel-power-peaking factor) value, used as a cali-
bration factor for the ROP/NOP detectors.
14 References
[Damario 2016] D. Damario, “The Fuelling Machine”, chapter in The Essential CANDU, edited by
W. Garland, 2016.
[Hébert 2012] A. Débert, D. Sekki and R. Chambon, “A User Guide for DONJON Version4”,
Technical Report IGE-300, École Polytechnique de Montréal, 2012.
[Irish 2002] J.D. Irish and S.R. Douglas, “Validation of WIMS-IST”, in Proceedings of the 23rd
Annual Conference of the Canadian Nuclear Society, Toronto, Ontario, Canada, 2002
June.
[Marleau 1999] G. Marleau, “DRAGON, Theory Manual”, Technical Report IGE-236, Institut de
Génie Nucléaire, École Polytechnique de Montréal, 1999 February.
[Rouben 1995] B. Rouben, “Description of the Lattice Code POWDERPUFS-V”, AECL-11357, 1995
October.
[Rouben 2002] B. Rouben, “RFSP-IST, The Industry Standard Tool Computer Program for CANDU
Reactor Core Design and Analysis”, in Proceedings of PBNC-2002 (13th Pacific Basin Nu-
clear Conference), Shenzhen, China, 2002 October.
[Rouben 2007] B. Rouben, “Review of the CANDU Time-Average Model and Calculations”, in
Proceedings of the 28th Annual Conference of the Canadian Nuclear Society, Saint John,
New Brunswick, Canada, 2007 June 3-6.
[Rozon and Shen 2001] D. Rozon and W. Shen, “A Parametric Study of the DUPIC Fuel Cycle to
Reflect Pressurized Water Reactor Fuel Management Strategy”, Nucl. Sci. Eng. 138, 1-25
(2001)
15 Acknowledgements
Richard Chambon
Ovidiu Nainer
Thanks are also extended to Bill Garland for expertly editing and assembling the final copy.
1
Author Biographies
Table of Contents
1 Dr. Robin A. Chaplin (chapters 1, 2, 8, 9, appendix Basic Electrical Theory)........................... 2

2 Dr. Willy Cook (chapters 14, 15).............................................................................................. 3
3 Ms. Diane Damario (chapter 20)............................................................................................. 3
4 Mr. Milan Gacesa (chapters 17, 18, 19) .................................................................................. 4
5 Dr. William J. Garland (editor in chief, prologue, chapter 7)................................................... 5
6 Dr. George Alan Hepburn (chapter 10) ................................................................................... 6
7 Dr. Jin Jiang (chapter 11) ......................................................................................................... 7
8 Dr. Derek Lister (chapters 14, 15)............................................................................................ 7
9 Dr. Guy Marleau (chapter 3) ................................................................................................... 8
10 Dr. Eleador Nichita (chapters 4,5) ........................................................................................... 8
11 Dr. Nik Popov (chapters 6, 7, 16)............................................................................................. 9
12 Dr. Benjamin Rouben (chapters 4, 5, 21) .............................................................................. 10
13 Dr. Victor G. Snell (chapter 13, 16)........................................................................................ 11
14 Mr. Mukesh Tayal (chapters 17, 18, 19) ................................................................................ 12
15 Dr. Edward J. Waller (chapter 12).......................................................................................... 13
©UNENE, all rights reserved. For educational use only, no assumed liability. Author Biographies – December 2016
1 Dr. Robin A. Chaplin (chapters 1, 2, 8, 9, appendix Basic Electrical

Theory)
Robin Chaplin obtained a B.Sc. and M.Sc. in Mechani-
cal Engineering from University of Cape Town in 1965
and 1968 respectively. Between these two periods of
study he spent two years gaining experience in the
operation and maintenance of coal fired power plants
in South Africa. He subsequently spent a further year
gaining experience on research and prototype nuclear
reactors in South Africa and the United Kingdom and
obtained an M.Sc. in Nuclear Engineering from Imperi-
al College of London University in 1971. On returning
to South Africa and taking up a position in the head
office of Eskom he spent some twelve years initially in
project management and then as Head of Steam
Turbine Specialists. During this period he was involved
with the construction of the 3 x 80 MW Ruacana Hydro Power Station in Namibia and the 2 x
900 MW Koeberg Nuclear Power Station in South Africa being responsible for the underground
mechanical equipment and civil structures and for the mechanical balance-of-plant equipment
at the respective plants. Continuing his interests in power plant modeling and simulation he
obtained a Ph.D. in Mechanical Engineering from Queen’s University in Canada in 1986 and was
subsequently appointed as Chair in Power Plant Engineering at the University of New Bruns-
wick. Here he has taught thermodynamics and fluid mechanics and specialized graduate level
courses in nuclear and power plant engineering in the Departments of Chemical Engineering
and Mechanical Engineering. An important function was involvement in the plant operator and
shift supervisor training programs at Point Lepreau Nuclear Generating Station. This included
the development of material and the teaching of courses in both nuclear and non-nuclear
aspects of the program. He has also been involved with the UNESCO sponsored Encyclopedia of
Life Support Systems (EOLSS) as Honorary Theme Editor and primary author of the theme on
Thermal Power Plants and has also assisted with the theme on Nuclear Energy and Reactors.
Altogether he has contributed some three dozen chapters to this major source of international
knowledge. As an adjunct professor at Waterloo University he has established and taught a
graduate course in power plant thermodynamics to engineers in the nuclear industry as part of
the UNENE masters program in nuclear engineering. He has served as Acting Chair and Chair in
the Department of Chemical Engineering at University of New Brunswick and has been a consul-
tant for Canadian Power Utility Services.
Author Biographies 3
2 Dr. Willy Cook (chapters 14, 15)

Willy Cook is an Associate Professor in the Department of Chemi-
cal Engineering at the University of New Brunswick (UNB). He is a
professional engineer with consulting and research interests in
power plant chemistry and corrosion control and teaches courses
related to these topics and global energy issues. Willy earned his
BSc and MSc degrees in Chemical Engineering at UNB in 1997 and
1999 where he completed option programs in Power Plant and
Nuclear Engineering. He worked as the research engineer and
laboratory manager for the Chair in Nuclear Engineering at UNB
until 2004 where he supervised the academic research and
industrial contracts undertaken by the Chair. He completed his
PhD in 2005 (UNB) and has since been actively involved in the
Generation IV International Forum (GIF) evaluating materials and
chemistry control regimes for a supercritical water-cooled reactor. The restart of the CANDU-6
plant at Point Lepreau following its refurbishment outage in 2012 coincided with a sabbatical
leave from UNB and Willy spent nine months during this time as a technical consultant for the
chemistry department at the station during the restart activities. In the Fall of 2014, he was
appointed as the Scientific Director of the Centre for Nuclear Energy Research, an Institute on
UNB’s campus that undertakes research and provides consulting to the power generating
industry and develops commercial products for corrosion measurement and control.
3 Ms. Diane Damario (chapter 20)

Diane is a registered professional engineer in the province of Ontario.
She earned a B.A.Sc. in Mechanical Engineering at the University of
Toronto in 1990. After completing her studies, she joined AECL at
Sheridan Park. She has been involved in fuel handling at Sheridan Park
ever since, transitioning to Candu Energy Inc., a member of the SNC-
Lavalin Group, when the commercial division of AECL was privatized in
2011. Diane's fuel handling experience has mainly been based on the
CANDU 6 station design, particularly for the Qinshan and Wolsong 2, 3
and 4 stations. Over the years, she has delivered training on fuel
handling at Sheridan Park and Chalk River.
4 Mr. Milan Gacesa (chapters 17, 18, 19)

Independent consultant - nuclear fuel and project
management. Milan earned his BASc and MAsc in
Mechanical Engineering at the University of Windsor
and is a registered professional engineer in the province
of Ontario. He worked at Westinghouse Canada Ltd. for
7 years, initially performing tests and developing de-
sign/analysis tools used for CHF and thermal-hydraulic
stability analysis, and later leading the nuclear fuel
design section where the design of C6 fuel was pro-
duced. At AECL he continued to work on C6 fuel design,
testing, and verification until its successful launch in the first generation C6 reactors. As manag-
er of the fuel branch Milan broadened the branch’s participation in all aspects of the CANDU
fuel industry, in particular fuel design, testing, development of analytical tools, assessment of
plant performance feedback (especially as related to defective fuel), and assistance to opera-
tors. As Engineering and Project Manager of CIRENE Technical Assessment (Italian BWR similar
to Gentilly 1) he led a group of selected AECL specialists to a successful completion of an inde-
pendent evaluation of CIRENE design and, in particular, safety. Following CIRENE, Milan com-
pleted a number of other projects, including an extensive multi-disciplinary study of alternative
fuel cycles in CANDU for a Japanese client, before his overseas assignment as Manager of
Engineering at KAERI (Republic of Korea) where he led an expat group of AECL senior technical
specialists and Korean engineers in completing the design packages for a number of C6 NSSS
systems. For the past 14 years Milan has provided consulting services to a number of clients,
including AECL where he worked for a number of years as the senior fuel specialist on the ACR
project. In addition to producing the fuel bundle configuration for ACR, the fuel team spear-
headed a number of methodology changes, including the adoption of fuel acceptance criteria,
development of design verification strategy and plan, and standardization of verification as-
sessments, which can be used on any future CANDU project.
5 Dr. William J. Garland (editor in chief, prologue, chapter 7)

Bill Garland is Professor Emeritus of Nuclear Engineering in
the Department of Engineering Physics at McMaster Universi-
ty (Ontario, Canada). He specialized in reactor physics and
thermalhydraulics. Since 2000, he has been Academic Direc-
tor of CANTEACH, the public on-line library for technical
information on CANDU reactors. On partial and full second-
ment from McMaster from 2004 to 2008 he served in various
capacities within the University Network of Excellence in
Nuclear Engineering (UNENE), namely Secretary / Treasurer,
Program Director, and Executive Director. Since retiring from
McMaster he served as UNENE President until September
2009. From October 2009 to September 2011 he was Academic Consultant to UNENE with
Knowledge Management as the primary duty.
He served as Department Chair in the Department of Engineering Physics at McMaster from
1988 to 1994 and was Director of MNR (McMaster Nuclear Reactor) for 1994 to 1995 leading up
to the decision to revitalize reactor operations. Subsequent to that he is took a lead role in
MNR safety analysis and operational support analysis.
From 1975 to 1983, he worked in the Canadian Nuclear Industry specializing in CANDU heat
transport system analysis and design (Darlington and CANDU 6). The major effort from early
1980 to mid 1983 was in heat transport system stability investigations leading up to the now
standard header to header interconnection. This included pre- and post-test analysis, coordina-
tion of AECL's process team, test planning, and liaison.
While on research leave at Harwell Nuclear Labs in England, Bill conducted knowledge engineer-
ing for heat exchanger selection and developed a heat exchanger selection computer code
which became a commercial product distributed through the international company HTFS. In
addition, computerized water property functions developed at McMaster are in use in about 13
countries by over 120 users in 51 institutions and major engineering firms worldwide.
Bill received the CNS Education and Communication Award in 2001 and became a CNS fellow in
2006.
William J. Garland, P. Eng., FCNS (B. Eng. Phy. 1970, M. Eng. Phy., 1971, Ph.D. Chem. Eng. 1975,
McMaster University, Canada)
6 Dr. George Alan Hepburn (chapter 10)

Alan Hepburn earned his B.Sc. in Electrical Engineering at the Universi-
ty of Glasgow in 1966. After a further four years study in the field of
automatic control, he obtained a Ph.D. in Electrical Engineering in 1970.
During his postgraduate work, he had the use of an early prototype
computer which had been developed to demonstrate that semiconduc-
tors were a practical replacement for vacuum tubes in digital comput-
ers.
Despite the assurance of the Canadian embassy that there were “no
jobs for Ph.D.s in Canada”, Alan emigrated to Canada in the fall of 1970, where he obtained a
position in a small company applying minicomputers to data acquisition.
Alan joined the Plant Computer Systems Branch at AECL in Mississauga in 1972. He was the first
employee of the branch to have had experience with computers prior to joining the company.
His first task at AECL was to write the Reactor Regulating System program for Bruce A. In 1974,
he was posted to Bruce “A” NGS, where he was AECL’s representative on the Digital Control
Computer (DCC) commissioning team.
Following his return to Sheridan Park in 1976, Alan continued his work in real time computer
systems development for Bruce B, the early CANDU 6’s, and Darlington, including involvement
in the development of the world’s first fully computerized shutdown systems for Darlington GS.
He became a section head, followed by branch manager of the Plant Computer Systems Branch,
then Department Manager of the Instrument and control systems Department at Sheridan Park.
In 1987, Alan left the nuclear industry and moved to the aerospace, where he managed the
Engineering Sciences group at Garrett Canada, primarily responsible for the development of
software for the environmental control systems on many of the western world’s larger civil and
military aircraft. After four years at Garrett, he transferred to TRW who were subcontracted to
the Systems Engineering group of CDC in Nepean to develop a major communications system
for the Canadian Army. Just two years later, the project was transferred to Calgary, and after a
further two years, TRW’s contract was cancelled.
Rather than move to Washington with TRW, Alan decided to see if there was still a position for
him back at AECL, and as a result, found himself working as a contractor on the development of
a modern distributed control system for the CANDU 3, and subsequently CANDU 9 reactors in
Saskatoon. This was followed by four years at Bruce A as AECL’s on-site representative on
replacement project.
In 2000, Alan moved to Chalk River to head the Control and Operations Technology Branch,
whose responsibilities included the early development work on a distributed control system for
the Advanced CANDU Reactor. He retired from this position in 2006.
Alan’s only teaching exposure has been as a flight instructor, in which capacity he is still active.
7 Dr. Jin Jiang (chapter 11)

Professor, Western University
An NSERC/UNENE Senior Industrial Research Chair Professor, in the
Department of Electrical and Computer Engineering, Western Univer-
sity in Canada. His research interests are in the areas of fault-tolerant
control of safety-critical systems, instrumentation and control of
nuclear power plants, and control of electrical power plants and
power systems.
He is a fellow of Canadian Academy of Engineering, a fellow of IET
(formally IEE), and a fellow of ISA. He is also a member of International
Electrotechnical Commission (IEC) 45A subcommittee to develop
industry standards on instrumentation and control for nuclear facili-
ties. He also works closely with International Atomic Energy Agency (IAEA) on modern control
and instrumentation for nuclear power plants.
8 Dr. Derek Lister (chapters 14, 15)

Derek Lister is Professor Emeritus in the Department of Chemical
Engineering at the University of New Brunswick in Fredericton, where
he continues to hold the Research Chair in Nuclear Engineering. He has
a BSc and an MSc in chemical engineering from the University of
Manchester and a PhD in physical chemistry from the University of
Leicester. He entered the nuclear industry after graduation as an R&D
Engineer with the English Electric Co. in the UK, where he worked on
Magnox fuel systems and was on the commissioning team for the
Hinkley Point A reactor. He joined AECL at Chalk River to develop
CANDU fuel but soon moved to the Chemical Engineering Branch to
work on coolant chemistry and corrosion, specialising first of all in activity transport. His re-
search broadened over the years to cover many aspects of primary and secondary coolant
chemistry and corrosion and in 1992, when Manager of the System Chemistry and Corrosion
Branch, he accepted the appointment of Industrial Research Chair in Nuclear Engineering at
UNB. His research still involves coolant chemistry and corrosion and besides his Chair research
he manages a number of research contracts dealing specifically with power-plant corrosion. He
serves on a number of national and international advisory committees. He is a Fellow of the
Chemical Institute of Canada and as a Member of the Institution of Chemical Engineers in the
UK he is a Chartered Engineer. He is a member of the Nuclear Institute and of the Canadian
Nuclear Society, which presented him with an Outstanding Contribution Award in 2011.
9 Dr. Guy Marleau (chapter 3)

Professor, Department of Engineering Physics, École Polytechnique
de Montréal.
Guy earned his BSc (1977) in physics from the University of Ottawa
and his MSc (1979) and PhD (1983) in physics from McGill University.
In 1983, he joined École Polytechnique de Montréal as a research
scientist where he became a professor of nuclear engineering and
engineering physics in 2006. From 2007 to 2013 he served as the
director on the Institut de Génie Nucléaire at École Polytechnique de
Montréal. For the last 30 years, he has been involved in the devel-
opment of reactor physics simulation codes being one of the co-developers of the well known
lattice code DRAGON that is used to solve the neutron transport equation for general 2D and 3D
geometry. Over the years, he taught several courses in nuclear engineering including reactor
statics, reactor simulation, radiation and radioprotection as well as nuclear technology. He
received in 2010 the WB Lewis medal from the Canadian Nuclear Association and the Canadian
Nuclear Society for his contribution to the development of advanced computational methods in
reactor physics.
courriel/email: guy.marleau@polymtl.ca, Tel: (514)-340-4711 x 4204
addresse: Institut de genie nucleaire, Ecole Polytechnique de Montreal, C.P. 6079, succ. Centre-
ville, Montreal, QC, H3C 3A7, Canada
10 Dr. Eleador Nichita (chapters 4,5)

Dr. Eleodor Nichita, Associate Professor, Faculty of Energy Systems
and Nuclear Science, University of Ontario Institute of Technology.
Eleodor holds a B.Sc. in Engineering Physics from the University of
Bucharest, a M.Sc. in Health Physics from Georgia Institute of
Technology, a M.Sc. in Medical Physics from McMaster University,
and a Ph.D. in Nuclear Engineering from Georgia Institute of
Technology. He is a Licensed Professional Engineer in the Province
of Ontario. He has extensive nuclear-industry experience, begin-
ning at the Institute for Nuclear Research in Pitesti, Romania, and
continuing at Atomic Energy of Canada Ltd., where he worked on
analysis and code development for the neutronic core-simulator, RFSP. In 2004 Eleodor joined
the University of Ontario Institute of Technology where he is currently an Associate Professor
and Director of the undergraduate Nuclear Engineering Program. He teaches courses in Reactor
Physics, Transport Theory and Medical Imaging, and performs research in Nuclear Reactor
Modelling and Simulation, Computational Particle Transport and Computational Medical Phys-
ics. Eleodor held the presidency of the Canadian Nuclear Society between 2009 and 2010.
11 Dr. Nik Popov (chapters 6, 7, 16)

Dr. Popov has over 35 years of experience in the nuclear industry in
Canada and internationally in various disciplines including thermal-
hydraulics, safety analysis, licensing, computer codes development
and validation, and severe accident analysis. He worked at the
University Cyril & Methodius at Skopje, Macedonia as lecturer and
assistant professor (1978-88). Then he worked 23 years at the Atomic
Energy of Canada (1988-2011), the last 3 years as the Director for
Safety and Licensing. He also worked at SNC-Lavalin-Nuclear as
Director of Safety and Licensing (2011-2014). He is the owner of
DENIPO Consulting Ltd., which provides nuclear engineering consult-
ing and educational services.
Dr. Popov was involved and managed various projects at AECL, such as the licensing of ad-
vanced CANDU reactor in Canada, USA, UK, and other countries. He was involved in interna-
tional projects at AECL, such as serving as the Canadian representative in the G-8 safety review
of the Kursk 1 RBMK nuclear power station in Russia (2001-2002); and leading the AECL team
that interfaced with the European Commission Panel safety review to obtain approval to pro-
ceed with the Cernavoda 3-4 CANDU project in Romania (2009-2010). At COG he was the lead
of the Severe Accident Working Group of the joint project JP-4426 working on the update of
the Severe Accident Management Guidance documentation (2012-2014).
Dr. Popov is an Adjunct Professor at McMaster University, Engineering Physics (since 2000),
where he teaches graduate courses at McMaster and the UNENE program (reactor thermal-
hydraulics design and reactor thermal-hydraulics analysis). He is also an Adjunct Professor at the
University Cyril and Methodius, Electrical Engineering, in Macedonia, where he teaches gradu-
ate courses (reactor engineering, nuclear energy industrial applications, and environmental
aspects of power engineering). Dr. Popov is a lecturer of an international team organized by the
University of Pisa, Italy (since 2006), delivering 3D-SUNCOP courses at annual international
seminars. He teaches best-estimate and uncertainty assessment, safety culture in safety analy-
sis, and development and validation of computer programs.
Dr. Popov participated in the preparation of over 200 publications and industrial reports.
Nikola K. Popov (B. Mech. Eng., 1976, M. Sci. Nucl. Eng., 1983, Ph.D. Mech. Eng. 1988, University
of Zagreb, Croatia).
12 Dr. Benjamin Rouben (chapters 4, 5, 21)

Ben Rouben holds an Hons. B.Sc. from McGill University in Physics
and Mathematics, and a Ph.D. in Nuclear Physics from the Massa-
chusetts Institute of Technology. After pursuing research at the
Université de Montréal for several years, he joined Atomic Energy
of Canada Limited (AECL) in 1975, where he worked in many areas
of reactor physics. From 1993 to 2006, Ben served as Manager of
the Reactor Core Physics Branch at AECL.
Ben was President of the Canadian Nuclear Society (CNS) in 1997-
98, and is currently its Executive Director. He was elected CNS
Fellow in 1999. He is also a member of the American Nuclear
Society (ANS), where he is active on the Reactor Physics Standards
Committee and on the ANS International Committee.
Ben retired from AECL in 2007 February. He is currently a consul-
tant, and an Adjunct Professor at McMaster University and at the University of Ontario Institute
of Technology. He teaches courses on Reactor Physics, Reactor Dynamics, Nuclear Fuel Man-
agement, Power Plant Operation, and Nuclear Concepts for Scientists and Engineers. He
teaches also in the M. Eng. Program offered by UNENE, the University Network of Excellence in
Nuclear Engineering, where he also serves as Secretary/Treasurer.
roubenb@alum.mit.edu
13 Dr. Victor G. Snell (chapter 13, 16)

Dr. Snell is the owner of VGSSolutions, a nuclear safety consulting
and education firm. He has 38 years of nuclear experience, largely
in the area of safety and licensing. As a member of the Senior
Global Assessment Team in 2010-11, he helped review the Global
Assessment of the NRU reactor. As Enhanced CANDU 6 Licensing
Manager for AECL in 2008-9, he was responsible for obtaining a
positive conclusion by the Canadian Nuclear Safety Commission
for the Phase 1 pre-project design review. For ten years, as Direc-
tor of Safety and Licensing at AECL, he had overall responsibility
for interaction with regulatory agencies on AECL's product line of
reactors. This included obtaining a favourable AECB report on
licensability of CANDU 9; initial licensing of Wolsong 2 with KINS in
Korea; negotiating licensing requirements for an unattended urban heating reactor (SES-l 0);
and managing licensing for CANDU 3U with the USNRC, to the point of submittal of an applica-
tion for Standard Design Certification. Other major projects included responsibility for pre-
project licensing, safety design and Probabilistic Safety Analysis of AECL' s Advanced CANDU
Reactor (ACR) design; responsibility (for two years) for the safety analysis and licensing of the
MMIR Project; participation in an IAEA-sponsored design review of a district heating nuclear
plant in Gorky (the first mission of its kind in the Soviet Union); and team leader of a design
review of CANDU & Chernobyl. In the nuclear education area, he is currently Programme
Director, University Network of Excellence in Nuclear Engineering (UNENE), responsible for
course development and deployment and administration of the nuclear M.Eng. programme. He
has over fifty publications on Reactor Safety and Licensing.
Contact: Email: vgssolutions@rogers.com, Phone: (905) 824-6236
14 Mr. Mukesh Tayal (chapters 17, 18, 19)

Mukesh is an independent consultant in
nuclear fuel. He has 37 years of nuclear
experience in areas of nuclear fuel, safety
analysis, stress analysis, and heat transfer. He
has developed a criterion-based approach to
qualify fuel designs which significantly increas-
es the robustness of design verification while
reducing its cost and schedule. He has per-
formed, and provided technical leadership to,
design, testing and analytical verification of
fuels such as 37-element fuel for C6, Bruce,
and Darlington stations; 43-element fuel for
ACR (Advanced CANDU Reactor); and 61-
element fuel for HAC (Highly Advanced Core) CANDU reactor. He has played key roles in devel-
oping several computer codes used by Atomic Energy of Canada Limited (AECL) for design
assessments of nuclear fuels. These include: ELESTRES code for thermo-microstructural-
mechanical performance of fuel under normal operating conditions; ELOCA-A code for accident
analysis of fuel; FEAT code for non-linear thermal analysis; FEAST code for non-linear stress
analysis; BOW code for non-linear lateral deflection; LONGER code for ridging and collapse;
BEAM code for rigidity and vibration; FEED code for diffusion; and GASOUT code for transient
release of fission gas. He has also developed science and models for several phenomena such
as environmentally-assisted cracking (stress corrosion cracking), bowing, longitudinal ridging,
and endplate fatigue. Mukesh has also worked on advanced fuel cycles such as thorium fuel
and enriched high-burnup fuel. Mukesh received the CNA/CNS award for Innovative Achieve-
ment in 2007 and has authored/co-authored well over 200 technical documents in the area of
nuclear fuel.
Mukesh Tayal, P. Eng., (B. Tech., Mech. Engg., 1970; M. Sc., Mech. Engg., 1973, Univ. of Sask.,
Canada).
15 Dr. Edward J. Waller (chapter 12)

Faculty of Energy Systems and Nuclear Science, University of
Ontario Institute of Technology
Ed earned his BSc in Physics and MScE in Chemical Engineering at
the University of New Brunswick (UNB) and his PhD in Nuclear
Engineering at Rensselaer Polytechnic Institute, New York (RPI). He
worked for over 15 years in industry for Science Applications
International Corporation, primarily in threat assessment, health
physics and applications of radiation. In 2003 he joined the Uni-
versity of Ontario Institute of Technology. He is currently an NSERC
Industrial Research Chair in Health Physics and Environmental
Safety. Ed is a Professional Engineer (PEng), Certified Associate
Industrial Hygienist (CAIH) and Certified Health Physicist (CHP). He
teaches radiation protection, health physics, environmental effects
of radiation, Monte Carlo methods and nuclear forensics at UOIT, and performs research in
areas of emergency response, radiation dosimetry, applied health physics, radiation safety, and
threat assessment.
1
Abbreviations
ACNS Advisory Committee on Nuclear Safety (Canada, historical)
ACR Advanced CANDU Reactor
ACRS Advisory Committee on Reactor Safeguards (USA)
AEC Atomic Energy Commission (USA)
AECA Atomic Energy Control Act (Canada)
AECB Atomic Energy Control Board (Canada)
AECL Atomic Energy of Canada Limited
AESOP Atomic Energy Simulation of Optimization (computer code)
ALARA As Low as Reasonably Achievable
ALARP As Low as Reasonably Practical (UK)
AOO Anticipated Operational Occurrence
ASDV Atmospheric Steam Discharge Valve
ASLB Atomic Safety and Licensing Board (USA)
ASSERT Advanced Solution of Subchannel Equations in Reactor Thermalhydraulics (computer
code)
ASTM American Society for and Testing Materials
BEAU Best Estimate with Analysis of Uncertainties
BECS Boiler Emergency Cooling System
BLC Boiler Level Control
BLW Boiling Light Water
BPC Boiler Pressure Controller
BSL Basic Safety Limit (UK)
BSO Basic Safety Objective (UK)
C/S Containment and Surveillance
CANDU CANada Deuterium Uranium – the generic name for Canadian heavy water power
reactors
CATHENA Canadian algorithm for thermal hydraulic network analysis (computer code)
CCP Critical Channel Power
CEAA Canadian Environmental Assessment Act
CFR Code of Federal Regulations (USA)
CHF Critical Heat Flux
CPR Critical Power Ratio
CRL Chalk River Laboratories
CRT Cathode Ray Tube
CSA Canadian Standards Association
CSDV Condenser Steam Discharge Valve
CSNI Canadian Standards for the Nuclear Industry; also Committee on the Safety of Nuclear
Installations
DBE Design Base Earthquake
DCC Digital Control Computer
©UNENE, all rights reserved. For educational use only, no assumed liability. Abbreviations – September 2014
DF-ET Drift Flux-Equal Temperature

DF-UT Drift Flux-Unequal Temperature
DIV Design Information Verification
DNB Departure from Nucleate Boiling
DSA Deterministic Safety Analysis
EA Environment Agency (UK)
ECI Emergency Coolant Injection (a subsystem of ECC)
EFPH Effective Full Power Hours
EIS Environmental Impact Statement
ENSREG European Nuclear Safety Regulators Group
EPR European Pressurized Reactor
EU European Union
EVET Equal Velocity Equal Temperature
EVUT Equal Velocity-Unequal Temperature
EWS Emergency Water Supply
FBR Feed, Bleed and Relief; also Fast Breeder Reactor
FMEA Failure Modes and Effects Analysis
FP Full Power
GAI Generic Action Item
GAR Global Assessment Report
GCR Gar Cooled Reactor
GDA Generic Design Assessment (UK)
GDC General Design Criteria (US)
HEM Homogeneous Equilibrium Model
HPECC High-Pressure ECC
HSE Health and Safety Executive (UK)
HTS Heat-Transport System (see also RCS)
HWP Heavy Water Plant
HYDNA Hydraulic Network Analysis (extinct computer code)
I&C Instrumentation and Control
IBIF Intermittent Buoyancy Induced Flow
ICRP International Commission on Radiological Protection
IIP Integrated Implementation Plan
INES International Nuclear Event Scale
INSAG International Nuclear Safety Advisory Group (IAEA)
IRRS Integrated Regulatory Review Service (IAEA)
LBLOCA Large Break Loss of Coolant Accident
LGR Liquid Graphite Reactor
LLOCA Large Loss-of-Coolant Accident
LMFBR Liquid Metal Fast Breeder Reactor
LNT Linear No Threshold (dose response hypothesis)
LOC Loss of Coolant
LOC/LOECC Loss of Coolant with Coincident Loss of Emergency Core Cooling
LOCA Loss of Coolant Accident
Abbreviations 3
LOP Loss of Pumping

LOR Loss of Regulation
LPECC Low-Pressure ECC
LWR Light-Water Reactor
MBA Material Balance Area
MCCR Ministry of Consumer and Commercial Relations
MCR Main Control Room
MCS Maintenance Cooling System
MDEP Multinational Design Evaluation Programme
milli-k Unit of reactivity for reactor physics
MPECC Medium-Pressure ECC
MPMO Major Projects Management Office
MSSV Main Steam Safety Valve
MUF Material Unaccounted For
MWe Megawatt electric
NDPB Non Departmental Public Body (UK)
NISA Nuclear and Industrial Safety Agency (Japan)
NPD Nuclear Power Demonstration
NPSH Net Positive Suction Head
NRA Nuclear Regulation Authority (Japan)
NRU National Research Universal
NRU National Research Universal (reactor)
NRX National Research Experimental Research Reactor
NUCIRC Nuclear Circuits (computer code)
OECD Organization for Economic Co-operation & Development
OH Ontario Hydro (now OPG)
ONR Office for Nuclear Regulation (UK)
OPG Ontario Power Generation
PNGSA Pickering Nuclear Generating Station A
PHTS Primary Heat Transport System
PHW Pressurized Heavy Water
PHWR Pressurized Heavy Water Reactor (CANDU)
PRESCON2 Pressure Containment (computer code)
PRV Pressure-Relief Valve
PSR Periodic Safety Review
PWR Pressurized-Water Reactor
QA Quality Assurance
R&D Research and Development
R&M Reliability and Maintainability
RAMA Reactor Analysis Implicit Algorithm
RB Reactor Building
RBMK Реактор Большой Мощности Канальный (High Power Channel-Type Reactor)
RCS Reactor Coolant System (see also HTS)
©UNENE, all rights reserved. For educational use only, no assumed liability. Acronyms – September 2014
rem röentgen equivalent man (radiation dose)

RIDM Risk-Informed Decision Making
RIH Reactor Inlet Header
ROH Reactor Outlet Header
ROP Regional Overpower (Protection System)
RSAC Reactor Safety Advisory Committee (Canada, historical)
RTD Resistance Temperature Detectors
S/C Suppression Chamber
SAM Severe Accident Management
SAPs Safety Assessment Principles (UK)
SCA Secondary Control Area
SCDF Severe Core-damage Frequency
SDM Safety Design Matrix
SDS Shutdown System
SER State Evaluation Report
SFAIRP So Far As Is Reasonably Practicable (UK)
SG Steam Generator
SGHWR Steam Generating Heavy Water Reactor (United Kingdom)
SIR Safeguards Implementation Report
SIS Systems Important to Safety
SOE Safe Operating Envelope
SOPHT Simulation of Primary Heat Transport (computer code)
SOR Shut-Off Rod
SRP Staff Review Procedure (Canada)
SRP Standard Review Plan (US)
SRV Safety Relief Valve
SSC Systems, Structures and Components
STUK Säteilyturvakeskus (Radiation and Nuclear Safety Authority, Finland)
TAG Technical Assessment Guide (UK)
TUEC Total Unit Energy Cost
TOFFEA Two Fluid Flow Equation Analysis (computer code)
TOR Tolerability of Risk (UK)
TSO Technical Support Organization
UVUEUP Unequal Velocity, Unequal Energy, Unequal Pressure (thermalhydraulic model)
UVUT Unequal Velocity Unequal Temperature (thermalhydraulic model)
VB Vacuum Building
VC Vacuum Chamber
WENRA Western European Nuclear Regulators’ Association
WRE Whiteshell Research Establishment
ZEEP Zero Energy Experimental Pile

The Essential Can Du

Uploaded by

Copyright:

Available Formats

The Essential Can Du

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Essential Can Du

Uploaded by

Copyright:

Available Formats

THE

CANDU (short for CANada Deuterium Uranium)

Editor-in-Chief: Wm. J. Garland, Professor Emeritus, McMaster University, Hamilton, Ontario

Library of Congress Cataloging in Publication data:

Includes bibliographical references.

This textbook is dedicated to the students of CANDU everywhere.

1 CANDU (short for CANada Deuterium Uranium)

• George Bereznai (University of Ontario Institute of Technology)

Ontario Power Generation (OPG)

2 First some basics

(called thermal neutrons) can initiate a

of uranium that occurs in nature. This is thermal neutron

that is mined from the ground is 0.7% U-

gamma radiation, fast (or energetic)

2.2 The fuel: the source of energy by the fission process

2.3 The moderator: slowing down those speedy neutrons

Coolant Contol Coolant

Fuel Rods Moderator Moderator Fuel Rods

Figure 2 Nuclear reactor components

2.4 The coolant: to take away the heat generated by fissioning

Figure 3 A Conceptual CANDU

Figure 4 Basic power reactor schematic

2.5 Control: staying within desired and safe limits of power

2.6 Shielding: providing protection from radiation

2.7 The system that pulls it all together

Figure 5 The overall CANDU nuclear plant

• The Virtual Nuclear Tourist at http://www.virtualnucleartourist.com/ - a very popular

Figure 3 Fission and absorption characteristics of uranium ........................................................... 5

1 Nuclear Reactor Principles

Figure 1 Range of stable and unstable nuclides

Figure 2 Binding energy per nucleon

1.2 Nuclear Reactor Principles

1/v Resonance Smooth

Figure 3 Fission and absorption characteristics of uranium

Figure 4 Nuclear reactor components

PWR Pressurized Water Reactor

2.2 Commercial Reactors

Figure 5 Diagrammatic cross section of a typical BWR

2.2.2 Pressurized water reactor (PWR)

Figure 6 Diagrammatic cross section of a typical PWR

2.2.3 Pressurized heavy water reactor (PHWR) (CANDU)

of a pressurized heavy water reactor is given in Figure 7.

Figure 7 Diagrammatic cross section of a typical PHWR (CANDU)

2.2.4 Gas cooled reactor (GCR)

Figure 8 Diagrammatic cross section of a typical AGR

Figure 9 Diagrammatic cross section of a typical RBMK (LGR)

2.2.6 Liquid metal fast breeder reactor (LMFBR)

3 World-Wide Commercial Power Reactors

3.2 Commercial Reactors in Service

Table 1 Nuclear reactors in service and planned*

BWR–Boiling (Light) Water Reactor

Argentina PHWR 2 1 935 692

Country Reactor Type Number Operational + Existing Capacity +

* Adapted from Nuclear News, March 2012

ers around the world.

3.3 Recent Past and Near Future

Reactor Type Number in % Total Capacity %

PWR 267 61 246 555 67

TOTAL 435 100 369 997 100

*Adapted from Nuclear News, March 2012.