IAEA-TECDOC-1569

Liquid Metal Cooled Reactors:

Experience in Design and
Operation

December 2007
IAEA-TECDOC-1569

Liquid Metal Cooled Reactors:

Experience in Design and
Operation

December 2007
 The originating Sections of this publication in the IAEA were:
INIS and Nuclear Knowledge Management and
Nuclear Power Technology Development Sections
International Atomic Energy Agency
Wagramer Strasse 5
P.O. Box 100
A-1400 Vienna, Austria

LIQUID METAL COOLED REACTORS: EXPERIENCE IN DESIGN AND OPERATION

IAEA, VIENNA, 2007
IAEA-TECDOC-1569
ISBN 9789201079077
ISSN 10114289
IAEA, 2007
Printed by the IAEA in Austria
December 2007
 FOREWORD

In 2002, within the framework of the Department of Nuclear Energys Technical Working
Group on Fast Reactors (TWG-FR), and according to the expressed needs of the TWG-FR
Member States to maintain and increase the present knowledge and expertise in fast reactor
science and technology, the IAEA established its initiative seeking to establish a
comprehensive, international inventory of fast reactor data and knowledge. More generally, at
the IAEA meeting of senior officials convened to address issues of nuclear knowledge
management underlying the safe and economic use of nuclear science and technology
(Vienna, 1719 June 2002), there was widespread agreement that, for sustainability reasons
for fissile sources and waste management, long-term development of nuclear power as a part
of the worlds future energy mix will require the fast reactor technology. Furthermore, given
the decline in fast reactor development projects, data retrieval and knowledge preservation
efforts in this area are of particular importance. This consensus concluded from the
recognition of immediate need gave support to the IAEA initiative for fast reactor data and
knowledge presevation.

To implement the IAEA initiative, the scope of fast reactor knowledge preservation activities
and a road map for implementation have been developed. The IAEA supports and coordinates
data retrieval and interpretation efforts in the Member States joining the initiative and ensures
the collaboration with other international organizations (mainly OECD/NEA) and eventually
establishes and maintains a portal for accessing the fast reactor knowledge base.

The IAEA assists Member State activities by providing an umbrella for information exchange
and collaborative R&D to pool resources and expertise within the framework of the TWG-FR
and the Agencys International Nuclear Information System (INIS) and Nuclear Knowledge
Management Section (NKMS). The IAEA collects and summarizes the scientific and
technical information on key fast reactor technology aspects in an integrative sense useful to
engineers, scientists, managers, university students and professors.

This publication has been prepared to contribute toward the IAEA activity to preserve the
knowledge gained in the liquid metal cooled fast reactor (LMFR) technology development.
This technology development and experience include aspects addressing not only
experimental and demonstration reactors, but also all activities from reactor construction to
decommissioning. This publication provides a survey of worldwide experience gained over
the past five decades in LMFR development, design, operation and decommissioning, which
has been accumulated through the IAEA programmes carried out within the framework of the
TWG-FR and the Agencys INIS and NKMS.

The IAEA appreciate the advice and support of the IAEAs TWG-FR members in the
preparation of the publication. The draft of the report, compiled by A. Rineiskii (consultant),
in cooperation with W. Mandl, A. Badulescu and Y.I. Kim of the IAEA, has been reviewed by
the TWG-FR Members of China, France, India, Japan, Republic of Korea and the Russian
Federation. The work was guided by the IAEA officers A. Stanculescu and Y. Yanev and they
are responsible for this publication.
 EDITORIAL NOTE

The use of particular designations of countries or territories does not imply any judgement by the
publisher, the IAEA, as to the legal status of such countries or territories, of their authorities and
institutions or of the delimitation of their boundaries.
The mention of names of specific companies or products (whether or not indicated as registered) does
not imply any intention to infringe proprietary rights, nor should it be construed as an endorsement
or recommendation on the part of the IAEA.
 CONTENTS

INTRODUCTION...................................................................................................................... 1

1. LIQUID METAL COOLED FAST REACTOR DEVELOPMENT

AND IAEA ACTIVITIES ................................................................................................ 2
1.1. Activities on fast reactors design and technology ................................................... 2
1.2. The joint research activities on LMFR.................................................................... 6
1.2.1. Sodium void effect ....................................................................................... 6
1.2.2. Intercomparison of LMFR seismic analysis codes:
comparison of experimental results with computer prediction ............ 9
1.2.3. Sodium mixing problems ........................................................................... 12
1.2.4. Core structural and fuel materials assuring high fuel burnup .................... 17
1.2.5. Concept of fuel resources and waste management .................................... 18
1.2.6. Core disruptive accident ............................................................................ 19
1.2.7. Acoustic signal processing for the detection of sodium boiling in
reactor cores or leak detection and location in steam generators ....... 20
1.3. Activities on advanced fast LMFR and technology .............................................. 20
1.4. Safety research issues for advanced LMFR development..................................... 25
1.4.1. Progress made since the past...................................................................... 26
References ...................................................................................................................... 27

2. PROTOTYPE FAST REACTOR................................................................................... 29

2.1. Design features and review of operating history................................................... 29
2.2. Review of the PFR steam generator design concept and operating history .......... 32
2.2.1. PFR steam generator: choice of the design concept .................................. 34
2.2.2. Review of steam generators operating history........................................... 37
2.2.3. The under-sodium leak in PFR superheater............................................... 40
2.3. Effects of sodium aerosol deposition in LMFRs................................................... 43
2.4. The primary circuit oil spill................................................................................... 46
2.5. Cracks in the PFR air heat exchangers .................................................................. 50
2.6. Sodium mixing problems and flow and mechanically induced vibrations ........... 52
2.7. The effect of neutron-induced distortion of core components .............................. 52
2.8. Fuel development .................................................................................................. 55
Bibliography ................................................................................................................... 56

3. PHENIX AND SUPER-PHENIX REACTORS............................................................. 57

3.1. Phnix reactor........................................................................................................ 57
3.1.1. Commissioning and design features .......................................................... 57
3.1.2. Intermediate heat exchangers operating experience .................................. 64
3.1.3. Steam generator operating experience ....................................................... 66
3.1.4. Secondary circuit operating experience ..................................................... 72
3.1.5. Pumps operating experience ...................................................................... 73
3.1.6. Negative reactivity shutdowns ................................................................... 74
3.1.7. Sodium aerosol deposit .............................................................................. 75
3.1.8. Sodium circuits, reactor and equipment inspection and renovation .......... 76
3.1.9. Protection for fires ..................................................................................... 79
 3.1.10. Plant statistics and conclusions .................................................................. 80
3.2. Super-Phnix reactor ............................................................................................. 83
3.2.1. Design features and commissioning history .............................................. 83
3.2.2. SPX steam generator: design and operating experience ............................ 88
3.2.3. Storage drum operating experience............................................................ 91
3.2.4. Primary sodium contamination .................................................................. 96
3.2.5. Argon leak from the sealing bell of the intermediate heat exchangers ...... 98
Bibliography ................................................................................................................. 100

4. BN-600 AND BN-350 REACTORS ............................................................................ 102

4.1. BN-600 pool-type LMFR.................................................................................... 102
4.1.1. Commissioning and design features ........................................................ 102
4.1.2. Equipment tests ....................................................................................... 108
4.1.3. Operating experience .............................................................................. 112
4.1.4. Reactor core: design and operation.......................................................... 121
4.2. BN-350 loop type reactor plant ........................................................................... 125
4.2.1. Design features......................................................................................... 125
4.2.2. General results of operation ..................................................................... 129
4.2.3. Steam generators and intermediate heat exchangers:
operating experience......................................................................... 130
4.2.4. Reactor core and refuelling equipment:
design and operating experience ...................................................... 133
4.2.5. BN-350 electromagnetic pumps............................................................... 140
4.3. Pool vs. loop type design: some knowledge gained from
constructions and operations...................................................................... 141
4.3.1. BN-600..................................................................................................... 143
4.3.2. Phnix....................................................................................................... 145
4.3.3. PFR .......................................................................................................... 146
Bibliography ................................................................................................................. 147

5. TECHNOLOGIES FOR SODIUM AND REACTOR COMPONENTS

MANAGEMENT AFTER SHUT DOWN OF THE BN-350 TYPE LMFR ............... 149
5.1. Introduction and summary................................................................................... 149
5.2. Pre-shut down radiological characteristics of the BN-350
coolant and equipment ............................................................................... 152
5.2.1. Primary sodium activity ........................................................................... 153
5.2.2. Activated equipment of the primary circuit in the area adjacent to
the reactor vessel cavity.................................................................... 153
5.2.3. Activity of the radionuclides in the primary cold traps ........................... 156
5.2.4. Activity of cooling system of reactor vessel cavity and
the primary circuit cells .................................................................... 156
5.2.5. Activity of the equipment of the secondary circuit.................................. 157
5.3. Purification and utilization of spent sodium coolant of primary and
secondary circuits....................................................................................... 157
5.3.1. Requirements to the coolant quality for its reuse in newly
constructed LMFR............................................................................ 158
5.3.2. Requirement for quality of the coolant to be used in industry................. 162
 5.3.3. Cleaning of the BN-350 spent sodium coolant from impurities .............. 163
5.4. Reprocessing of the radioactive sodium wastes into
safe condition for disposal ......................................................................... 165
5.4.1. Burning in air ........................................................................................... 166
5.4.2. Reprocessing using steam and water ....................................................... 166
5.4.3. Gas-phase oxidation of sodium................................................................ 167
5.4.4. Solid state sodium oxidation .................................................................... 167
5.4.5. Reprocessing of radioactive alkaline wastes using liquid state
oxidation method.............................................................................. 167
5.5. Washing the residual sodium deposits and decontamination of reactor circuits,
equipment and structural materials ............................................................ 168
5.5.1. Input data for the choice of methods........................................................ 169
5.5.2. Methods for removal of coolant from undrainable parts of reactor and
equipment ......................................................................................... 169
5.5.3. Methods for washing and decontamination of circuits and equipment ... 171
5.5.4. Technology for washing and decontamination of
reactor circuits and equipment.......................................................... 172
5.6. Radiation conditions in the primary circuit cells, activity of reactor
block elements ........................................................................................... 176
5.6.1. Radiological conditions of dismantling works in the primary
circuit cells........................................................................................ 176
5.6.2. Radiological characteristics of activated in-vessel structures, reactor
vessel and reactor cavity................................................................... 179
References .................................................................................................................... 191

6. LEAD-BISMUTH COOLED (LBC) SHIP REACTOR .............................................. 194

6.1. Introduction ......................................................................................................... 194
6.2. Operating experience with LBC ship reactor plants ........................................... 194
6.2.1. The reactor core meltdown ...................................................................... 194
6.2.2. Radioactive coolant leaks caused by damage to the primary pipelines ... 195
6.2.3. Radioactive coolant leaks owing to corrosion damage of the
steam generator................................................................................. 196
6.3. R&D on LBC reactor technology........................................................................ 198
6.4. Conclusions ......................................................................................................... 200
References .................................................................................................................... 201

7. LMFR CHARACTERISTICS AND THEIR DEMONSTRATION

DURING THE FINAL STAGE OF OPERATION...................................................... 203
7.1. Passive characteristics with regard to reactor shutdown ..................................... 203
7.2. Passive characteristics with regard to emergency heat removal ......................... 205
7.3. Demonstration of safety characteristics with experimental LMFR..................... 210
7.3.1. Rapsodie reactor....................................................................................... 210
7.3.2. EBR-II reactor.......................................................................................... 219
7.3.3. The FFTF reactor ..................................................................................... 224
Bibliography ................................................................................................................. 230
8. DOUNREAY FAST REACTOR: DESIGN FEATURES,
EXPERIMENTS DURING THE FINAL STAGE OF OPERATION ......................... 231
8.1. Design features and main operaton results.......................................................... 231
8.2. In-core-coolant boiling experiments ................................................................... 235
8.2.1. Introduction.............................................................................................. 235
8.2.2. Objectives of the experiment ................................................................... 235
8.2.3. Design of the instrumented vehicles ........................................................ 236
8.2.4. Operation.................................................................................................. 240
8.2.5. Analysis of the results and conclusions ................................................... 241
References .................................................................................................................... 246

9. SODIUM COOLED FAST REACTORS DEVELOPMENT AND

OPERATING EXPERIENCE: LESSONS LEARNED IN THE
PAST AND CHALLENGES FOR THE FUTURE...................................................... 247
9.1. Mastering of sodium coolant technology ............................................................ 247
9.2. Demonstration commercial fast reactors operating experience:
achievements and setbacks ........................................................................ 250
9.2.1. BN-600..................................................................................................... 250
9.2.2. Super-Phnix (SPX)................................................................................. 251
9.3. Challenges for the future ..................................................................................... 252
9.3.1. Increase of safety margins........................................................................ 252
9.3.2. Simplification of NSSS structure and operation
and maintenance (O&M).................................................................. 255
9.3.3. Steels stabilized by titanium .................................................................... 257
9.3.4. Steam generators: principal design & structure materials ....................... 258
References .................................................................................................................... 259

ABBREVIATIONS................................................................................................................ 261

CONTRIBUTORS TO DRAFTING AND REVIEW ........................................................... 263

 INTRODUCTION

Fast reactors have been under development for about 50 years, and several countries have
important fast breeder reactor development programs. Ten test fast reactors with thermal
power ranging from 8 to 400 MW(th) and six commercial size prototypes with electrical
output ranging from 250 to 1 200 MW(e) have been constructed and operated. Two reactors
are currently being operated with a thermal efficiency of 4345%, that is the highest value in
the nuclear power practice. Great strides have been made in fast reactor technology in the
past, which encourage future development. The closed fuel cycle has been demonstrated and
an effective breeding ratio was experimentally confirmed. In total, fast reactors have gained
approximately 300 reactor-years of operation. Fuel burnup in excess of 130 000 MWd/t has
been reached at several reactors, and major steps towards commercial fast reactor designs
have been made.

The worldwide investment already made in the development and demonstration of the unique
liquid metal fast breeder reactor (LMFR) technology exceeds US$ 50 billion [4, 5]. Research
on LMFRs during the last decades has significantly improved our understanding of LMFR
safety. The achievement of the past safety research has been effectively used to develop a
system of safety analysis methods which were used to evaluate the safety characteristics of
the existing and advanced fast reactors. It is predicted that liquid metal cooled fast breeder
reactors currently being designed can achieve a very high degree of safety. A well-validated
way forward to commercial utilization of fast reactors has been established. This way is
generally consistent with other studies, and indicates that the goal of competitive fast rectors
may be within reach.

From the 1980s onward, and mostly for economical and political reasons, fast reactor
development in general began to decline, particularly in countries characterised by advanced
market economies, relatively slow growth in primary energy consumption and good
availability of oil and gas. This has resulted in the premature shutdown of some test and
prototype fast reactors and the slowing down or closing of reactor development programs, and
the retirement of many of the key developers.

The report documents the knowledge in LMFR design and technology to preserve and
disseminate it until the large economic need for fast reactors becomes apparent. The
information presented will serve for saving funding required for R&D work in the future
owing to preserved data, and will help countries, which plan to embark on their own fast
reactor development programs (e.g. some South and East Asian countries). The report intends
to provide the public with unbiased information on fast reactor development.

Section 1 presents the state of the art of LMFRs development and IAEA activities. Sections
24 provide design features and operating experience of demonstration and commercial size
nuclear power plants with sodium-cooled fast breeder reactor. Section 5 provides information
on activation characteristics of the primary coolant, reactor and components; treatment and
disposal of the spent sodium; removal of the residual sodium deposits and decontamination
after shut down of the BN-350 reactor a typical loop type LMFR. Section 6 provides an
overview of lead-bismuth cooled ship reactor operating experience and heavy metal
technology development. Sections 7 and 8 provide information on passive principles of fast
reactor emergency shutdown and heat removal, and on demonstration of safety characteristics
through tests performed in experimental fast reactors during their final stages of operation.
Section 9 provides an analysis and assessment of the advantages and disadvantages of sodium
as coolant, giving due consideration to the advances in the technology and design of sodium
components.

1
1. LIQUID METAL COOLED FAST REACTOR DEVELOPMENT AND IAEA
ACTIVITIES

1.1. ACTIVITIES ON FAST REACTORS DESIGN AND TECHNOLOGY

Studies and design work on fast reactors have been carried out for about 50 years. Nuclear
experimental electricity was first generated on 20 December 1951 by 0.2 W(e)
EBR-I LMFR at Argonne National Laboratory (ANL) in the USA. The former USSR fast
reactor BR-5 was the first reactor in the world using sodium as a coolant and plutonium oxide
as a fuel [1].
Sodium cooled fast reactors have been operated successfully in several countries to produce
energy and the sea water desalination, and to demonstrate the fuel breeding. The first in the
world demonstration nuclear power plant with fast reactor, BN-350 of 750 W(th) power
was commissioned on 16 July 1973 and has been operated for over 25 years (its design life
time), providing 100 000 tons/day fresh water and 150 W electricity for a large city Aktau
(Kazakhstan) and the adjoining industrial region in the desert. The first fuel assemblies
containing breeding plutonium recycled into new mixed oxide (MOX) fuel were loaded in the
French reactor Phnix and in the UK prototype fast reactor (PFR) cores in 1980 and in 1982,
respectively, thereby closing the reactor's fuel cycle. The French fast reactor Super-Phnix of
1 200 W(e) power, commissioned in May 1986) demonstrated the first large size
LMFR [2].
Energy production with fuel breeding is the main goal of fast reactor (FR) development to
ensure long-term fuel supply. Fast reactors are also being investigated to reduce the actinide
content of nuclear waste, and to take advantage of their high thermal efficiency.

The 65 W(th)/25 W(e) China experimental fast reactor (CEFR) is under construction.
This is the first step in the Chinese fast reactor engineering development. Ninety percent of
the concrete constructions, including the main building, have been completed:
hundreds components have been installed in the building. First criticality is foreseen in 2008.
As a second step in the Chinese fast reactor technology development effort, a 600 MW(e)
China prototype fast reactor (CPFR) is presently under consideration. The role of minor
actinide transmutation is also being evaluated taking as reference for the CPFR [3].

The French liquid metal reactor technology has demonstrated a number of positive examples
of designs, project realizations and experience in LMFRs construction and operation:
experimental reactor Rapsodie (40 W(th) power, 1967-1983), prototype reactor Phnix
(255 W(e) power, commissioned in December 1973), the large size LMFR Super-Phnix
(1986-1998). The plutonium produced in the Phnix was used as the fuel for its core.
The cumulative amounts of FR fuel reprocessed in France are about 30 tons. A breeding ratio
1.16 was experimentally confirmed in the Phnix. The Phnix reactor has been operated for
~ 100 000 hours at a temperature of 560C of the reactor hot structures with thermal
efficiency of 45.3% (gross), that is the highest value in the nuclear power practice; an average
burnup was increased from 50 000 MWd/t to 100 000 MWd/t, with maximum burnup
exceeding 150 000 MWd/t. These levels were reached with 8 cores of fuel which was
166 000 fuel pins.

At Phnix, after completion of the plant renovation program, power operation was resumed on
15 June 2003. The role of Phnix as an irradiation facility was further strengthened to
compensate for the premature shutdown of Super-Phnix in 1998, particularly in support of
the CEA R&D program in the context of line 1 of the December 30 1991 law on long-lived

2
radioactive waste management (CAPRA-CADRA programme). The first experiment, called
SUPERFACT, led to the incineration of minor actinides (neptunium and americium) was
carried out. Since 2003, the reactor Phnix power has been limited to 350 MW(th),
145 MW(e) on two primary/secondary loop operations. The reactor will be operated for
6 irradiation cycles of 120 EFPD each. As regards R&D, CEA has launched a comprehensive
R&D program to study alternative technologies for future nuclear energy systems including
based on a gas cooled FR with on-site closed fuel cycle. Many of the long-term options
investigated are believed to be of generic interest, and to offer the chance of developing high
performance fuels and materials. Decommissioning work that started in 1999 is underway at
the Super-Phnix. The last fuel subassembly has been unloaded on 19 March 2003.

With France in the leading role, the European fast reactor (EFR) design has been completed.
This is synthesis of the extensive experiences from France, Germany and the United Kingdom
with large pool-type oxide-fuelled reactors. One of the outstanding achievements of the
EFR program has been to make firm and reliable cost estimates. The construction of a reactor
to the EFR design may not be possible in the near future, but a well-validated way forward to
commercial utilization of fast reactors has been established. This way is generally consistent
with other studies, and indicates that the goal of competitive fast rectors may be within reach.

The development of LMFR in Europe has been delayed. However, alternative fast reactor
applications are being developed in France, Germany, the UK and other countries, namely the
transmutation of long-lived nuclear waste and the utilization of surplus plutonium [1-5].

The 500 W(e) power prototype fast breeder reactor (PFBR) is under construction in India.
Manufacturing technology development of the key plant components was completed.
Development of technology of low doubling time fuels and structural materials capable of
sustaining high neutron fluence has already been initiated and work is going on satisfactorily.
The Indian fast reactor development program is built based on the experience accumulated
with the small-size [40 W(th)/13 W(e)] fast breeder test reactor (FBTR) located at
Kalpakkam, which is operational since 1985. Important works including PFBR shielding
experiments, testing of transfer arm in air, boron enrichment, post-irradiation examination of
FBTR fuel after 125 GWd/t burnup, structural integrity testing, and reprocessing of carbide
fuel are being carried out [3].

The potential of natural uranium resources in India, estimated to be around 50 000 tons, is
negligible [about 1 billion tons of coal equivalent (btce)] if utilized in an once-through cycle
and the capacity will also be limited to about 10 GWe. If the same uranium along with the
plutonium generation in PHWR is invested in FBR, the resource potential can be enhanced to
180 btce and the capacity also can be increased to 250 GWe. FBRs thus forming the second
stage of the nuclear power programme of the country. FBRs in India will be deployed on
U-Pu cycle for rapid growth of nuclear power capacity and to generate enough nuclear fuel
simultaneously for deployment of Th-U cycle in the third stage of the program. Thorium is
abundantly available in India and the resource level of 320 000 tons is estimated to be
equivalent of 1000 btce. Therefore one of the major alternatives for India is nuclear energy.

The fast reactor development program is a key part of the Japanese policy for greater energy
independence. The feasibility studies on commercialised fast reactors cycle systems are in
progress. The Japanese R&D are being focused on the design of the candidate concepts and
on fundamental tests of key technologies [3]. The prototype reactor MONJU of 280 W(e)
power was stopped temporarily due to a leak in the non-radioactive secondary heat transport
system, that occurred in December 1995 during the 40% power pre-operational testing phase.
In the Japanese program, it was clarified that MONJU is at the core of the fast reactor

3
research activities, and steps are taken to resume its operation as soon as possible.
Considerable effort has been put into activities aiming at regaining public understanding and
acceptance. The local governor of Fukui made pre-consent for plant modification work of
Monju on 7 February 2005. And the Japan Nuclear Cycle Development Institute (JNC) started
the preparatory work and the plant modification work on 3 March and 1 September 2005,
respectively. In Japan the experimental fast reactor Joyo has shown excellent performance for
more than 26 years. The pre-service inspection for the JOYO MK-III upgrade by the relevant
Ministry was completed on 27 November 2003. The JOYO reactor has successfully
completed and tested the plant and core modifications for the MK-III upgrade program, and
rated power operation was started in 2004. The upgraded MK-III core provides a significantly
enhanced irradiation testing capability compared to the MK-II core. Initial critically of the
MK-III core was achieved on 2 July 2003, which was followed by the successful operational
demonstration up to the rated thermal power of 140 MW. Functional and performance testing
verified the design parameters. The utilization plan for future fuels and materials
developments and safety testing in the JOYO MK-III core has been developed [2, 3].

In the Republic of Kazakhstan, the fast breeder reactor BN-350 at Aktau commissioned in
November 1972, was finally shut down in April 1999. The general plan for the BN-350
decommissioning was developed in close cooperation with Russian, US, and UK specialists
within the framework of a relevant project. Operations on primary sodium drainage (510 tons,
~ 19 hours) into special tanks have been completed in December 2003. The project EAGLE is
under way since 2000 under a contract between the National Nuclear Centre of Kazakhstan
and JNC. The project comprises the preparation and conduct of out-of-pile and in-pile
experiments designed to address the key safety issues relevant to eliminating or mitigating the
re-criticality potential during a postulated core-disruptive accident in future commercial
sodium cooled FRs [3].

Republic of Korea's LMFR program consists in the development of basic design technologies.
During Phases 1&2 (1997-2001) of the program, basic technologies and the conceptual design
of KALIMER-150 of 150 MW(e) capacity has been developed. Basic key technologies and
the advanced concept KALIMER-600 with a capacity of 600 MW(e) is being developed
during Phase 3 from the year 2002 to 2004. During this phase, efforts were concentrated
on the development of basic key technologies and on the establishment of advanced concepts
with emphasis on proliferation-resistant core design, and on the enhancement of economics
and safety. In 2003, the preliminary KALIMER-600 design concept was established and
several experiments were performed for the validation of computer codes and models. A core
without blankets, but maintaining the capability of transuranics (TRU) self-recycling is being
developed. From the evaluation of the reactor vessel size and decay heat removal capacity
(based on the natural circulation of sodium and air) was retained as the most favourable
passive design concept. Also in 2003, a conceptual study on ultra-long life cores with power
densities higher than the conventional ones has been performed [3, 6].

Russian experience gained on operation of test, prototype and semi-commercial LMFRs

(BR-5/10, BOR-60, BN-350 and BN-600) is rather good. The first fast reactor with sodium
coolant and plutonium fuel (BR-5/10) was created in 1958 and was in operation for over
44 years. The BN-600 reactor was connected to the grid in April 1980 and full power was
reached in October 1981. Reactor operation is stable: its load factor is 75-77%, turbine
efficiency being ~ 43%. Up to end of 2004, the total on-power operation of BN-600 reactor
plant time amounted to ~ 170 000 h, and ~ 91 billion kWh of electricity was generated.
Current efforts with regard to the fast reactor development in Russian Federation are directed
towards increasing safety margins and improving economics. The detailed design of

4
commercial size fast reactor BN-800 [800 W(e)] was completed, and the license was issued
for its construction. According to the revised Programme for nuclear power development in
the Russian Federation for the period 1998-2005, and for the period until 2010, the start-up
of the 4th power unit equipped with a BN-800 reactor at the Beloyarskaya NPP (BN-600) site
is scheduled for 2010. The next important step in the area of fast reactors in Russian
Federation is the development of design proposals for a power plant with a sodium cooled fast
reactor BN-1800 [1800 W(e)]. Further activities in the FR area in Russian Federation
include:

(1) Justification of life extensions for BN-600;

(2) Justification of a hybrid, as well as full MOX core design for BN-600 to incinerate
weapons-grade plutonium;
(3) Design of the BOR-60 experimental reactor modification, including its replacement by
the sodium cooled BOR-60M plant;
(4) Development of advanced and innovative FR designs with enhanced safety with
mixed oxide fuel, two circuit reactor concepts with sodium reactors coolant and
gas-turbine cycle [both large size reactors (1200-1600 MW(e), and small/medium size
modular and transportable co-generation reactor ATES-BN-300/100, 300 MW(e) plus
100 MW(th) for district heating]; and
(5) Design studies of FR with alternative coolants: supercritical water, lead
(BREST-OD-300 and BREST-1200), and lead-bismuth eutectic (SVBR-75/100)
[2, 3, 7].

Until 1967, the major problem of damage to cladding materials was embrittlement. However,
in 1967 the evidence of considerable void swelling taking place in austenitic stainless steels
irradiated to high fast neutron fluences, was firstly detected in the UK Dounreay experimental
fast reactor DFR (1962-1977). This phenomenon has since then tended to dominate the
attention in the development of cladding and duct alloys. A high nickel alloy was developed in
the UK as reference cladding material. In the 250 W(e) prototype fast reactor (PFR),
commissioned in January 1975, large numbers of MOX fuel pins reached high burnup with an
irradiation dose in excess of 130 displacements per atom (dpa). These results have been
confirmed and surpassed by irradiation in Phnix to more than 160 dpa [3]. The first fuel
assemblies containing own plutonium recycled into new MOX fuel were loaded in the PFR
reactor core in June 1982, thereby closing the reactors fuel cycle. In March 1994, the
Dounrey reprocessing plant had treated a total of over 23 tons of MOX spent fuel with the
highest burnup in the fuel about 18%, providing the technical feasibility of MOX fuel
reprocessing via a Purex-cycle, with recovery of over 99.5% of the plutonium. This high
recovery was also reflected in the low amounts of plutonium in the liquid and solid waste
streams from the plant. The amount of radioactivity discharged to the environment was
always about an order of magnitude less than the licensed limits. The plant is subject to IAEA
and Euratom safeguards.

In the UK, at present there is no Government sponsored FR R&D programme except for the
UKAEA PFR decommissioning work at Dounreay. However, a BNFL-funded FR R&D
programme involving BNFL/NNC/SERCO Assurance (formerly AEA Technology) is
pursued in the CAPRA-CADRA collaboration with France, Germany, and Belgium, covering
reactor systems for plutonium and minor actinides burning sodium cooled FRs. The UK
contribution is focused on key skills areas: nuclear methodology and core design, fuel
performance and fuel cycle modelling. Recently, emphasis is shifting to gas cooled FRs and
to accelerator driven system (ADS) analyses. The UK is contributing to core design, thermal
hydraulics design, and fuel design and performance of various FR and ADS concepts [2, 3, 8].

5
In the USA, the sodium-cooled integral fast reactor (IFR) project has been developed by the
Argonne National Laboratory (ANL) based on the use of a ternary U-Pu-Zr fuel alloy for its
core loading. The IFR concept was integrated by General Electric into a full plant design of a
300 W(e) advanced liquid metal cooled reactor (ALMR). The plutonium is not separated
from higher actinides; these are recycled together in the reactor and never leave the reactor
site. All IFR designs are based on full actinide recycling using a pyrochemical processing and
fuel fabrication plant co-located with the reactor complex. In fact, by using a U-Pu-10%Zr
alloy and ferritic-martensitic HT9 cladding as duct, a burnup of about 20% has been achieved
in EBR-II. All irradiation results achieved in the EBR-II and the Fast Flux Test Facility
(FFTF) reactor have demonstrated reliable performance of metallic fuel and the potential to
achieve high burnup in prototypical fuel elements cooled by liquid sodium. These ideas were
borrowed for some other reactor concepts. The ALMR was designed to provide high
reliability for the key safety functions, including reactor shutdown, heat removal, and
containment. These functions can be achieved by passive means; thermal expansion,
temperature effects on neutron absorption, natural circulation of sodium coolant, and natural
air circulation [8]. However, in the U.S. from the 1980s onward, and mostly for economical
and political reasons the fast reactor development began to decline. By 1994, the Clinch River
breeder reactor (CRBR) project had been cancelled, and the two fast reactor test facilities, the
FFTF and EBR-II had been shutdown.

In the U.S, there have been substantial activities in continuing the development of
technologies related to advanced nuclear energy systems, including advanced fuel cycles, fast
reactors and transmutation. The two main programs that cover the relevant activities in
advanced fuel cycles and fast reactor and transmutation are the advanced fuel cycle initiative
(AFCI), and the Generation IV nuclear energy systems program [2, 3, 9].

1.2. THE JOINT RESEARCH ACTIVITIES ON LMFR

The IAEA has a long standing program to foster international information exchange and
cooperative research and development in the field of LMFRs. This program has been carried
out during the past 35 years under the expert guidance of a Technical Working Group on Fast
Reactors (TWG-FR), formerly International Working Group on Fast Reactors (IWG-FR).
The TWG-FR is the only global forum for the review and discussion of LMFR programmes,
and is composed of leading specialists in national programs for LMFR development.
The TWG-FR was established in order to suggest to the Agency R&D activities that meet
Member States needs and to serve as an international forum for exchanging information and
performing collaborative research on fast reactor development. The TWG-FR deals with all
aspects of fast reactor technology, including fuel, coolant and components. Some important
results achieved through activities carried out within the framework of the TWG-FR are
summarized below.

1.2.1. Sodium void effect

In the event of coolant loss (by boiling or gas intrusion) traditional large LMFR cores show a
significant reactivity increase. Some experts consider that in the LMFR, an increase of the
coolant temperature should initiate an expansion of the absorber rod guide structure, of the
fuel in the axial direction and of the core grid plate in the radial direction, resulting in
negative reactivity coefficients counteracting the positive sodium void coefficient in large
LMFR. However, taking into account that the cooling disruptions and sodium boiling might
be on a time scale much shorter than the time scale of the passive negative feedbacks, there is
a strong incentive to reduce the positive sodium void coefficient in large LMFR cores. The
idea of the IAEA/CEC benchmark calculation 1990/94 [10] was to investigate the capability

6
of reducing the sodium voiding feedback reactivity of an axially heterogeneous core where a
sodium plenum is introduced above the core instead of the upper axial blanket.

The benchmark has shown that the overall void reactivity effect of the reference
2 100 MW(th) MOX-fuelled reactor core with postulated voiding configuration might be
close to zero. Future investigations were required to determine the differences in severe
accident response, in order to estimate the improvement in overall safety that could be
achieved from a reduction in the sodium void worth for reactor cores. Knowledge of in-core
sodium boiling and voiding physical phenomena was necessary for the determination of the
reactivity insertion rate.

There is a feature of contemporary fast reactors which have attracted particular attention on
the grounds of safety: positive reactivity transients which may be exacerbated by the fact that
the coolant void reactivity coefficient is under some circumstances positive.
An IAEA/EC 1994/98 project assessed the capability of reducing the sodium void feedback
reactivity of the homogeneous core by introducing a volume of sodium or plenum
immediately above the core, as was proposed by Russian specialists in advanced reactors.
It has been established by analyses done in various countries in the framework of the
Coordinated Research Project (CRP) that, in the event of an accident in which the sodium is
overheated, boiling in the core generates vapour which expands rapidly and voids not only the
core where it causes a positive reactivity change, but also the plenum where it causes a
negative change so that the overall change is less positive or even negative. This approach has
been adopted in the BN-800 fuel assembly design (Fig. 1) [11].

Main advantages of the innovative BN-800 type core design are to be seen in providing an
additional inherently activated safety margin of preventing fuel pin failure or local boiling in
the domain of operational and severe transients to be considered in the design basis. These
features complement well the large margin to fuel pin failure achieved already with the
hollow pellet fuel pin design and a clad material providing ductility even under high dose
loads. In the beyond design basis accident, domain some clear advantages of the innovative
core design have been identified:

Unprotected reactivity initiated accidents most probably lead to an early reactor shutdown
either due to pre-failure in-pin fuel relocation and/or due to a rapid fuel dispersal after fuel
pin failure in a few subassembly groups. Linear ratings at failure conditions are most
probably high, i.e. at about 1000 W/cm and more. Evaluations of the long term coolability
of the established core configuration after reactivity initiated accidents were not part of
this comparative exercise. They need careful consideration to evaluate potential
consequences of a thermally induced subassembly to subassembly propagation.
In case of unprotected loss-of-flow (ULOF) accidents, the main advantage of the as
specified innovative core design is that it is hardly possible to approach or exceed prompt
criticality in the initiating phase of the transient. At the end of most of the calculated event
sequences, core configurations that needed transition phase analyses were established.
Release of thermal and/or mechanical energy cannot be predicted without performing
appropriate analyses taking representative results of the initiation phase as initial
conditions.

7
 FIG. 1. BN-800 core fuel assembly with MOX-fuel (dimensions in mm, from the top:
boron screen (50), sodium plenum (300), core (880), blanket (350), gas (670) [11].

Conclusions drawn from this comparative exercise hold for the specified case set-up. They
need to be reviewed when some of the design features are changed or when more detailed
evaluations lead to different input data like:

Magnitude and/or spatial distributions of reactivity feedback coefficients of core

materials;
Reactivity feedback effects due to radial core expansion;
Fuel pin mechanical properties; and
If more rapidly developing consequences of control rod drive line expansion could be
demonstrated.

8
It is recognized that there are possibilities for improvement of the analyses and/or for
optimization, especially when a more realistic core design would be considered. However, the
comparative exercise has shown as well that consequences of these type of modifications need
to be analyzed carefully and in detail on a case to case basis. The use of more sophisticated
and experimentally validated theoretical models would be helpful to improve the reliability of
results. Evaluation of the impact of the as specified core design features on the core behaviour
during operational transients was not part of this exercise as well as stability analyses. This
would have needed other theoretical approaches to evaluate the potentially involved
problems.

1.2.1.1. Methods and codes for transient analysis

Results of this comparative exercise have shown as well that theoretical approaches chosen by
India with their PREDIS code package provides comprehensive results for single phase
analyses but they use simplified approaches for two-phase flow. Fuel pin mechanics is not yet
modelled in transient.

The Russian GRIF-SM code package with the complementary CANDLE-code package
provides results for ULOF-type transients up to molten clad relocation. However, it is
strongly recommended to couple a transient fuel pin mechanics code package to the system, to
develop fuel pin failure criteria considering special features of the BN-800 fuel pin design and
to extend the capabilities of the code system to describe core material relocation phenomena
after fuel pin failure or break-up. The different code versions of the SAS4A-code family
available in Japan, France and Germany allow evaluation consequences of accident initiators
leading to core destruction along all stages of the initiation phase up to hexcan melting on the
basis of experimentally qualified models. Even these code systems undergo continuous
improvement.

In France, the pre-failure in-pin fuel relocation model EJECT is approaching completion with
qualification and in Japan coupling to space time kinetics methods is far advanced. Thus,
more improved analysis tools will become available in future which provide the possibility of
re-evaluating the current results and to follow continuously the impact of new and/or
optimized design features of innovative core designs on the results of accident analyses to be
considered in the beyond design basis accident domain [12]. The IAEA/EC program on
evaluation of benchmark calculations on a fast reactor core with near zero sodium void effect
was one very good example of international cooperation. The participation included Germany,
India, Italy, Japan 1 (PNC, Mitsubishi, Hitachi, Toshiba), Japan 2 (Osaka University), Russian
Federation, U.K., and the USA.

1.2.2. Intercomparison of LMFR seismic analysis codes: comparison of experimental

results with computer prediction

One of the primary requirements for nuclear power plants and facilities is to ensure safety and
the absence of damage under strong external dynamic loading from, for example, earthquakes.
The use of seismic isolation for structures has been gaining worldwide acceptance as an
approach to aseismic design. Seismic isolation of important buildings such as nuclear power
plants would result in reducing in seismic induced load and, hence lead to economical
structural design. Fast reactors operate at high temperature which induces high thermal
stresses during transients. Hence the thickness of the structures needs to be minimized to limit
the thermal stresses, which contradicts the conventional requirements. This design approach
was pursued by adopting seismic isolation as is studied for the ALMR (Fig. 2) and other
LMFRs.

9
 FIG. 2. ALMR: nuclear steam supply system [9].

Therefore, the IAEA through its advanced reactor technology development programme
supports the activities of Member States to apply seismic isolation technology to LMFRs. The
application of this technology to LMFRs, and other nuclear plants and related facilities would
offer the advantage that standard designs may be safely used in areas with a seismic risk. The
technology may also provide a means of seismically upgrading nuclear facilities. Design
analyses applied to such critical structures need to be firmly established, and the CRP
provided a valuable tool in assessing their reliability.

The IAEA has sponsored two Coordinated Research Projects (CRPs) aimed at establishing the
reliability of analytical methods and computer codes applied to predicting the behaviour of the
reactor core and base-isolated reactor block structures to earthquakes. The studies under the
first CRP were useful for the verification and improvement of the reactor core seismic
analysis methodologies on the basis of comparison of experimental and numerical results
[13]. The second CRP was set up following the good performance of base-isolated
buildings [14].

High damping rubber bearings (HDRBs) provide a simple and economical isolation system.
They possess the low horizontal stiffness needed and are capable of safely withstanding the
large horizontal displacements imposed on them during an earthquake. The need for
additional dampers is avoided. In the HDRBs damping is incorporated into the rubber
compound. Figure 3 shows a diagrammatic section of an HDRB.

10
 FIG. 3. Sketch of the 1:8 scale prototype of the ALMR high damping isolation bearing;
HDRB deformed under a compression and shear strain [14].

Numerical simulation of rubber bearings by code ABAQUS gives satisfactory results as far as
material properties are evaluated properly and a suitable strain energy density function is
selected. Rubber material can be characterized by using ABAQUS by two important forms of
strain energy density functions: polynomial and Ogden. For the cyclic loading tests, numerical
simulation using a strain function of polynomial formulation gives better agreement with the
HDRB test. However, for the ruptures test, where the displacement is larger than the cyclic
loading test, simulation using Ogdens formulation gives better results.

The achieved results confirmed that finite element (FE) methods are useful tools for both the
detailed analyses of elastomeric bearings and their design for the better, they allow for a
considerable reduction of the number of tests to be performed. The FE analysis of the whole
of the rubber bearing carried by participant shoved that the analysis of a single layer of the
bearing can be used to predict the horizontal deformation of the bearing by scaling up the
results. This is significant in the sense that it reduces the computational time greatly. Further
this model can be used effectively to validate the material behaviour. However, more detailed
three-dimensional FEM is necessary to analyze the stress distribution of the isolator or to
evaluate the behaviour of the bearing at very large deformation.

When the same input data were used, all codes provided predictions of the horizontal
force-deformation characteristics of all isolators consistent with the test data. All predictions
of vertical behaviour in the absence of horizontal displacement were consistent with the test
data. A significant deviation of the numerical prediction of vertical displacement when
compression loading is combined with shear was found for some types of rubber bearings.

11
This deviation is attributable to the modelling of rubber, as almost incompressible, in the
constitutive equations.

The modelling of lead proved to be a problem for all teams as the material is deformed in
shear within the isolators, while the codes require as input data in tension. Continuing
research is needed on the accurate prediction of isolator hysteresis.

The main recommendation derived from the results of the CRP is that the study of isolated
nuclear structures should be continued and extended to non-seismic extreme load conditions.
The refinement of the characterization of hyper-elastic behavior of the elastomer is needed to
predict multi-directional response under combined loading. The modelling of the flexibility of
the reinforced plates and connecting plates should be improved. Investigation of the impact on
material characteristics of the special environmental conditions of nuclear facilities is needed.
Investigation of the finite element prediction of isolator failure mechanisms is needed.
Simple, accurate, reliable models for the isolator response over a wide range of multi-
directional deformation is essential for accurately predicting floor response spectra and other
dynamic design quantities [14].

Future research work should also look at the development of alternative seismic protective
technologies such as passive, semi-active and active control for the seismic protection of
nuclear facilities and components. The dynamic method to test large scale structures has been
validated for base-isolated civil structures and should be extended to isolated nuclear
facilities. The influence of vertical ground input on the response of all internal components of
an isolated nuclear structure should be investigated.

Fifteen Institutes from India, Italy, Japan, the Republic of Korea, Russian Federation, U.K.,
and the USA cooperated in this CRP.

1.2.3. Sodium mixing problems

During reactor operation, temperature fluctuations in the coolant close to a structure may
occur in many areas such as core outlet zone, lower part of hot pool, free surface of pool,
secondary circuit, and water/steam interface in steam generators. In certain conditions, these
temperature fluctuations can lead to thermomechanical damage to structures. In 1992
extensive cracking was found in a control rod guide tube that had been removed from the core
of the UK Prototype Fast Reactor (PFR).

High-cycle thermal fatigue was found to be the cause of the cracks in the connecting pipe and
the middle-stage heat exchange shell at the Tsuruga-2 PWR (Japan) in 1999: two coolant
flows lower temperature main flow inside the inner cylinder of the HE and higher
temperature bypass flow outside the inner cylinder were mixed. Repair of the damage
interrupted the reactor operation programme. Consequently, knowledge of temperature
fluctuations and induced thermomechanical damage to structures is essential to properly
support operation maintenance of a nuclear reactor during the plant lifetime. In an LMFR,
several areas of the reactor are also subject to this problem.

During normal operation, sodium at low temperature flows into the main pipe of the
secondary circuit. A small pipe, connected with a tee-junction to the main pipe, discharges
sodium into the main pipe at a temperature which is higher than in the main pipe. In certain
conditions, temperature fluctuations in the coolant close to a structure caused by thermal
striping can lead to thermomechanical damage to structures. This issue has been encountered
in the Phnix reactor in the secondary loop, where initial crack in a tee-junction zone was

12
detected during a campaign of inspections. In 1993, at the BN-600 reactor a sodium leak on
the purification loop of primary circuit was observed. Metallurgical expertise showed that it
was due to thermal fatigue caused by a fluctuating mixing of hot and cold sodium. There are
often difficulties in calculations, because of the complexity of the phenomena involved.

The Technical Working Group on Fast Reactors (TWG-FR) recommended the IAEA to
organize a Specialists meeting on Correlations between material properties and
thermohydraulics conditions, fluid temperature fluctuations and induced thermomechanical
damage in LMFR, to identify common trends in the interpretation of experimental and
analytical work, and the influence on design features.

The IAEAs specialists meeting held in Aix-en-Provence, France in November 1994, noted
that various computer codes for thermomechanical analyses had been and were being
developed in some countries. Significant progress has been made in substantiating the
phenomenological basis of the material design codes available to reactor designers, but some
problems still remain to be harmonized.

Great advances have been made in the last few years to improve calculation methods for
predicting fluid temperature fluctuation amplitudes and frequencies, involving large-eddy
simulation and direct solution of the Navier-Stokes equation. Within a short time, when
improved computing capacity which at present is being implemented, has become available, it
should be possible to provide information on fluid temperature at all times and at all
important locations. When this level of detail in thermohydraulic calculations has been
reached, all the requirements for calculating stresses in structural materials will have been
met.

However two problems need subsequent development and harmonization. Firstly, the question
of appropriate boundary conditions between more approximate fluid mechanic codes used to
predict overall flow properties and codes to predict the fine detail of thermal fluctuations in
critical regions has to be harmonized. Secondly, adequate experimental validation including
measurements of suitable detail and precision has to be provided.

The above mentioned specialists meeting also stressed that the best way to improve and
validate calculation codes and methods would be to organize benchmark calculation
comparison in the frame work of the TWG-FR. The TWG-FR following the recommendation
of Member States, at its 28th Annual Meeting in May 1995, proposed to perform benchmark
analyses in order to verify and validate the thermohydraulic and thermomechanical codes and
analytical methods using experimental data.

Institutes from a number of Member States have a great motivation to improve engineering
tools and predictions techniques concerning the characterization of the thermal striping
effects, in which numerical models have a major role.

Two possible benchmark problems were discussed at the Specialists Meeting in

Aix-en-Provence to verify and validate the thermomechanical codes using experimental and
analytical data: the square channel with transverse jet, and the t-junction of the LMFR
secondary circuit. The former would test thermohydraulic calculations and experiments while
the latter would test thermohydraulic and mechanical calculations. Design analyses applied to
thermal striping phenomena need to be firmly established, and the benchmark exercise on
T-junction of LMFR secondary circuit was approved by the IAEA in 1996 in the scope of the
CRP on Harmonization and validation of fast reactor thermomechanical and
thermo-hydraulic codes and relations using experimental data [20]. The physical

13
phenomenon chosen here deals with the mixture of two flows of different temperatures which
may induce temperature fluctuations that results in fatigue damage of the pipe wall.

Eleven institutes from France, India, Italy, Japan, the Republic of Korea, Russian Federation
and the UK participated in this CRP.

Work has been done at Framatome-Novatome (France) to provide the participants with
experimental data to harmonize and validate codes and methods by comparison of predictions
with test results. A set of French experimental data was made available to all the countries
participating in the CRP. The participating countries applied their codes and methods to
analyze these data. The input data provided come consequently from the actual operation of
the reactor. Necessarily, because of the complexity of these data, some have been simplified
where possible (i.e. when it was certain that this simplification would not influence the
results). Also, the comparison of thermohydraulic results with the actual phenomenon was
possible, owing to an instrumentation installed on site during the campaign of inspection. It
was agreed that each participant may concentrate his efforts to one or several benchmark
problem areas. Geometrical characteristics are:

main pipe D = 494 mm, t = 7 mm

small pipe D = 68 mm, t = 8.5 mm
circumferential weld located 160 mm downstream from the tee-junction.

Operating conditions: no transient to be considered (only nominal steady state); total duration:
90 000 hours.

Through cracks of about 100 mm length were observed on the circumferential weld (160 mm
from the centre line of the branch) at almost symmetrical locations on either side of the
meridional line. These cracks were noticed after grinding of the external weld bead. The pipe
was cut in air to observe the nature of cracks on the inner surface. An immediate appearance
of a white spot around the tee (1st plume) was also noticed. While the white spot was due to
the constant wetting of hot sodium, the black spot may be due to oxidation due to contact of
air in the zone where hot and cold sodium mixing takes place. The cracks appeared on the
black spot in the weld adjacent to heat the affected zone originated from the inner surface.
Information was provided on the mass flow rates, pressures and membrane stresses. Under the
operating conditions temperatures of the sodium in main pipe were 320C with flow rate of
2 850 m3/h and the corresponding values for sodium in branch pipe are 410C and 25 m3/h.
Additionally, 15 thermocouples, mounted upon the external diameter of the main pipe, have
given both mean and fluctuating temperatures for relevant positions. In addition to these
measurements, a lot of visual inspections and metallurgical examinations have been
performed and were available. They made possible to make comparisons with
thermomechanical evaluations. After the completion of thermalhydraulic and thermal
calculations, each participant to the benchmark was required to:

Characterize the mixing area, which means locate, define the dimensions of the
mixing area, defining the parts of the pipe subjected to the mixing phenomenon;
Provide isothermal lines (mean temperatures) on the parts of the pipe as defined
previously;
Provide the temperature signals (temperature variations as a function of time) on inner
and outer skins of the pipe at points significant, as well as the points noted a in Fig. 1
on inner and outer skins; (the points noted in Fig. 1 locate thermocouples, which are
fixed on outer skin of pipes);

14
 Provide the results of a spectral analysis of the metal temperatures at the points
defined above.

The CRP results and recommendations are given in Ref. [15].

1.2.3.1. Thermohydraulics modelling

The models used in this benchmark show that the pseudo direct Navier-Stokes
simulation is an alternative way in case of no subgrid scale model available, and is
possible with standard codes.
The steady-state approaches need a priori assumptions on the frequencies which are
therefore not a result of the calculations. Moreover, the finite volume (or finite
difference) implementation with precise discretization schemes is the most used.
The boundary conditions prescribed are mainly steady-state, with spatial distributions
in some cases and particular care is taken for the treatment of the wall attenuation.
The domain size depends on the kind of grid. Non-structured meshes allow to model
larger domains and a part of the hot pipe. Local refining is also useful for large
domains. The number of computational cells varies significantly, but one can retain the
order of 100 000 cells. The physical duration of the simulation remains short and
insufficient as regard to the damaging low frequencies.

1.2.3.2. Comparison with Phnix measurements and between thermohydraulics computations

Different types of flow behaviours are obtained: computation giving flattened jets
provides temperatures close to the Phnixs ones for location just behind the
tee-junction. For further locations, mean temperature results are sparse. The closest
results to experiments correspond to computational domains covering the whole pipe
and made of fine meshes.
An important dispersion is observed for the fluid peak to peak values. The low values
are due to too diffusive calculations (k- run or finite element run with 1st order in
space).
The longer runs (as regard to the physical time) provide the lower frequencies (this
means that the results depends strongly on this time). In all the cases, low frequencies
measured on the Phnix are not found, which implies a too high attenuation by the
metal wall. The maximum of fluctuation is found further downstream from the tee for
the cases where steady boundary conditions are the closest to the tee junction.
A reasonably good matching with measurements but on the meridian line is obtained
for some cases, but the maximum measured peak to peak value of 19 K is not found
by any calculation.

1.2.3.3. Thermomechanical assessments

In spite of the different codes used, the maximum stress range values obtained are
consistent. The prediction of cracking is strongly dependent on the fatigue strength
reduction factor applied in the method. It appears that if this coefficient is too severe,
the parent metal is found cracked which is not observed in reality.
The geometrical discontinuity formed by the weld bead has to be considered.
However, the actual value in a sharp discontinuity is not easily assessable by classical
FEM (finite elements method). In that case a d approach could be a good alternative.
The crack propagation results are diverging due to the different formulations used for
the stress intensity factor variation and due to the presence or not of a mean stress
effect.

15
1.2.3.4. Range of frequencies, sampling

The range of the damaging frequencies from the wall thickness should be determined
firstly:
(i) Frequencies lower than this band do not produce sufficient T across the wall,
(ii) Frequencies higher than this band cannot penetrate the wall,

The physical time of computation is deduced from the lower bound of the range,
considering that this physical time must cover at least 10 periods of this low
frequency. The time step of the computation must also be able to catch the higher
bound of frequency. [There can be other and more severe constraints (numeric) for the
time step)].

1.2.3.5. Domain and boundary conditions

The main objective is to provide realistic boundary conditions. Some phenomena cannot be
modelled by a local thermalhydraulic domain. Hence, the boundary conditions should include:

Secondary flows (e.g. swirl flow);

Low frequency variations of temperature and/or velocity.

A parametric study can be a good way to state on the necessity of modelling those
phenomena.

The following recommendations are proposed for elaborating the thermohydraulic model:

Use of an extended domain (mainly upstream) plus prescription of low frequency

phenomena at boundary conditions. Mesh must remain very fine in the region of
temperature fluctuations: non structured grids and local refinement are therefore very
useful;
An alternative way is the use of realistic boundary conditions from experiments or
other computations which can allow the use of a smaller domain;
If the physical time required (by the range of frequencies of interest) is too long and
not compatible with the CPU time, catch high frequencies on the small domain and
during a relatively short physical time by assuming to disconnect high and low
frequencies, add low variations of temperature to the computed signal of temperature
for the thermomechanical analysis;
The low frequencies (if in the range of interest) may appear in the computation as a
combination of the two neighbouring frequencies in case of fluctuating flow near the
wall. But, if they depend on a global behaviour (loop scale or plant itself), should be
got from measurements (e.g. correlation with flow rate signals). In case of a pipe, the
domain must include the full pipe (not half or part of it).

1.2.3.6. Thermohydraulic simulation

Transient simulation using a Large Eddy Simulation model (or as a pseudo direct
Navier-Stokes simulation if no subgrid model is available) is recommended. The
discretization schemes must be at least of order 2 in space. Order 2 in time is also
better but not completely needed.
Care must be given to the transient behaviour of the first computational mesh (mesh
adjacent to the wall) in association with transient heat transfer coefficient with induced
filtering of high frequencies. The effect of small geometrical singularities may not be

16
 negligible, but this problem is of secondary priority and cannot be solved before the
problem of attenuation on a flat plate.

1.2.3.7. Thermomechanical calculations

Mean stresses shall be considered in the fatigue damage assessment. The mean
stresses influence also the crack propagation. A parametric study is recommended to
quantify this influence and the manufacturing residual stresses have to be included in
the analysis.
A weld can be characterized by 3 parameters:
(i) The weld bead which is a geometrical discontinuity;
(ii) A different material from the parent metal; and
(iii) The presence of residual stresses.

These parameters influence the time to initiate and propagate a crack and must be defined
precisely. In particular, the factor 1.25 applied on the fatigue rupture curve to represent a
material effect is too low to cover all the consequences induced by the as-welded weld. It is
therefore recommended to take care of the representation of the weld in thermomechanical
assessments:

Combination of major cycles with thermal striping cycles is necessary to get the total
fatigue damage.
The time to propagate a crack is sensitive to formulation chosen for the stress intensity
variation. The strain controlled formulation is preferred to the stress controlled one for
this problem.
A 3D crack could be assessed in relation to the size of the mixing zone involved (here
the size of the hot spot).
The sensitivity of the defect depth on the threshold of the stress intensity factor should
be taken into account. The d approach using the random signal if possible could be a
good alternative.
Beyond 70% of the thickness, an analysis of the instability of the remaining ligament
shall be performed.

1.2.4. Core structural and fuel materials assuring high fuel burnup

At an IAEA technical meeting on the Influence of high dose irradiation on advanced reactor
core structural and fuel material [16], it was stressed that high fissile materials burnup is an
aim of all current designs because it reduces reprocessing costs and losses of radioactive
materials in reprocessing. During more than 40 years of intensive multinational development,
a significant experience has been accumulated on fast reactor MOX fuel pins as follows:

In Europe, more than 7 000 pins have reached burnup values of 15 at%. In addition,
some experimental pins (with solid or annular pellets) have attained burnup levels of
23.5 at% in PFR and 17 at% in Phnix;
In the U.S., more than 63 500 pins with solid pellets have been irradiated in FFTF
under prototypical conditions with more than 3 000 pins at 15 at% burnup and with a
maximum burnup level of around 24.5 at%;
In Japan, 64 000 pins with solid pellets have been irradiated in JOYO and foreign fast
reactors, with a maximum burnup level of around 15 at% in JOYO and 15 at% in
FFTF;
In the former USSR, presently in Russian Federation, a great experience was gained
with vibrocompacted MOX fuel. A record high burnup level of about 35 at% has been

17
 successfully reached with an experimental subassembly in BOR-60 (6 fuel pins) while
about 260 standard fuel pins have attained burnt levels of 25-30 at%. More than
4 000 fuel pins with pellet MOX fuel were irradiated in BN-350 and BN-600;
maximum burnup in BN-600 was 11.8 at%.

Experts concluded that the cladding rather than the fuel or the wrapper material, provides the
greatest limitation in reaching high displacements per atom (dpa) levels and thereby impedes
high fuel burnup. Major degradation problems are void swelling for austenitic and, to a lesser
extent, embrittlement at low temperatures for martensitic and ferritic-martensitic steels.
Maintenance of desirable properties is directly coupled to the maintenance of a stable
microstructure against the action of neutron-induced displacements. The most important
requirements for such stability are well-defined specifications and well-controlled production
methods. There are three paths toward achievement of high fuel burnup:

Use of low nickel austenitic steel;

Martensitic and ferritic-martensitic alloys; and
High-nickel nimonic PE16 alloy.

Of these, the lower-nickel austenite is inherently unstable during irradiation and eventually
must swell. If fuel burnup of > 20% is to be achieved in fast reactors, the combination of
austenitic cladding together with one of the other two swelling-resistant materials for
wrappers may pose a problem.

1.2.5. Concept for fuel resources and waste management

The scope and role of nuclear energy for long term development will be dependent upon its
ability to properly utilize uranium resources, to minimize the overall environmental impact,
and to contribute to sustainable energy supply. Thermal reactors (e.g. LWR and HWR)
operated on once-through uranium fueling extract about 1% of the potential energy of natural
uranium, while the bulk of uranium mined ends up in enrichment tails and spent fuel.
Advanced thermal reactor cycles can double this extracted energy to about 2%. Because of
this relatively low energy yield, nuclear power based only on thermal reactors should be
considered as an intermediate step towards sustainable nuclear energy development. It is
understood that sustainable use of nuclear power can be achieved only with fast reactors. By
using weapons-grade and separated plutonium and plutonium accumulated in spent fuel of
thermal reactors, fast reactors could allow the use of the large amount of depleted uranium
from enrichment tails. It is expected that hundreds of tons of weapons-grade plutonium will
be declared surplus to military needs.

Thus there is a sufficient amount of fissile materials for introduction of commercial fast
reactors. Corresponding reprocessing capacity for extraction of plutonium and minor actinides
from spent fuel and MOX fuel fabrication capacity is needed. In the long run, the efficiency
of uranium energy extraction may be increased up to 60-80%, i.e. the available fissile
resources could be stretched with LMFRs by a factor of 60-80. Development and deployment
of nuclear reactor and fuel cycle systems for complete extraction of energy from every
uranium atom fed into nuclear system would imply that:

Using only the stock of depleted uranium would, if converted to fissile plutonium, be
sufficient to supply the world with energy for hundreds of years; and
Only shorter-lived radioactive waste fission products would leave the system as an
outflow.

18
It should be noted that none of the advanced reactor systems being developed entirely meets
these features, but LMFRs are well suited for still further extracting energy from the 238U and
actinides. Thus, in fast reactors a multiple recycle could achieve total destruction of
transuranics (plutonium and minor actinides) and selected fission products so that all energy
is extracted from uranium atom(s) and only shorter-lived fission products leave the system for
returning to the earth's crust. Depending on their core geometrics and compositions, fast
reactors at a given power rating and core size can increase, maintain or decrease the inventory
of transuranics. Using this flexibility the loading of fast reactors can variously be
configurated/composed to produce a transuranic conversion ratio (CR) of less or more than
1.0. If CR > 1 the reactor system would become a breeder and generate fissile materials in
response to increased nuclear fuel (power) demand. If CR < 1, the fast reactor would become
a burner and could decrease the fissile materials (and actinides) stock.

Long-lived radioactive materials are produced by the operation of reactors of all types and,
with the exception of the very few that have commercial applications, have to be treated as
waste. By the year 2010, it is estimated that there will be more than 300 000 tons of spent
fuel, including 3 000 tons of plutonium and of the order of 100 tons of 237Np and americium
[16]. The most intensely radioactive high-level wastes (HLW) from irradiated fuel are either
the spent fuel itself, if it is not reprocessed, or the waste streams from the reprocessing plant.
The main contributors to the high radioactivity are the fission products and the isotopes of
elements beyond uranium in the periodic table (apart from plutonium). The latter are often
referred to as minor actinides (MAs).

The long-lived fission products and MAs set severe demands on the arrangements for safe
waste disposal, because it is necessary to ensure that they are kept isolated until they have
decayed to activity levels at which they pose no danger to the health of people and other
living organisms. In some cases this requires secure containment for many millennia. In most
countries the policy is to construct waste repositories, which will ensure the adequate
protection of the environment for foreseeable future. However, the opinion is gaining
influence that it is not right to impose on future generations the obligation to care for the
waste products of the present day. If it is eventually decided that it would be better to destroy
(or incinerate) the HLW rather than store it, fast reactors will play an important role.
MA isotopes can be utilized more efficiently in fast than thermal reactors, because most have
a lower ratio of capture to fission for fast neutrons. The IAEA report addresses this approach
in more detail [18].

1.2.6. Core disruptive accident

The safety level in an advanced LMFR plant could be further enhanced also by improving
reliability of the safety system, by installing passive safety features, and by a plant
simplification. However, even given a passive safety regime, one can envisage extremely low
probability but plausible scenarios in which the bases envisaging passive safety response are
violated (e.g. strong earthquake, and manufacturing flaw stochastic failure). Therefore, core
disruptive accident (CDA) will remain subject to investigations and discussions because of
recriticality potential and loss of second barriers (vessel) integrity cannot be excluded from
consideration. The advanced reactor design, as well as experimental and analytical research
on CDAs, should be directed to show achievability of in-vessel retention of the debris
generated in the non-energetic initiating phase by demonstrating in-vessel coolability,
subcriticality and vessel integrity. As noted at the IAEA/TWG-FR Technical Committee
Meeting on Material-Coolant Interaction and Material Movement and Relocation in LMFR
[19], the safety analysis of CDAs is supported by three inter-related elements of safety

19
research: identification of key phenomena, improved understanding through in- and out-of-
core experiments, and development of computer codes for safety analysis. Several system
analysis computer codes have been developed to evaluate CDAs. For example, computer
codes SAS4A- and SIMMER series are used as a system of tools for whole core accident
analysis.

1.2.7. Acoustic signal processing for the detection of sodium boiling in reactor cores or
leak detection and location in steam generators

As all highly-rated heat exchangers, e.g. steam generators, have a risk of failure, the
probability of leaks has to be taken into account in the design and operation. The IAEA
Coordinated Research Project (CRP) on Acoustic Signal Processing for the Detection of
Sodium Boiling or Sodium/Water Reaction in LMFR covered acoustic leak detection as well
as acoustic boiling noise detection.

It used experimental recordings of leak noise from test rigs and of background noise from
operating steam generators, and showed that a 1 g/s leak can be detected within 1 s with high
reliability and a false indication rate of the order of one in 25 years. A passive acoustic leak
detection system can make use of the same detectors as a plant monitoring system to detect
vibration. Active acoustic leak detection uses the changes to the acoustic properties of the
steam generator caused by the gas liberated.

The cloud of bubbles formed increases the acoustic attenuation of sodium. This can be
detected by means of acoustic pulses transmitted along the length of the steam generator
between the tubes. High frequency pulses of around 10 kHz are transmitted from one
transducer, along the tube bundle, to a receiving transducer with a pulse repetition rate of
around 1 Hz. If there is a significant fall in amplitude of the received pulses, it is concluded
that a leak is present. An advantage of the system is that it is fail-safe and self checking [20].

1.3. ACTIVITIES ON ADVANCED FAST LMFR AND TECHNOLOGY

The experience gained with fast reactors has revealed no fundamental problems with reactor
physics, operation of various equipments and sodium technology. This experience will be
useful in upgrading the design of next generation LMFRs. The objectives of the development
of advanced LMFRs are to achieve better economics relative to alternatives, to find optimal
solutions for the back-end fuel cycle and to achieve a high degree of safety and reliability.

The final goal relating to the objective of achieving of a very high degree of safety is to
design a reactor system which, not only during normal operation but also in case of an
accident, could exclude any radiological impact that would require evacuation of the public.
The following aspects are considered among the most important [21]:

Assure stability of the reactor core under all modes of normal and abnormal operating
conditions. Minimization of excess reactivity and sodium void effects, the reactor's
strong negative power and reactivity feedback with increased temperature, the large
margin to reactor coolant boiling at its operating temperatures, low pressure system
with large thermal inertia and sufficient safety margins are key aspects;
Assuring the heat removal from the core and the reactor under all upset conditions, by
assuring the ability to limit coolant temperatures below boiling, and fuel element clad
temperature below prescribed limits without the need for rapid operator action, i.e.
having a reasonably long grace period;

20
 Take advantage of passive safety systems to provide safety related functions without
reliance on operator action or on external mechanical and/or electrical power signals;
Minimize the burden on the operator of a nuclear power plant. Extensive consideration
is being given to man-machine interfaces.

In considering the objective to increase safety and reliability margins, advanced LMFRs are
being designed using all previous experience of being both good and not so good. Notable
examples of the innovative safety characteristics are: passive decay heat removal systems;
passive (inherent) reactor shutdown and stabilization by thermal and reactivity response
characteristics of the reactor even under extremely unlikely accident conditions.

As to the argument that some LMFRs faced reliability issues, it should be noted that this
would be true for any other reactor line at the initial stage of development, as LMFRs are. Full
industrial development of fast reactors has not been completed yet. It is simply too early at the
prototype stage of development to more general view of LMFR technology and economics,
particularly in the present situation where antinuclear environmentalists trend to discourage
investors from entering this field with the required resources. Other reactor technologies,
including water cooled reactors, achieved high reliability and lower generating costs when
their respective large scale introduction had taken place. One cannot say that this will not
happen in the case of LMFRs.

The early development of experimental and prototype liquid metal fast reactors was
conducted to a large extent on a national basis. However, for advanced LMFRs, international
cooperation begins to play a greater role and the Agency promotes international cooperation
interalia in its development. Especially for R&D incorporating innovative features,
international cooperation can play an important role allowing a pooling of resources and
expertise in areas of common interest to help to share the high costs of development.
To support the IAEA's functions of encouraging LMFR development, the IAEA promotes
technical information exchange and cooperation between Member States with fast reactor
development program, offers assistance to Member States with an interest in exploratory or
research program, and publishes reports available to all Member States interested in the
current status of LMFR development. Experience gained from R&D, operation and
construction of fast reactors has been reviewed periodically by the TWG-FR.

The role of the IAEA as a truly international forum for cooperation and exchange of
information cannot be overemphasized. About 130 various specialist meetings, seminars and
symposia have been organized so far by the IAEA on the advice of the TWG-FR (formerly
IWG-FR) and about 3 900 specialists from more than 25 countries have participated in these
events. Some examples of topics of Specialists and Technical Committee Meetings held on
the advice of the TWG-FR/IWG-FR include:

Sodium combustion and its extinguishment-techniques and technology;

Steam generators for LMFBRs, fuel failure mechanisms;
Fuel and cladding interaction;
Properties of primary circuit structural materials including environmental effects;
Leak detection and location in LMFBR steam generators;
Sodium fires, design and testing;
LMFBR steam generator integrity and reliability with a particular reference to leak
development and detection;
Repair of FR steam generators: acoustic/ultrasonic detection of in-sodium water leaks,
Propulsion reactor technology for civilian applications;

21
 Sodium removal and disposal from LMFRs in normal operation in the framework of
decommissioning;
Operational and decommissioning experience with fast reactors;
Conceptual design of advanced fast reactor;
Absorber materials control rods and design of shutdown systems for advanced liquid
metal fast reactor;
Influence of low dose irradiation on the design criteria of fixed internals in fast
reactors;
Use of fast reactors for actinide transmutation, and
Creep-fatigue damage rules for advanced fast reactor design.

Research on LMFR during the last decades has significantly improved understanding of
LMFR designs, technology and safety. Pool type design concept was chosen for all small,
medium and large size advanced reactors: CEFR (China), PFBR (India), KALIMER
(Republic of Korea), BN-800 and BN-1800 (Russian Federation), and EFR-1500 (European
Fast Reactor, Fig. 4), except for DFBR-660 and JSFR (Japan) which uses the loop-type
concept.

1-hung safety vessel; 2-debris tray; 3-core; 4-IHX; 5-above core structure; 6-secondary loop;
7-primary pump; 8-hot collector; 9-cold collector; 10-grid plate; 11-core support structure;
12-spring bearing; 13-anchered safety vessel
FIG. 4. EFR: cut of reactor.
The pool reactor placed in a guard vessel has been shown to have very attractive safety
characteristics, resulting to a large extent from a liquid metal cooled reactor being a low
pressure system with large thermal inertia. This type of design practically excludes
unfavourable consequences of failures in the external radioactive systems and loss of reactor
coolant [22]. One of the outstanding achievements of the EFR programme is to make firm and

22
reliable cost estimates. Construction of a reactor to the EFR design may not be possible in the
near future, but a well validated way forward to commercial utilisation of fast reactors has
been established. This way is generally consistent with other studies, and it indicates that the
goal of competitive fast rectors may be within reach (Fig. 5).

FIG. 5. Generating cost comparison EFR vs. advanced PWR.

Findings of recent work have led to improvement in LMFR designs and safety. For
example, an advanced design of the system: main vessel - safety vessel - vault was developed
for EFR (Fig. 6).

FIG. 6. EFR: safety vessel and vault [22].

23
The main vessel is completely surrounded by a leak tight safety vessel anchored to the surface
of the concrete vault. A layer of metallic insulation covers the inside surface of the safety
vessel, which reduces both the heat losses and the thermal cycle skin fatigue. A layer of
sodium resistant concrete is provided between the safety vessel and the structural concrete of
the vault. The structural concrete is kept cool by the vault cooling system. The top entry loop
type design was selected for the Japanese DFBR because of the following considerations [23]:

Major primary components such as the intermediate heat exchanger (IHX) and the
pumps are outside of the reactor vessel, and this facilitates maintenance and repair;
The system has flexibility to introduce such innovative technologies as the
electromagnetic pump integrated component, which is needed for commercialization
of the FR; and
Experience gained at the prototype MONJU must be fully utilized. Considering that
the top entry system is quite a new concept, the conceptual design study, the
evaluation study of commercialization prospects and the water hydraulic tests using
models of thermalhydraulic properties specific to the top entry system were
conducted.

In progress of the loop type LMFR design development, Japan is now in a position to embark
on an in-depth study of an advanced plant configuration - a compact loop type LMFR design
JSFR (Fig. 7).

FIG. 7. Birds eye view of JSFR-1500 NSSS [3].

24
To achieve the economic target, several innovative technologies and LMFR design
improvement measures have been adopted [3]. The reduction of plant material is
accomplished by adopting the following technologies:

Shortening the piping length and reduction of the number loops by adopting 12 Cr
steel which has low thermal expansion with high strength;
Development of integrated intermediate heat exchanger (IHX) with mechanical pump.

Although there are differences in conceptual design approaches, a number of common topics
could be identified among the conceptual designs. These include improvements with regard to
safety margins, design simplifications and cost reduction.

Safety margins improvements include the consideration of core catcher design for excluding
the recriticality, cooling capabilities, passive backup reactivity shutdown, decay heat removal
systems and a strong inherent negative reactivity feedback with temperature rise. Economic
competitiveness should be improved by the following design efforts: the optimization of the
number of cooling loops and equipments, and achievements of reducing its weights as well as
the building volumes, findings of structural materials, fuel technology and core design to
achieve high burnup, and limiting the number of safety graded systems.

The development of simple, reliable, efficient and flexible systems and components is a
primary objective in the design of advanced fast power reactors. Systems and components for
developmental design (advanced reactor designs which range from moderate modification of
existing designs to entirely new design concepts) in general require much extensive testing
and demonstration to verify component and system performance. Key issues are scaling
effects for simulating plant configuration, design life and interactions among different
systems.

1.4. SAFETY RESEARCH ISSUES FOR ADVANCED LMFR DEVELOPMENT

The TWG-FR has paid much attention to reactor characteristics relevant to LMFR safety.
Some examples of topics of Specialists and Technical Committee Meetings held within the
framework of the TWG-FR (formerly IWG-FR) include:

Operational safety of sodium circuits;

Reliability of decay heat removal systems;
Role of fission products in whole core accidents;
Demonstration of structural integrity under Normal and Faulted Conditions;
Sodium fires and prevention;
Evaluation of radioactive materials release and sodium fires in fast reactors;
Passive and active features of LMFRs;
Material-coolant interaction and Material Movement and Relocation;
Reactor core with near zero sodium void Effect; and
Primary coolant pipe rupture event in liquid metal cooled fast reactors.

The meetings with above mentioned topics cover a wide area of LMFR technology and have
demonstrated fruitful international cooperation. A series of international conferences,
symposiums and topical meetings on fast reactor safety under the sponsorship or in
cooperation with the IAEA have been held in the past: Aix-en-Provence (1967), Beverly Hills
(1973), Seattle (1979), Lyon (1982), Guernsey (1986), Snowbird (1990) and Obninsk (1994).

25
1.4.1. Progress made since the past

Minimization of the risk to the public by further improving the safety level of advanced
nuclear systems is one of the ultimate goals of safety research. Although large progress has
been made since the time that the first prototype LMFRs: the 750 MW(th) BN-350, the
250 MW(e) PFR, the 255 MW(e) Phnix, and 600 MW(e) BN-600 were put into operation
more than two decades before, the search for improvements in performance and reliability by
design measures has called for continued R&D in the broad area of the LMFR technology
with a special emphasis on nuclear safety. To draw the findings of the international meetings
on fast reactors and related fuel cycle [25] and advanced reactor safety [23, 25], the following
advances have been made:

The amount of available data on LMFR technology has been expanded significantly.
Experimental design and analytical work were carried out for all important
components and systems including steam generators, sodium pumps, intermediate heat
exchangers, sodium and aerosol leak detection, fuel failure detection, hydrogen
detection after failure of steam generator tubes, sodium boiling detection, temperature
measurement. The most important achievements have been reached on liquid metal
technology and mixed oxide fuels.
A large number of tests and analytical work has significantly improved the
understanding of passive and natural safety characteristics of the LMFR. Two types of
passive safety features of the LMFR for prevention and mitigation of core disruptive
accident (CDA) are under consideration for advanced LMFRs: heat removal from the
core by natural convection; and strong negative reactivity feedback mechanisms to
control and/or restrict the core power in emergency situations. To demonstrate the
functioning of the decay heat removal (DHR) systems, an experimental and theoretical
programme has been carried out with in-core and out-of-core tests in water and
sodium in differently scaled systems, including full-scale DHR systems of the EFR.
At least 25 experimental rigs have been built in Germany, France, U.K., U.S., Russian
Federation, and India to study the characteristics of different DHR systems. As
concluded at the TWG-FR Specialists Meeting on Evaluation of DHR by Natural
Convection [27], the existing experimental data and the analytical work show that the
decay heat removal by pure natural convection is feasible. Concerning the objective of
passive safety, DHR by pure natural convection is an essential feature to enhance the
reliability of DHR. The rationale of safety strategy for advanced LMFR is to employ
completely mechanical and physical laws rather than engineered systems whose
reliability is subject to human action. Passive reactivity shutdown systems have been
developed and demonstrated for the BN-800 (suspended by coolant flow) and PRISM
(gas expansion module in the core). A design study of enhancement of absorber rod
drive line expansion was made for the EFR. Above a certain (switching) temperature
its thermal expansion is about 3 times more than a natural thermal expansion relative
to the core [28].
Three-dimensional codes to allow calculations of steady states or thermal-hydraulic
transients concerning the hot plenum and cold collector, fuel subassemblies and
secondary loops have been improved as a consequence of a large number of separate
effect tests and detailed numerical studies. Subsequent developments continue to
decrease uncertainties, particularly for gas entrainment phenomena, temperature
fluctuations for thermal striping and core outlet area modelling.

Research on liquid metal-cooled reactors during the last decades has significantly improved
our understanding of LMFR safety. The achievement of the past safety research has been

26
effectively used to develop a system of safety analysis methods which were used to evaluate
the safety characteristics of the existing and advanced fast reactors. It is predicted that LMFRs
currently being designed can achieve a very high degree of safety. However, in spite of all the
progress made on LMFR technology and in particular on safety development, the quest for
excellence calls for further work. The IAEA gives support within the framework of its Statute
and the available means, on topics to be briefly addressed below. As to the challenges for
future advanced LMFR, it is important to realize that all modern nuclear power plants employ
the defense-in-depth safety strategy which relies upon maintaining the structural integrity of
the three principal barriers preventing release of radioactive fission products: the fuel
cladding, the primary system boundary and the containment. Failure of these barriers
primarily results from mechanical and thermal loads, and safety research and nuclear power
design studies show that significant margins have been provided to avoid such failures. For
LMFRs there are a number of safety issues which influenced fast reactor licensing and safety
analysis and some of them had been discussed above.

REFERENCES

[1] INTERNATIONAL ATOMIC ENERGY AGENCY, Status of Liquid Metal Cooled

Fast Breeder Reactors, Technical Reports Series No. 246, IAEA, Vienna (1985).
[2] INTERNATIONAL ATOMIC ENERGY AGENCY, Status of Liquid Metal Cooled
Fast Reactor Technology, IAEA-TECDOC-1083, IAEA, Vienna (1999).
[3] Meeting of the Technical Working Group on Fast Reactors Meeting, 1014 May
2004, Vienna, Austria.
[4] MARTIN, L., et al., Life extension of Phenix NPP, paper presented in Mtg. on
Operational and decommissioning experience with fast reactors, 1115 March 2002,
Cadarache, France.
[5] GUIDEZ, J., MARTIN, L., Phenix: Thirty years of operation for research,
reactor renovation overview and prospect, Paper presented in Int. Conf. Fifty
years of nuclear power-the next fifty years, 27 June2 July 2004, Obninsk, Russian
Federation.
[6] INTERNATIONAL ATOMIC ENERGY AGENCY, Fast Reactor Database,
IAEA-TECDOC-866, IAEA, Vienna (1996).
[7] SARAEV, O.M., Operating experience with Beloyarsk fast reactor BN-600,
Unusual Occurrences During LMFR Operation, IAEA-TECDOC-1180, Vienna
(2000)101116.
[8] CRUICKSHANK, A., JUDD, A.M., Problems experienced during operation of the
prototype fast reactor, Dounrey, 19741994, Unusual Occurrences During LMFR
Operation, IAEA-TECDOC-1180, IAEA, Vienna (2000) 942.
[9] INTERNATIONAL ATOMIC ENERGY AGENCY, Status of Liquid Metal Cooled
Fast Reactor Development, IAEA-TECDOC-791, IAEA, Vienna (1994).
[10] INTERNATIONAL ATOMIC ENERGY AGENCY, Evaluation of Benchmark
Calculation on a Fast Power Reactor Core With Near Zero Sodium Void Effect,
(Final Report Supported by the IAEA and the Cec, 19921993) IAEA-TECDOC-
731, IAEA, Vienna (1994).
[11] KIRYUSHIN, A.I., et al., BN-800-next generation of Russian sodium reactors, paper
presented in Int. Conf. Innovative Technologies for Nuclear Fuel Cycle and Nuclear
Power, 2326 June 2003, Vienna, Austria.
[12] INTERNATIONAL ATOMIC ENERGY AGENCY, Transient and Accident
Analysis of a BN-800 Type LMFR with Near Zero Void Effect, IAEA-TECDOC-
1139, IAEA, Vienna (2000).

27
[13] INTERNATIONAL ATOMIC ENERGY AGENCY, Intercomparison of Liquid
Metal Fast Reactor Seismic Analysis Codes Volume 3: Comparison of Observed
Effects with Computer Simulated Effects on Reactor Cores from Seismic
Disturbances, IAEA-TECDOC-882, IAEA, Vienna (1995).
[14] INTERNATIONAL ATOMIC ENERGY AGENCY, Verification of Analysis
Methods for Predicting the Behaviour of Seismically Isolated Nuclear Structures,
IAEA-TECDOC-1288, IAEA, Vienna (2002).
[15] INTERNATIONAL ATOMIC ENERGY AGENCY, Validation of fast reactor
thermomechanical and thermohydraulic codes, IAEA-TECDOC-1318, IAEA, Vienna
(2002).
[16] INTERNATIONAL ATOMIC ENERGY AGENCY, Influence of High Dose
Irradiation on Advanced Reactor Core Structural and Fuel Material, IAEA-
TECDOC-1039, IAEA, Vienna (1998).
[17] KUSTERS, H., et al., The Nuclear Fuel Cycle for Transmutation: a Critical Review,
paper presented in Int. Conf. Evaluation of emerging nuclear fuel cycle system,
Global 1995, Versailles, France (1995).
[18] INTERNATIONAL ATOMIC ENERGY AGENCY, Use of Fast Reactors for
Actinide Transmutation (Proceedings of a Specialists Meeting, Obninsk, Russian
Federation, 2224 September 1992) IAEA-TECDOC-693, IAEA, Vienna (1993).
[19] Technical Committee Meeting on Material-Coolant Interaction and Material
Movement and Relocation in LMFR, 1994, O-Arai, Japan.
[20] INTERNATIONAL ATOMIC ENERGY AGENCY, Acoustic Signal Processing for
the Detection of Sodium Boiling or Sodium-water Reaction in LMFRs,
IAEA-TECDOC-946, IAEA, Vienna (1997).
[21] INTERNATIONAL ATOMIC ENERGY AGENCY, Objectives for the Development
of Advanced Nuclear Plants, Vienna, IAEA-TECDOC-682 (1993).
[22] ELECTRICITE DE FRANCE, Project Rapide 1500 MW, published by EdF, France
(1984).
[23] INTERNATIONAL ATOMIC ENERGY AGENCY, Conceptual Designs of
Advanced Fast Reactors, IAEA-TECDOC-907, IAEA, Vienna (1996).
[24] Meeting of the Technical Working Group on Fast Reactors, 1216 May 2003,
Daejon, Republic of Korea.
[25] Current Status and Innovations Leading to, Promising Plants, paper presented in Int.
Conf. Fast Reactors and Related Fuel Cycles, FR91, 28 October-1 November 1991,
Kyoto, Japan.
[26] International Topical Meeting on Advanced Reactor Safety (ARS 94), 1721 April
1994, Pittsburgh, USA.
[27] IAEA Specialists Meeting, 2223 February 1993, O-arai, Japan.
[28] INTERNATIONAL ATOMIC ENERGY AGENCY, Absorber Materials, Control
Rods and Designs of Shutdown Systems for Advanced Liquid Metal Fast Reactors,
IAEA-TECDOC-884, IAEA, Vienna (1996).

28
2. PROTOTYPE FAST REACTOR

2.1. DESIGN FEATURES AND REVIEW OF OPERATING HISTORY

The primary circuit of the prototype fast breeder reactor (PFR) was a pool type and had an
inventory of 900 tons of sodium. The equipment was stationary fixed on the upper plate from
which the reactor vessel is suspended. This design requires a complex thermal expansion
compensation system of skirts, inside which tube bundles of heat exchangers were inserted.
For this purpose the PFR reactor has bellows operating at high temperature and being
in-maintainable. Compensation of pressure pipelines thermal expansions was achieved by
means of the bends. In view of that several pipelines connect the pump to the pressure
plenum.

One of the main thermo-mechanical problems concerning the top support reactor design was
to create acceptable operational temperature conditions for the upper bearing plate, from
which the vessel is suspended and on which the equipment is installed (pumps, heat
exchangers, rotating plugs, etc.). For thermal shielding of the bearing plate internal surface in
the PFR reactor multi layer steel foil insulation was used, which turned out to be complicated
and expensive.

Experimental investigations and analysis of heat and mass transfer from the coolant surface to
the upper plate carried out at the design stage revealed a number of interesting effects. These
investigations showed that there is a great temperature drop between the sodium surface and
the plate (200-300C). The main part of this drop falls at the region of near-sodium boundary
layer and the rest part - at the near-the-plate boundary layer. In the central part of the gas
plenum vigorous mixing results in uniform temperature.

The experiments showed an active process of aerosol formation due to sodium evaporation
from the surface, its consideration near the surface and sodium particles fog formation in the
gas plenum. Thus, in the gas plenum a complex heat transfer takes place as a result of
convection, aerosol mass transfer and radiation.

The main mechanism of sodium surface and cover plate heat transfer is thermal radiation.
Its intensity to a considerable extent is determined by the emissivity of the sodium surface,
cover plate and sodium fog. With allowance for specific features of heat and mass transfer in
the gas plenum the analysis showed that at the sodium surface temperature of 520C the
temperature of the plate surface may be 350450C. Changes of surface effective emissivity
from 0.03 up to 0.1 results in increasing heat flux from 700 to 1250 W/m2, changes of optical
thickness within the range of 0.110 have a weak effect on the heat flux value and they
influence the gas temperature only at high values of the plate effective emissivity. The
radiation-induced heat flux is of prevailing importance. The sodium fog significantly
(2.5 times) reduces the heat flux density thus screening the plate from the sodium surface
radiation.

On the basis of the experimental investigations and analysis it was concluded that with
allowance for fog screening properties in the reactor gas plenum the thermal insulation design
could be simplified or due to ascertaining data on radiating properties of the hot sodium
surface and that of the cover plate surface.

Figure 1 shows a cross-section of the primary circuit. Heat from the 600 MW(th)-rated core
was transported by sodium primary coolant pumped by three electrically-driven (1 MW)
mechanical pumps to six intermediate heat exchangers (IHXs), which were grouped in three

29
non-adjacent pairs and each pair was connected with the relevant secondary circuit. Sodium
entered the core at a temperature of 400-430C; the core temperature rise was about 160C.

The primary sodium was contained in a primary vessel of 18/8/1 stainless steel, 12.2 m in
diameter and 14.0 m deep, surrounded by a guard vessel made from medium carbon boiler
grade steel. Sodium purity was controlled by means of an external cold trap loops.

FIG. 1. Elevation through PFR primary circuit.

Secondary sodium flowed through the shell side of each IHX and transported heat to the
steam generators consisting of an evaporator, a superheater and a reheater. There were three
secondary circuits, each containing about 75 tons of sodium which was circulated by a
mechanical pump similar to the primary sodium pumps and each coupling a pair of IHXs.

30
Superheated steam from the three circuits flowed to a common header to drive a 300 MW
turbo-alternator. The main feed was via a 100% duty steam-driven pump with 10% electric
and 10% steam-driven pumps for start-up and post-trip use (later, a 50% capacity auxiliary
electrically-driven pump was installed as back-up). Appropriate water conditions were
provided by a full-flow polishing plant, and the feed-heating by sets of low pressure direct
contact and high pressure tube units. The condenser was cooled by seawater.

The reactor core and its surrounding blanket were made up from an array of hexagonal MOX
subassemblies, 142 mm across flats. The assemblies were of a size appropriate to a full-scale
commercial reactor, and provided a core 910 mm high and about 1550 mm in diameter.
Control was exercised through five boron carbide absorber rods, and a further five similar
rods were available to shutdown the reactor.

A radial breeder (blanket) surrounded the core and was itself bounded by stainless steel
reflector assemblies to improve neutron economy. Outboard of the core and blanket was a
graphite shield which essentially eliminated neutron activation of major removable
components such as the primary pumps, valves and the IHXs, the secondary sodium, and the
primary vessel itself. Special loops filled with eutectic sodium-potassium alloy (liquid at
room temperature) as coolant were provided to reject decay heat from the primary coolant via
air-cooled heat exchangers to the atmosphere after reactor shutdown if the steam generators
were not available.

Fuel could be transferred from an adjacent preparation facility, the irradiated fuel cave (IFC),
to a storage rotor within the primary vessel while the reactor was operating. This rotor
reduced the time required for refuelling operations and, when irradiated fuel was discharged
from the rotor to the IFC, reduced the number and complexity of the transfer flask movements
because irradiated fuel removed from the core could be left to cool in the rotor before being
moved to the IFC after the reactor had resumed operation.

Transfers between the core and the storage rotor used a vertical lift pantograph charge
machine working through a single rotating plug in the reactor roof; for such moves the reactor
had to be shut down to allow the charge machine to be installed. Fuel discharged from the
rotor after irradiation was first stored under sodium in the IFC until either it had been
examined and returned to the reactor for further irradiation or was cool enough to be prepared
for reprocessing (including steam cleaning to remove all traces of sodium) and then moved to
a buffer store to await transfer to a reprocessing plant, also located on the site. The time from
start of construction to filing of the primary circuit had been seven years compared with the
four years planned, the delay being due principally to difficulties experienced in the welding
of the reactor vessel roof. The nuclear power plant construction cost was about 40 million
pounds (1974), with an additional 5 million pounds for the installation of a 170 km long high
voltage transmission line to connect with the main grid.

Commissioning had proceed without major problems though a water test of the test of the
sodium side one of the circuit had revealed a gas entrainment problem requiring modification
of all shut-of-circuits, and bearing problems on one of the primary pumps and two of the
secondary pumps. The approach to criticality began in February 1974 and was first achieved
on March 1974. Physics parameters for the core and for the reactivity effectiveness of, and
interaction between the control and shut-of rods agreed with prediction within the expected
uncertainties. The hot dynamic test was completed in June 1974.

31
The operating history of the PFR power plant can be divided into two phases. For the first ten
years (1974-1983), electrical output was limited, mainly because of a series of leaks in the
steam generator units, and the highest load factor in any year was 12% in 1978 (Fig. 2). After
1984, with the steam generator weld problems dealt with, plant performance improved, in the
final year of operation the load factor was about 57%.

FIG. 2. PFR annual load factors 1974-1994

(1994 for three months' operation only-before decommissioning), %.

In 1985, PFR was able to operate, for the first time since the commissioning period, with a
full set of steam generator units. In the second decade of operation there was one major
outage, in 1991/92. In this period, unto 1991 the reactor and primary circuit equipment were
responsible for only a very small fraction of unplanned outage time; on 25 June 1991, a
leakage of oil from a bearing of one of the primary pumps into the primary sodium led
interruption of reactor operation for 18 months. PFR was started up for the last time on
14 January 1994. Figure 2 shows a histogram of annual load factors. The major incidents and
unforeseen events over the 20 years of operation were as follows:

Series of steam generator leaks;

Sodium aerosol deposits in control rods;
Major oil leak into the primary circuit;
Cracks in the air heat exchangers of the decay heat rejection loops;
Sodium mixing and vibrations problems.

2.2. REVIEW OF THE PFR STEAM GENERATOR DESIGN CONCEPT AND

OPERATING HISTORY

PFR had three separate forced-circulation steam generators, each of which contains an
evaporator, a superheater, and a reheater. Figure 3 shows the general arrangement and Fig. 4
shows a superheater in more detail.

32
FIG. 3. PFR steam generators.

FIG. 4. PFR superheater.

33
2.2.1. PFR steam generator: choice of the design concept

All three heat exchangers are of U tube configuration in which the tubes were welded to
tubeplates. There were no other welds in the tubes. All three tube bundles can be removed
from their shells by breaking a flanged joint situated at tubeplate level that they are of
different configurations. In the evaporator the tubes are arranged so that there is a simple
diametrical baffle across the shell to provide counter-flow; the other two units (superheater
and reheater) achieved the same objective with annular baffle and annular steam headers.

A free sodium level was established below the tubeplate. The main idea was, that the gas
spaces which contain all the tube-to-tubeplate welds, to protect welds by keeping the
sodium away from them as well as to improve the sensitivity of detection of the leaks in the
tube-to-tubeplate welds. The UK experts believed that limitation of welds to the gas space
was essential in PFR, where one of the main objectives was to prove the single wall
concept.

The general philosophy of achieving reliability and high availability as well as leading to
simple design required detailed consideration of possible manufacturing and fabrication
problems. It was realised at the outset that the tube-to-tube-plate joint would be a critical area.

Three types of the tube-to-tubeplate joint and parameter tolerances set because of material and
tube size differences between the units were considered at the design stage (Fig. 5).
An explosive welding was rejected because the UK designers considered that it was
impossible to remove the sodium side crevice.

FIG. 5. Steam generator tube/tubeplate weld style.

34
Therefore two types of tube/tube-plate weld: stub/butt weld and butt/fillet weld
were selected for detail consideration (Fig. 6 a-b). Both welds rely on internal bore welding
using tungsten inert-gas welding without the addition of filler wire. Mixed Nb + Ti-stabilized
2 1/4Cr-1Mo tubing was welded.
a b

'stub/butt butt/fillet

FIG. 6a,b. PFR steam generator tube/plate weld joints.

Some experts believed (and this was confirmed by PFR SGs operation) that 'stub/butt weld
(Fig. 6a) provides a favourable mass balance for welding, the post-weld heat treated after
manufacture can be readily provided and the joint gives maximum ligament efficiency.
Initially single welds were made using a 200C preheat and a post-weld heat treatment of
30 min at 700C, but subsequent trials showed that welds could be made with satisfactory
profile and quality without any preheat.

Further tests were done on test blocks to simulate the conditions in service. No preheating was
used and radiographically acceptable and gas-tight welds were produced. There was a
tendency to produce porous welds, and in the main these could be satisfactorily repaired by
re-running. Specimens cut from these welds showed that the ultimate tensile strength was
32.2 tons/in2 at room temperature, but this was reduced to 26.2 tons/in2 at 480C. Other
mechanical tests e.g. bend and fatigue tests gave satisfactory results.

In parallel with the development work on the stub/butt weld, work was undertaken to
investigate an alternative geometry. It was found that the machining costs for the stub/fillet
weld type of joint were considerably cheaper than for the stub/butt' and thus additionally
there was no possible weakness due to the cross-grain of the machined stub. On the other
hand, this joint has the disadvantage of relatively poor ligament efficiency compared with the
stub/butt joint and the joint geometry was not ideal for either optimum coolant flow
conditions or for subsequent non-destructive examination

The butt/fillet weld was used for the design (Fig. 6b). The welding conditions were closely
controlled with particular attention being paid to fit of components, location of electrode, arc
length and rate of heat input. Tolerances and fit are rigidly controlled by inspection whilst the
electrode location can be preset. Welding parameters were controlled automatically within the
limits required to produce consistent welds. A similar type of welding control has been

35
applied to fuel-element assemblies where the accuracy of control was such that service
failures have been restricted to 1 in 30 000.

The design of each evaporator unit required the welding of 498 U-tubes 39-ft -in. long into a
7-ft 4-in. diameter tube plate. Furnace stress relief was not considered practical due to the
difference in section and the handling problems of a U-tube bundle with such a thick header.
Furthermore, a suitable local stress-relieving treatment was not considered possible and tests
were initiated to assess the performance of the unstressed relieved welds.

One of the main disadvantages with the butt /fillet' weld compared with the stub/butt' weld
was the difficulty in carrying out an adequate non-destructive test. A high degree of quality
control and reliability was maintained due to the automatic control of the important welding
variables. Additionally, with this type of joint the weld profile is a good indication of weld
quality and care was taken to ensure that the welding parameters were controlled within
acceptable limits. Special optical devices have been developed for visual examination of both
weld surfaces.

A permanent record could be kept of the bore examination by means of a TV camera and a
video recorder. The camera and endoscope were mounted in a housing which locates the
endoscope at the centre of the weld and allows focusing and both axial and rotational
movement. The lighting, focusing and drive-direction were controlled remotely by means of a
miniature transistorized console. The picture was displayed on an 11-inch monitor and stored
on magnetic tape. A sound channel was used to add a commentary to identify the weld and to
draw attention to any anomalies. The play-back is of good quality and cracks are clearly
defined.

Despite this high degree of quality control, development work aimed at finding a reliable non-
destructive testing technique to a successful conclusion. Two systems were possible: one an
ultrasonic technique using twin finely tuned shear-wave probes and the other radio graphic
using a small thulium-170 isotope source. The success of radiographic technique depends on
the production of small 1-mm1-mm right cylinders of thulium-170 sources, at the initial
source strength of 3 1/2 Cr and the development of a film cassette capable of being wrapped.

Using this new improved technique crack indications were observed in the tail-off region of
the weld. Electro-polishing of the outer weld profiles and careful sectioning showed that these
defects always occurred on the outside surface of the weld and were typically 0.010 in. long
and 0.002 in. deep. Initially it was thought that these cracks were due to the presence of small
quantities of Nb-rich eutectic giving low hot ductility on the rerun during the tail-off of the
weld, but subsequent tests showed that significantly greater quantities of eutectic than that
present in the weld did not affect ductility. Nevertheless, modification of the startup and tail-
off techniques eliminated the tendency to produce these fine cracks. Experience in the
manufacture of 2 000 defect-free welds confirmed the ability to satisfactorily weld the
stabilized steel in these geometries.

Thulium radiography was used for general weld soundness, visual examination ensured the
correct weld profile and a check was maintained on a percentage of the welds and by
destructive examination of production-control welds to ensure that the modified tail-off
technique has eliminated the micro-cracks referred to above.

36
2.2.2. Review of steam generators operating history

First criticality of the PFR reactor was achieved on 3 March 1974. In October 1974, during
early steam commissioning, a leak was detected in superheater 3.

Up to 1976 there were failures in gas-space leaks in one evaporator, two superheaters and one
reheater. These early failures were believed to have been due to manufacturing faults.

A total of 37 gas-space leaks were experienced in PFR SG units in the period 1974 to 1984
with 33 of these occurring in evaporators, 3 in superheaters and 1 in a reheater. All the
gas-space leaks originated at the welds between the tubes and the tubeplates.

In the case of the austenitic superheaters and reheaters, the leaks gave rise to considerable
concerns about the design. Although both the damaged superheaters continued in use up to
1986, having had the leaking tubes plugged, one of the superheaters and the reheater had
suffered from caustic stress corrosion cracking of the tube plate caused by the products of the
sodium-water reaction. The superheater was salvaged by grinding out the cracks and
thoroughly washing the tube plate with hot sodium to remove the reaction products.

Damage to the reheater tube plate was so extensive that the tube bundle was scrapped. It was
replaced by a plug in the empty reheater vessel until a replacement was fitted in 1984. For this
period the plant had to be operated with reduced reheat capacity. Following the early failures
of tube-to-tubeplate welds in the two superheaters and the reheater no further failures
occurred in the austenitic units until 1986, when a superheater tube leaked while the unit was
being pressurised with steam prior to being put on line. This incident is described below.

In the period 19841987 all the six austenitic tube bundles were replaced by new tube bundles
(Fig. 7a). The design benefited from the early experience of caustic stress corrosion following
the leaks in the austenitic units.

This unfavourable experience, so early in the operating life of the steam generating plant, led
to a decision to order a complete replacement set of superheater and reheater tube bundles of
modified design to be fabricated in 9Cr-lMo ferritic steel. As in the original units, the new
units were to have no under-sodium welds in the steam tubes, but, in addition, the need for
difficult tube-to-tubeplate welds, and, indeed, for any welds separating the steam and sodium
environments, was eliminated by passing the steam tube through a sleeve projecting above the
tubeplate, removing the joint to the steam header from the sodium environment.

Each sleeve was brazed to the steam tube and welded to a seal plate (Fig. 7b). The new design
also incorporated improvements to reduce flow induced vibration of the tube bundles. These
units were delivered to site in 1984, at which time it became possible to replace the missing
reheated and thereby restore the full complement of steam generator units. The replacement
work was completed by 1987.

37
 Fig. 7a Fig. 7b
FIG. 7a,b. PFR replacement reheater tube bundle, reheater thermal sleeve.

The 33 leaks experienced in the ferritic steel evaporator units were relatively benign as ferritic
steel is not so subject to caustic stress corrosion, but the effect on availability while leaking
tubes were being repaired was considerable. The evaporator gas space leaks were all
associated with cracking of the tube-to-tubeplate welds. These were hard and had high
residual stresses because there was no post-weld heat treatment. None of the evaporator leaks
gave evidence of wastage damage to the neighbouring tubes, probably because they were
detected early by the installed gas-space hydrogen detection system. This was based on
katharometers and was very sensitive, being capable of detecting leaks as small as 0.1 mg/s.
The leaks were repaired by plugging the affected steam tubes. Nevertheless it appeared that
one leak would, after a few days or weeks of further operation, cause others. It was concluded

38
that residual caustic reaction products in the gas space above the sodium caused further
cracking of welds and initiated more leaks after an incubation period. Sodium flooding of the
tubeplate at a temperature in excess of 400C for periods in excess of 24 hours had some
success in removing reaction products. It was also required to wash out sodium hydrides
which could lead to false hydrogen detection signals when they dissociated at high operating
temperatures. This did not cure the problem completely, however, and eventually it was
concluded that washing with hot sodium did not remove caustic material from the roots of
pre-existing fine cracks in the welds, so that in the presence of the residual stresses corrosion
continued and the cracks grew to give rise to further leaks.

With leaks having occurred in all three evaporators and a sodium-side origin of cracking in
evidence, it was decided that the most practicable solution was to by-pass the suspect tube-to-
tubeplate welds by using a sleeve. Each sleeve was a precisely machined cylindrical tube of a
9Cr-lMo steel inserted into the tubeplate and extending downwards into the top 75 mm of the
tube. The upper end of the sleeve was explosively welded to the tubeplate and the lower end
was braised into the inside of the tube (Figs 8 and 9).

FIG. 8. FIG. 9.

FIGs 8 and 9. PFR evaporator weld repair sleeve and explosive weld examination.

39
Following extensive laboratory trials, four experimental sleeves were first fitted to operational
evaporators towards the end of 1980. A further 11 evaporator weld leaks occurred in 1981 and
a further 41 sleeves were fitted to by-pass defective and suspect welds. In parallel, work was
in progress to examine whether the sleeving technique could be applied on a routine basis.
A trial installation of 200 sleeves was conducted on the spare evaporator tube bundle in the
latter part of 1982. A decision was then made to sleeve all tube-to-tubeplate junctions in the
three evaporators.

The problem was finally solved by the fitting of sleeves which spanned the original welds, as
shown in Fig. 8. In all 3 000 sleeves were fitted over a 14 month period. The work was
completed early in 1984 after a 14 month campaign. While sleeving was underway the station
operated on a single circuit. Following sleeving no further problems were experienced with
the evaporators.

The type of direct tube-to-tubeplate weld (the butt/fillet' weld) adopted initially at PFR,
which could not be heat treated after manufacture, should be avoided in future fast reactors.
Austenitic steels are in question for LMFR steam generators because of the high risk of
caustic stress corrosion damage following even small leaks.

2.2.3. The under-sodium leak in PFR superheater

An under-sodium leak occurred in PFR superheater 2, one of the original units made from
austenitic steel (Fig. 3), in February 1987. It provided valuable information on the behaviour
of sodium water reactions in an operating steam generator and led to a complete reassessment
of the design basis steam generator accident for subsequent fast reactors.

On 27 February 1987 PFR was being operated at full power when a sodium-water reaction
trip was caused by the rupture of a bursting disc on the stem side of superheater 2. This
initiated a dump of the steam and sodium in the secondary circuit and automatic shutdown of
the plant. Shutdown to a safe state took approximately 10 s, as designed. It was confirmed
shortly after the incident that a large under-sodium leak had occurred in superheater 2.

Figure 3 shows one of the original PFR superheaters. After the sodium circuit had been
cleaned to remove reaction products, the superheater tube bundle was removed from its vessel
in a nitrogen-filled bag and examined. This revealed that between two tubes support grids one
of the six baffle plates forming the central sodium inlet duct had become detached, and the
remaining 5 plates in this region were deformed. Considerable distortion of steam tubes could
be seen through the aperture left by the missing baffle plate.

An incident at the PFR steam superheater, however, made to treat more attentively tubes
behaviour analysis in the large water-into-sodium leak zone in a steam generator within a time
interval following the shock load effect. As a result of a leak occurrence in one tube in the
process of the accident 40 tubes had a through damage a guillotine type rupture. According
to the data from material research, the temperature in the reaction zone increased up to
1 300C. In addition to 40 tubes having through damages, about 500 tubes were high
temperature (in excess of 800C) reaction products, on 70 tubes swelling was revealed.

As a result of the analysis of the whole set of data the British specialists presented the
following version of the accident development. The initial leak occurred on the tube no. 16
(Figs 10 and 11).

40
 FIG. 10. The failed tubes in PFR superheater 2.

FIG. 11. The location of fretting marks on the PRR superheater 2 central baffle.

The cause was a crack due to wall wear by about 70% of its thickness. The leak was not
detected because of a lack of hydrogen-in-sodium monitoring, and the system of hydrogen
monitoring in the gas space proved to be too delayed in time (it operated after a rupture of the
bursting disc). Such an initial leak could continue for several hours at a flow rate of 0.001 g/s
and a few minutes at a flow rate of 1.0 g/s. Then the leak passed into an intermediate one and
during about 20 s was at a level of 0.5 kg/s. The consequence of this leak was a guillotine
type rupture of a few adjacent tubes weakened by abrasion wear due to vibration and by their
being in the high temperature zone. The cause of the rupture was a high steam pressure
(13 MPa).

41
The rest of the tubes were damaged in the process of steam generator shut down when after
the cessation of water and steam flow rate the heat removal from the reaction zone was
stopped. The process was aggravated by an inadequate design of the sodium and reaction
products removal system. Altogether, about 150 kg of water penetrated into sodium. Thus it
has been found that at an initial large leak, under some conditions the secondary failure of
heat exchange tubes is possible because of metal overheating.

Experiments carried out by USA specialists have confirmed the above fact. Thus, one of the
tests was carried out at an initial leak of 454 g/s. The first of the secondary leaks appeared on
the target tube in 16 s and was caused both by corrosion and by a rupture under the effect of
internal pressure. The traces of corrosion wastage were found on one more tube. All the other
failures (a total of 23 tubes were failed) were of an overheating-rupture character. It was
recognized that most of the secondary failures were the result of the reaction between steam
leaking from the failed tubes and residual sodium after its draining. An important result of the
experiments is also that despite a considerable number of damaged tubes the acoustic waves
of pressure were determined by the initial large leak.

On the basis of a generalized analysis of the results of the experimental validation studies
carried out on the EFR design, it has been found that the effect of a tube wall failure by the
overheating rupture mechanism is possible at an appearance of the secondary (or of the
primary) leak with a flow rate more than 80 g/s. And in this case the time from the appearance
of this large leak up to a failure of adjacent tubes depends on the leak size. It was shown that
with an increase of a leak from 100 to 4 000 g/s the above time interval reduces from 50 to
4 s. Thus, when developing an approach to the choice of a design basis leak value, the
following should be taken into account:

(1) The rate of hydrodynamic processes in the sodium circuit (and, therefore, the circuit
strength) is determined by the rate of an increase of an initial large leak. It is
expedient, therefore, as an initial leak to take that appearing at an instantaneous
guillotine failure of at least one tube.

(2) Steam generator tube bundle damage zone (the number of heat exchange tubes), as
well as the amount of water penetrating into the sodium circuit (and, therefore, the
scope of repair work) depend on the SG design features and on its safety system
characteristics. It is desirable to aim at the most rapid possible termination of the
sodium-water interaction reaction after the appearance of the secondary (or the
primary) large leak.

Flow induced vibrations were at the origin of the most important under-sodium leak at PFR. Analysis
of the event and examination of the superheater concerned has shown that flow induced vibration of a
single tube resulted in the initial leak, which thereafter developed rapidly. All the original reheater and
superheater tube bundles were replaced with new and much improved units.

The under-sodium leaks demonstrated that it is possible for a large number of tubes to fail due
to overheating in a period of a few seconds. The incident led to a reassessment of the design
basis accident for the steam generators of both PFR and EFR. In the case of PFR the design
basis-accident was changed from a single double-ended guillotine fracture to 40 double-ended
guillotine fractures spread over a period of 10 s.

42
2.3. EFFECTS OF SODIUM AEROSOL DEPOSITION IN LMFRs

Most of fast reactors have to some extent experienced problems related to sodium aerosol
depositions. At BN-350 and BN-600 some deposits were observed in the gap between the
large rotating plug and the reactor vessel roof, inducing difficulties of rotation. At KNK-II
and Phnix, deposits have been found in the control rods between the shielding piston and the
guide tube, causing difficulties with their insertion.

At PFR, the possibility of sodium aerosol deposition in those parts of the absorber drive
exposed to the primary vessel gas blanket was recognized in the design and an argon purge
flow was provided in these areas to prevent aerosol transport. However, it was not completely
effective and the removal of small deposits on the magnet faces in the drive mechanisms was
a routine procedure at PFR. This led to plant trips on a number of occasions due to absorbers
dropping off their magnets during power operation.

The question of difficulties with the PFR control rods has been examined in two phases.
During the first phase (up to late 1985) the results gave no cause for concern. Since 1985, the
reactor operated at high power and high temperatures and greater problems were encountered.

It was believed at the beginning that sodium deposition was a main cause and after
investigations confirmed this. Particular features to note are the roof insulation region, the
weight sensing equipment and the dashpot. Initially, it was thought that the main issue
concerning the freedom of rod movement would be bowing caused by neutron induced
voidage (NIV) coupled with thermal effects. The computer code Peeble predicted maximum
friction of 25 kgf. However actual measured maximum friction was 40 kgf and the distribution
of measured friction with rod position was not in good agreement with the Peeble prediction.

PFR had 5 shut-off rods (normally fully raised) and 5 control rods inserted to control power.
The rods were essentially identical B4C assemblies supported by electromagnets (Fig. 12).

On a trip all ten rods dropped. Magnet current, apparent rod weight, rod release time and time
of flight were measured by installed instrumentation

At all shutdowns and after plant trips the electromagnet pick-up and drop-off currents were
measured. These were the minimum magnet currents at which the absorber could be raised
and at which it dropped off after being raised. On the basis of these figures a decision was
made on whether the magnet faces had to be cleaned before return to power. If required the
drive and magnet assembly were removed by simple bagging techniques and the magnet face
was cleaned in an argon purged glove box. The extension rod face was cleaned in situ using
commercial Scotchbrite cleaning pads, again making use of a simple bagging technique.

After late 1985, when the high power operation started a sudden increase of friction was
measured with values going up to 80 kgf. At this time the possibility of sodium aerosols in the
cover gas and deposition on the keys and key ways in the upper part of the mechanism were
discussed. As an explanation it is suggested that sodium has deposited in the positions
indicated in Fig. 13.

43
 FIG. 12. The PFR absorber rods.

Deposits in these positions could give the distributions of friction measured, largely as a result
of the detailed design of the keys which are attached to the Latch/Delatch tube and which
engage in key ways in the extension rod. In 1988 a special glove box was made which
allowed examination of the liner tubes and the extension rods. Examination of a number of
rods confirmed that sodium deposits were present but in smaller quantities than expected and
confined to the keyway of the extension rods. None were found on the liner tube as originally
hypothesised. The sodium was soft and easily removed. Although the absence of deposits
other than in the keyways was surprising, when they were removed the friction of the restored
rods to normal. It took some 40 effective full power days (efpd) of operation for friction levels
to begin to rise noticeably.

A difficult problem in operation was posed by sticking in the shutdown systems of KNK-II
(Fig. 14).

44
 FIG. 13. Location of sodium aerosol deposits on a PFR absorber rod mechanism. FIG. 14. Schematic representation of the shutdown rods in the
KNK-II shutdown systems.

45
In December 1986, a control rod of the primary shutdown unit for the first time was found to
stick while the reactor was shutdown. The cause was found to be sodium aerosols plated out
during prior handling steps, when the rod actuating equipment had not been swept with gas.

In December 1988, deposits were found on a rod of the secondary shutdown system; they
impaired the mobility of the component, but not the shutdown function. Probably the fact that
the primary system had been opened for maintenance purposes a number of times before had
caused the quality of the cover gas to deteriorate and thus produced the deposits. On January
1991 the scram at 15% power reactor operation was initiated by a sudden absorber movement
after an obstruction in movement had been overcome. This blockage in the primary shutdown
system again was caused by depositions in the rod actuating equipment in a phase in which
the cover gas quality had been insufficient.

When the quality of the cover gas is insufficient, the sodium in the rod actuating equipment
was oxidized to sodium oxide whose dough-like consistency impeded lifting movements of
the equipment. This blockage in the primary shutdown system was caused by depositions in
the rod actuating equipment in a phase in which the cover gas quality had been insufficient.

Movable parts within the primary reactor envelop, might be exposed to cover gas carrying
considerable amounts of sodium aerosols, have to be protected by structural elements like
expansion bellows or double seals with cover gas in between.

If narrow gaps between fixed and moving parts within the reactor vessel gas plenum are
unavoidable and venting gas has to be applied, measures for the flow and quality surveillance
of such venting gas are essential. In case of poor gas quality, caused by impurities like
hydrogen, oxygen or methane, a possibility to switch over to clean gas from the liquid gas
storage tank has to be installed. Depending on certain circumstances the gap has to be chosen
such as to minimize possible convection of sodium aerosols.

Solving the problem of aerosol deposition is a difficult one because it is related to gas
convection and to the temperature of the different structures. In the frame of the new project,
an important effort is being made for the development of a high performance calculation code.
If needed, design provisions should be incorporated such as tight annular spaces for the
rotating plugs, large clearances between moving parts, heating devices to avoid solidification.
Also, gas injections should be limited to a minimum in order to avoid oxide formation and
deposits on moving parts. The experience with short-circuit devices for the detection of
sodium leaks is good. Nevertheless, an important effort is still necessary to improve their
sensitivity and reliability.

One solution is to use metallic (ferritic) reactor roof in which the component penetration are
machined, the tolerances are more controlled, so smaller clearances are possible. It gives a
further benefit of lower a sodium aerosol and heat transfer from the cover gas.

2.4. THE PRIMARY CIRCUIT OIL SPILL

The PFR primary sodium circulation system is shown in Figs 1 and 15. Figure 16 shows
details of a primary sodium pump (PSP). The sodium from each PSP flows through filters and
a stop valve to the diagrid, and thence to the fuel subassemblies. Each subassembly has a filter
at its inlet. Figure 15 shows the relationship between the pump and subassembly filters. In
1974 primary sodium pump 2 (PSP 2) was removed from the reactor for modifications to its
instrumentation and was noticed to be heavily contaminated by a black sooty deposit. In the
same year the charge machine was removed, revealing that its immersed surface was black

46
with adherent tarry lumps. During this period of operation some 65 L of oil had been lost
from the pump upper seal oil systems, part of which is believed to have entered the reactor
vessel. When the reactor was taken critical no effects of the oil were observed. It is suspected,
however, that as a result of the spill the filter on PSP 2 valve failed due to high differential
pressure because it became blocked by oil-sodium reaction products (O in Fig. 15). It is also
thought that partial blockage of the pump casing overflow pipe was to lead to the major
problem in 1991 (Fig. 16).

FIG. 15. PFR primary sodium pump and its associated valve and filter assembly.

47
 FIG. 16. Detail of a PFR primary sodium pump housing.

At the end of 1991, PFR had been shut down since 29 June when the reactor had been
manually tripped following observation of overheating of the top bearing of PSP 2.
An extensive examination of the circumstances which led to the leakage of primary pump
bearing oil into the primary sodium circuit, has concluded that it involved two stages, the first
on 25 June 1991 and the second on 29 June 1991. Prior to these events:

There was a maximum indicated level of 17 L of oil in the pump drains tank above a
in layer of sodium/oil sludge;
On 24 June 1991 the main argon gas blanket cover gas flow to the reactor dropped to
zero due to blockage or partial blockage of the two absolute filters and/or the drain
line associated with the aerosol filter;
On 25 June 1991 efforts were made by the operators to restore the cover gas flow by
venting the system through the gaseous effluent system.

48
Venting the reactor cover gas system to the effluent system reduced the cover gas pressure in
the pump support vessel, allowing sodium to rise up the annulus between the pump drive shaft
and the drains tanks inner wall and enter the drains tank. This sodium would displace oil to a
level above the upper lip of the drains tank. Restoration of the cover gas blanket pressure on
25 June then forced the 17 L of oil down the annulus between the pump shaft and the drains
tank into the pump support vessel sodium. Over the next few days, a further 5.5 L of oil which
were added to the seal oil sump could have reached the sodium in the pump vessel. A sudden
loss of 12.7 L of oil from the pump lubrication system sump occurred on 29 June. This
activated the loss-of-sump oil level alarms in the control room and led to the observed
overheating of the top bearing which initiated the decision to manually trip the reactor.
Examination showed no oil external to the system leading to an estimated total of 35.2 L of oil
ingress into the primary sodium.

The remainder of 1991 had been spent in cleaning the primary sodium and in making
preparations for examination of the three primary pump filters and the inlet filters on some of
the fuel assemblies which had shown temperature increases at the time of the spillage.
Removal of the pump filters was a major task which had never been attempted before.

All three primary pump filters were removed for examination. Deposits incorporating some
carbon were found on all three. Some difficulty was experienced in refitting them because of
distortion of some thermocouple guide tubes and it was necessary to cut these away. The
success of the overall operation showed that the opacity (and, as it was the primary circuit, the
activity) of the sodium coolant is not an insurmountable impediment to maintenance work.
Examination of a number of fuel assemblies also showed deposits on the inlet filters of those
which had shown temperature increases prior to shutdown but none on assemblies which had
shown no temperature increases. However, the deposits were slight (about 0.5 g) and mainly
of sodium. It was improbable that these were sufficient (even allowing for the possibility of
some loss during handling) to cause the observed temperature rises and it is possible that
roughening of the cladding surfaces by adsorption of oil degradation products might have
been a contributory factor.

In parallel with these out of reactor examinations of fuel assemblies, a device was developed
which would measure flow through each of the remaining fuel assemblies and through some
of the blanket assemblies. Use of this device showed acceptable flows through all of them.
Work on the reactor was supported by laboratory studies to examine the effects of
temperature, sodium and irradiation on both new and degraded oil. These included tests on the
PROTVA rig at the IPPE at Obninsk where oil was injected into sodium upstream of a PFR
fuel assembly inlet filter, and studies on a National Nuclear Corporation (NNC) water rig to
examine the blockage-forming potential of material similar to that observed in PFR.

A major effort was required to remove all three valve and filter assemblies from the reactor
for examination. These were the longest components in the reactor vessel, at 12 meters, and
required considerable care in handling. Examination showed that at least one panel of each
valve filter had failed, and oil-related debris was found on all the filters. A number of the fuel
subassemblies which had showed outlet temperature rises during the incident were removed,
and oil-related debris was found on their inlet filters and wrappers. The result of the oil
ingress was an 18-month shutdown while PSP valve and filter assemblies were removed and
new filters were fitted. The pump seal oil systems were modified to prevent any further
possibility of oil ingress, and alarm and trip systems were added to prevent blockage of the
pump filters in order to protect the subassembly filters.

49
Very fine particles of carbon were found in primary sodium samples after 1974. These are
believed to have come from the 1974 oil ingress, and it appears that in the long term oil debris
breaks down into finely-divided carbon particles which are dispersed in the sodium, pass
through the filters, and circulate without obvious effect.

Oil ingress into the primary circuits of an LMFR is undesirable because of the potential
release of methane gas through the core causing reactivity effects, and possible blockage of
the subassemblies by solid carbon debris. In the case of PFR no reactivity effects were seen,
possibly because the oil was retained in the pump cone for a prolonged period and broken
down slowly without the formation of large bubbles. In the long term oil bearings are
probably best avoided. The EFR design was changed following the PFR oil ingress incident
by the introduction of the innovative features of magnetic bearing and ferro-fluid seals to
eliminate oil completely and remove the potential hazard of its ingress into the sodium.

2.5. CRACKS IN THE PFR AIR HEAT EXCHANGERS

In the decay heat removal system of PFR, leaks were detected in the air heat exchange (heat
exchange between the NaK circuit and the atmosphere). These were associated with
anomalous temperature differences between tubes in the heat exchanger, due to aspects of the
design together with difficulties in achieving filling with NaK. PFR had three thermal syphon
decay heat rejection loops (Fig. 17).

FIG. 17. Schematic diagram of a PFR thermal siphon decay h eat rejection loop.

The ten tons of NaK were contained in this system. Each consisted of a NaK filled loop
connecting a heat exchanger coil positioned in the main reactor vessel adjacent to an
intermediate heat exchanger to an air heat exchanger (AHX) on the roof of the reactor
containment building. In the event of loss of electric power supplies each loop was capable of
removing 1.5 MW of decay heat from the reactor by natural convection.

50
Each AHX was equipped with 2 fans connected to emergency diesel power supplies, which
could enhance the decay heat removal to over 4 MW per loop. When the rector was operating
normally heat removal was limited by dampers which restricted the airflow to the AHXs.
Although the thermal syphon system operated well by 1984, it had become apparent that the
AHXs suffered from a systematic fault leading to failures and leaks. Each AHX consisted of
forty serpentine parallel tubes welded to pulled tees in two headers, as shown in Fig. 17.
The tubes were finned along the straight lengths but plain at the bends, which were clamped
together and supported. Further rigidity was provided by cleats which were welded to the tops
of the fins on adjacent tubes. Flow of NaK was from the top down. Leaks were occurred at the
welds between the tubes and the pulled tees in the headers. As an interim measure operational
constraints were imposed as the frequency of failures could have invalidated the risk analysis
in the safety report, and hence jeopardised the authorisation to operate the plant. Meanwhile
the AHXs were heavily instrumented with strain gauges and thermocouples to identify the
cause of the problem and indicate a solution.

The measurements indicated that the problems occurred essentially because the AHX tubes
were in parallel, and were horizontal with no fall to ensure good filling. When the AHXs were
filled, gas locks were occurring at the pipe bends. The gas-locked tubes remained cold, and as
a result oxide impurities could be precipitated causing permanent blockages. Because of
temperature differences between a cold blocked tube and the adjacent hot tubes to which it
was clamped, large stresses were imposed. As a result the weakest point in the system, the
weld between the tube and the header, was stressed, suffered cracking and eventually leaked.
Replacement AHXs (RAHXs) were manufactured to an improved design which avoided the
problem of gas locks and afforded greater toleration of loss of flow in individual tubes. The
new RAHXs were fitted in 1986/1987. Operation was trouble free until 1996, when the
thermal syphons were finally emptied for decommissioning. The following changes were
made and are shown in Fig. 18.

A 2 slope was given to the tubes to give better venting and drainage;
Each tube was given individual support;
The tube-header connections were reinforced;
Larger diameter headers were fitted to give better NaK distribution.

FIG. 18. The original and replacement PFR thermal syphon air heat exchangers.

51
In 1984 a common mode failure problem in the PFR thermal syphon AHXs was jeopardising
the plant authorisation. A rapid research, development, manufacturing and installation
programme solved the problem by 1986. It is notable that the design feature essential to
solving the problem was the inclusion of a simple 2 slope on the tubes.

2.6. SODIUM MIXING PROBLEMS AND FLOW AND MECHANICALLY INDUCED

VIBRATIONS

At PFR, whenever it was necessary to operate on less than three secondary circuits, there was
a small flow of sodium at core outlet temperature past nominally shut sleeve isolation valves
in the intermediate heat exchangers which are not in service. This provides the potential for
mixing of sodium flows at core inlet and outlet temperatures and hence possible thermal
cycling damage to the heat exchangers and containing pod structures. These effects were
extensively studied in laboratory rigs and resulted in constraints on the core temperature rise
when operating on less than three circuits. The performance of the intermediate heat
exchangers has been examined both theoretically and experimentally to establish a
satisfactory position.

Flow induced vibrations, together with mechanically induced vibrations have been an
important cause of unplanned outage. Two of the PFR failures (one primary and one
secondary pump) occurred during early commissioning. The primary pump failure resulted
from incorrect finish machining of the shield plug, causing shaft distortion and damage to the
hydrostatic bearing. Two secondary pump failures, the latest in April 1984, were due to
hydrostatic bearing seizure caused by detachment of part of the bearing surface or a
foreign body in the sodium. Table 1 summarizes the experience with sodium pump
performance at PFR.

TABLE 1. SUMMARY OF THE EXPERIENCE WITH SODIUM PUMP

PERFORMANCE AT PFR

Operating hours Failures Failures

Primary Secondary Primary Secondary
405 965 231 960 1 2

2.7. THE EFFECT OF NEUTRON-INDUCED DISTORTION OF CORE

COMPONENTS

The core is divided into modules of six subassemblies, each with an associated control rod or
support position. The diagrid is provided with a hole at each one of these module positions
into which a pressure chamber, described as the fuel element carrier, is fixed by a single bolt.

The fuel element carrier is shop-fabricated and can be removed from the reactor relatively
easily should it be damaged. In the central position of the fuel element carrier, a support post
has been fixed which is usually referred to as the leaning post and it carries some support pads
at a position level with the bottom of the axial breeder of the core where the damage flux is
relatively modest. In each carrier, there are six upper and lower bearings which carry the
support spike of the six subassemblies. The lower bearing is offset outwards about 1/8 in. and
the bottom spike of the subassembly consists of an upper bearing fixed rigidly to the
sub-assembly envelope and a lower bearing mounted on an elastic cantilever.

Figure 19 shows schematically the support system for the PFR core. As the subassembly is
lowered into the fuel element carrier, the lower bearing of the subassembly is pushed across

52
by the offset lower bearing of the fuel element carrier and this forces the subassembly against
the pad provided on the leaning post. The pad loads are set high enough to prevent the
subassembly vibrating due to the coolant forces generated at the subassembly exit. The lower
bearing offset is, therefore, determined by this required load and the sum of all the tolerance
effects, such as sub-assembly bow. The upper bearing is provided with a step a few
thousandths of an inch high to ensure that the subassembly pivots correctly about this bush.

FIG. 19. PFR fuel subassembly support system.

Radiation damage resulting from the high neutron fluxes and operating temperatures of a fast
reactor can give rise to dimensional changes in core components. The mechanisms involved
are swelling caused by neutron-induced voidage (NIV) and radiation creep. These phenomena
affect core components by causing axial extension, bowing in transverse gradients of neutron
flux or temperature, and dilation. NIV was first detected during post-irradiation examination
of components from the Dounreay fast reactor (DFR) in 1965. PFR had been designed in
1963 without taking account of the need to accommodate the effects of NIV.

In consequence calculation routes had to be developed to predict the distortion of PFR core
components so that they could be managed in such as way that operation would not be

53
impeded. In particular it was essential to be able to ensure that no core component was at risk
of becoming so distorted that it interfered with the movement of the absorber rods or could
not be removed. The calculations, including the important effects of interaction between
components, were based on empirical material deformation rules obtained from post
irradiation examination of irradiated components. They were successful in guiding operations
except when problems arose due to unexpectedly rapid growth of particular materials.

It was necessary to predict the bowing of fuel subassemblies in order to prevent handling
problems. The operating limit was 14 mm bow at the subassembly shoulder. Bows beyond
21 mm at the subassembly shoulder would have presented difficulties when it came to
extraction from the core. Subassemblies were routinely rotated through 180 part way through
their residence in the core in order to correct the bowing.

Immediately before refuelling operations in PFR three sweep arms were employed to ensure
that there were no obstructions above the core which would prevent rotation of the rotating
shield. At the start of a reload in 1988 the sweep arms were found to contact or partially
contact objects in two core positions. These positions were identified as containing
subassemblies with cold-worked EN58B steel wrappers, with a calculated dose of greater than
60 displacements per atom (dpa).

Using a special tool the heights of all subassemblies of the same material were checked.
A distinct trend of rapid increase of growth at doses above 50 dpa was revealed, although not
all subassemblies were affected. Two subassemblies in particular, JRA and GYN, had
measured growths of about 40 mm. As a result all components with predicted doses likely to
exceed 50 dpa by the end of the next run were removed from the core. Considerable difficulty
was experienced in handling the severely distorted components and special tools had to be
manufactured for their extraction and removal from the reactor.

Although the absorber rods and associated components in PFR were manufactured from
nimonic PE16, an alloy known to be subject to low swelling, it was important to ensure that
NIV distortion would not prejudice operation of the system or hinder rod drop in a scram.
The major cause for concern was distortion of the guide tube in which the absorber rod
moved, either by NIV bowing or by pressure on it from adjacent bowed fuel subassemblies.
In addition to the calculations, regular exercising of the absorber rods over their full stroke
gave assurance that no such problems were arising.

On only one occasion there were observable effects in the operation of an absorber. In this
instance a shut off rod developed unusually high and increasing friction at the top of its stroke
while being exercised. Although the rod operated correctly during subsequent trips, indicated
that the problem was probably caused by interaction of the guide tube with an adjacent
distorted fuel subassembly.

The subassembly was discharged, and PIE confirmed the analysis. It was another
subassembly clad in cold-worked EN58B, with higher than expected swelling. The allowed
doses for EN 58B was reduced to prevent further problems of this sort. Large differences in
NIV swelling rates could occur in different batches of the same material.

This led to handling problems in the case of components made of cold-worked EN58B.
Materials chosen later in the lifetime of PFR, such as nimonic PE 16, had considerably lower
swelling rates. Components manufactured from the ferritic steel FV 448, which was under test
at the time of PFR closure, had extremely low swelling rates. NIV distortion was not expected
to be life-limiting for this material.

54
2.8. FUEL DEVELOPMENT

The first fuel assemblies containing plutonium from PFR recycled into new oxide were
loaded into PFR in June 1982, thereby closing the reactor fuel cycle. In March 1994, the
reprocessing plant had treated a total of over 23 tons of oxide fuel, recovering more than
3.5 tons of plutonium, from 239 fuel assemblies, 144 mixer breeders and 10 radial breeder
assemblies. The highest burnup in the fuel was 17.6%. Typical cooling times were
270-360 days, but the shortest cooling time was 136 days when some short-cooled fuel was
reprocessing experimentally to study movements of iodine species. The plant has been
operated since 1980 in campaigns to keep pace with fuel discharges from PFR. The highest
throughput in any year so far was 4.79 tons in the period from April 1993 to March 1994.

The PFR fuel reprocessing plant proved the technical feasibility of oxide fuel reprocessing via
a Purex cycle, with recovery of over 99.5% of the plutonium. This high recovery was also
reflected in the low amounts of plutonium in the liquid and solid waste streams from the
plant. The amount of radioactivity discharged to the environment was always about an order
of magnitude less than the licensed limits. The plant is subject to IAEA and Euratom
safeguards.

One of the principal tasks of PFR was to demonstrate a reliable, safe and robust fuel capable
of routinely achieving a high burnup target. Successful completion of this task was one of
PFR's major achievements. The major advances made in the second decade of PFR operations
resulted from the introduction of Nimonic PE16 as the reference cladding alloy. The fuel
assembly discharged at 15.9% burnup in 1986, mentioned earlier, gave the first indications of
the benefit of this change. The pins showed maximum diametral increases of only 1%, with
uniformly low diametral change profiles showing little pin-to-pin variability, compared with
the 5-8% (maximum) diametral changes and highly peaked profiles showed by earlier
examinations of pins clad in cold-worked M316 steel and irradiated to half the exposure.

Destructive examination of the pins indicated that the fuel column was stable, that internal
corrosion was low and that there was no evidence of any fuel/clad mechanical interaction
resulting from containment of a high burnup swelling fuel in a non-distending cladding tube.
This confirmed that higher burnups were probably feasible. Measurements of the PE16
wrapper showed trivial length increase, across-flats distension and bowing, and revealed no
potentially life-limiting changes. By 1990, irradiations of PE16 clad fuel pins in driver
assemblies and in experimental subassemblies (Sas) had achieved more than 17 and 21%
burnup, respectively. Even with displacement doses of the order of 130 displacements per
atom (dpa) no life-limiting features could be identified in either the wrappers or the cladding.

Meanwhile materials irradiation experiments had shown ferritic/martensitic steels to be

particularly resistant to void swelling. The high temperature mechanical strength of these
materials seemed to preclude their use as pin cladding. The application as wrapper materials
seemed to be practicable. Accordingly, in the late 1980s driver fuel assemblies with wrappers
of the ferritic/martensitic steel FV448 and pins clad in Nimonic PE16 were introduced into
PFR. By the end of the reactor's working life, over 20 such assemblies had been loaded and
seven of these had exceeded the 15% burnup (110 dpa) target; one had achieved the then
world record, for a mixed-oxide driver charge assembly, of 19.8% burnup (155 dpa). Statistics
illustrate the undoubted success of the fuels development program in PFR. Approximately
98 000 pins were irradiated and, of these, over 40 000 exceeded the original 7.5% target
burnup. The introduction of PE16 as cladding allowed over 2 400 pins to attain burnups in
excess of 15% with about 320 of these having successfully exceeded 20% at the end of
operations. The peak burnup (in lead pins irradiated in an experimental SA) was 23.2%.

55
 BIBLIOGRAPHY

ADAM, E.R., GREGORY, C.V.G., A Brief History of the Operation of the Prototype Fast
Reactor at Dounreay, the Nuclear Engineer, 35 (1994) 112117.
ANDERSON, A., et al., LMFR Steam generators in the United Kingdom, paper presented in
Mtg. Maintenance and Repair of LMFBR Steam Generators, 48 June 1984, O-Arai, Japan.
BOARDMAN, C.E., HUI, M., NEELY, H.H., Test results of sodium-water reaction testing in
near prototypical LMR steam generator, paper presented in Specialists Meeting Steam
Generator Failure and Failure Propagation Experience, 2628 September 1990,
Aix-en-Provence, France.
CURRIE, R., LINEKAR, G.A., EDGE, D.M., The under-sodium leak in the PFR superheater
2 in February 1987, paper presented in Specialists Meeting on Steam Generator Failure and
Failure Propagation Experience, 26-28 September 1990, Aix-en-Provence, France.
DUMM, K., EFR-600 MW Straight tube steam generator. The strategy towards the definition
of a Design Basis Accident, paper presented in Specialists Meeting Steam Generator Failure
and Failure Propagation Experience, 2628 September 1990, Aix-en-Provence, France.
INTERNATIONAL ATOMIC ENERGY AGENCY, Status of liquid metal cooled fast
breeder reactors. Technical Reports Series No.246, IAEA, Vienna (1985).
INTERNATIONAL ATOMIC ENERGY AGENCY, Status of liquid metal cooled fast reactor
technology, IAEA-TECDOC-1083, IAEA, Vienna (1999).
INTERNATIONAL ATOMIC ENERGY AGENCY, Fast reactor database, IAEA-TECDOC-
866, IAEA, Vienna (1996).
INTERNATIONAL ATOMIC ENERGY AGENCY, Problems experienced during operation
of the prototype fast reactor, Dounrey, 1974-1994, IAEA-TECDOC-1180, IAEA, Vienna
(2000).
JUDD, A.M., et al., The Under-Sodium Leak in the PFR Superheater 2, February1987,
Nuclear Energy, 31 (1992) 221-230.
KANTREY, L., Engineering components for sodium-cooled fast breeder reactor, paper
presented in Specialists Meeting Steam Generator Failure and Failure Propagation
Experience, 26-28 September 1990, Aix-en-Provence, France.
LUDWIG P, W., Conclusions from the sodium - water Reaction experiments performed with
straight Tube bundle model for steam Generator with respect to the Calculation Method of the
Accident Design Pressure, paper presented in Study Group Meeting Steam Generators
LMFBRs, 14-17 October 1974, Bensberg, Federal Republic of Germany.
SACANO, K., Large-leak sodium-water reaction analysis for steam generator, paper
presented in Study Group Meeting Steam Generators LMFBRs, 1417 October 1974,
Bensberg, Germany.
TAYLOR, D., Prototype fast reactor heat-transport system, paper presented in Symp.
Sodium-cooled fast reactor engineering, 2327 March 1970, Monaco.
WATKINS, B., et al., The use of stabilized steels in the PFR circuit, paper presented in Symp.
Sodium-cooled fast reactor engineering, 2327 March 1970, Monaco.

56
3. PHENIX AND SUPER-PHENIX REACTORS

3.1. PHENIX REACTOR

3.1.1. Commissioning and design features

A description of the French demonstration prototype sodium cooled fast breeder reactor
Phnix has been given in an earlier IAEA publication and a comprehensive listing of design
and operational parameters is presented in another. Therefore, it is not proposed to give an
extended description of the plant here, three figure show the general arrangement of the
reactor: isometric view, the elevation through Phnix primary circuit and flow diagram,
describes below.

The reactor plant Phnix with a nominal ~ 255 MW(e) power rating (565 MW(th)), was
firstly connected to the electricity grid on 13 December 1973; the nominal power was reached
on 12 March 1974, 18 days ahead of plan.

The nuclear power plant (NPP) construction cost was 800 million franc (approximately
3.2 billion franc 2000, that is equivalent to ~ 512 million Euro). The planned budget was
exceeded by less than 10%. The NPP was generally operated at the power tolerated by
reactor and equipment, with comparatively high load factor. Phnix has currently provided
about 100 000 hours of grid-connected operation representing 3 900 equivalent full power
days at operating temperatures of 560C for the reactor hot structures. The plant has achieved
the objectives of demonstration of fast breeder reactor technology which were set at the time
of construction, including the following significant achievements:

An average burnup was increased from 6%1 heavy atoms (h.a.) to 11% h.a. at the
center of the core and to 13.5% h.a. at the periphery of the core. These levels were
reached with eight cores of fuel which was 170 000 fuel pins; experimental pins and
subassemblies have reached maximum burnup exceeding 17% h.a. and an irradiation
dose of 150-160 dpa (displacement per atom) on the fuel cladding;
The measurement of the breeding ratio made at the time of dissolution of the fuel
evacuated from the plant, gave a true value of 1.16 (projected 1.13);
The fuel cycle, based on mixed oxide fuel and PUREX reprocessing, has been closed
and the first fuel subassembly made with reprocessed plutonium was loaded in the
reactor in January 1980. About of five of Phnix cores (~ 25 tons) were reprocessed.
The successful and regular operation at the highest temperatures and nominal power
until 1990 resulted in validation of the pool concept option and did much knowledge
regarding the high temperature design and structural material of fast reactors;
The highest in the nuclear power engineering practice gross/net plants thermal
efficiency of 45.3/42.3% during the periods of stabilized operations with nominal
parameters. On the average, gross/net plants thermal efficiency is equal to 40/38%
owing to operations at 2/3 the rated power now and then.

From 1992, the role of Phnix as an irradiation facility has been emphasized, particularly in
support of the CEA R&D programme in the context of line 1 of the 30 December 1991 law on
long-lived radioactive waste management. The first experiment, called SUPERFACT, led to
the incineration of minor actinides (neptunium and americium). This programme was further
strengthened in 1998, to compensate for the shutdown of Super-Phnix. It involves

1
Planning at the design stage the high burnup was a big risk owing to stainless steel fuel pins cladding swelling
observed in the Rapsodie in 1971.

57
transmutation of minor actinides and long-lived fission products. Since 1993, the reactor
power has been limited to 350 MW(th), 145 MW(e) on two secondary loop operations.

3.1.1.1. Reactor block (Fig. 1a,b)

The Phnix reactor block is of an integrated (pool) design except for a few auxiliary circuits.
The entire primary sodium system, containing 800 tons of radioactive sodium, is enclosed in
the main reactor vessel. The reactor block is suspended to the slab via 21 hangers. These
hangers have three welds, which are difficult to access.

The reactor vessel dimensions: inside diameterheight: 11.8212 m, wall thickness: 15 mm is

under the upper slab, 15 m in diameter and 1.5 m high, on which the equipment is installed.
The slab (~ 800 tons: 200 tons metal and 600 tons concrete) is installed on 22 ball- and -
socket type pads. The vessel cover plate is under high temperature conditions, similar to the
BN-600 reactor, but in this case it is unloaded. The slab bearing plate operates under low
temperature because between the plate and the reactor cover there is insulation and besides
this zone is cooled. The reactor vessel was manufactured on site workshop using 15 to 25
mm cynical sheets after prefabricated elements were examined in detail.

The pumps and heat exchangers are located on movable sliding support, sealing of the pump
and heat exchanger penetrations is carried out by means of the bellows. A peculiar feature of
this reactor top design is a massive cover plate, of 60 mm thick and 12 000 mm in diameter
having significant thermal inertia. It restraints a power growth rate or the reactor starting up
and may be a cause of some thermal stresses occurring in the junction of the vessel cylindrical
part and the horizontal cover.

The sodium in the vessel is separated into two zones:

Hot pool at the core outlet where the hot sodium flows into the intermediate heat
exchangers;
Cold pool taken from a peripheral annular space between the primary tank and the
wall of the main reactor vessel, which contains the three main circuit circulating
pumps and six heat exchangers, suspended from the upper slab;
The leak tightness of the penetrations of the reactor vessel by the IHX is provided by
an argon seal.

A number of other devices are located in the main vessel: the fuel transfer arm, the six control
rods, neutron flux detectors, thermocouples, failed fuel detection and location devices, the
core acoustic detection system components, etc. An argon gas atmosphere is maintained
above the sodium surface to prevent any contact with air.

The main vessel is closed at the top by a flat roof with openings for pump and heat exchanger
pipes. It is associated with the cylindrical seating of a rotating plug in the slab penetrations
forming the top of the reactor block. An outer guard vessel surrounds the main vessel. It has
the double function of containing any sodium escaping by leakage, and preventing a drop in
the sodium level of the main vessel which might affect core cooling.

58
 1

5
6

1-manhole ( =5 00mm), 2-movable ladder, 3-catwalk, 4-fixed ladder, 5-

conical skirt, 6-cable tracks
FIG. 1a. Phnix reactor block: isometric view. FIG. 1b. Elevation through Phnix primary circuit.

59
3.1.1.2. Heat transfer circuits (Fig. 2)

Owing to an intermediate circuit between the reactor and the steam generator, it is very likely
to prevent an accidental interaction between the radioactive primary sodium and the
water/steam in the electricity generating system. In such nuclear steam supply system design,
the incident-secondary sodium-water reaction could be classified as a chemical incident in the
non nuclear components.

The three primary sodium pumps are variable speed units (150 to 970 rpm) delivering about
950 kg/s at 825 rpm, which is their normal service speed. The circulating sodium enters the
core at 400C and moves from there, at 560C, to six intermediate heat exchangers which are
connected in pairs with three independent secondary loops.

Sodium must be kept very pure to prevent corrosion of the steel piping and plugging of circuit
components. It is purified by cold traps operating on the principle of precipitation of any
oxide in the sodium at low temperature.

Secondary sodium, which is not radioactive, is circulated by a mechanical pump with a flow
delivery of 700 kg/s. It enters the intermediate heat exchangers (IHX) at 350C and leaves at
550C. Each secondary loop is connected to a steam generator consisting of an evaporator,
superheater, and reheater, in 12 modules for each stage.

3.1.1.3. Reactor core

The reactor core, in which most of the reactor power is generated, is surrounded by a fertile
blanket and neutron shielding to prevent activation of the secondary sodium flowing through
the intermediate heat exchangers. The fuel is uranium dioxide mixed with plutonium dioxide
(UO2-PuO2). It is contained in 103 subassemblies, each containing 217 pins, which in turn
consist of a stack of sintered oxide pellets, 5.5 mm in diameter, enclosed in a stainless steel
cladding.

The pins are assembled in clusters in a stainless steel outer shell, which also contains the
upper and lower fertile blanket pins (depleted uranium oxide) and the upper neutron shielding.
The radial blanket is composed of depleted uranium dioxide pellets measuring 12.15 mm in
diameter, in 90 assemblies of 61 pins each. The structural components of these subassemblies
are identical with those of the fissile subassemblies, with sodium flow through the spike
inserted into the diagrid.

The first fuel load consisted of 50% MOX and 50% enriched UO2, during the reloads the
proportion of MOX fuel increased and after four years of operation were 100% MOX.

During the first two years operation was interrupted mainly by fuel reloading and by minor
incidents including the core seven fuel pin failures. The pin failures were easy to locate by
wet sipping at the subassembly outlets and less than three days loss of energy production.

60
 FIG. 2. Phnix flow diagram.

61
3.1.1.4. The emergency cooling system

An outer leak jacket, which was made of carbon 18 mm thick cynical shells aims at
containing any radioactive products that might escape from the main vessel in the event of an
accident. This containment is cooled by a water circuit which maintains the concrete of the
reactor block at a low temperature, and which is capable of acting as a standby cooling circuit
for decay heat removal after shutdown, should all the secondary sodium circuits being out of
service.

Under faulted conditions the decay heat is removed from the double wall reactor tank to a
third, so called safety tank, which has a water-filled tube coil system on its outer surface.
The 12 cooling sub-circuits with a total length of 4 300 m of pipes were welded on the outside
surface. To improve the heat transport between the double tank and the safety tank it was
decided to replace the nitrogen by helium totally or partly. The emergency cooling system
comprises two independent circuits, each supplying water to half the tubes and coils, three
pumps, and two heat exchangers.

The two circuits are isolated from one another by a single valve. Of the three pumps two are
normally operating, with the third on standby. In the event of loss of off-site power, the pumps
are powered by the diesel generators after 3 seconds. Heat removed by the coolant is
transferred to raw water circulating in heat exchangers. Lately, during inspection, repair and
safety upgrade the two independent heat exchange circuit (air/water) for decay heat removal
in the event of loss normal means have been installed and tested; a 1 500 kW air cooler and
720 kVA stand alone added.

The emergency cooling circuits have a secondary permanent function: during normal
operation, treated water is circulated through them to remove heat from the cylindrical part of
the safety vessel and its lower head, the roof slab including penetrations, the primary
containment concrete penetrations for auxiliary system pipes, and the upper part of the vessel.
The following major incidents or unforeseen events are shown in Table 1:

Series of leaks of secondary circuit sodium in the IHX;

Sodium-water reaction in the steam generators;
Negative reactivity trips;
Cracking of welded joints on certain parts on the main secondary pipes and some
components, particularly in austenitic steel like 321 stabilized with titanium over
about 30 years operation in three positions:

19731990 Demonstration of fast reactor technology and closed MOX fuel cycle;
19901993 Investigation after negative reactivity shutdowns; and
1993present time Renovation, test and operation with limited reactor power,
350 MW(th), 145 MW(e) on two secondary loop.

62
 TABLE 1. THE MAIN EVENTS AND THE GRID-CONNECTED OPERATION TIME IN THE RELEVANT YEARS

Main events 74 75 76 77 78 82 83 84 86 88 89 90 98 00 03

IHX secondary XX X X XX X X X
circuit sodium leaks

X2
1 1 2
Steam generator X X X X2 X2
leaks X1 X1 X2

Secondary circuit X XX X XX X X3
main pipe sodium X4
leaks

Negative reactivity X X X X
shutdowns XX

Grid-connected 81 68 54 24 67 62 63 71 80 72 31 Investigation after reactivity

operation time, % transient, renovation, test
and operation

X1-water leaks into the evaporator box space through the sub-headers shell wall;
X2- sodium-water reaction;
X3-leak in the bellow of the sodium purification system valve;
X4-leak in the electromagnetic pump of the steam generator hydrogen detection circuit.

63
3.1.2. Intermediate heat exchangers operating experience

The intermediate heat exchanger (IHX) transfers heat from the primary radioactive coolant to
the secondary non-radioactive coolant while maintaining a physical barrier between two.
The IHXs are connected two by two to a secondary cooling circuit and suspended from the
upper part of the slab. The design of the heat exchanger is shell and straight tube, counter
flow, with the primary coolant on the shall side; 2 279 tubes (outer diameterwall thickness:
141.0 mm, 5 300 mm length, austenitic steel 316) are fixed onto lower and upper tube plates
by expansion and welding. At the top pf of the IHX, the second sodium enters a central tube
and flows to the bottom of the distribution box with a convex bottom welded to the lower tube
plate.

In the early years of operation the only serious loss of power generation was caused by the
need to repair the IHX as a result of leaks of inactive secondary sodium. Leaks were detected
in the annular inter space between the cold inlet and hot outlet ducts in two separate instances.
Three successive incidents affected IHX: on 11 July 1976, on 3 October 1976 and on
31 August 1977 led to complete shutdown of the plant for about ten months to allow for
disassembly and examination of the damaged IHX. The NPP restarted and operated at two
thirds of its rated power with four repaired/modified IHX and dummy devices while repairs of
the IHXs.

In March 1984 a new secondary sodium leak occurred in the IHX. The leak was found by
detection of the sodium in the annular space separating cold and hot secondary sodium. The
leak was very small and it was possible to continue the operating cycle for almost two
months, at the end of which it was decided to remove the IHX and to replace it with a spare
one.

Replacement was carried out in July 1984, but in December 1984 the same defect appeared on
this IHX, forcing the plant to operate at two-thirds power for nine months while repairs were
made. It was concluded that these two successive leaks were not generic defects but were the
result of two repairs made in 1977 after the first series of IHX incidents.

In October 1988, a leak of secondary sodium into the annular interspace of the IHX was
detected. The IHX was removed and replaced by a spare new one. This new exchanger was
equipped with 66 thermocouples, some of which were on the primary side of the tube bundle.
As before, the cause of the leaks was found to be an improper mixing of the secondary
sodium at the outlet of the tube bundle: an important radial temperature gradient existed
between the outer and the inner sodium flows, the former being hotter. As a result, mechanical
constraint developed, leading to overstressing of welds in parts of the secondary sodium duct.

Modification to a revised design with improved sodium mixing capabilities (by a mixing
device) and flexible design elements at the secondary outlet were necessary for each
intermediate heat exchanger (Fig. 3).

The removal and modification one-by-one of the Phnix IHX was perhaps rather difficult
maintenance task foreseen on pool type fast reactors. Its accomplishment, with the minimum
of reactor downtime, represents a conspicuous success and demonstrates the importance of
foreseeing events at the design stage, making adequate provision of equipment and space for
the repair, planning and learning from previous operation.

64
 before repair after repair before repair after repair

FIG. 3. PHENIX intermediate heat exchanger (IXH): places of failure, repair and modification.
[(a)-before modification; (b)- after modification (flexible design elements and a flow-mixing device in the sodium header at the tube plate outlet)].

65
Two IHXs were experienced leak from the tube bundles during the test and renovation period:
in November 1998 (reactor in operation) and in November 2000 (reactor shut down).
The cracks were found in the tube inner wall in the expansion zone, at the rolling area under
the upper tube plate. Experts concluded, that the cause of the cracking was corrosion under
stress owing to presence of soda or polluted secondary sodium hydroxide, or sodium during
the draining phases for these IHXs (several openings for work on the secondary loop between
1995 and 1997). The damage in the Phnix IHX was due to a design error leading to
deformation and stresses because of the different heat expansion of the external and internal
rings at the sodium exit to the secondary system; the different thermal dilatation of two plates
cause the crack2.

Apparently, for a future IHX design a variable flow distributed should be provided inside the
IHX tubes with a higher secondary or/and lower primary flow on the outer rows to improve
the temperature distribution in the tube bundle; a mixing device could be also helpfully at the
secondary outlet. Feedback from these incidents and developed technology was very valuable
for next French fast reactor Super-Phnix.

3.1.3. Steam generator operating experience

The steam plant is composed of three steam generators (SGs); each with its own independent
secondary sodium system. The SG is modular in each stage, which made it easier to replace.
Each SG comprises of three stages: evaporator stage, superheater stage, and reheater stage.
Phnix SG specifications are shown in Table 2.

TABLE 2. PHENIX SG SPECIFICATIONS

Item Evaporator Superheater Reheater

2
Heat transfer area, m 1226.3 1214.6 1218.5
Outer diameter and wal1 thickness of tube, mm 284 31.83.6 42.42.0
Tube length, m 60.8 26.9 21.09
Steam water inlet temp., C 246 375 308
Steam outlet temp., C 376 512 512
Steam outlet pressure, bars 175 168 34
Sodium inlet temp., C 475 550 550
Sodium outlet temp., C 350 475 475
Sodium flow rate, kg/s 740 423 317
Steam, water flow rate, t/h 250 250 223

Each evaporator is associated with a reheater and superheater; stage is made up of 12 modules
in parallel. A module consists of a shell containing seven steam tubes surrounded by
sodium (Fig. 4).

2
INTERNATIONAL ATOMIC ENERGY AGENCY, Status of Liquid Metal Cooled Fast Breeder Reactors.
Technical Reports Series No. 246, IAEA, Vienna, (1985), p.65.

66
 FIG. 4. Phnix SG design.

The SG is designed to be cooled with water or air during reactor shutdown. The hot sodium
from the intermediate heat exchangers is distributed first between the superheater and reheater
stages, and then the entire flow enters the evaporator stage. On the steam side, the stages are
in series; the reheater stage is supplied by steam at partially reduced pressure from the turbine
high pressure cylinder outlet.

The evaporator is made from ferritic steel 2.25 Cr-1 Mo whereas the superheater and reheater
are made from austenitic steel 321 H. The superheater and reheater modules are S-shaped, or
two S-shaped modules in the case of evaporators. Each module is composed of seven tubes
placed inside a shell with water/steam flowing in the tubes and sodium flowing outside the
tubes in the opposite direction. The steam generators are fitted with water leak protection
system and detection devices to detect water or steam leaks in the sodium: hydrogen detects in
the argon blanket of the expansion tank of each secondary circuit.

Main operating parameters under rated conditions are shown in Table 3.

67
TABLE 3. PHENIX MAIN OPERATING PARAMETERS UNDER RATED
CONDITIONS

Characteristic Value
Sodium inlet/outlet temperature, C 550/350
Sodium flow rate in each SG, kg/s 740
Water inlet temperature, C 246
Water flow rate in each SG, kg/s 210
Steam outlet temperature of the evaporator, C 376
Superheated steam temperature, C 512
Superheated steam pressure, bar 165
Reheated steam temperature, C 512
Reheated steam pressure, bars 34

Four water leaks were happened in the evaporator inlet of the SGs between November 1975
and September 1976: subheader under frame was wearing out that was provoked by
perturbations generated by the water flow distribution orifice plates in the evaporator tubes.
The 252 orifices plates were replaced by new ones with improved design between October
1976 and February 1977. In 1982/83 leaks in reheater modules led to a period of operation on
two circuits. On 29 April 1982 a reheater leak was detected by the hydrogen detection system;
the secondary sodium side as well as the water side was dried out. The pressure of the
nitrogen, which ought to have replaced the water, did not reach the required value of 0.7 MPa
owing to the failure in the opening of a check valve in the steam generator. This allowed large
quantities of sodium during the refill of the secondary system to penetrate into the water
system. Twelve reheater modules and a related steam piping were polluted by the
sodium-water reaction products. Finally, 6 L of sodium reached the air and caught the fire.
This was extinguished within 2 minutes. The chronological order of the next three leaks
occurrence:

Leak no.2: 16 December 1982 on reheater module 12 of SG 1;

Leak no.3: 15 February 1983 on reheater module 12 of SG 3;
Leak no.4: 20 March 1983 on reheater module 11 of SG 1.

In the case of leak no. 4, it should be noted that the no. 12 modules were new or had been
removed. The following remarks may be made for a leak no. 1, 29 April 1982: SG isolation
dry out was requested by the operator with about 5 minutes delay; this explains the large
amount of wastage on the adjacent tube and the onset of wastage on the module shell. As
pointed out before, after dry out, nitrogen injection into the reheater was not carried out due to
a faulty non-return valve, and the large amount of sodium entered the steam tubes and the
reheater connection pipes.

Only the first sodium-water reaction led to complications, with sodium contaminating the
steam-side of the reheater modules. The radiographies detected holes in two tubes ~ 2 cm2
that damaged the modules shell. Nevertheless, the incident had no consequences for plant
safety. When the leaks occurred, the SG had ~ 5 0000 hours of operation under rated
conditions. Subsequent to this first leak, signal processing by the computer was improved for
greater rapidity and operating instructions were also reviewed. These improvements proved
beneficial during leaks nos 2, 3 and 4. Remarks for leaks nos 2, 3 and 4 are as follows:
equipment and operators reacted well. Operator response time was only a few seconds.

68
The SG isolation-dryout sequence performed well. The number of failed tubes were one, and
two for leaks nos 2, 3 and 4 respectively. The four leaks showed several common points. The
fault occurred:

On the reheater stages on the bends (Fig. 5);

A short time (1 to 5 days) after a power plant start-up;
Three times on module 12, once on module 11 (module 12 inlet pipe is the first of the
reheater steam inlet main pipe);
At a butt weld beads of the tubes and on one of the two welds on the hottest tube part.

FIG. 5a

FIG. 5b
FIG. 5a,b. Phnix SG: location of leaks.

69
After expert analysis of various possible causes such as tube vibration, blocked differential
expansion, chemical corrosion, etc., the most probable appeared to be the penetration of water
into the reheater during the power plant start-up phase when the turbine bypass system is
operating: when water abruptly turns to steam in the reheater tubes, that caused thermal shock
which affect their resistance. The sodium outlet temperatures of the modules showed fast
fluctuations with an increasing amplitude from module no. l to no. 12 (Fig. 7). These water
induced thermal shocks and temperature variations caused major thermal stresses where the
steel is thicker and generated a through crack. It was concluded that the failures were owing to
a combination of operating condition and the quality of some welds.

Two main precautions were taken to prevent reoccurrence of the fault:

At the design stage by specifying a maximum excess thickness of weld beads;

During operation by introducing a procedure of purging and preheating the steam
pipes.

The corrective actions concern after SGs leak:

1983: Time gain in hydrogen detection signal processing

Introduction of double insulation of the water supply pipe to the evaporator stages;
Addition of a second nitrogen injection system to each SG stages;
Addition of a second decompression valve to the evaporator stages;
Systematic sodium dump of the affected circuit five minutes after the trip. This delay
corresponds to the minimum time needed to homogenize the sodium and to avoid local
sodium hydroxide concentrations;
Dryout of the other SG five minutes after the dryout of the damaged SG. This
operation prevents any steam ingress into the damaged circuit in case of leaking
isolation valves.

1984: Processing of the hydrogen detection signals by two dedicated computers auctioning
the rapid shutdown and the isolation and dryout of the SG.

1985: Installation of the GENEVA monitoring system. As well as the permanent signal
monitoring this computer allows all the routine tests to be carried out (e.g. calibration by
injection of hydrogen, control of the diffusion rate); development of an experimental acoustic
detection system.

1988: Installation of an induction heater in circuit No. l evaporator stage; the modular
SG design facilitated the location of the faulty module and minimized the outage while taking
a module out of service.

The cost of four sodium-water reaction incidents in steam generators was a total of six months
of reactor shutdown and nine months of reactor restricted operations at two-thirds power.

A fifth sodium-water reaction in the reheater of the SG no. 1, module no. 12 took place on
13 September 2003 during operation on power level of ~ 300 MW(th): about two kg of water
came into contact with the sodium. As in the previous cases, the failure was located at the
upper bend of a reheater module and were due to a combination of operating condition and an
initial manufacturing defect in the weld. A wastage in the shell across the tube hole was
observed (Fig. 6a,b). The damage module was replaced with a new one.

70
 FIG. 6a. Phnix SG tube failure, - a fifth sodium-water reaction (September 2003).

FIG. 6b. Phnix SG: sodium outlet temperature curves for reheater modules 1 and 2.

71
3.1.4. Secondary circuit operating experience

Each of three independent secondary cooling systems, which transfers heat from the primary
system to the steam generator and isolates the steam supply from the activated primary
sodium, composing two intermediate heat exchangers, pump, buffer and expansion tanks,
auxiliary circuits and connecting pipe work (outer diameter wal1 thickness: 5106.0 mm;
material: hot leg-stainless steel 321, cold leg-stainless steel 304). The secondary sodium flows
inside the intermediate heat exchanger tubes and outside the steam generator tubes.

Minor secondary sodium leakages were detected three times, in September 1974, and then in
March and July 1975. Leakage was traced to a joining weld on a large diameter pipe
delivering sodium to the reheater part with 450 mm control butterfly valve. Approximately
20 L of sodium was lost on the first two events and only about one L the second events with
no sodium fire outside. Repairs were ineffective and the valves in all three systems were
replaced by diaphragms.

In March 1986, a leak occurred on the secondary circuit buffer tank. The repair of the tank
needed a difficult operation because the sodium aerosols had corroded the tank wall that
required welding a new metal on the spherical tank wall. Later three tanks were replaced by
new ones 316 steel.

On 5 May 1986 during full power operation, a leak occurred from one of the main pipes of
secondary circuit. It appeared at the entrance of the main pipe carrying sodium into the
superheater part of the steam generator and was detected by the leak detection system. Around
50 kg of sodium leaked and froze in the thermal insulation casing surrounding the pipe.
No local signs of leakage were visible. Regular patrolling was stepped up in the area until the
next scheduled shutdown on 19 May 1986.

The circuit was drained and the insulation removed from the pipe. Several tens of kilograms
of mixture of sodified sodium and insulation material were found. The leak was located in a
weld seam joining the 500 mm diameter tee fitting to the inlet header in the steam generator
unit. The tee fitting consists of two AISI 321 half shell castings 19 mm thick, while the
AISI 321 sleeve is 7 mm thick.

Relatively marked corrosion was observed on the wall of the tee fitting, especially around the
perimeter of the solidified sodium mass. Corrosion damage locally reached several
millimetres. Dye penetrant, gamma radiography and metallographic examinations were
conducted in-situ prior to removal of the tee fitting. The results indicated that the crack
~ 120 mm ran along the weld seam in the heat-affected zone on the inlet sleeve side of the
steam generator. The intergranular crack extended on either side of the upper generatrix, and
measured 120 mm in length. After removal of the tee fitting, the following metallurgical
examinations were carried out:

Macrographic examination of the inner and outer cracked surfaces;

Micrographic examination in planes perpendicular to the weld seam at different
positions in the crack;
Fractographic examination of a 2 mm thick cross section through the weld seam;
Ferrite and carbon content analysis.

72
In addition, the area of the tee fitting that had been corroded by the sodium-insulation mixture
was dimensionally checked by an ultrasonic thickness measurement device. The results
suggested that the cause of the leak was a delayed reheat cracking mechanism liable to appear
in Ti stabilized austenitic stainless steels submitted to mechanical stresses3.

Concerning the external corrosion of the tee fitting, dimensional measurements indicated that
the zone in contact with sodium covered about 1 m2. The most highly corroded area was
around the perimeter of the contact zone, and the maximum corrosion depth reached 7.4 mm
out of an initial thickness of 19 mm. The age of the leak was estimated at a few thousand
hours. The tee-piece was replaced with a new one specially made. The corresponding parts of
secondary loops 1 and 2 were also examined, but no defects were found in the tee fittings. The
plant was operated at two-thirds of rated power with only two secondary loops until the end of
August 1986, pending delivery of a replacement tee fitting. The tee fitting was replaced. The
main secondary pipe leak detection system was improved by continuous monitoring of the
electrical isolation of the preheating pyrotenax.

In the event of a sodium leak the heating element was generally damaged by corrosion
products. Monitoring the electrical isolation of the two pyrotenax units provides an additional
means of detection. The units are located on the pipe midplane: in case of leakage from the
upper portion of the weld, they are affected more quickly than the leak detection wire along
the lower generatrix. An operating error caused a sodium leak in October 1988.

The Phnix plant, like the other sodium cooled reactors, had significant experience with
sodium leaks. Since the facility had started up, there had been some twenty leaks,
approximately one per year of power operation: the leaks in IHX and SG- the most damaging
caused the largest production loses. The aim of leak before break (LBB) argument in liquid
metal cooled fast reactors safety analysis is to show significant lines of defence which allows
rejecting in the residual risk an unacceptable of main vessels and piping. The important
conclusion which could be derived from the Phnix secondary circuit leaks is as follows: the
applicability of LBB argument should be confirmed in terms of stress analysis, risk of
corrosion, leak detection efficiency.

3.1.5. Pumps operating experience

At Phnix, vibrations were detected on one primary pump and were shown to be due to a
design defect which allowed the hydrostatic bearing bush to become separated from the shaft
as a result of expansion during thermal transients. Modifications were made to four primary
pumps including the spare. A similar failure has occurred in 1987 on a secondary pump.
Table 4 summarizes the Phnix sodium pump performance.

3
This is a defect triggered in the root pass of a weld in the immediate neighbourhood of the contact zone, and
which spreads radially between the grains when under stress in service. This only occurs with certain materials
such as 321 type stainless steel, as it caused by hardening of the steel due to tine precipitations of titanium
carbide inside the crystalline structure. The hardening causes the plastic deformation capacity to be transferred to
the periphery of the grains. For this to happen, the following conditions have to be met:
- a high operating temperature (> 475C for 321 steel);
- a geometrical discontinuity at the weld root;
- strain hardening at the weld root showing significant shrinkage;
- heavy local load, which may be due to welding stresses; and
- a defect in the weld root (e.g. a small shrinkage crack) [Jean-Francois Savage, Phenix 30 years of history:
the heart of reactor, CEA/EDF, Valrno-BP 17171-30207 Bagnols-sur-Ceze cedex], p.220.

73
TABLE 4. PHENIX SODIUM PUMP PERFORMANCE

Operating hours Failures

Primary Secondary Three-primary,
472 830 390 200 one-secondary

3.1.6. Negative reactivity shutdowns

On three occasions in summer 1989, the reactor was stopped by automatic emergency
shutdown, the negative reactivity threshold (-10 pcm) being exceeded. This reactivity
variation was very fast: first a minimum after 50 ms followed by an increasing oscillation, and
then a decrease, caused by the control rod drop, 200 ms after the start of the transient.

The first two events were thought to be spurious (a neutronic chamber fault) and the reactor
was restarted. The normal plant instrumentation did not allow proper recording of the
transient so following the second trip special instrumentation was installed. After the third
trip, the reactor was shut down in order to identify the cause of the events. It was found that
the phenomenon could be explained by gas entrainment through the core, after accumulation
under the diagrid.

The void coefficient explained the transient loss of reactivity. After some reactor
improvements related to this explanation, the reactor was allowed to restart at the end of 1989,
but the event occurred again in September 1990, after 182 EFPD of operation.

Investigation of the negative reactivity shutdowns which had occurred since the reactors
commissioning showed that two trips, in April 1976 and in June 1978, are similar to those in
1989 and 1990.

An expert committee was then set up, and an extensive study of all the possible phenomena
was started. Also it was decided to fit the plant with special monitoring equipment including
fast recording systems', and to perform tests. Around 200 data were concerned. Tests on
vessel and component mock-up were also planned. The tests were performed with the reactor
shut down, critical at zero power (since October 1991) and at 350 MW(th) power (for around
12 days - February 1993). In the same time, checks were performed on the plant, especially on
the reactor, its components and auxiliaries. Reactor tests were very satisfactory. They proved
the good behaviour of the instrumentation, and data are now stored as reference of steady
state power and emergency shutdown conditions.

By the end of 1993, studies had not led to a clear explanation of the phenomenon: false
reactivity variations (a neutronic mask between core and neutronic chambers, or a spurious
signal) are thought to be impossible. Among "real" reactivity variations, a sodium void effect
or variation of the relative displacement of fuel and control rods are also thought to be
impossible. There remains only a radial core volume variation, the origin of which
(a pressure wave) has not been found. These studies confirm that the reactor safety was not
affected. Operation with special instrumentation seems to be the only way to understand the
origin of these negative reactivity trips.

It was concluded that the phenomenon was harmless and could have had at least two possible
causes: signal interference or mechanical movement by the subassemblies in the core.

74
3.1.7. Sodium aerosol deposit
Phnix has two types of control rod mechanisms (this principle being adopted in order to
minimize the probability of common mode failures). One type of mechanism, which is
equipped with bellows protecting it against aerosols, has never experienced any problems in
17 years of operation.

However, during the test and renovation time a leak in the bellows of the two control rod
mechanisms was detected. This bellows located between the sleeve and the rod ensure the
sealing between the sodium or the argon of the reactor cover gas and the argon inside the
mechanism, so as to limit the presence of aerosols inside the mechanism.

The other mechanism has experienced two series of problems, both linked with sodium
aerosols, although neither of these problems prevented the control rods from being normally
inserted when required to shut down the reactor.

The first series of problems occurred in January 1974, a few months after the first criticality.
Following a normal scram, two rods were found 15 mm above their lower position.

After extraction of the mechanisms, it was discovered that the cause of the blockage was the
fresh argon flow of 600 L/h which was permanently being entered into the annular space
between the rod and the sleeve (oddly enough in order to prevent any upward motion of
eventual sodium aerosols in this annular space). The argon outlet in the hot pool was under
sodium. So the argon flow had two consequences:

(a) Depression of the sodium level in the annular space between extension rod and sleeve;
and
(b) With the argon flow oxygen impurities were transported and space was higher than
before, returning to the overall level inside the reactor; thus all the moving part of the rod,
going though the lower bearing, would be permanently bathed in the hot sodium.

The three mechanisms (of the second type) were cleaned and remounted and no other
problems occurred until 1981. In 1981, one of the same types of mechanism showed new
anomalies in its functioning, which were first detected during systematic measuring of the
drop time of each rod. One rod exhibited longer than normal drop time (100 ms).

A series of tests was initiated. At first it was thought that the dash-pot bearing the whole
mechanism at the end of the course was responsible, until it was realized that the cause was
some friction force acting during the stroke of the rod. After dismantling it was observed that
some aerosols had been deposited along the extension rod, being formed near the labyrinth
with a clearance of 2 mm in diameter and spread over the length of the tube during the raising
and lowering of the mechanism over seven years of operation (Fig. 7).

75
 FIG. 7. Phnix: deposit of oxides on control rod mechanism.

This explains why the problem occurred progressively. It was decided this time to install on
each mechanism a permanent device to measure with great accuracy the force needed to
exercise the control rod.
3.1.8. Sodium circuits, reactor and equipment inspection and renovation

The Phnix reactor had been shutdown in September 1990 owing to the negative reactivity
trips. Some times the reactor was in transient conditions at low power to maintain staff
competence in reactor operation. Then the reactor was often kept in a state which was ready to
start up. By this time the plant had accumulated about 100 000 hours, whereas it was designed
for 140 000 hours of power operations. The operator proposed to start the plant back up.
But each new request added a new safety issues. The delayed reheat cracks on the secondary
circuit were some of them. Several objectives motivated the additional control, repair and
manufacturing new components. At Phnix, renovation works as well as inspection,
maintenance and repair activities have been done during 1993-2003 (J. Guidez and L Martin:
Int. Conf. Fifty years of nuclear power- the next fifty years, Moscow/Obninsk, Russian
Federation, 27 June-2July 2004):

Special inspections of the welds of the secondary piping system, steam generator,
reactor core support structures and the upper internal structures of the reactor block;
The repair of the secondary piping system and steam generator;

76
 The portioning of the secondary sodium circuit in the steam generator to improve
protection against large sodium fires;
The seismic reinforcement of the plant building;
The additional of safety control rod to the reactor;
The installation of two redundant seismic resistant emergency cooling circuit and
anti-wip system on the high pressure pipes;
The defective equipments were repaired, replaced or left as is and justified by a
non-propagation analysis for the defects, which backed up the non-destructive testing
done on the site.

3.1.8.1. The secondary piping system

At the end of 1993, since the beginning of the inspection campaign in 1990, 1 200 meters of
welds have been inspected, 4 000 gammagraphy films have been produced and interpreted.
All circular welds, longitudinal welds and elbows, and 10% of the longitudinal welds of the
straight parts were inspected by gammagraphy, due penetrant test, and where was possible by
ultrasonic techniques. The pipes made of 304 steel were the first to be examined, for they
were the longest and parts of them had been located inside the reactor and steam generator
buildings. Very few small, non evolving manufacturing defects were revealed, and some ten
welds were repaired.

In 1992 two cracks were observed on a 304 steel pipe downstream from the hydrogen
detection return due to thermal stripping associated with the temperature fluctuations. The hot
leg pipes around the steam generator up to including the buffer tanks, steam generator sodium
headers and modules were made of 321 steel stabilized by titanium to improve the mechanical
strength at high temperature.

From the analysis of these inspections and the metallography expertise of the dismounted
welds, the following conclusions were drawn: 20 circular 321 stainless steel piping welds
were found to contain cracks, in a total of 60 which were dismantled and 220 which were
inspected; the cracks were located in the straight parts of elbows and tee-junction with
thickness variations. Cracks were observed on 12% of the welds operating at 550C and 3%
of the welds operating at 475C.

Non-destructive control revealed that cracks emerged at the root of the weld of the thermally
affected zone. An ultrasound indicator detected the crack, but did not determine precisely the
depth of the defect. Related expert studies showed that it was difficult to guarantee that crack
under the operating stresses would not grow to a depth up to a thin wall of the pipe, which in
Phnix is a 6 mm thick. The piping and SG portions of the 321 type steel were replaced by
new 316 steel, which have been successfully used at SPX.

3.1.8.2. Steam generators

During testing of the reheater in SG 2, a leak appeared on the connection weld of five
modules on the steam inlet side. This fault occurred in a well known manner and may be
explained as follows: the welding operation between the steam collector, made of chromesco1
ferritic steel, and each of the reheater modules, made of AISI 321 austenitic steel stabilized
with titanium, initiated microcracks in the weld seam by hot cracking.

In 1982 during a sodium-water reaction, sodium penetrated into the steam part of the modules
subsequent to the malfunctioning of a check valve. This sodium was not completely
eliminated during cleaning, especially in the microcracks that were present.

77
Inspection revealed cracks in some of the secondary circuit components made of 321 stainless
steel, and conclusions repairs have been made. Several defects of the relaxation cracking type
were displayed, which led to systematic replacement of such secondary piping and
the SG sodium headers. A significant crack was detected following the examination of a
SG no. 2 module. Studies for repair were undertaken in early 2001. Repairs were made on the
321 steel superheaters and reheaters (48 units in total) of SG no. 1 and no. 3. They involved
the sodium inlet cones and the first hot bend in each module (Fig. 8) and control of the cold
bend.

FIG. 8. Phnix steam generator repair and control.

The technological operations were as follows:

Disassembly of the modules;

Cleaning for residual sodium removal;
Repair and control of some non replaced parts;
Inspection of the new welds;
Leak resistance and hydraulic tests for each module;
Reassembly in the steam generator.

3.1.8.3. Reactor block

Four non-destructive tests have been carried out at the Phnix reactor:

Ultrasonic inspection of the core support conical skirt (pos. 5 in Fig. 9a) welds under
liquid sodium at ~ 150C in period September-October 1999. Five holes specially
were provided for these operations. The work was automated owing to a hot and
irradiation environment.
Inspection of the top portion and the handling shells of the main reactor vessel using
small transducers.
Inspection of the tubeplate in IHX using a photothermal camera.

Remote visual inspection (by periscopes) of the core cover plug and internals structures of the
reactor in March-April 2001 with the partial (420 m3 of ~ 900 m3) draining of the primary
sodium to the level of the fuel subassembly heads.

78
 1

1-manhole ( = 500 mm)

2
2-movable ladder
3-catwalk
3
4-fixed ladder
5-conical skirt
6-cable tracks

5
6

FIG. 9a

FIG. 9b
FIG. 9a,b. Phnix reactor conical skirt welds US inspection.

3.1.9. Protection for fires

During the test and renovation new studies and significant works on protection of the steam
generator building against large sodium fires have been undertaken. The latter include
(J Guidez and L Martin: Int. Conf. Fifty years of nuclear power-the next fifty years,
Moscow/Obninsk, Russian Federation, 27 June-2 July 2004):

The separation by steel insulated walls and doors firebreak of zones so as to limit the
spreading of a large fire; two separation steam generator cells reconstructed (Fig. 10)
to resist a major sodium fire with temperature on order 1 000C for 30 minutes;
The portioning or the housing of cable trays and building steel structures;
The ventilation and the smoke cleaning circuits;
Installation of multi-sampling detection circuit.

79
Most of the buildings were modified: offices and the control room, the handling, the reactor,
the SG and auxiliary circuits buildings. A vast modernization programme at the Phnix NPP
has been undertaken.

Sodium zone partitioning

FIG. 10. Protection of the Phnix steam generator building against large sodium fires.

3.1.10. Plant statistics and conclusions

The French prototype fast breeder reactor until 1990 operated with a comparatively high
operational characteristics and demonstrated the industrial feasibility of sodium cooled reactor
technology (Figs 11 and 12).

FIG. 11. Phnix operating histogram.

80
FIG. 12. Phnix power plant-operational chart 1974-2004.

81
The main production data of Phnix are shown in Table 5.

TABLE 5. PHENIX MAIN PRODUCTION DATA

Characteristic Value

Effective full power days, EFPD 3 900 EFPD

Gross electrical energy production, GWh 22 424 087
Load factor (since commissioning in July 1974), % ~ 50
Number of irradiated subassemblies 829
Number of irradiated pins 166, 521
Burnup (maximal), h.a. 17%

In the period 1974-1989 the plant was connected to the grid ~ 70% of the time, 25% of which
was at reduced power. On the average the periods of stabilized operations is equal to 40%
owing to the long periods operations at two-thirds of the rated power (following the events
encountered on the SGs, IHXs and secondary sodium pipe leaks).

Many judgments were made during the design of Phnix NPP with a much more limited data
base than is available to those viewing the plant retrospectively. With the exception of the
choice of 321 steel stabilized by titanium to improve the mechanical strength at high
temperature and IHX design, these judgments are generally vindicated by expert evaluations
conducted on the materials and the reactor components, after 100 000 hours of operation with
projected modern parameters. In the future sodium cooled reactor designs the using of the
321 steel stabilized by titanium are probably best avoided.

The Phnix plant, like the other sodium cooled reactors, had significant experience with
sodium leaks. Since the facility had started up, there had been more than twenty leaks,
approximately one per year of power operation.

The methodology developed by team of designers, researchers and operational staff to extend
reactor lifetime, the development and realization of special investigation/inspection and
renovation programs have resulted in significant progress for R&D and greatly increased
expertise from which the entire nuclear program will benefit. Phnix contributions continue as
it provides the first experience in the transmutation of long-lived nuclear waste and the
utilization of surplus plutonium in the real reactor conditions.

The Phnix reactor has been restarted in June 2003 after five years of being off line for a
major inspection, repair and safety upgrade. The cost of the upgrade work (~ 3 million
hours-men) was 250 million EUR. The Phnix reactor will be used to irradiate actinide
transmutation experimental rods. There are 12 transmutation projects to be carried out
between 2003 and the scheduled closure date of the reactor in 20094.

4
Nucl. Eng. Int., August 2003, p. 6.

82
3.2. SUPER-PHENIX REACTOR

3.2.1. Design features and commissioning history

The Super-Phnix (SPX) plant was derived from Phnix is of a pool type. Changes were
either necessitated by the increased size of the reactor, or they were to achieve definite in
economic and safety performance. For instance:

The primary sodium coolant purification units were within the main vessel;
There were four helical steam generators (SGs), each with a power of 750 MW(th);
Slightly larger subassemblies, designed to achieve a higher burnup;
Design of the main vessel and roof were simplified;
A dome over the upper part of the vessel was added to provide containment.

Main design and economical data are presented Tables 6 and 7, respectively. The
cutaway view of the SPX reactor is shown in Fig. 13. The comparison of the SPX-1200 with
the 1400 MW(e) PWR plants Paluel-1 and -2, built at the same period, showed that the cost
per unit power installed of SPX approximately a factor of 2.7 higher than a PWR (Table 6)5.

TABLE 6. SPX SPECIFICATIONS

Item Value
Power (thermal/electric), MW(th)/ MW(e) 3000/1200
Thermal efficiency, % 40
Inner diameter/height of main vessel, m 21/19.5
No. of loop (prim./sec.) 4/4
No. of main pump (prim./sec.) 4/4
No. of IHXs 8
Sodium inventory (prim./sec.), tons 3 500/1 500
Sodium flow rate (prim./sec.), t/s 44.24/43.27
Primary sodium temp. (hot leg/cold leg), C 545/395
Secondary sodium temp. (hot leg/cold leg), C 525/345
Loop concept classical
Steam temp. at turbine inlet, C 487
Steam press. at turbine inlet, bar 177
Steam flow rate, kg/s 4340
Feed water temperature, C 237
Type of steam-water cycle Steam reheating
system
No. of SG per loop 1 once-through SG
l
Total mass, tons 194

5
M. RAPIN, Fast breeder reactor economics, presented in the Royal Society Meeting on the Fast-neutron-
breeder fission reactor, London, U.K., 24-25 May 1989; R. CARLE, Detailed design studies demonstrate major
improvements in economics. Nucl. Eng. Int., February 1988.

83
TABLE 7. COST POWER GENERATION BY SPX AND PWR- P4, centimes/kWh [1]

Cost category PWR-P4 SPX-1

[1 400MW(e)] [1 200 MW(e)]

Capital investment 8.5 22.6

Fuel 5.3 10.0
Operating and maintenance (O&M) 3.2 5.0
Power generation costs 17.0 37.6
Capital investment SPX/PWR-P4 2.66
Power generation costs SPX-1/PWR-P4 2.222

FIG. 13. SPX: cross-section of the primary circuits.

84
That is why, when designing the next European LMFR EFR, the problem of choosing an
optimum type of arrangement, reactor and equipment design was raised again. The primary
goal was clearly to cut the costs on the basis of the design and construction for SPX.
The important step in this direction was the power rise from 1 200 to 1 500 MW(e).

Primary sodium coolant is entirely enclosed in the main stainless steel vessel which contains
the core, and in which are installed four primary pumps and eight intermediate heat
exchangers. The main vessel is closed above the free level of sodium and argon cover gas by
the slab which contains in its central section two eccentric rotating plugs and the core cover
plug which supports the control rod drive mechanisms and the core instrumentation. It is
surrounded by the safety vessel, welded to the slab, which is itself topped by a metallic dome.

This dome can resist a pressure of 3 bar at 180C. The safety vessel and the dome make up
the primary boundary, and the reactor building in reinforced concrete constitutes the
secondary boundary. The main design data are given in Table 3.

Successive sodium filling of the storage drum, two secondary loops, the reactor block and
lastly the two remaining secondary loops, took from June to December 1984. Filling the
reactor block took 2 months, from 23 August to 31 October; after prior heating to about
150C by circulation of hot nitrogen. These filling operations were preceded by supply and
on-site storage of 5 650 tons of sodium, transported by 291 tankers.

Isothermal tests started in January 1985 with an initial build-up in temperature limited to
395C following the appearance of a phenomenon of vibration of the thermal shield in the
main vessel during increased primary pump speed vibration of part of the SPX reactor internal
structure occurred at certain flow rates, caused by small waves on the sodium surface exciting
one of its resonant modes.

These vibrations were investigated using the MIR (visual examination by camera and
displacement measurements through the MIR reactor inspection robot) vehicle for
in-service inspection in the interspace between main and safety vessels, by adapting the
focused ultrasonic transducers normally intended for the inspection of the main vessel
weldments and by mock-up under water, rapidly attributed this phenomenon to a
hydrodynamic coupling of the fluid and the structures, caused by vessel coolant sodium flow
through the spillway.

Hydraulic studies showed that the sodium level at the main vessel cooling weir was a critical
factor. By changing the flow adjusting gags in some 20 fuel subassemblies and thus
increasing the flow rate over the weir, it was possible to reduce the fall over the weir, taking
vibrations into the quiescent region for all normal operating conditions.

The first subassembly was loaded on 20 July and first criticality was achieved on 7 September
1985, with a core made up of 325 fissile subassemblies, as planned. In October 1985, 33 fuel
subassemblies were added to constitute the core for power build-up; it was with this core that
zero power testing was carried out at no more than 30 MW(th): control rod worth, neutron
flux distribution and fission rates, calibration of neutron channels, measurements of reactivity
and feedback coefficients. The first nuclear steam was received on 31st December 1985.
The SPX plant was connected to the grid on 14 January 1986, full power was reached on
9 December 1986 and operated, but not without difficulty. Upon successful start in January
1986, SPX reactor operation was interrupted by incidents (the operating and administrative
history of SPX is illustrated in Fig. 14), rated at level 1-2 of INES.

85
86
53 months normal operation (production and maintenance): 4.5 years
25 months recovery after incidence: 2 years

administrative delay 4.5 years

54 months
(including 24 months of public hearing)
11 years in total
January May March April July March July August September December
86 87 88 89 90 91 92 94 95 96

1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997

Total

17 months 15 months 5 months 16 months 53 months

Na purification Ar leak in exchanger
Storage drum
10 months 8 months 7 months 25 months
Administrative delay Administrative delay Public hearing Administrative delay

13 months 16 months 24 months 1 month 54 months

FIG. 14. The operating and administrative history of SPX.

During 11 years of SPX NPP existence, it was in operation for 4.5 years (energy production,
correction of malfunctions, and design modifications). Special repair procedures for
elimination of consequences of three incidents took 2 years; and administrative and social
procedures concerning license issue for reactor start-up took 4.5 years, because of the nature
of the accidents, as well as a very strong anti-nuclear social movement in the post Chernobyl
period, and a sodium fire at the Almeria facility that coincided with reactor start-up tests and
power increase. During 4 cycles (4.5 years), the SPX NPP produced 7.9 billion kWh (half of
this energy was produced in 1996), i.. a reactor was operated with load factor about 20%;
101 events and incidents have occurred at the reactor plant from the start of fuel loading (July
1985) till its closing down (July 1998). Out of these, 66 did not concern fast reactor specific
features and, among 35 incidents concerning LMFR, 16 were related to sodium. Seven events
were rated at level 1, and two events were rated at level 2 by the INES.

Three incidents marred the commissioning procedure and caused lengthy delays. In March
1987 there was a sodium leak from the used fuel storage drum. Two years after the incident
plant was granted the permit to resume operation and reached its nominal power on 16 June,
1989. In June-July 1990 there was an air leak into an auxiliary circuit which caused extensive
contamination of the primary sodium.

The plant has been technically ready to be restarted at the end of June 1991. Partly as a result
of this experience the safety of the plant was thoroughly reviewed and modifications to
improve the response to secondary sodium fires were made.

A public enquiry on renewal of the operating license was held and reported positively, and in
1994 the plant was granted the permit to restart. This was further delayed by a small leak of
argon from the sealing bell surrounding one of the intermediate heat exchangers. This justified
shutdown of the reactor in December 1994. Corrective action was taken and in the latter part
of 1995 and through 1996 power was gradually raised to the full-power level. Since
24 December 1996 the reactor scheduled shutdown followed by legal cancellation of its
operation license. It was made very clear that the reason of the shutdown was in no way with
safety problems, but with economy: the government said that when uranium appears now
durably cheap there is no need today to operate an industrial fast reactor prototype which
cost more than expected.

The longest outages of the SPX-1 reactor were caused not only by and not because of using
sodium as a coolant. Sodium bears relation neither to the fall of the turbine hall roof caused
by snow load with extensive damage to the steam plant, nor to the vibration of water/steam
piping system resulted from the water hammer. There were also some other incidents in the
BOP, which have been reported at the international meetings. The frequency of the events was
7.8 per year over the entire period considered. As noted in the overall frequency of the events
is similar to that observed on the PWR reactors, i.e. around 8 events declared per year,
covering all after commercial start-up [reported at the IAEA meeting on Unusual occurrences
during LMFR operation, 9-13 November 1998]. A large social response on any incident
existed because in the time of SPX commissioning anti-nuclear sentiments raised by
Chernobyl in some social groups and governmental structures were too strong. With
downturns in energy demand and political changes in 1998 the French Government finally
confirmed to discontinue it operation.

As a whole, the final operating experience of SPX-1 was incomplete but not as negative as
sometimes reported: over eleven years of existence it has been operating during four and half
years producing 7.9 billion kWh (half in 1996). Experience feedback on large components

87
remains significant in spite of the short operating period. Primary and secondary pumps total
more than 60 000 hours on main motor, and the continuous improvement of maintenance
operations has allowed an increase in reliability and availability. As far as the steam
generators are concerned, the sodium/water reaction detection systems have been improved
on the basis of validated calculation codes through experience. Numerous draining and filling
operations (more than 30 for the secondary loops and more than 20 for the decay heat removal
emergency circuits) have allowed validation of the corresponding procedures. Knowledge of
primary circuit behaviour has in fact been improved thanks to natural convection tests which
showed that it was established in the core in about 5 minutes.

3.2.2. SPX steam generator: design and operating experience

The biggest change in SPX design was undoubtedly that to the SG. Of all the fast reactor
design features this seems was the one that was the least well established and that showed the
biggest difference in style and unit size between the various designs proposed by the major
fast reactor countries.

The equipment of a large LMFR with economical means requires the use of SGs with high
self power. The only experience in the world was SPX where helical alloy 800 units with
750 MW(th) power were installed and very successfully operated. Each 750 MW steam
generator (SG) was a once-through unit comprising a vertical tube bundle, whose tubes were
wound helically around a cylindrical support, and an outer cylindrical shell, penetrated by
inlet and outlet thermal sleeves for each tube. This system avoids thick tubular plates (and
associated problems) and resists the thermal shock during sodium transients. Table 8 shows
the main design and Fig. 15 shows a cutaway view of SPX SGs, respectively.

FIG. 15. SPX steam generator.

88
TABLE 8. SPECIFICATION OF SPX SG

Item Value
Type of SG unit Once-through
Thermal power, MW 750
Total length, m 22.4
Shell outside diameter, m 2.88
Shell material Z 3 CND 18.12
No. of heat transfer tube, per unit 357
Length of heat transfer tubes, m 92
No. of welding on one tube 7
Out. diameter and wall thickness of tube, mm 25/2.6
Material of tube Incolloy 800
2
Heat transfer area, m / SG 2.565
No. of tube layers 17
Diameter of inner tube layer, m 1.17
Diameter of outer most tube layer, m 2.61
Steam/water inlet temperature, C 237
Steam/water outlet temperature, C 490
Steam water flow rate, kg/s 339.6
Steam/water outlet pressure, bar 184
Steam/water pressure loss, bar 349
Sodium inlet temperature, C 525
Sodium outlet temperature, C 345
Sodium flow rate, t/s 3.273

The SPX steam generators were of a new advanced design. Whereas the BN-600, Phnix and
PFR steam generators comprise a number of small (modular solution), and divided into three
sections - evaporator, superheater and reheator this solution was not adopted for the
SPX plant. Here, for further simplification, the evaporator and superheater were combined in
a single unit. There was no sodium reheat, because the need for SG reliability increase and
cost reduction by means of decrease of the number of units and sodium/steam heat transfer
surface made it necessary using steam as the heating medium (extracted from the HP turbine,
Fig. 16).

The advantages of the large diameter helical design are extremely good space utilization and
good heat transfer properties. The tubes joined together in a manifold outside the steam
generator vessel. This has the advantage that none of the walls containing sodium, apart from
the tubes themselves, and certainly none of the sodium-containing welds, carries a high
primary stress. Some disadvantages are the comparative complexity, and therefore cost, of the
manifolds to connect several hundred tubes, and the difficulty of gaining access to individual
tubes for inspection or repair and provision that has to be made for inspection during
operation. In some SG designs (e.g. BN-600 and PFR) it was possible to avoid welds in the
length of the tubes, where inspection is particularly difficult, but with helical tubes such welds
are unavoidable.

89
90
1

16
10 14, 15
17 18 19
11

20

6
23
2
21
7 12
24

3 8 13

4
25

9 22

1-turbine; 2-start up tank; 3-flash tank; 4, 5-water treatment plant; 6, 8-sodium inlet/outlet; 7-steam generator; 9, 22-HP, LP feed heaters; 10, 11-steam drum,
auxiliary boiler; 12, 13-feed water pumps (12-turb. Driven); 14, 15-deaerator-feed water; 16-reheator/moisture separator; 17, 18, 19-HP, LP1, LP2 turbine;
20-bypass; 21-blowdown tank; 23-condenser; 24, 25-extraction pumps, water treatment

FIG. 16. SPX steam-water flow scheme.

As known, in the earlier years of the development of fast reactors there was a high incidence
of leaks in the steam generators. Since leaks in sodium-heated steam generators are
intolerable, however small they may be, whenever a leak occurs the unit in question has to be
shut down and isolated for repair. This could be done without serious diminution of the power
output only if there are a large number of small power separate units, any one of which can be
isolated without reducing the total power significantly. The outstanding example of the
advantages of the modular approach to steam generator design is afforded by the French
Phnix and Russian BN-600 plants. This has three secondary sodium circuits, each with
12 (Phnix) and 8 (BN-600) separate steam generator modules, and each of these consists of
separate evaporator, superheater and reheater sections, making a total of 108 (Phnix) and
72 (BN-600) separate heat exchangers. Partly because of this the availability of the above
mentioned plants has been consistently high. However, these operational advantages are
achieved at the expense of higher metal content and costs, and of complication of the
SG layout. The metal content and cost of the high self power SG are reduced significantly by
applying the modular approach to steam generator design (Table 9). It should be pointed out
that some of leaks occurred where the tubes are attached to the tubeplate (BN-600 modules)
and where the tube are welded to tube (Phnix SG- on the bends).

TABLE 9. SGs COMPARISON

Item Phnix SG BN-600 SG SPX SG

Power (thermal/electric), MWth/MWe 186.3/85 490/200 750/300

Mass of SG, tons 225 600 194

Specific metal content, tons/MW(e) 2.64 3.0 0.65

Number of welds per 1 MW(e) 16. 0 27 8.3

(tube to tube) (tubes to (tube to tube)
tube- plate
attachments)

Each of the SPX tubes was ~ 92 m long. There were seven welds per tube, so SPX-1 had
about 10 000 welds in comparison with ~ 4 000 in Phnix. The flow tubes were butt welded
by the TIG (tungsten inert gas) process with welding metal. The operation was automatic and
100% radiographic control was carried out. A 45 MW SG prototype was built and had
undergone very extensive operating tests for several years at the EdF test center. The welding
and weld quality control procedures were developed and checked at this SG. That is why the
SPX SGs were operated successfully; past errors were not repeated and good design solutions
were incorporated. The outstanding success of the SPX operation has undoubtedly been the
demonstration of reliable operation of SGs with high self power (750 MW(th)).

3.2.3. Storage drum operating experience

3.2.3.1. Description and role

The SPX fuel handling equipment was designed for a maximum residual power of about
28 kW per subassembly (Figs 1719). In the initial design of fuel handling operations, the fuel
storage drum played a double role:

91
 It ensured the transfer of all new and spent fuel subassemblies during reactor loading
and unloading operations;
It ensured the storage in sodium of spent fuel subassemblies to allow the decay of their
residual power from 28 to 7.5 kW before cleaning and storage in the on-site water pool.

FIG. 17a

FIG. 17b

FIG. 17a,b. SPX new and spent fuel handling systems.

92
 FIG. 18

FIG. 19

FIGs 18, 19. SPX storage drum and leak location.

The drum played the role of a storage buffer between the reactor and the water storage pool; it
allowed reducing loading and unloading time to a minimum, and thus the duration of reactor
outages. The storage drum is part of the fuel handling line.

93
To facilitate handling of the fuel between the storage positions and to reach the required
capacity of fuel assemblies, two-level storage was provided; the maximum residual power
from fuel assemblies stored was 3 MW, which was removed through two redundant sodium
loops (Figs 17 and 18).

The storage drum consists of a 9 m diameter and 13 m high vessel made of 15 D3 ferritic steel
with a low molybdenum alloy. It contained about 800 m3 (700 tons) of sodium at a nominal
temperature of 200C. This sodium was cooled or heated by sodium circuits circulating in a
bundle of tubes fixed on the inner side of the vessel. The storage drum is equipped with a
safety vessel. Both vessels were separated from each other by a 150 mm gap, and were
located in a 2 m thick concrete pit.

In the initial plant project, it was envisaged to remove the fuel subassemblies to a
reprocessing plant after a period of one year in the sodium-filled storage drum. It was also
planned to renew 50% of the core after a half-cycle of 320 EFPD. The storage drum could
receive 409 subassemblies with a decay heat of up to 28 kW. Removal of the subassemblies
was carried out when the decay heat was less than 7.5 kW.

3.2.3.2. The incident

On 8 March 1987 a leak was detected in the space between the storage drum vessels by the
leak detection system. Several methods were implemented to confirm the leak, and assess its
importance. The balances for the levels of the storage drum and the storage tank showed that
there were 20 m3 of sodium between the vessels.

The leak was confirmed and the first investigations to locate the leak and size of the crack
started. This work took place between 27 August and 9 September 1987. Infrared
thermograph, xenon and helium detection in the inter-vessel space were used. The localization
of the leak had been carried out during the fall of the sodium level as a result of emptying the
storage drum in August 1987. The leak was found to be along one of the lower support-plates
for the storage drum cooling circuit (Fig. 19).

At the time of the leak, the storage drum contained only virtually radiation-free subassemblies
whose decay heat was negligible. The most serious risk, although very improbable, was, in
the short term, a leak on the safety vessel made of the same material as the storage drum
vessel, not knowing the cause of the leak. The unloading operations of the storage drum were
finished on 31 July 1987. The fuel and fertile sub-assemblies were then, after approval, put
back in the reactor, and the slightly irradiated dummy subassemblies were at this stage stored
in special containers constructed for the purpose.

After emptying the sodium completely, and when the storage drum had cooled down,
investigations consisting in particular of X-rays through the two drums showed up that other
plates were affected by the same type of not-yet-penetrating fault. It became obvious that
local repair would be impossible, and that it was necessary to anticipate human access into the
drum. This meant an air-fill of the main drum and its cleaning.

3.2.3.3. Causes of the accident

The leak was caused by a horizontal crack around 60 cm long on the lower angle welding
bead which secured a plate. After the drum sodium drainage and the first investigations,
identical faults to those observed on the plate at the origin of the leak were found on similar
plates. Numerous other observations made in-situ showed that cracks of similar type, but

94
nevertheless less severe, existed not only in the vicinity of the plate-supports but also on the
structure welds of the drum.

The reuse of the initial sodium drum after repair proved to be impossible and it was necessary
to define replacement. When the drum was removed, it was found out that long (several
meters) cracks had also formed in the constituting weld beads of the main vessel.
Considerable effort was expended on researching into the cause of the fault, in order to
determine the follow-up necessary (reuse or not of the other facilities in 15 D3 ferritic carbon
steel grade), and to assess possible risk on other structures or components.

The destructive examination samples taken at the beginning of 1988 showed that cracking
was very probably due to:

The existence of start sites (micro-cracking) in zones of high hardness;

Embrittlement by hydrogen, and
The cracks developed disruptive zones under the influence of residual welding stresses
close to the elastic limit of the material.

3.2.3.4. Dispositions chosen for installation repair

After weighing up a considerable number of solutions to solve the problem posed by the
occurrence of the leak, it appeared from October 1987 that the choices were between two
categories of solutions:

The first solution consisted of reconstructing the two geometrically very similar
drums, but in austenitic steel;
The second solution consisted of abandoning the decay storage function of the drum
(this function being therefore assured in the reactor itself), and maintaining only the
function of subassembly transfer.

It was the last solution that was finally chosen, in March 1988, particularly because it allowed
personnel to resume more rapidly operation in normal conditions. The new device was called
the fuel transfer station (PTC). The implementation of this solution required dismantling the
fuel storage drum. This was started by the conversion of the residual sodium into sodium
carbonate through a controlled additional of water vapor and carbon dioxide gas.

This operation (began in August 1988) allowed placing the vessel in contact with atmosphere
and were performed with pressure suits. So as to ensure compatibility of site-work with plant
operation, a drum work zone (zone de travail barilla, ZTB) was created to isolate the repair
area from the rest of the reactor building, not only from the point of view of ventilation but
also detection, protection, fire risk, health physics, and handling operations. In July 1989, the
dismantling was over.

The schedule for the realization of the PTC shows that it will be available at the beginning of
1992: the in-situ assembling of the new stainless steel chamber started in December 1989.

Reactor operation was not possible without a fuel subassembly discharge route during the
PTC construction. Consequently a special flask was designed to transfer the subassemblies
directly from the reactor to the fuel cave.

In 1982 the necessity for fast reactors in France was less acute and therefore there was less
need for a dedicated reprocessing centre. This led NERSA to decide on the on-site

95
construction of temporary storage for several spent cores. This was referred to as the fuel
storage pool building (APEC). Then in 1988, when repair of the storage drum turned out to be
impossible, NERSA chose to eliminate it altogether and replace it with a gas-filled transfer
chamber.

This modification, which was made possible through the existence of the fuel storage pool
building, in turn led to modification of the management mode of the core which was based on
frequency one. The core was renewed entirely after a cycle of 2 to 3 years (640 EFPD).
Replacing the subassemblies at a later stage required 7 to 8 months delay, including an initial
period of 2 months, for decay of the first subassemblies to a level of 7.5 kW.

The APEC and the PTC make up two links of the handling line. Construction of the APEC
covered the period 1984 to 1989, and that of the PTC lasted two years (1990/91). The PTC
was a facility that implements a fuel removal process with no intermediate storage between
the reactor core and the water filled pool. This facility was very different from the ferritic
steel fuel storage drum.

The subassemblies were placed by the transfer machine in a sodium-filled container which is
carried by the pot in the A-frame. In the argon-filled fuel transfer station under argon, the
container carrying the subassembly is placed on a pivot arm which transfers it to the handling
line. The pivot arm has a third position to receive the new subassemblies to be loaded.

After washing, the subassemblies are placed in a transfer shuttle which can receive three
subassemblies. The APEC offers storage for approximately 1 700 subassemblies between the
pool with a capacity of about 1 400 subassemblies and a hall which can house about
300 subassemblies in casks under gas atmosphere. Unloading capacity of the handling line is
about four to five subassemblies per day.

3.2.4. Primary sodium contamination

After plant shutdown to permute diluents (7 September 1989 to April 1990) and shutdown for
works as a result of detection of leak on the feed water purification circuit (28 April 1990 to
31 May 1990), the reactor was moved into critical mode and reached its nominal power on
11 June 1990. During all this period up to the temperature build-up (12 June 20% nominal
power), sodium clean lines were monitored by the operating team on recording delivered by
the plugging indicators using the usual operation methods: plugging temperature, general
slope of the curve. Monitoring confirmed that there was a good level of cleanliness; in
particular, the 110C level lasted around six hours.

That phase of operation did not last very long. Already during startup, elevated plugging
temperatures were found in the primary sodium system and continued to rise excessively. In
normal operation, the plugging temperature must be under 120C. The technical specification
required: between 120 and 150C to shut down within one month, above 150C to shut down
immediately. The criteria were established to limit the corrosion rate and to avoid impurity
crystallization during fuel transfers and maintenance outages at 180C.

On 10 June, during power escalation, the plugging temperature raised and stabilized at 140C.
This corresponded to a small and normal contamination of the sodium caused by maintenance
during the outage. Normally this is a temporary situation and the plugging temperature
decreases after a while. On 20 June, a second inflexion point appeared on the flow curve. This
point corresponded to a plugging temperature of 180C.

96
Indeed, an accurate analysis of the flow curve highlighted several inflexion points
corresponding to several plugging temperatures. These plugging temperatures can be
indicative of different types of impurities contained in the sodium. According to the chemists,
the plugging temperature of 180C could be disregarded because it was giving indications of
hydride (and not oxide) that would be quickly eliminated through the stainless steel
intermediate heat exchanger tubes to the secondary system thanks to hydrogen permeability.
It was then decided to take into account the plugging temperature of 140C only, and so to
remain in power.

On 26 June, one of the two filter cartridges of the integrated primary purification system
clogged. On 30 June, the second filter cartridge also clogged. As the purification system was
unavailable, the unit was planned to be shut down on 3 July to replace the saturated
cartridges, and indeed shutdown occurred automatically due to an electrical defect. After
reactor fast shutdown, it was maintained isothermal at 250C, ~ 40C above the oxygen
saturation temperature in the primary sodium when the rate of impurities was maximum
(15 ppm oxygen) to avoid crystallization in the coldest parts of the reactor. Meanwhile,
investigations were carried out to identify the origin of this contamination.

The argon taken from the primary circuit is then channeled by a compressor through a
measurement chamber (Fig. 20).

1-compressor bode
2-main rod
3-intermediate ring
4-top of connecting rod
5-head
6-closures
7-compression chamber
8-neoprene working diagram
9-neoprene safety diagram
10-safety chamber
11-control unit

FIG. 20. Compressor of the SPX gas system.

97
Investigations of the reasons indicated that the membrane of one compressor had been
defective for a long time already, and the argon cover gas therefore had been exposed to
atmospheric air. About 600 L/h of an argon and air mixture entered the reactor argon circuit
downstream of this circulation pump.

There was no system to measure the purity of the argon cover to provide an alarm for the
operator in case of contamination of this neutral gas. The amount of sodium oxide produced
was estimated to amount to 350400 kg. Subsequent cleaning of the contaminated sodium
took several months and was managed by the in-plant cold traps.

To remove these impurities personnel operated with the integrated primary cold traps, and the
two new cartridges operated clogged in about 4 weeks retaining about 100 kg of oxide. After
temperature increase in the primary circuit in order to dissolve residual oxides on the sodium
surface and on cold metallic parts, we were free of sodium purification at the end of January
of 1991 without clogging the second pair of new cartridges. The safety valve on the primary
argon circuit, which was also polluted with oxides have also been replaced.

The other corrective actions concern:

Review of the argon circuit design to avoid air intakes, and in particular setting up
instrumentation on the compressors for detecting membrane failures;
Installation of a chromatograph for periodic monitoring of the argon cover gas;
Review of the technical specifications in order to give more accurate criteria, based on
the duration of low temperature on the flat on the plugging temperature recorder and,
on the unplugging temperature more representative of the saturation one;
Estimation of induced risks of corrosion, plugging and aerosols.

Besides, this incident gave the opportunity of further knowledge on the chemistry of sodium
(new tests on the Cadarache loop), on the behaviour of integrated cold traps and on the
workings of plugging indicators.

3.2.5. Argon leak from the sealing bell of the intermediate heat exchangers

The argon supply of the bell gas cavity was carefully monitored by the operator. A significant
increase in the frequency of bell re-inflation on one IHX was observed in December 1994.

The leak was plugged by installing a metal sleeve applied and held in position by permanent
deformation (expansion by passing hydraulic pressure) using two crimping on either side
of the leak on the inner side of the pipe (pipe co-expanded with the sleeve). The equipment
used to install the sleeve was then left in the piping, and was subject periodic monitoring.
Corrective action was taken and in the latter part of 1995 the plant was ready to operate.

The seal between the intermediate heat exchangers (IHX) and the inner vessel for the SPX
was provided by a gas (argon) seal sealing bell surrounding the IHX (Fig. 21).

98
FIG. 21. Gas (argon) seal sealing bell of the IHX.

99
 BIBLIOGRAPHY

ALANCHE, J., et al., Phenix steam generator sodium/water reaction incident, paper presented
in Mtg. Steam generator failure and failure propagation experience, 26-28 September 1990,
Aix en Provence, France.
ASTY, M., et al., MIR Inspects Superphenix, Nucl. Eng. Int. 31-381 (1986) 35.
BARBERGER, M., et al., The Creys-Malville plant fast breeder station; test and start-up,
Proc. Am. Pow. Conf., Chicago, Vol. 48 (1986) 679-682.
BROCKMAN, K., et al., The influence of sodium aerosols on fast reactor operation, Proc. Int.
Conf. Fast Reactors and Related Fuel Cycles, FR91, 28 October-1 November 1991, Kyoto,
Japan, Vol. 4, (1991) 6.3-1 - 6.3-11.
CREYS-MALVILLE, Nuclear Power Station, Construction of the World's First Full Scale
Fast Breeder Reactor, Nucl. Eng. Int., (1978) 43-60.
ELIE, X., CHAUMONT, J.M., Operation experience with the Phenix prototype fast reactor,
Proc. Int. Conf. Fast Reactors and Related Fuel Cycles, FR91, 28 October-1 November 1991,
Kyoto, Japan,Vol. 1, (1991) 5.1-1 - 5.1-10.
GOURDON, J., et al., Superphnix physics, Nucl. Sc. and Eng. 106.1 (September 1990) 1-97.
GUIDEZ, J., MARTIN, L., Phenix: thirty years of operation for research, reactor renovation
overview and prospect, paper presented in Int. Conf. Fifty years of nuclear power - the next
fifty years, 27 June-2 July 2004, Obninsk, Russian Federation.
INTERNATIONAL ATOMIC ENERGY AGENCY, Status of Liquid Metal Cooled Fast
Breeder Reactors. Technical Reports Series No. 246, Vienna (1985).
INTERNATIONAL ATOMIC ENERGY AGENCY, Status of Liquid Metal Cooled Fast
Reactor Technology, IAEA-TECDOC-1083, Vienna (1999).
INTERNATIONAL ATOMIC ENERGY AGENCY, Fast Reactor Database, IAEA-
TECDOC-866, Vienna (1996).
LACROIX, A., et al., Experience gained from 1200 MWe Superphenix FBR operation, Proc.
Conf. Fast Reactors and Related Fuel Cycles, FR91, 28 October1 November 1991, Kyoto,
Japan, Vol. 1 (1991) 2.2-1 - 2.
MARTH, W., The story of the European fast reactor cooperation, KfK 5255,
Kernforschugszentrum Kalsruhe GmbH, Karlsruhe, Germany (1993).
MARTIN, L., et al., Life extension of Phenix NPP, paper presented in Mtg. Operational and
decommissioning experience with fast reactors, Cadarache, France, 1115 March 2002.
MARTIN, L., Leak before break operating experience from European fast reactors, Proc. Int.
Conf. Fast Reactors and Related Fuel Cycles, FR91, 28 October1 November 1991, Kyoto,
Japan (1991) Vol. 1, 5.4-1 - 5.4-14.
MERGUI, A., Commissioning the World's First Commercial Scale FBR at Creys-Malville,
Nucl. Eng. Int. 33-406 (1988) 20-24.
MERGUI, A., et al., Experience of the 1200 MWe Superphenix FBR Operation, Proc. Am.,
Pow. Conf., Chicago, 1990.
PEROTTO, G., et al., Repair of the Creys-Malville fuel storage drum, Proc. Int. Conf. Fast
Reactors and Related Fuel Cycles, FR91, 28 October1 November 1991, Kyoto, Japan,
(1991) Vol. 4, 6.1 - 6.1-11.

100
RAHMANI, L., DECHELETTE, S., BANDINI, C., SPX significant events and whether it
would have happened on EFR, Unusual Occurrences During LMFR Operation, IAEA-
TECDOC-1180, IAEA, Vienna (2000).
SAVAGE, J.-F., Phenix 30 years of history: the heart of reactor, CEA/EDF, CEA Valrno-BP
17171-30207 Bagnols-sur-Ceze cedex, France, 2004.
TARBY, S., et al., L'atelier pour l'vacuation du combustible de la centrale de Creys-Malville,
paper presented in Int. Symp., 1985, Lyon, France.
VERRIERE, Ph., et al., Maintenance and repair of LMFBR steam generators, paper presented
in Mtg. Maintenance and repair of LMFBR steam generators, 48 June 1984, O-Arai, Japan.

101
4. BN-600 AND BN-350 REACTORS

4.1. BN-600 POOL-TYPE LMFR

4.1.1. Commissioning and design features

The nuclear power plant (NPP) BN-600 has been operating since 1980 as the
Beloyarsk-3 power plant. The NPP construction cost was ~ 312 million Rubles [1980]
[approximately 620 million US$ (1980)]. The planned budget was exceeded by less than 5%.

First criticality was reached on 26 February 1980. The basic result of the physical startup in
March 1980 (213 low (21%) enrichment fuel subassemblies (FSAs), 143 high (33%)
enrichment FSAs and 13 permanent reactivity compensators) showed that the measured
physical characteristics of the reactor were correspondent with the design values.
Measurement of sodium flow through each FSA was carried out two times: before and after
the power startup of the reactor.

The primary circuit hydraulics was investigated, at zero reactor power, both under
steady-state conditions and simulated emergency conditions. All loops of secondary circuit
were filled with sodium in February 1980. The investigation showed that the hydraulic
resistance of the loops was two times below the design values. Power startup began on
5 April 1980. In mid of June power 40%, in mid of August 80% and on 18 December the
reactor power reached the nominal power level 1470 MW(th) and 600 MW(e).

The design of the reactor is generally on the pool (integral) concept, with all the main primary
circuit components placed in one vessel: the core, the radiation shielding, the main circulation
pumps, the intermediate heat exchangers, and also the piping. The integral concept enables
very compact reactor unit, which is reliable and safe as regards cooling of the core and
confining radioactivity.

Pipelines with high temperature coolant, operating under stress are excluded, also are the
cumbersome electric heating cables and the sealed concrete cells for location of the primary
equipment. Less metal is used for the components, and the amount of construction work is
greatly reduced. The surface area of load-bearing walls separating radioactive sodium from
the external environment is largely reduced. Absolute leak-tightness of the main primary
circuit pipes is not required, as leaks are confined within the reactor vessel.

The reactor vessel is of simple cylindrical shape, 12.8 m in diameter and 12.6 m high, with no
nozzles below the sodium level. The low cover gas pressure in the reactor (~ 0.4 kg/cm2)
enables the large-sized reactor vessel to be made with a small wall thickness (30-40 mm).
As experience has shown, this kind of vessel can be assembled on-site, from individual
factory-produced parts, with minimal problems. It may be mentioned that the geometric
dimensions of the BN-600 reactor vessel are almost the same as those of foreign pool type
reactors-Phnix and PFR - although BN-600 plant power is some 2.5 times higher. This has
been achieved as a result of using rationalized designs for the reactor and the equipment

The distinctive feature of the BN-600 reactor is bottom support of the reactor vessel which
gives, in the designers' opinion, certain structural and technological advantages compared
with alternative option of top-suspended reactors. Through a support ring welded at the point
where the cylindrical wall joins the base, the vessel is seated on foundation roller supports.

102
Inside the vessel on the support ring is a rigid box-type support structure carrying the pressure
chamber with headers and the core, the breeder blanket, the spent assembly store, and also the
in-vessel radiation shielding, pumps and intermediate heat exchangers.

Thus all loads from the weight of the vessel, the in-vessel structures and the mass of sodium
(~ 700 tons) are transferred to the foundation of the reactor well. The main advantage of this
support layout is the load removal from the upper part of the reactor, which operates at the
highest temperatures.

In the neck of the conical reactor cap there is a rotating plug with an eccentrically positioned
rotating column in it, which carries the safety system drives and the two refuelling
mechanisms. By combined rotation of plug and column, the refuelling mechanisms and grab
system can be aligned with any recess in the core, the radial breeder or the in core storage. At
the same time the rotating plug and column constitute a component part of the upper
biological shielding of the reactor. Freezing liquid seals are used to prevent leaks.

To guarantee core cooling even in the event of hypothetical seal failure of the reactor vessel,
the vessel itself and the adjoining pipe runs are enclosed in a sealed safety casing. Even
complete filling of the space between vessel and casing will not cause a break in sodium
circulation around the primary circuit. On the outside the safety casing is coated with thermal
insulation. The whole reactor unit is contained in a concrete well covered from above by a
protective plate.

Primary circuit components - the heat exchangers and pumps are installed in cylindrical
support sleeves passing through corresponding necks in the reactor vessel cap. Connection
of the sleeves to the vessel and the safety casing is effected using a bellows compensating
the difference in temperature expansions (the relevant temperature difference < 100C).

Heat removal from the core takes place via a three-circuit layout: there is sodium in the
primary and secondary circuits, and water and steam in the tertiary circuit (Figs 1 and 2).

The primary heat exchangers are situated behind the radiation shielding, in an area of low
neutron fluxes, so the secondary sodium reaching the steam generators has virtually no
activity. The secondary circuit also provides shielding of the reactor against the consequences
of possible faults in the steam generators. Coolant circulation takes place via the three parallel
primary circuit loops, each of which incorporates two heat exchangers and one vertical
immersion-type centrifugal pump with direct intake from the reactor tank.

The pumps with a delivery of about 10 000 m3/h each, are situated in the cold part of the
circuit, after the heat exchangers, which ensures low and stable temperature conditions for
their operation. Sodium at a temperature of ~ 380C is fed by the pumps to the pressure
header, from which it is distributed around the assemblies of the core and radial breeder
according to the maximum heat release in them. Part of the coolant goes to cool the reactor
vessel (~ 1000 m3/h), the spent assembly store and the in-tank shielding. Emerging from the
assemblies at a mean temperature of 550C, the sodium passes into the upper space of the
reactor, from where it flows out via six intermediate heat exchangers.

The secondary circuit sodium, at a temperature of 320C, is fed down a central fall pipe into
the lower space of the heat exchanger, from where it rises up through tubes, becoming heated
to 520C by the heat given off by the primary circuit coolant. The ascending movement of the
coolant in the core of the reactor is chiefly selected on the basis of the conditions for the
development of natural circulation.

103
There are no shut-off valves in the primary circuit. Shutting down one loop is carried out by
forced closure of a non-return valve with the corresponding primary circuit pump halted. For
the power plant to operate at intermediate power levels with parameters close to nominal, a
controllable drive is used in the pumps, providing a smooth variation in the rotation speed
(and the sodium flow rate) in the range 25100% N nom.

1, 2-core, fuel assembly, 3-primary pump, 4-intermediate heat exchanger (IHX), 5-central column,
6-control rod drive mechanism, 7-loading-unloading elevators, 8-neutron channel, 9-neutronic
measurement chambers, 10-reactor supports, 11-reactor vault, 14-rotating plug,
15 neutronic protection, 16-refuelling cell

FIG. 1. BN-600 reactor block.

104
 1-core, 2-primary pump, 3-intermediate heat exchanger, 4-secondary pump, 5-buffer tank, 6-steam generator,
7-turbine, 8-generator, 9-condensers, 10-condensate pumps, 11-low pressure heaters, 12-deaerator,
13-feed electric pumps, 14-high pressure heaters

FIG. 2. Principle heat diagram of the power plant BN-600.

There are no shut-off valves in the primary circuit. Shutting down one loop is carried out by
forced closure of a non-return valve with the corresponding primary circuit pump halted. For
the power plant to operate at intermediate power levels with parameters close to nominal, a
controllable drive is used in the pumps, providing a smooth variation in the rotation speed
(and the sodium flow rate) in the range 25100% N nom. The core is made up of a compact
grouping of around 400 hexagonal assemblies with an across-flats size of 96 mm and a
spacing interval of 98 mm. An assembly (subassembly) comprises 127 pins of 6.9 mm
external diameter contained in a casing tube with a perforated spike at the bottom for a feed of
coolant from the pressure header, while at the top there is a head for gripping by the refuelling
mechanism during assembly transportation; 27 holes in the core are occupied by the control
and safety system rods of the reactor.

The fuel charging of the core and the efficiency of the reactivity compensation system ensure
continuous operation of the reactor for a period of some 150 days (at full power).
The duration of a core run is around 450 days, so that at each refuelling 1/3 of the assemblies
which have achieved the design burnup are replaced. Around the core assemblies are
assemblies which are identical in external configuration, forming the radial breeder.
The elements of these assemblies are filled with waste uranium dioxide. The axial breeders
consist of the same material. These are situated in the same claddings as the pins. Behind the
radial breeder there the positions for the in-pile store. It is used to cool assemblies removed
from the core before they are removed from the reactor. The BN-600 reactor is designed to
operate on both uranium and plutonium fuel, and a gradual transition to breeder mode by
making up the reactor with plutonium bred if it is possible. Fuel subassembly refuelling takes
place with the reactor shut down using a set of mechanisms comprising. Figure 3, as an
example, shows of the loading/unloading system of the BN-600, which is in operation since
1980. The latter comprises:

105
 Three of an eccentric arrangement rotating plugs with two in-pile refuelling
mechanisms (close and distant relative to the reactor core axis) installed on the small
plug, which carry out replacing of assemblies inside the reactor;
Two drums for new and spent fuel assemblies;
A spent fuel-to-washing cell transfer mechanism;
Fuel transfer and washing cells;
Two inclined loading-unloading elevators, which transport the assemblies from the
reactor to the transfer box of the handling and transport channel and back;
An assembly transfer mechanism situated in the transfer box, which transfers
assemblies from the elevators to the handling and transport channel and back.

The sodium level in the reactor vessel is such that transportation of spent assemblies inside
the reactor takes place under a layer of sodium. This prevents any danger of inadmissible
heating of fuel being moved under conditions where there is no forced coolant circulation.
Subsequent cooling of spent assemblies in the in-pile storage with organized cooling enables
the level of heat emission in the fuel to be reduced substantially, which simplifies subsequent
handling of it in the external handling and transport channel. A rectangular shape was adopted
for the BN-600 reactor building after comparing different versions (Fig 3 ad).

1-reactor, 2-reactor hall, 3-secondary pump, 4-crane, 5-ventilation system, 6-water pool, 7-irradiated fuel transfer
cask, 8-sodium storage tank, 9-electric heating control system, 10-turbine hall, 11-control and protective system,
12-steam generator (SG), 13-crane (for SG)

FIG. 3a. Nuclear island and turbine building layout-elevation.

106
 FIG. 3b. The BN-600 power plant building.

FIG. 3c. BN-600 reactor central hall

(on the top-secondary pumps, bottom-reactor protective dome (red).

107
 FIG. 3d. Turbine hall of the BN-600 NPP.
4.1.2. Equipment tests6

A special attention was given to verify the correctness of the design principles used in the
BN-600 and the functional qualities of the equipment. The primary circuit hydraulics tests
were performed on a 1:6.8 scale reactor model with operating pumps (Fig. 4).

FIG. 4. Hydraulic model of the BN-600 reactor.

6
For more detailed information see: LEIPUNSKII, A.I. et al, The BN-600 fast reactor; BUDOV, V.M., et al.,
A NPP BN-600 - the plant for the near future; Reports presented at the Nuclex-75, Basel, 1975.

108
The data produced in this way enabled checks and changes to be made with the results of
calculation studies, on such basic issues as:

The distribution of the coolant flow rate through the heat exchangers in various reactor
operating modes with one loop not operating;
The variation in the coolant level in the reactor vessel and the pump stand pipe under
various reactor operating modes;
The absence of any mechanical entrainment of gas by the coolant from the reactor
vessel into the circuit, etc.

On a separate rig a study was performed of the coolant distribution through assemblies in the
pressure headers. The influence on the flow rate through the assembly of the position of the
holes in the assembly spikes was determined. Tests and final adjustments to the refuelling
mechanisms were performed on sodium rigs with circulation of hot sodium. These tests
particularly envisaged the performance of a number of studies in artificially created
accident situations, such as failure of the control system, complete loss of power to
mechanisms, high oxide content in the sodium, etc. Comprehensive checking of the refuelling
system together with the control system was performed on a control assembly rig which
comprised: metal structures simulating the vessel and reactor transfer cell, rotating plugs,
central column with control and safety system drivers, ramps, mechanisms for refuelling and
transfer of assemblies, dummy subassemblies, a simulated inlet plenum and electrical
equipment and movement systems. On the rig, a number of subassembly refuelling operations
were carried out, both in normal mode and in conditions simulating of all types of events. The
latter included:

Loading and unloading of deformed subassembly with a 10 mm bend, with

subassembly rotated through 60, 120 and 180;
Loading and unloading of subassembly, with misalignment of axis of refuelling
mechanism relative to subassembly head axis by 10 mm;
Checking possibility of disconnection and removal of refuelling machine from
jammed subassembly, raised to a height of 500 mm.

Under all these conditions the refuelling mechanisms were found to function satisfactorily.
The refuelling mechanisms, elevators and fuel transfer mechanisms in the cells have been
operated without any disturbances; above 40 reactor reloads have been carried out since the
power unit BN-600 was put into operation. The total operating age in terms of double strokes
of the in-reactor refuelling mechanism amounts to 36 100, for the elevators to 9 060 and for
the ex-reactor fuel transfer mechanism to 29 200. These exceed the respective design values.
Based on results of audits of the refuelling mechanisms, their operating life has been
extended.

4.1.2.1. Sodium rig for testing of drivers and control and safety system rods

Experimental final design work and full life tests were performed on the control and safety
rod drive and their rods using a special sodium rig at a temperature of 550-580C (Fig. 5).
The control and safety drive prototypes with absorber rods worked about 16 000 h on the rig,
during which time the following operations were carried out:
Automatic power regulation drive mechanism - 7 500 raising and lowering;
Safety rod drive mechanism - 650 drops in fast scram mode and 600 raising and
lowering in slow emergency protection mode;
Shim rod drive mechanism - 2 600 raising and lowering.

109
These lengthy tests confirmed the reliability of all the control and safety rod drive
mechanisms and the conformity between the main working characteristics and the design
characteristics.

FIG. 5. Sodium rig for testing of drives, control and safety system rods for the BN-600.

4.1.2.2. Intermediate heat exchanger (IHX)

The IHX shell-and-tube sodium/sodium heat exchanger has 4 974 OD 161.4 mm tubes with
expansion bends. In order to minimize the temperature difference between the bundle and the
central tube, a double-tube arrangement was introduced. The annular gap is gas-filled. The
perforated plate at the below part of the IHX was used to reduce the temperature difference
between central tube and bundle. The distribution of the coolant flow rate through the heat
exchangers in various reactor operating modes, including modes with one loop disconnected
were performed on a 1:6.8 scale reactor unit model (Fig. 6). The study of IHX thermal
hydraulics was performed on the mock-up tube bundles with the limited number of tube rows
(one-, three-row bundle).

110
 1, 2-secondary Na inlet/outlet; 3-shielding block; 4,9-primary Na inlet/outlet; 5-seeling element;
6-central downcomer tube; 7,8-heat exchange tube; 10-lattice

FIG. 6. BN-600 intermediate heat exchanger.

The low cycle fatigue and fracture investigations done on tube-tubesheet joint IHX nozzle
model and IHX to IHX support shell joint. For the tube-tubesheet joint, two models with
19 and 37 tubes respectively were studied. The test results for heating rate 50-70 K/h hold
period of 3 h at 550C and cooling rate with of 5-6.5 K/s had not indicated any fatigue
damage.

The tests on nozzle model with the heat rate 60-90 K/h hold at 580-600C for 3 h and cooling
in time 30-40 s up to 400C revealed no crack initiations in the all surfaces of 2 test models.
The calculated peak to peak stream values was 0.62% which can permit up to 3 000 cycles.
The last aspect was on the creep fatigue damage on IHX support shell joint. The calculations
showed that the margin on number of load cycles is less than the permissible value. Hence
tests were conducted on the flat model, with 12Cr18 Ni steel.

The test temperature was 490C. The calculation and experiments comparison showed a
good coincide. Accordingly crack initiation in the weld bead zone was observed after

111
1 100 cycles by acoustic emission method for the clearance of 0.5 mm between the shells. If
this width is increased to 2 mm, cracks were seen only after 2 000 cycles. All six IHXs are
operated since 1980 without any faults or troubles.

4.1.2.3. The pumps

The pumps underwent various model and full scale tests on rigs, where there was the
possibility of simulating widely (bearing) varying operating conditions were carried out to
guarantee the required high level of operational reliability. When the pumps were being
designed, much thought was given to the hydraulic side, using computers.

Models of the pumps were created, and these were subjected to various hydraulic tests with
water. Some of the pump units (shaft seal, bearings) were tested on special rigs, which
enabled design changes to be made to them promptly and various designs to be tried out.
Besides the usual testing of the functional properties of the pumps and their drives in all
possible operating conditions, the full scale water tests also included checking of the pumps
under extreme conditions, e.g. during operation under cavitation conditions, with
non-regulation start-up, with delivery overload, etc. Rigs for testing pumps with hot sodium
enabled the hydraulic and electro-mechanical characteristics of the pumps and their parts to be
checked under various thermal conditions. Successful long term life tests on pumps confirmed
the correctness of the designs selected for them.

The strength of the highly loaded reactor vessel parts was tested using models. In particular, a
1:10 scale model of the roof slab was made. The basic principle of the modeling work was
equivalence of the stress level of the model to that of the real-life version. On a rig
simulating a real pipe with normal pump, studies were performed of the vibration and
vibration-strength characteristics.

Construction work and the manufacture and assembly of equipment for the BN-600 began in
1969. The reactor plant building was made ready for equipment assembly in 1973. Assembly
of the main equipment commenced in 1974. By mid 1975 welding of the reactor vessel was
practically complete. While the vessel was being assembled, the secondary circuit pumps and
the component handling and storage equipment were also being put together. Assembly
work was completed in 1979. The power station with BN-600 reactor was commissioned in
April 1980.

4.1.3. Operating experience

Up to 1 January 2004 the reactor plant total on-power operation time amounted to 164 000 h,
and about 88 000 GWh of electricity was generated. The power histogram and load factor for
the unit operation are given in Figs 7 and 8.

It may be mentioned that the BN-600 is operating as a base load plant, with the reactor not
participating in the regulation of load and frequency in the power system. The reactor power
was therefore, nominal [600 MW(e)], 2/3 of nominal during operation without one of the heat
transfer loops is not available, or zero during shutdown. The time spent by the BN-600 in
each of these three states is characterized by the following data:

Nominal (600MW(e)): 68%

2/3 power: 11%
Zero power: 21%.

112
2001

2002
Nov

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Dec

FIG. 7. BN-600 operating histogram (reactor is being shutdown two times per year for refuelling,
1998-the rotating plug repair).

113
 100

90

80 83.53
80.29 79.89
78.19 77.35
76.60 75.89 76.32 76.43
74.11
70 71.76 72.75 72.48 73.46 72.97 73.23
Load factor, %

69.83 70.31
65.91
60
56.00
50
50.10
47.93

40

30
27.50

20

10

0
80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

00

01

02
19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

20

20

20
Year

FIG. 8. BN-600 load factor (1998-the repair of the rotating plug).

The unit load factor during initial years was as follows: 1980-27.5%; 1981-50.1%; 1982-56%.
With experience, the load factor rose steadily, shutdowns of the unit are now mainly caused
by routine preventive maintenance and refueling. The average load factor for the whole period
of the reactor plant operation on full power equals to 70%, in 2001- 79.89%. The highest load
factor (83.5%) was reached in 1992. As to reliability indicators the BN-600 power unit is
among the best NPPs in Russian Federation. The BN-600 control system is so constructed that
if any of the three heat-transfer loops are shut off the unit remains at 67% power. From 1982
on work was carried out by operational personnel to produce an algorithm for resumption of
operation of a disconnected loop without reactor shutdown, with power reduced to
around 20%.

There have been some 69 disconnections and connections of loops over nearly 22 years of
operation. Analyses showed, if the two-loop operating mode had not been envisaged, load
factor losses over the operating period which has elapsed would have been about 7%. This in
fact took place during the early years of operation. Experience in operation of the BN-600
(and also of other three-loop fast reactors) showed that the need to work with two loops arises
when equipment fails because of so-called chronic faults which require long-term repair (or
replacement). Under BN-600 conditions, such faults were found in primary circuit pumps,
steam generators and turbines in the initial phase of operation. After scrapping of the faulty
steam generator modules and adjustments to the pumps, the BN-600 operated with a minimal
number of loop disconnections. One may thus conclude that for serially-produced reactor
installations with fully-developed equipment, and one turbine for reactor, and steam
generators with high self power there is no special need to envisage stopping/switching of the
loop without reactor shutdown, especially if this means additional capital cost and a safety
margins decrease. The greatest losses associated with unplanned shutdowns (equipment and
system failures) at the beginning stage of reactor operation have been connected with the
primary circuit pumps.

114
Thus, in BN-350, there were disturbances in the pumps at start up after a long-term outage,
caused by solidification of sodium in the gap between the pump shaft and its casing. There
were also difficulties resulting from an increase in the sodium leakage flow from the impeller
delivery due to temperature variations in transients, which prevented the sodium level rising
or varying in the pump tank. During the initial period of operation of each of the reactors
certain technical problems were solved. This concerned adjustment of the systems and
components, bringing the reactor to nominal (rated) power, and feedback of operating
experience.

4.1.3.1. Primary pumps

Some problems in the operation of the BN-600 primary pumps arose in the process of coming
up to power in 1981 during a routine stage of increasing reactor power. When the pump
rotation speed was increased, increased bearing vibration was noticed. It was found that the
frequency-controlled motor coupled to the synchronous-rectifier drive induced torsional
pulsations of 68% of the nominal torque. When the torque pulsations coincided with a
natural frequency of shaft vibration resonances arose resulting in adverse consequences, such
as cracks in the shafts and failure of the couplings (section 14 in Fig. 9). During 1982-83
strain measurements were made to evaluate the pump shaft stresses and prohibited zones in
the shaft rotation frequency range were determined.

1-check valve, 2-lower scroll, 3-impeller, 4-upper scroll, 5-hydrostatic bearing, 6-shaft, 7-cover, 8-cooler,
9-level gage, 10-motor base, 11-face seal, 12-check valve drive, 13-radial-thrust bearing, 14-coupling, 15-motor
FIG. 9. BN-600 reactor primary pump.

115
In a number of cases the pump and the corresponding loop had to be disconnected, putting the
reactor into the 2/3 power operating mode. During repairs to the shutdown pump, breakages
were found in the half-coupling teeth and springs connecting the pump shaft to the electric
motor shaft. Analysis of the damage to these parts led to the conclusion that they were not
strong enough. These defects were eliminated by replacing the connecting couplings by
stronger ones. During subsequent operation of the primary circuit pumps, areas with high
amplitude of torsional vibrations of the pump shaft were found in the rotation speed range of
925-940 rpm. The dynamic deformation constant in the resonance zone did not exceed
20-25% of the static load on the pump.

The torsional vibration frequency coincided with the pulsation frequency of the
electromagnetic moment of the drive and the mechanical vibrations of the pump unit itself.
All these phenomena were, however, observed in an area other than the working rotation
speed of the pump. The removable parts of the primary circuit pumps have been replaced.
When the unit is at nominal power, the pumps operate in unregulated mode, i.e. in
squirrel-cage rotor mode. The use of advanced shafts and couplings and changing to a steady
mode of pump operation after attaining the preset reactor power level have eliminated any
failures of the reactor coolant pumps since 1985. There have been no problems with operation
of the secondary circuit pumps. Long term operating experience has allowed the factors which
limit the lifetime of the pumps to be identified. Cavitation-erosion wear of the impeller blades
is now the single factor requiring periodical removal of the primary coolant pump internals to
replace the impellers. The operators were not satisfied with the design lifetime of the face
shaft seals, which was limited by degradation of the rubber sealing collars (lose of elasticity
and cracking). The seals were replaced with spares, as a rule, during planned maintenance
outages. The design of the seals currently in use has been modified significantly compared
with the original design and has the following advantages:

Rubber sealing rings are used instead of collars - the possibility of misalignment of the
sealing surfaces is reduced by improvement of the items which transfer the torque to
the rotating sealing rings;
A wear-resistant graphite-to-graphite sealing pair is used instead of graphite-to-
steel;
The oil-cooling surface in the seal is increased.

On the basis of operating experience and resulting from the improvements made in design the
operating life for the pump internals has recently been increased significantly up to 50 000 for
the primary pumps and to 100 000 for the secondary pumps.

4.1.3.2. Steam generators

In the earlier years of the development of fast reactors there was some incidence of leaks in
the steam generators. Since leaks in sodium-heated steam generators are intolerable, however
small they may be, whenever a leak occurs the unit in question has to be shut down and
isolated for repair. This can be done without serious diminution of the power output only if
there are a large number of separate units, any one of which can be isolated without reducing
the total power significantly. The outstanding example of the advantages of the modular
approach to steam generator design is afforded by the Russian BN-600 plant. It can be seen on
Figs 10 and 11, and its main characteristics are shown in Table 1. This has three secondary
sodium circuits, each with 8 separate steam generator modules, and each of these consists of
separate evaporator, superheater and reheater sections, making a total of 72 separate heat
exchangers. At least partly because of this the availability of BN-600 has been consistently
high.

116
 1-main steam removal, 2-reheated steam removal, 3-feed water supply, 4-superheater, 5-evaporator, 6-reheater, 7-sodium removal, 8-sodium supply, 9-steam to superheater,
10-steam removal from evaporators, 11-steam to reheater, 12-sodium expansion tank, 13-steam-sodiumreaction products dump

FIG. 10. BN-600 modular stem generator layout in compartment design line. FIG. 11. BN-600 modular steam generator.

117
TABLE 1. MAIN CHARACTERISTICS OF THE BN-350 AND BN-600 SGs

Steam Tube Number Tube Tube Sodium Water/ Water/

generator material of tubes diameter, spacing, temp., steam (steam)
module per mm mm C pressure, temp.,
bundle bar C

BN-350 2 1/4 Cr-1Mo 816 322 (333) 44 350-420 45 250-260

evaporator

BN-600 2 1/4 Cr 1Mo 349 162.5 28 320-450 140 220-340

evaporator

BN-600 Cr18Ni9 239 162.5 33 450-520 135 350-505

superheater

BN-600 Cr18Ni9 235 252.5 33 450-520 28 505

reheater

Of the 12 cases of failure of the seal between circuits in the steam generator in only 2 cases
was it necessary to shut down the corresponding heat-removal loop, but the reactor; the total
duration of such shutdowns being around 400 h. In the other cases, the presence of valves
enabled the sodium and water/steam section to be cut off with the heat removal loop in
operation, virtually without reducing the reactor power. The modules which had suffered
leaks between circuits were removed and replaced. Studies showed that the most likely cause
of the water-into-sodium leak was manufacturing faults. This situation occurred in the
superheater modules (7 leaks) and reheater modules (4 leaks) which are made of austenitic
steel.

Studies revealed that the point of the original leak was in the area of the upper tube plates.
The suggested cause was the development during operation of initial defects at the point of
junction between tube and tube plate. In all the events irrespective of a leak size, no increase
in pressure in the expansion tank up to the alarm setting has been observed and the
operational power plant safety has remained within the limits. The precise design concept of a
'Large leak' event and the steam generator emergency protection system operation algorithms
were validated by the analysis of the leak behaviour. In 10 events the failed modules have
been replaced with new ones and in 2 events the failed modules have been restored and put
into operation again. The restored evaporator module has accumulated 7 280 trouble-free
running hours after the intervention and has been replaced with the new one for material
testing. The restored reheater module has accumulated 53 854 running hours after the
intervention and has been removed out of service on indications of water-sodium reaction
having totally accumulated 68 366 running hours. Thus the design concept of the
maintainability of the steam generator modules if they have been disconnected at the early
phase of the 'Small leak' event development, has been also validated.

Analysis of the incidents that had occurred led to the following conclusions:

The monitoring system worked satisfactorily under these conditions and in most cases
enabled the time of the leak, the section it was in and even the module to be
determined promptly;
The confinement system also operated satisfactorily. It ensured a controlled dumping
of argon-hydrogen mixture without the maximum design pressure being exceeded. At

118
 the same time problems were encountered in implementing the algorithm for rapid
creation of nitrogen counter-pressure on the tertiary circuit side following evaporation
of the steam-and-water mixture from the affected module. Operating experience
revealed the following possibilities:
Continuing operation of the loop with the section disconnected without reducing the
loop power;
Disconnection of the affected section without shutting down the reactor, and even
without disconnecting the loop;
Connecting up a repaired loop without shutting down the reactor.
To give the modules better functional ability, the results of analysis of affected modules were
used to develop and carry out a number of measures, particularly the installation of electric
heating of the entry sections of steam lines of the superheaters and reheaters. Generally
speaking, the sectional/modular design of the steam generator, fitted with the appropriate
shut-off and stop gate valves, has proved acceptable under operating conditions, though
relatively expensive.
4.1.3.3. Central rotating column
After 34 years of operation there has been some tendency for the rotation torque to increase
because of deposition of sodium oxide. The cause has been identified and modification to
correct the problem was done. During recent several years, growing increase of torque of the
central rotating column (CRC) has been observed within some angle of its rotation. By 1997,
it reached as high as about 60% of permissible value determined by the strength and capacity
of its drive. In 1997, CRC bearing assembly was inspected by drilling special opening and
using endoscope and some amount of sodium was detected there. This phenomenon was
interpreted as a result of sodium vapor transfer from the reactor cover gas via the gap between
the CRC and rotating plug (RP) and subsequent accumulation of sodium and its compositions
in the CRC bearing assembly. Works on partial withdrawal of the CRC from the reactor
vessel were scheduled in order to provide access to the bearing assembly for the sodium
removal, and appropriate preparations were made. This task was unique, since no procedures
of such a kind had ever been implemented on this type reactor in any country. During
inter-maintenance interval, the program and the process of the CRC withdrawal from the
reactor vessel were developed, and all necessary devices were designed and manufactured.
CRC withdrawal confirmed that almost 100% of the bearing assembly surface was covered
with sodium deposits filling the gaps between the bearing balls and cages. Besides, high
density sodium deposits filled the gap between CRC and RP in the areas of seizure. Both
cages and balls were withdrawn, sodium residues were removed from the bearing races and
CRC/RP gaps, and new balls and cages were installed. Currently, obtained results and
possible sources of this event are analyzed. The main conclusion is that owing to the
comprehensive work, CRC availability has been restored and also experimental data of great
importance for the fast reactors development and operation have been obtained.

4.1.3.4. Sodium leaks

In 20 cases the amount of poured sodium did not exceed 10 kg. In the rest 7 cases, the amount
of sodium released were 30, 50, 100, 300, 300, 600, and 1000 L respectively. All leaks were
detected in proper time by detection systems. All leaks took place in the auxiliary systems and
caused only one reactor shut down. Table 2 gives the main characteristics of sodium leaks in
BN-600 reactor. Over a total operation period (here means 24 years) 27 leaks into the
environment of which 5 events involved radioactive sodium leaks have occurred. Fourteen
events have involved passive burning of sodium.

119
TABLE 2. SODIUM LEAKS

System Number Quantity, No of Na

L burnings
Reactor 0 -
Intermediate heat exchanger 0 -
Storage drum 0 -
Primary auxiliary systems 5 -
Gas purification 1 0.1 -
Sodium purification system 4 0.3; 3; 0.2; 1000 1
Secondary circuit 18 -
Main pipelines 0 -
SG valve seals 4 1; 300; 30; 10 3
SG leak detection system 1 2.0 1
Drain and blow-off lines 10 0.2; 1; 10; 600; 300; 6
100; 0; 0; 1;
0000000000000000.0;
0.0;1.0
Sodium storage 3 1.0; 0; 0 -
Sodium reception system 4 10.0; 50.0; 10.0; 0.0 3

TOTAL 27 ~ 2 500 14

The main causes of sodium leakages were:

Poor quality repair: 8 events;
Latent defects of manufacturing and mounting: 6 events;
Depletion of equipment lifetime due to inadequacy of the design: 7 events;
Equipment design imperfections: 4 events;
Human errors during operation: 2 events.
On 7 October 1993 the largest leak happened on the pipeline for removal sodium from the
cold trap. The total amount of sodium escaped during the event was assessed to be
approximately 800 kg. The first symptoms of the emergency were short circuits in the main
and stand-by electric heating systems. Following actuation of the automatic alarm system by
rising radioactivity in the exhaust ventilation air duct from the primary sodium purification
system pipeline room, the supposed leaky section was isolated by the valves and the fire
ventilation system was activated. These operated normally. Despite these measures a buildup
of radioactive releases from the vent stack was registered as well as a deterioration of
radiation conditions in the reactor building rooms, including the main control room. The
release of radioactivity outside the plant was only 10 Ci. As a result, the primary sodium
purification system was isolated completely and the reactor was shut down and put into the
scheduled maintenance outage state a week earlier the assigned date. There was no
overexposure of the plant personnel and no contamination on or off the site was registered.

120
All this gave grounds for classification of the event as an anomaly, i.e. level 1 to the
International Nuclear Event Scale. After completion of the scheduled refuelling and the
necessary repair the reactor was returned to power on 24 October 1993. The probable cause of
a through crack in the joint between tubes with sodium flows of different temperatures could
be high thermal stresses due to temperature cycling in combination with stresses caused by
thermal expansion of the tubes. Despite the negligible radiological consequences it was
concluded from analysis of the event that in the next BN type reactors external primary
sodium pipework systems have to be excluded by arranging them inside the reactor vessel, or
(if impossible) by providing complete jacketing in order to have two leak-tight protective
barriers.

The largest leakage of secondary sodium happened in May 1994 in a drain pipeline
(ID 48 mm). Approximately 600 L of sodium were lost, but only about thirty kilograms
burned. The remaining sodium was retained and smothered with extinguishing powder in the
catch system. In both leaks the protective systems were effective. The damage was not
extensive and repairs were affected quickly. Of the 27 cases of leaks in only one case was it
necessary to shut down the reactor;

4.1.4. Reactor core: design and operation

Due to the core fuel failures in the period from 1983 till 1987 the reactor was shut down six
times for unplanned refuelling. Core design improvement became essential to provide
operating reliability and safety of the reactor plant. Examination of the failed fuel revealed
stress-induced corrosion of the annealed austenitic steel cladding as one of the main causes of
early failure. The cladding was damaged mainly in the peripheral region of the core. It was
due to the very unfavourable operating conditions for peripheral fuel assemblies. Because of
reshuffling and rotation in the course of operation, the fuel rod linear heat rating and cladding
temperature rose to 54 kW/m and 710C respectively at the end of a fuel cycle. In the
advanced core M design (the first core modification) the following changes had been
implemented to improve conditions for fuel operation:

The core active height was increased from 75 to 100 cm, decreasing the fuel rod
maximum linear heat rating to 47.2 kW/m;
Reshuffling and rotation of the fuel assemblies were eliminated;
Swelling-proof cold-worked austenitic steel was used for the cladding and for the fuel
assembly ducts.
By the end of 1987 the reactor core was completely assembled with advanced fuel
subassemblies. Loss of cladding integrity events virtually terminated resulting in substantial
reduction in fission product activity in the reactor gas plenum. The increase of caesium
nuclide concentration in the primary system had also stopped.

In 19901992 the reactor core was changed to the advanced Ml design (the second core
modification, Fig. 12). Ferritic steel was used in the new duct design and boron-modified
cold-worked austenitic steel for cladding. Fuel burnup in the core has reached 10% h.a. with a
fuel cycle length of 160 effective power days (efpd).

121
 FIG. 12. BN-600 core.

Development of advanced radiation-resistant steels was (and is) the main problem in the
attainment of higher fuel burn up. Considering the present status of this problem, a fuel
burnup of 12% h.a. (90 dpa) is believed to be quite realistic for an advanced reactor core fuel
subassemblies (Figs 13 and 14). This would give two options for the core refuelling pattern:
to increase the refuelling interval from 160 to 190 efpd or to change to a refuelling pattern
replacing 114 FAs at a refuelling interval of 145 efpd.

In either case the core reactivity margin will have to be increased, e.g. through expansion of
the medium fuel enrichment zone at the expense of the adjacent FAs of the low
enrichment zone. If the medium enrichment core zone were to be expanded to the limit of one
FA row, the heat rating of the fuel rods would remain at an acceptable level ~ 48 kW/m.

The core of the first type was designed for quite high performance (original design, Figs 13
and 14, Table 3). However, as the initial operating experience showed, it turned out to be
unreliable. Even during the first fuel cycles cladding loss of integrity events started, increased
swelling of fuel assembly ducts and some control rod items was observed, as well as loss of
ductility of the control rod guide tube material.

122
TABLE 3. BN-600 REACTOR CORE DESIGNS EVOLUTION
Performances Reactor core type
1 M M1
1. Reactor thermal output (max.), MW 1470 1470 1470
2. Reactor core diameter, mm 2058 2058 2058
3. Active core height, mm 750 1000 1030
4. Axial blankets height, mm
Upper 400 300 300
Lower 400 380 350
5. Number of different fuel enrichment zones 2 3 3
6. Fuel enrichment (U-235), %
LEZ 21 17 17
MEZ - 21 21
HEZ 33 26 26
7. Number of FAs in core zones
LEZ 209 136 136
MEZ - 94 94
HEZ 160 139 139
8. Core fuel cladding OD wall thickness, mm 6.90.4 6.90.4 6.90.4
9. Fuel rod length, mm 2400 2400 2400
10. Fuel rod gas plenum length, mm 808 653 653
11. Number of fuel rods in FA 127 127 127
12. Duct width across flats wall thickness, mm 962 962 962
13. Core structural materials:
Cladding* EI-847 EI-847 ChS-68
Duct Cr16Ni11Mo3 Cr16Ni11Mo3Ti **
14. Fuel rod maximum linear heat rating, kW/m 54.0 47.2 47.1
15. Fuel rod cladding peak temperature, C 700 700 700
16. Maximum fuel burnup, % h.a. 7.2 8.3 10
17. Maximum radiation dose to cladding, dpa 43.5 53.9 75.0
18. Fuel operating life, eff. d 200/300 300/495 480
19. Core fuel cycle, eff. d 100 165 160
20. Fuel inventory in core, kg 8260 11630 12090
21. Average fuel burnup, MWd/kg U 42.5 44.5 60.0
22. Temp./power reactivity effect, %, k/k -1.4 -1.3 -1.3

* EI-847 - Cr16Ni15Mo3Nb - austenitic steel, ChS-68 - Cr16Ni15Mo2Mn2TiB - austenitic steel;

** Cr12MoBnVB-ferritic-martensitic-steel.

123
124
FIG. 13. BN-600 reactor core fuel subassembly FIG. 14. BN-600 radial breeder subassembly.
(first type - the core active height 75 cm).
4.2. BN-350 LOOP TYPE REACTOR PLANT

4.2.1. Design features

The first-in-the-world demonstration commercial liquid metal cooled fast breeder reactor
BN 350 was designed and built by the former USSR, at present in the Republic of
Kazakhstan. The BN-350 reactor plant is located near the city Aktau (former Shevchenko),
Mangyshlak Region of Republic of Kazakhstan on the shore of the Caspian Sea (Mangyshlak
peninsula). It was designed and constructed as a two-purpose pilot-industrial power plant (for
electricity generation and heat production for seawater desalination) by organizations of the
former Soviet Union under the supervision of the Ministry of Atomic Energy (MINATOM).

The BN-350 sodium cooled fast reactor has been operated by the Mangushlak Power
Generation Company (MAEK) as a part of the industrial complex including the large capacity
co-generation power plant and the distillate production works providing electricity and fresh
water for Aktau and the adjoining industrial region, where natural fresh water resources were
unavailable. A particularity of the plant was that the reactor for electricity production was
physically separate from the desalination plant.

BN-350 has a dispersed (loop) arrangement of the primary circuit components, i.e. the
primary sodium pumps, intermediate heat exchangers, and valves are housed in separate
compartments (cells) and are connected with the reactor and by interconnected pipelines
(Figs 15 and 16). The temperature expansions of the pipelines are accommodated by the
bends. The reactor plant includes the following main components:

Fast sodium cooled reactor, six primary loops;

Six intermediate (secondary) loops;
steam generators,
Refuelling complex (integrated mechanical system);
Primary and secondary sodium purification system;
Automated process control system, including the reactor control and protection system
(CPS);
Diagnostic systems for monitoring the operating state of the safety related components
and systems.

During power operation core heat removal and transport to the working medium
(steam/water) are provided by a three-circuit flow scheme (Fig. 17). The primary circuit is
composed of six intermediate heat exchangers (IHXs), six primary sodium pumps (PSPs), and
sodium pipelines with gate and non-return valves. The pressure chamber with the core diagrid
and the upper mixing plenum above the core are the common sections of primary sodium flow
path.

The volume of the radioactive sodium in the primary circuit is equal to about 550 m3,
including 150 m3 in the reactor vessel, 170 m3 in the pipelines of the primary six loops, 6 m3
in six PSPs, 16 m3 in six PSPs overflow tanks, 30 m3 in five cold traps and ~ 130 m3 in six
IHXs.

125
126
1-reactor; 2-primary pump; 3-heat exchanger 1-reactor; 2-primary pump; 3-heat exchanger

FIG. 15. Vertical longitudinal section of BN-350 reactor. FIG. 16. BN-350 Nuclear island: vertical and horizontal section.
Sodium flow is distributed from the diagrid into the core and the radial blanket fuel
assemblies. A portion of the primary sodium (250 t/h) is removed from the pressure chamber
through throttles and utilized for cooling the reactor vessel and its outlet nozzles. There is a
capability to isolate each primary loop from the reactor using two gate valves on the suction
and pressure pipelines of the circuit. On the pressure pipeline of each loop downstream of the
primary sodium pump (PSP) a flap-type check valve is provided eliminating coolant
backflow in the event of a PSP trip in one loop when the other PSPs are operative. Basic data
for the BN-350 reactor plant are given in Table 4 for a nominal power of 750 MW.

The BN-350 reactor includes: the reactor vessel which contains the core diagrid with neutron
reflector and a set of core and blanket fuel assemblies; the reactor refuelling system; the above
core structure with CPS drive mechanisms and in-core instrumentation guides (Figs 1719).
The band and the core diagrid are made of Cr18Ni9 steel. Weight of the band with the
shielding subassemblies is about 35 tons, weight of the core diagrid being about 17 tons.

1-reactor vessel; 2-core diagrid; 3-reactor core; 4-reactor well liner; 5-lateral shield; 6-upper-stationary shield;
7-elevator, 8-refuelling mechanism; 9-FAa transfer mechanism; 10-fuel transfer cell; 11-protective dome;
12-control rod drive mechanism; 13-above core structure; 14-rotating plugs

FIG. 17. BN-350 reactor.

127
128
1-set of ionization chambers; 2-reactor core; 3-sodium outlet pipe; 1-sodium pressure chamber; 2-sealing ring; 3-diagrid; 4-neutron reflector;
4-elevator; 5-channels for additional ionization chambers 5-elevator lower support; 6-emergency protection rod sleeve;
7-reactivity compensation rod sleeve; 8, 9-throttling sleeves and devices

FIG. 18. Reactor plan view (cross section). FIG. 19. Core diagrid.
TABLE 4. BN-350 BASIC OPERATING PARAMETERS

Item Value
1. Reactor thermal output, MW 750
2. Primary sodium temperature at reactor inlet/ outlet, C 288/437
3. Sodium flow through reactor, t/h 141000
4. Secondary coolant temperature at SG inlet/outlet, C 420/260
5. Sodium flow in secondary loop, t/h 3400
6. Number of operating loops (plus one-reserve) 5
7. Main steam temperature / pressure, C/ MPa 405/4.5
8. Steam flow, t/h 1070
9. Maximum electric output of power unit, MW 125150
10. Distilled water output per day, t 100 000
11. Maximum neutron flux in core, n-cm-2 s-1 61015

The diameter of the cylindrical part of the reactor vessel is 6 000 mm, and the wall thickness
is 30 mm. In the middle section of the vessel there is a support belt by which the reactor is
located on 16 roller bearings arranged on a support shell of 5 850 mm diameter transmitting
the reactor weight load onto the foundations. The reactor vessel is enclosed in a guard vessel.
Thermal shielding, main reactor vessel, reactor guard vessel and supporting ring are
manufactured of Cr18Ni9 steel. Weight of thermal shielding of the reactor vessel is about
53.9 tons, and weight of the main vessel with the guard vessel and supporting ring is about
101.5 tons. The rotating plugs, ~ 2.5 m thick, have multilayer structure formed of steel sheets,
graphite and thermal insulation. Under the plugs the shielding is located made of Cr18Ni9
steel sheets of total weight of ~ 94 tons.

Outside the reactor vessel (in the reactor cavity) there are 7 650 mm high 3-grade steel lateral
shielding consisting of 160 mm thick plates, located inside iron ore filling shielding. Iron ore
filling is placed in the 3-grade steel reinforcing cage, having 20 mm thick walls; total weight
of the iron ore filling is ~ 1 303 tons. The next to it is the concrete covered with 20 mm thick
3-grade steel sheet lining (Figs 18 and 19). Under the vessel similar shielding structure is
provided. The total weight of 3-grade steel shielding is ~ 344 tons.

4.2.2. General results of operation

The BN-350 plant history is as follows (Fig. 20):

19651971: construction period;

29 November, 1972 first criticality of the reactor;
16 July 1973: power startup of the reactor. The extended start up was due to loss of
integrity events in four evaporators (detected by the appearance of hydrogen in the gas
plenum) when the SGs were filled with water;
end of 1973 February 1975: SGs repair; 19731975: operation at power levels up to
300 MW(th);
from March 1975: operation at 650-750 MW(th) for electrical power [~ 150 MW(e)]
generation and sea water desalination (~ 100 000 tons of desalinated water per day);
from January 1996 to June 1998: operation at 420 MW(th), 50 MW(e), producing
45 000 tons of distilled water per day;
April 1999: final shutdown.

129
 FIG. 20. BN-350-a freshwater source in the desert: overall survey.

For more than twenty five years, the operation of the power unit associated with the
BN-350 reactor have promoted the exploration of the new industrial region of Kazakhstan
which is rather rich in natural resources.

4.2.3. Steam generators and intermediate heat exchangers: operating experience

The BN-350 steam generators consist of two super-heaters with U-shaped tubes and two
evaporators with re-entrant tubes inside which water flows under natural convection and
partial evaporation conditions. Main data for the BN-350 SG (evaporator) is given in Table 1.
In general, the major plant components had been tested prior to assembly, and had resulted in
reliable operation in the main. However, it had been decided not to test the steam generators
for reasons of expense. The initial period of reactor plant operation was characterized by
unreliable operation of the SGs. Numerous loss of integrity events occurred in the re-entrant
evaporator tubes.

Metallographic examination of a great number of re-entrant tubes showed the presence of

microcracks in the tube-to-bottom weld joints (Fig. 21). Mechanical deformation of the tube
bottoms during cold stamping were acknowledged as the most probable cause of the
microcracks. Growth of the microcracks could occur under the effect of internal stresses
arising during welding the bottoms to the tubes and under cyclic thermal loads during
evaporator operation.

130
 1-tube sheet; 2-sodium; 3-downcomer tube; 4-heated outer tube; 5-lower weld seam (cracks);
6-bottom of Field tube; 7-boiler water; 8-steam-water mixture outlet

FIG. 21. BN-350 evaporator re-entrant tube.

After repair of the re-entrant evaporators when outer tubes of 322 mm (ODwall thickness)
were replaced by 333 mm tubes with machined bottoms, reactor plant operation was
continued with five loops at thermal power of 650 MW. The superheaters had caused no
problems, but when some of the peripheral tubes were found to be vibrating they were
plugged as a precautionary measure. The availability of stand-by components (six loops) and
the ability to operate with different numbers of loops (from three to five) provided for stable
operation of the reactor plant and production of electricity and fresh water. As a result, repair
and replacement work on the steam generators (as necessary) were carried out without
shutting down the reactor.

One of the repaired evaporators failed later due to a large leak of water to sodium. It was
dismounted and replaced with the micromodular SG Nadjozhnost-1 of Czech fabrication
(1980). Another SG of the same type Nadjozhnost-2 was put into operation in 1982 instead
of one of the re-entrant SGs which had been operated reliably since the reactor start up. In
general, the plant operation demonstrated the reliability and high operating performances of
the re-entrant tube SG design taking into account particularly their behavior in water to
sodium leak events.

131
The thermal-hydraulic disadvantage of the design was revealed by natural convection flow
instability during emergency residual heat removal. The use of Nadjozhnost-type steam
generator eliminates this disadvantage and in addition allows the removal of heat from the
SG by air flowing outside the modules under forced or natural convection conditions.
In January 1989, both these SGs failed due to stress-induced corrosion on the steam-water
side of the tubes promoted by non-uniform thermal-hydraulic conditions in the tube bundles.
After repair the SGs were put again into operation in 1993.

The reactor was designed for thermal output of ~ 1000 MW, but in the early periods of
operation its power level was limited by unsatisfactory operation of the steam generators.
To give of less demanding conditions for the operation of evaporators with re-entrant tubes a
maximum power of 130 MW(th) was specified for each SG. For the Czech-produced SGs
Nadjozhnost the allowable power level was set somewhat higher at 200 MW(e).
In addition, from an experiment carried out in 1976 on plant emergency cooling with loss-of-
normal power it was found that the available capacity of the steam/water system under these
conditions restricted reactor to 750 MW(th). This power level was not exceeded during
subsequent power operation periods, taking into account the power limitations imposed by the
SGs as well. During its operating life the reactor has operated at various power levels.

The average load factor in respect to allowed power levels was 85%. Reduction in the load
factor was due to outages for refuelling and planned maintenance of the equipment. The
reactor was normally shutdown for scheduled refuelling and maintenance two or three times
per year with outage durations of 20-30 days. Horizontal tube-and-shell IHXs with three
modules connected in series, made of U-shaped tubes were used in the BN-350 (Fig. 22).

1, 2-primary Na inlet/outlet, 3, 4-secondary Na inlet/outlet

FIG. 22. BN -350 intermediate heat exchanger.

The IHX of each loop consists of two sections connected in parallel both for primary and
secondary coolant flows. The IHX is located in a suction loop upstream of the primary
coolant pump, while in the secondary circuit it is in a pressure loop downstream of the
circulating pump. The IHX tube bundles can be removed if necessary and replaced with new
ones. The most stressed units in the IHX are the fixing joints for the tube module covers and
for the frame which stiffens the flat walls of the IHX body. Measurements of temperatures and
stresses in various items of the IHX were carried out during reactor plant operation. On this
basis requirements were formulated to limit the rate of the IHX heating-up in steps of 10%

132
specified power with delays of 510 h in each step. By 1998 the IHXs have operated more
than 180 000 h at various power levels without any disturbances and failures.

Radiation doses to the BN-350 personnel have been associated mainly with repair and
maintenance activities for sodium components in the primary circuit boxes. Radiation
conditions in the rooms (except the reactor well) depended substantially on the state of fuel
rods in the core. In the initial period of the reactor plant operation when the first type core was
used numerous cladding failures caused a significant rise of fission products activity in the
primary system. In 1979 the gamma radiation dose rate on the surface of the sodium
equipment at reactor shutdown reached 8.90 Sv/s, 80% of which was attributed to caesium
nuclides. Radiological conditions were improved significantly after completion (in 1979) of
the core change over to the second type fuel design. This was provided not only by reduction
in the number of failed fuel rods, but also by clean up of caesium from the primary coolant
using a special trap with a graphite absorber. Since 1984 the gamma radiation dose rate in the
primary circuit rooms during outages has not exceed of 1.5 Sv/s.

Due to the loop design of the reactor primary system the reactor well is virtually inaccessible
even at outages because of high radiation-induced activity of the reactor vessel structures.
Transport of gaseous fission products into the upper part of the control rod drive mechanisms
and to the oil system of the reactor coolant pumps turned out to be one of the reactor
operating peculiarities, which degraded radiological conditions in the reactor servicing zones.
The major source of the reactor radiological impact to the environment was gaseous
discharges from the equipment air cooling system and from the reactor plant rooms through
the vent stack. Improvements in the fuel design and the associated reduction in the number of
failed fuel rods in the core (which became quite rare in later operation periods) resulted in
radioactivity of the plant discharges to the atmosphere being determined by the radiation-
induced activation of air in the reactor well cooling system. Daily release of gaseous nuclides
was 0.55-0.74 TBq while that for aerosols was 1.910-6 TBq. Observations over many years
of radioactivity of the flora and fauna, and radiological conditions in the local populated areas
and in the sanitary restricted zone around the NPP, showed that those characteristics affected
by natural and man-induced radiation sources corresponded to background radiation levels.

4.2.4. Reactor core and refuelling equipment: design and operating experience

The reactor central part contains a set of core and blanket fuel assemblies (FAs) guide sleeves
for CPS rods and neutron reflectors installed in the diagrid (Figs 23 and 24). The diagrid is
attached to the sodium pressure chamber. The internal plenum of the diagrid is divided into
two chambers: high and low pressure. The sodium from the high-pressure chamber is
distributed for cooling the core FAs, control rods and FAs of the inner blanket. From the low
pressure chamber the coolant goes for cooling FAs of the outer blanket. Core FAs transferred
to the internal store around the outer blanket periphery are cooled by natural convection of
coolant in the reactor vessel. During the initial period of the reactor operation until 1979,
when the first design core (fuel rod of 6.1 mm OD) was used, large number of fuel failures
(loss of clad integrity events) occurred. These fuel failures caused a significant increase in
fission fragment activity in the primary circuit and consequently resulted in deterioration of
radiation conditions in the reactor plant cells. Therefore the second design of core fuel
assembly was developed with fuel rods of 6.9 mm OD. This advanced core provided for
increased fuel burnup and more reliable operation of the fuel rods, mainly due to the
following improvements:

133
 The gas plenum height in the fuel rod was increased at the expense of integration in
one clad tube (6.90.4 mm) of core and axial blankets material (fissile and fertile) and
reduction of the lower blanket height;
The fuel assembly duct material Cr18Ni10Ti (austenitic steel) was replaced by
stabilized austenitic steel Cr16Ni11Mo3 in a heat-treated state;
The coolant pressure in the middle plane inside the duct was diminished by
approximately 35% resulting in a decrease in duct deformation by radiation-induced
creep;
The power distribution over the core radius was flattened by the incorporation of a
medium fuel enrichment (21%) zone between the existing core zones with low
(17%) and high (26%) enrichment of fuel, resulting in a decrease of the fuel rod
specific heat rating.
The measures mentioned above and reactor operation under power conditions resulted in a
significant reduction in the number of defective fuel rods in the core. Further increase in fuel
burnup was achieved through utilization of the ferritic steel EP-450 for ducts and improved
austenitic steel in cold worked state for cladding.

FIG. 23. BN-350 fuel subassembly and rod.

FIG. 24. BN-350 blanket subassembly and rod.

134
The BN-350 refuelling system consists of two components:

Complex of in-reactor refuelling mechanisms (Fig. 25);

Complex of out-of-reactor refuelling systems (Figs 26 and 27a,b).

FIG. 25. BN-350 refuelling mechanism-mode of operation.

135
 FIG. 26. Fresh subassemblies loading principle.

FIG. 27a. Spent subassemblies unloading principle.

136
 FIG. 27b. Spent subassemblies unloading principle.

The design parameters of BN-350 core and blanket assemblies are presented in Tables 5 and 6,
respectively. The first complex is a constituent of the reactor and includes: two rotating plugs
(large and small), refuelling mechanisms and loading-unloading elevators. This complex
provides for loading and unloading of the main core items and their rearrangement inside the
reactor. The rotating plugs are disposed eccentrically to each other. On the small rotating plug,
the refuelling mechanisms are mounted eccentrically. By rotating the plugs the refuelling
mechanism is guided to any position in the core, radial blanket and in-vessel store, or to any
control rod. The refuelling mechanism simultaneously with reshuffling of FAs rotates them for
alignment with the hexagonal cells in the core. The complex of out-of-reactor refuelling
systems includes:

Fresh FAs drums;

Fuel transfer cell transfer mechanism;
Spent fuel drum;
Washing cell fuel transfer mechanism;
Loading-unloading elevator plug lifting mechanism.

137
 TABLE 5. DESIGN AND PARAMETERS OF BN-350 DRIVER ASSEMBLIES

138
Assembly type
Parameter
Type I Type II Type III
Years of usage 1974-1980 1978-1990 1987-1998
Hexagonal duct material 12-18Cr-10Ni-Ti 08-16Cr-11Ni-3Mo 12-13Cr-2Mo-Nb-V-B
Duct flat-to-flat distance wall thickness, mm 962 962 962
Number of fuel elements 169 127 127
FE length, mm 1140 1790 2440
FE diam. wall thickness, mm 6.10.4 6.90.4 6.90.4
Cladding material 16Cr-15Ni-3Mo-Nb 16Cr-15Ni-3Mo-Nb 16Cr-15Ni-2Mo-2Mn-Ti-V-B
Fuel column height, mm 1060 1060 1000
Gas plenum length, mm 25 250 720
Burnup, % 5.6-5.8 13 13
Dose (displacements per atom) 50-60 85 90

TABLE 6. DESIGN PARAMETERS OF BN-350 BLANKET ASSEMBLIES

Assembly type
Parameter
Type I Modernized
Years of usage 1974-1992 1992 onwards
Hexagonal duct material 12-18Cr-10Ni-Ti 08-16Cr-11Ni-3Mo
Flat-to-flat distance wall thickness, mm 962 962
Number of fuel elements (FEs) per assembly 37 37
FE column height, mm 2492 2492
FE Cladding diameter wall thickness, mm 14.00.4 14.00.4
Cladding material 16Cr-15Ni-3Mo-Nb 16Cr-15Ni-3Mo-Nb
Fuel pellet outer/inner diameter, mm 13.0/0.0 13.0/0.0
Fuel length, mm 2410 1580
Gas plenum length, mm 18 18
UO2 enrichment (235U, %) 0.4 0.4
Spent FAs and other items are transported into sockets in the washing cell and then
transferred into the spent fuel water pool for decay. The fuel transfer cell with its transfer
mechanism and the washing cell adjoin the reactor. New fuel assemblies are loaded into the
reactor from the new fuel storage drum located beneath the refueling cell. New fuel loading
operations have not posed any difficulties during the reactor operating life. The design intent
was that irradiated fuel assemblies were to be transported from the fuel transfer cell to the
washing cell through the spent fuel storage drum located in a tank filled with sodium-
potassium alloy beneath the junction between the cells.

In the BN-350 reactor, there was an incident (1976) causing the fuel storage drum failure. The
cover (plug) of the drum included concrete filler. In the course of heating, vapor of
crystallization water from the concrete penetrated into sodium-potassium alloy, filling the
drum, and, upon interaction with this alloy, chilled the drum. BN-350 personnel developed
special technology to dissolve the formed conglomeration using water-oil emulsion, and as a
result of its implementation, fuel subassemblies stored in the drum were set free and placed in
the water pool. A special lead-shielded flask was designed and manufactured for transporting
spent fuel assemblies from the transfer cell to the washing cell.

It was noticed during operation that increased forces were required for rotation of the shield
plugs. Probable causes were sodium vapor condensation in gaps or nonuniform heating of the
hydraulic seals. Plug operation became normal after increasing the Pb-Bi eutectic working
temperature to 200C and temporary interruption of the plug air cooling system in the process
of heating up the hydraulic seals. The refuelling mechanisms, elevators and fuel transfer
mechanisms in the cells have been operated without significant disturbances. Minor
disturbances were remedied through replacement and modernization of individual items.

To improve the reliability of the mechanisms a system to control installation and withdrawal
forces was incorporated. During the entire period of reactor operation (till October 1995)
56 planned refuelling cycles have been fulfilled. The time spent for one fuel assembly
replacement was one hour. The total number of loading-unloading operating cycles of the
elevators and fuel transfer mechanisms was ~ 3 200. Minor disturbances were remedied
through replacement and modernization of individual items. To improve the reliability of the
mechanisms a system to control installation and withdrawal forces was incorporated.

The first-in-the-world demonstration commercial liquid metal cooled fast breeder reactor
BN-350 was designed and built by the former USSR in the Republic of Kazakhstan for an
estimated 20 year life-time, which was reached in 1993. Considering the importance of the
reactor for the region, both as a source of desalinated water and electricity, it was the intention
of the Kazakhstan Government to extend the useful life of the reactor and continue to operate
it as long as safety conditions permitted.

In 1995, as a result of its limited experience in decommissioning, and its limited financial
resources, the Republic of Kazakhstan requested IAEAs assistance in improving safety for
the extension of the life of the reactor and in preparation of the proper technological
procedures for the eventual decommissioning. To this end, the IAEA organized a mission to
review the safety status of the reactor and a Consultancy (CS) on Harmonization of
international assistance to ensure stable operation during the remaining life-time and the
development of an effective decommissioning programme for the BN-350 fast reactor in
October 1996.

A composite programme for the prolongation of operation, and safety assurance of the
BN-350 plant, which included modernization and replacement of some reactor systems, was

139
developed by design organizations of the Russian MINATOM and the Kazakh Atomic Energy
Committee (KAEC). Some features to improve reactor safety were installed; but financial
difficulties delayed full implementation. Taking into account the large expenditures needed
for fulfilment of the programme, a Government decision to shut down the BN-350 plant was
taken in April 1999: that is, after ~ 25 years of reliably supplying fresh water and energy to
Aktau (former Shevchenko) and the adjoining region, a large city and industrial centre in the
desert, that had quickly grown following construction of the reactor.

4.2.5. BN-350 electromagnetic pumps

Recently the investigations on electromagnetic pumps in fast reactor secondary circuits have
been in progress. In the former USSR this investigation was carried out as applied to the
BN-350 plant. Electromagnetic pumps (EMP) are widely used for liquid-metal coolant
pumping over the booster circuits of fast reactors. A certain positive operating experience was
provided concerning the main circuits of the research reactors BR-10, EBR-II and the
secondary circuit of the BOR-60 reactor plant. The electromagnetic pumps fabrication and
operation experience for the flow rate up to ~ 1000 m3/hr provides a possibility of activities
on designing higher-power pump with the flow rates of several thousands m3/hr.

The development of these pumps was supported by a system of theoretical calculations and
experimental and design studies performed in D.V. Efremov's NIIEFA Institute in
St. Petersburg (Russian Federation). A cylindrical induction pump with the metal flow turn
through 180 was adopted for operation in the main fast reactor circuits. This design provides
the dismounting and assembling of the inductor related to its repair or replacement without
circuit loss-of-tightness.

The secondary and primary circuit EMP with the loop configuration are set in the circuits
pipelines through, inlet and outlet connections and do not need dipping into sodium. The
inductor coil cooling-down is substantially simplified due to this. The pump of this
design-TSLIN-3/3500 was developed and then fabricated for 3500 m3/h flow sodium
pumping, under the pressure 0.3 MPa and the temperature 350C in one of the loops of the
BN-350 reactor secondary circuit. Its key design parameters are: number of poles - 12; pole
pitch - 0.156 m; channel height - 26 mm, average diameter of channel 0.95 m, overall
dimensions: height - 5 m, diameter - 1.8 m. The total test data were conducted at a liquid-
metal rig at the sodium temperature 300C in 1986; processing was completed in 1987. The
nominal operating conditions are specified by the following parameters:
Flow rate Q = 3600 m3/h
Developed pressure P = 0.3 Mpa
Intake power ~ 1 MW
Voltage V = 650 V
Current I 3000 A at f = 50 ha (c.p.s)
Cos = 0.3
The experimental values are fairly close to the calculated ones (the calculation error is as high
as 9%, for the efficiency at the same flow rates - 3.5% absolute).The thermal tests of the
pump have shown, that the maximum temperature of coil measured by thermocouples on the
coil surface is 280C, and the average temperature of coil measured in the resistance method
is ~ 240C. The corresponding calculated temperatures are ~ 295 and ~ 260C, respectively,
that very close to the experimental ones. The mentioned operating temperatures of coil with
the utilization of material brought to a commercial level provide the required lifetime with a
high degree of reliability. In addition, a series of tests were performed to confirm inherent

140
requirements of operation in fast reactors. In the course of these tests the excessive inlet
pressure (top section of the channel) changed from 1.1 to 0.35 kg/cm2 at the flow rate
3750 m3/hr (the rate in the active section of the channel is ~ 13.5 m/s). In this case no
deviations in the pump parameters, vibration values were observed. In. the whole it was
established, that the pump satisfies the requirements on non-cavitational operation in the
nominal and partial operating conditions. The general view of the fabricated pump is
represented below. Its key design parameters are: number of poles - 12; pole pitch - 0.156 m;
channel height - 26 mm, average diameter of channel - 0.95 m, overall dimensions:
height -5m, diameter -1.8 m.
4.3. POOL VS. LOOP TYPE DESIGN: SOME KNOWLEDGE GAINED FROM
CONSTRUCTIONS AND OPERATIONS

Some experts believe that future sodium cooled fast reactors will have a pool type
arrangement of the radioactive circuit. It was chosen now in all countries dealing with these
reactor types except for Japan. The advantages of the pool type design become especially
apparent at present when safety requirements for all nuclear power plants become more and
more stringent. The Russian Federation is the only country where relatively large power loop
type (BN-350) and pool type (BN-600) reactors have been designed and successfully operated
(BN-350 in Kazakhstan) for a long time.

Experience has shown that a loop type design of the primary circuit is simpler and faster in
construction. However, this design type has some inherent specific difficulties in solving
safety problems. The main of them are assured preclusion of coolant leakage from the reactor
at large ruptures of the primary circuit and high reliability tightness of equipment and piping
cells to preclude radioactive sodium burning at small leaks. The problem of reliable protection
of both the reactor and personnel at radioactive coolant leaks is one of the most important for
the loop type design. As a result of long-time, almost 25-years operation of the BN-350
loop-type reactor in Kazakhstan interesting information was obtained which was published in
technical magazines and discussed at international conferences and meetings. For example, on
the basis of the BN-350 experience one can comment upon the maintainability of loop-type
designs. It was believed that due to localization of specified parts of valves the reduction of
coolant activity during reactor on-power operation can be attained.

However, operating experience has shown that in large power reactors two combined
hazardous factors are acting: an increase of the piping diameter and an increased distance of
valves along the pipe from the reactor. So, in the BN-350 the valves in the cells are situated
near the dead supports of piping, the pump and heat exchanger, i.e. nearly 40 m long
non-separated piping is between the reactor and the valve. Such layout is explained by
arrangement considerations and is related with large weight of valves, considerable vertical
displacements of reactor nozzles and maintainability requirements. After the reactor
commissioning, the assumes were confirmed of a vigorous natural convection to be developed
in 600 mm diameter piping connecting the reactor and the heat exchanger that caused heat and
radioactive sodium transport from the reactor along the pipe. Moreover, this convection
proved to spread also over the intermediate heat exchanger and was developed with a small
reverse flow in the loop.

Special studies have shown the stable natural convection to develop in 2-3 hours after shutting
the valve. Good reproducibility and agreement (within a measurement error) of temperature
variation curves for various loops were indicating of an objective and non random character
of the process. These results are confirmed by measurements of background in the cells of
shut off loops at bringing the reactor to power. Due to convection, the value of the radiation

141
flux in cells with cut-off loops differs little from those in operation. The value of convection
will grow with increased coolant temperature rise in the reactor and the pipe diameter. This
phenomenon characteristic of large diameter piping with horizontally arranged expansion
bends somewhat decreases the advantages of the loop type layout from the viewpoint of
carrying out repairs at the loop with the reactor in operation. The fabrication and installation
experience of the BN-350 and BN-600 reactor vessels

One of the main elements of the reactor is its vessel. The specific conditions of its operation
and the need to ensure its high reliability and fault-free operation for a long period of time
place stringent requirements upon the quality of the material, welding, manufacturing
precision and vessel monitoring methods. Therefore, from the economic and technical point
of view it is most advisable to build the vessel completely at the factory. This version is
possible for vessels only on condition that they are transported by a waterway which connects
the manufacturing site and the vessel installation site that is not always desirable from the
viewpoint of the plant construction site selection.

The vessels of reactors such as the BN-350 (H = 13 000 mm and D = 6 000 mm) and the
BN-600 (H = 13 000 mm and D = 12 900 mm) were divided into large blocks in order to
transport them by rail to the site. In the process the vessels were divided into elements which
not only satisfied the requirements of railroad packaging, but also made it possible to properly
assemble and weld them on site with minimum expenses. Thus, at transporting by rail, a loop
type vessel has no special advantages over a pool type vessel due to the necessity of dividing
them into installation blocks. However, the fact that the dimensions of the BN-600 vessel
were increased by more than twice in comparison to the BN-350 made it necessary to divide it
into a considerably larger number of installation blocks, and this somewhat complicated the
technology and increased the vessel installation time.

On the basis of analyzing the experience of manufacturing, assembling, welding and installing
the equipment of the primary circuit and reactor vessel of the BN-350, as well as the
BN-600 reactor vessel, we can draw several conclusions:

(1) In manufacturing the assemblies of the vessel at the factory with their subsequent
assembling and welding on site, it is absolutely necessary for the manufacturing
industries to carry out a complete trial assembly of the vessel for checking all joints
and geometric parameters, especially in connection with the fact that sometimes there
are deviations from the drawing for the two layouts under consideration manufacturing
is a unified process;
(2) The manufacture, assembly, welding and installation of the vessel requires the use of
special equipment which should be designed to be used in assembling and welding,
both at the factory and during installation on site; this ensures greater precision in the
assembled structure and greatly reduces the expense for vessel manufacture;
(3) The primary circuit equipment (pumps, heat exchangers, fuel handling system, etc.)
should be delivered for installation completely assembled after undergoing the
necessary series of checks in the shops and being subjected to easily removable
preservation, the installation operations must be kept to a minimum and be as simple
as possible, and they must not require internal adjustment of the equipment. In
general, the experience of the on-site fabrication of the BN-350 and BN-600 vessels
and of their long time operation is quite satisfactory;
(4) The reactor plant operating experience has demonstrated quite acceptable
maintainability of the pool design of the LMFR.

142
There have been no additional difficulties in repairing primary pumps on BN-600 compared
with BN-350. Repairs of the reactor equipment items and sodium systems, including such
complicated operations as replacement of the reactor coolant pumps, rearrangement and
jacketing of primary sodium systems, extraction of distorted items from the reactor core, etc.
had virtually no influence on the duration of planned outages, which were defined, as a rule,
by inspections and preventive maintenance of the steam-water system components.

There have been six leaks in the austenitic stainless steel 316 Phnix intermediate heat
exchangers (IHXs). The decision was taken to remove all 6 IHXs and modify them. The
removal and modification one-by-one of the Phnix IHXs was perhaps the most difficult
potential maintenance task foreseen on pool-type fast reactors. Its accomplishment, with the
minimum of reactor downtime, represents a conspicuous success and demonstrates the
importance of foreseeing events at the design stage, making adequate provision of equipment
and space for the repair, planning and learning from previous operation. In spite of the
restricted number of the pool type fast reactors under operation there is a great diversity of
their design incarnation. Two ways of the pool type reactor vessel fixing are known:
suspension from the upper plate and bottom supporting. There exist different ways of fixing
pumps, IHX's and core. The important part of the pool type reactor is the inner vessel, which
separates cold and hot coolant inside the vessel. A number of other alternative solutions are
known. Below considered those which have a considerable effect on the reactor economic and
engineering characteristics.

One of them is the way of reactor vessel fixing. The pool type PFR, Phnix and Super-Phnix
reactor vessels are known to have an upper suspension and only two reactors, operating
BN-600 and BN-800 that is under construction, have their vessels supported in the bottom.
The vessel bottom support has some advantages which were discussed at the international
conferences and meetings. The preliminary opinion was that this type of support is preferable
for the middle power reactors (with the vessel diameter of about 1215 m). The design
versions using the bottom support have been considered in other countries, e.g. in the U.S.
and Japan.

The Bechtel Corporations making use of the bottom support significantly enhances seismic
stability and allows designers to reduce wall thickness of the reactor main and guard vessels.
This corporation has evaluated cost savings as a result of using bottom support versus top
suspension of the reactor vessel. It is necessary to state that the way of the vessel fixing
directly influence the reactor cover plate design and fixing of the in-vessel components. Once
again the European countries and Russian Federation designing demonstration commercial
reactors encountered the problem of selecting the optimal way of the vessel being fixed and
the way of smoothing out the load in the in-pile components. Thus the experience gained
should be analyzed. The different ways of fixing reactor vessels and in-pile devices carried
out in demonstration reactors are characterized below.

4.3.1. BN-600

It is characterized by fixing all components: the core and neutron shield, pumps, heat
exchangers on the low supporting ring and rotating plugs on the reactor roof. Through
special supports the load is transferred to the foundation. The supporting ring with all the
equipment is installed and the upper part of the reactor operates at different temperatures.
Therefore such a design reveals a complex problem of matching thermal expansions of pumps
and heat exchangers to those of the reactor vessel and that of sealing vessel penetrations. The
BN-600 reactor (cross section of pump and IHX) is shown in Fig. 28.

143
 FIG. 28. The BN-600 reactor (cross section of pump and IHX).

In the BN-600 reactor these problems are known to have been solved by means of the bellows
used in the cylinders where pumps and heat exchangers are installed. If necessary, the bellows
can be changed. Cooling the vessel and especially its upper part decreases the difference in
lateral displacements of the reactor vessel head and supporting ring, where pumps and heat-
exchangers are installed, thus improving the bellows operation conditions.

Supporting structures on which pumps, heat exchangers and pressure chamber are fixed
operate at the same temperature, so there is no need in using compensators between them,
being a great advantage of this design. This way of vessel fixing has confirmed a good
performance during BN-600 operation. As some negative peculiarities insignificantly
impeding its operation we can point out the following:

Position of the central column (with control rods) and the core depends on the vessel
head temperature;
Significant vertical displacements of the upper part of IHXs and secondary pipes.

144
4.3.2. Phnix

The vessel of the Phnix reactor is suspended from the upper plate on which the equipment is
installed. In this design, similar to the BN-600 reactor, the reactor vessel cover plate is under
high temperature conditions but in this case it is absolutely unloaded. The bearing plate
operates under low-temperature conditions because between the plate and the reactor cover
there is an insulation layer and besides this zone is cooled.

The pumps and heat exchangers are location movable sliding supports, sealing of the pump
and heat exchanger penetrations is carried out by means of the bellows. A peculiar feature of
this reactor top design is a massive cover plate having significant thermal inertia. It restraints
a power growth rate or the reactor starting up and may be a cause of some significant thermal
stresses occurring in the junction of the vessel cylindrical part and the horizontal cover. The
sodium in the main vessel is separated into two zones:
Hot pool at the core outlet where the hot sodium flows into the intermediate heat
exchangers;
Cold pool taken from a peripheral annular space between the primary tank and the
wall of the main reactor vessel, which contains the three main circuit circulating
pumps and six heat exchangers, suspended from the upper slab.
One of the main problems concerning the top support reactor design is to create acceptable
operational temperature conditions for the roof, from which the vessel is suspended, and on
which the equipment is installed (pumps, heat-exchangers, rotating plugs, etc.). For thermal
shielding of the roof internal surface in the PFR reactor multi layer steel foil insulation was
used, which turned out to be complicated and expensive. The roof was a very critical item in
the construction sequence of the top suspension designs. That is why an another design the
roof design is again considered, e.g. a simple solid ferritic steel roof of 0.85 m thick was
adopted for EFR plant. Phnix: top support reactor vessel is shown in Fig. 29.

FIG. 29. Phnix: top support reactor vessel.

145
4.3.3. PFR

The equipment is stationary fixed on the reactor roof from which the reactor vessel is
suspended. This design requires a complex thermal expansion compensation system of skirts,
inside which tube bundles of heat exchangers are inserted. For this purpose the PFR reactor
has bellows operating at high temperature and being inmaintainable. Compensation of
pressure pipelines thermal expansions was achieved by means of the bends. In view of that
several pipelines connect the pump to the pressure plenum. In the BN-800 reactor the bottom
support is used as well. But the supporting ring is of a much smaller diameter, the load
transfer set-up from the supporting ring to the vessel supporting assembly is changed, to
enhance seismic stability roller bearings are replaced with the annular welded skirt. Elevation
through PFR primary circuit is shown in Fig. 30.

FIG. 30. Elevation through PFR primary circuit.

146
 BIBLIOGRAPHY
ASHURKO, Y.M., et. al., Fast Reactor Operating Experience Gained in Russian Federation:
Analysis of Anomalies and Abnormal Operation Cases, Unusual Occurrences During LMFR
Operation, IAEA-TECDOC-1180, IAEA, Vienna (2000).
BAGDASAROV, YU.E, et al., BN-800 reactor A new stage of fast reactor development,
Fast Breeder Reactors: Experience and Trends, paper presented in Int. Symp. Fast Breeder
Reactors and Future Trends, Lyon, France, 2226 July 1985.
BOLGARIN, V.I., et al., Experience and organization and performing repair works on main
equipment of BN-350 reactor, paper presented in Int. Symp. Fast Breeder Reactors and Future
Trends, 22-26 July 1985, Lyon, France.
BUDOV, V.M., et al., A NPP BN-600 the plant for the near future, paper presented in
Nuclex-75, Basel, 1975.
BUDOV, V.M., et al., Intermediate heat exchangers design and experimental testing, paper
presented in Int. Symp. Design, Construction and Operating Experience of Demonstration
Liquid Metal Fast Breeder Reactors, Bologna, Italy, 1014 April 1978.
GOLAN, S., et al., The bottom supported fast breeder reactor vessel - an alternative
approach to seismic accommodation and reduced cost, Proc. Fourth Int. Conf.
Liquid metal engineering and technology, 1721 October 1988, Palas des Papes, France,
Vol. 1, 110-1 - 110-11.
INTERNATIONAL ATOMIC ENERGY AGENCY, Status of Liquid Metal Cooled Fast
Breeder Reactors, Technical Reports Series No. 246, IAEA, Vienna (1985).
INTERNATIONAL ATOMIC ENERGY AGENCY, Status of Liquid Metal Cooled Fast
Reactor Technology, IAEA-TECDOC-1083, IAEA, Vienna (1999).
INTERNATIONAL ATOMIC ENERGY AGENCY, Fast Reactor Database, IAEA-
TECDOC-866, IAEA, Vienna (1996).
INTERNATIONAL ATOMIC ENERGY AGENCY, Conceptual Design of Advanced Fast
Reactor, IAEA-TECDOC-907, IAEA, Vienna (1996).
KAMANIN, YU.L., et al., BN-600 reactor plant safety insuring. Considering its operating
experience, paper presented in Int. Fast Reactor Safety Meeting, 1216 August 1990,
Snowbird, Utah, USA.
KIRUSHIN, A.I., et al., Experience of BN-600 reactor plant safe operation as a part of the 3rd
power unit of Beloyarok NPP, paper presented in 4th Annual Scientific & Technical
Conference of the Nuclear Society Nuclear Energy and Human Safety, NE-93,
28 June-2 July 1993, Nizhny Novgorod, Russian Federation (in Russian).
KIRUSHIN, A.I., et al., Evolution of BN-600 Reactor Core, paper presented in 4th Annual
Scientific & Technical Conference of the Nuclear Society Nuclear Energy and Human
Safety, NE-93, 28 June2 July 1993, Nizhny Novgorod, Russian Federation (in Russian).
KOCHETKOV, L.A., et al., Main results of operation of nuclear power stations with BN-350
and BN-600 fast reactor, paper presented in Int. Symp. Fast Breeder Reactors Experience
and Future Trends, 2226 July 1985, Lyon, France.
KOCHETKOV, L.A., et al., Operating experience on fast breeder reactors in the USSR, paper
presented in Int. Conf. Fast Reactors and Related Fuel Cycles, 28 October1 November 1991,
Kyoto, Japan.

147
KOCHETKOV, L.A., RINEISKII, A.A., Advanced fast reactors for nuclear power of the
nearest future, paper resented in Int. Conf. Fast Breeder Systems: Experience Gained and Path
to Economic power Generation, 1317 September 1987, Richland, Washington, USA.
LEIPUNSKII, A.I., et al., The BN-600 fast reactor, paper resented in Nuclex-75, Basel, 1975.
MAKOVSKY, A.A, et al., Comparative analysis of the arrangement and design features of
the BN~350 and BN-600 reactors, paper presented in Int. Symp. on Design, Construction and
operating experience of Demonstration Liquid Metal Fast Breeder Reactors, Bologna, Italy,
1014 April 1978.
MITENKOV, F.M., et al., Control rod mechanisms for sodium fast reactor control and
protection systems. Atomizdat, Moscow (1980) (in Russian).
NEVSKY, V.P., MALYSHEV, V.M., KUPNYI, V.I, Experience of design, construction and
start-up of the power unit with BN-600 in Beloyarsk NPP, Atamaya Energia, Vol. 51, No. 5,
(1981) 292296.
OSHKANOV, N.N., BN-600 operating experience and BN-800 design, paper presented in
Int. Conf. Fifty years of nuclear power the next fifty years, Moscow/Obninsk, Russian
Federation, 27 June2 July 2004.
OSIPOV, S.L., et al., Peculiarities of radioactive emission products distribution in the BN-600
reactor, paper presented in CIS/European Fast Reactor Information Exchange Meeting on
Sodium Technology, 1416 December 1992, Siemens, KWU, Bergisch Gladbach, FRG.
OSIPOV, S.L., et al., Radiation influence of BN-600 fast reactor power unit on the
environment, paper presented in 4th Annual Scientific & Technical Conference of the Nuclear
Society Nuclear Energy and Human Safety, NE-93, 28 June2 July 1993, Nizhny
Novgorod, Russian Federation (in Russian).
RINEYSKII, A.A, KAMAEV, A.A, A new safety approach in the design of LMFBRs, paper
presented in ANS Top. Mtg. Safety of Next Generation Power Reactors, 15 May 1988,
Seattle, WA, USA.
RINEISKII, A.A., et al., Commercialization of fast reactors, paper presented in Int. Conf. on
Fast reactors and related fuel cycles, 28 October-1 November 1991, Kyoto, Japan.
SARAEV, O.M., Operating experience with Beloyarsk fast reactor BN-600, Unusual
Occurrences During LMFR Operation, IAEA-TECDOC-1180, IAEA, Vienna (2000).
STEKOLNIKOV, V.V., et al., Operation experience of sodium-water steam generators in the
USSR and prospects for their development, paper presented in Int. Symp. Fast Breeder
Reactors Experience and Future Trends, Lyon, France, 2226 July 1985.
TROYANOV, M.F., RINEISKII, A.A., State of work on fast reactors in the USSR, paper
presented in 22nd Annual Meeting of the IWG-FR, 1821 April 1989, Vienna, Austria (in
Russian).
USYNIN, G.B., KUSMARTSEV, E.V., Fast Reactors, Energoatomizdat, Moscow, 1985 (in
Russian).

148
5. TECHNOLOGIES FOR SODIUM AND REACTOR COMPONENTS
MANAGEMENT AFTER SHUT DOWN OF THE BN-350 TYPE LMFR

5.1. INTRODUCTION AND SUMMARY

Two consultancy meetings on Decommissioning of the Kazakh BN-350 power plant (Vienna,
Austria, October 1996; Obninsk, Russian Federation, February 1998) were convened by the
IAEA. These meetings brought together a group of experts from France, Russian Federation,
Kazakhstan, the UK, and the USA to exchange information on, and to review current
technical knowledge and experience in the management of radioactive coolant and reactor
components following closing of liquid metal fast reactors (LMFR), as well as their design
features relevant for decommissioning procedures.

In response to the Kazakh Atomic Energy Committee (KAEC) and the necessity to preserve
the knowledge in LMFR decommissioning, the IAEA collected and summarized of the
technical information presented at the above mentioned meetings7. This chapter provides
general and detailed information on activation characteristics of the primary coolant, reactor
and components; treatment and disposal of the spent sodium; removal of the residual sodium
deposits and decontamination of the BN-350 reactor a typical loop type LMFR.

When decommissioning the BN-350 reactor, there arises the problem of the utilization of
primary sodium with a specific activity of ~ 300 MBq/kg and of secondary sodium with a
specific activity of ~ 0.5 Bq/kg. In principle, sodium can be used in the nuclear industry for
newly constructed LMFRs, and in the chemical, metallurgical, engineering industries as well.
In the former case, some limitations both on radioactive and non-radioactive impurities are
imposed upon sodium. The rules applying to the transport of radioactive materials limit the
radiation rate on the container surface to 0.55 Sv/s and at a distance of 1 m from it to
0.028 Sv/s. On the basis of calculations performed it has been shown that when using tank
cars for sodium transportation these requirements are met by sodium with a specific activities
of 22Na 6105 Bq/kg and 137Cs 3.7105 Bq/kg. Taking into account the initial sodium
activity of these isotopes of 3107 Bq/kg and 2.5108 Bq/kg, respectively, the BN-350
primary sodium needs to have the caesium nuclides reduced by a factor of ~ 700; to reduce
Na22 activity, decay storage duration of ~ 15 years is required. Another approach to coolant
activity limitation is connected with non-admission of NPP personnel over-irradiation in the
course of equipment maintenance at start-up and adjustment activities on a new reactor with
reused coolant. Calculations have shown that these requirements are met by sodium with Na22
activity 4.5106 Bq/kg and with 137Cs activity 2.5106 Bq/kg. To meet these less stringent
requirements, the spent sodium needs to have the activity from caesium nuclides reduced by a
factor of 100, and must be stored for 6 years to reach the required 22Na level.

For primary sodium utilization in the chemical industry, more rigid requirements are posed:
reduction of caesium nuclides by a factor of 20 000 is required. For inactive impurities the
following requirements are posed: carbon 30 ppm (particles per million), oxygen 10 ppm,
nitrogen 10 ppm, potassium 200/1 000 ppm (primary/secondary circuit, respectively), and
hydrogen < 0.5 ppm. The spent coolant from BN-350 satisfies these impurities, requirements,
except for potassium that is 400 ppm in the primary circuit and 2 000 ppm in the secondary

7
E.R. Adam (UK AEA) and G. Heusener (FZK, Germany) served as Chair persons of the Consultants Meetings;
L. Kochetkov, A. Tsykunov, A. Karpov, V.P. Matveev, Yu. Nalimov, E. Popov, V. Shirin, and V. Yarovisyn (all
from the IPPE, Russian Federation) made relative calculations and provided written contributions at the Obninsk
Meeting.

149
circuit. Cleaning of the potassium impurities is possible by means of a single distillation at
250C or of dilution with pure sodium.

The sodium from the BN-350 secondary circuits can be used in a new LMFR without
purification.

Overall only a small amount of so-called sub-standard sodium, ~ 20 tons of BN-350

radioactive waste sodium coolant, is unsuited for re-use, and after reprocessing must be
dispatched for burial. It includes sodium strongly polluted with the products of interaction
with air or water, with fire extinguishing means, and sodium removed from cold traps and
cesium traps. The analysis of the methods for processing this sodium for safe burial has
revealed that the methods of waste sodium processing by means of sodium injection into
water or alkali developed in France and the UK cannot be used for the processing of the sub-
standard sodium. The development of a new safe technology is required for which the
experimental tests should be initiated. In order to identify the best procedure to process the
sub-standard sodium it is recommended to test the following four methods of sodium
processing: burning in air, gas-phase (O2 or CO2), solid-phase (metal oxides), and liquid-
phase (with water under vacuum). The experimental studies performed at the Institute of
Physics and Power Engineering (IPPE) on a laboratory scale confirmed the feasibility of these
methods for the processing of the BN-350 sub-standard sodium radioactive waste.
Semi-industrial tests will allow the development of the technology to be used at the BN-350
reactor.

IPPE has carried out the experimental substantiation of R&D works on new technology
development aiming at rendering harmless medium level liquid radioactive waste (LRW).
This new technology has been developed based on the absorption extraction of caesium and
strontium radionuclides, and the formation of a new type of cement, namely geocement, by
means of radioactive sorbent immobilization into the mineral-like alkaline and earth-alkaline
hydroaluminosilicates. Studies have been pursued on the following:

Principal absorbtion laws of caesium and strontium radionuclides by the natural

zeolite-clinoptilolite both under steady state and transient conditions from low and
medium LRWs, and
The immobilization of the sorbent, containing 137Cs, 134Cs and 90Sr into the geocement
stones.

It has been shown in the studies that clinoptilolite, modified by sodium ions, is an effective
sorbent for the extraction of 137Cs, 134Cs and 90 Sr radionuclides from the medium LRW. In
addition, the clinoptiolite has been manifested to be capable, provided that the prescribed
parameters of the process are observed, of assuring high quality LRW cleaned from cesium
and strontium (cleaning degree up to 99.98%, distribution coefficient up to 1106 mL/g) and
the transfer of medium LRW to the non-radioactive category with respect to 137Cs, 134Cs and
90
Sr content, with their concentration lower than that for 90Sr permissible in water. In the
process of immobilization of the sorbent containing cesium and strontium radionuclides by
grouting under normal humidity conditions, samples of geocement were obtained with the
sorbent content up to 40 mass.%, high moisture resistance 106-107 g/cm2 day value (close to
that for the glass-like materials), and high compressive strength (10-25 MPa).

The process has been developed on the basis of accessible, low price natural materials, such
as clinoptilolite as a sorbent, and metallurgical slag, clays and solutions of compositions of
alkaline elements as components of the binding system. The stability of the slag-alkaline
geocement, obtained as a result of hydration solidification of the binding system - the

150
analogues of the rock forming minerals, provides readiness for environmentally safe disposal
or long-term storage. The proposed method permits to convert cesium and strontium
containing radionuclides LRW of magnitude into the solid form with more than two orders of
magnitude reduction in volume, and the consumption of the clinoptilolite being relatively low
(2-3 kg/m3). After the cesium and strontium radionuclides are extracted from the LRW salt
solutions, the radioactivity of these solutions is decreased down to zero or to an insignificant
level (if some other radionuclides are present), thus simplifying considerably their further
management.

The total volume of the remaining in the circuits after draining the sodium from the
primary and secondary circuits and the NaK drained from the cold traps has been estimated to
be ~ 100 tons. Considerations have been given how to remove this remaining sodium, and
ensure remaining volumes allowing the safe conduction of washing and decontamination
operations of the circuits and equipment (i.e. 150 L in the reactor vessel, and 15 L in the cold
trap). The main method for the residual sodium removal in the Russian Federation is the
usage of the steam-gaseous procedure. This procedure has been successfully employed three
times in washing of the BR-10 reactor primary circuit. Preliminarily, for the primary and
secondary circuits it is recommended to carry out vacuum distillation of sodium. For the
decontamination of the primary circuit there are solutions and procedures that have proved
their effectiveness at carrying out decontamination of equipment:

0.5% KMnO4 at 70C; 5% HNO3 + 1% H2C2 O4 at 70C.

It is recommended to perform the cleaning of the cold traps using the technology developed
by IPPE that had been already employed at reprocessing of the BN-350 sodium from the
secondary circuit cold traps.

A considerable amount of data on the evaluation of accumulated activity in BN-350 structural

materials as well as in the reactor cavity has been produced using the relevant computer code.
Steel structure activity is dependent on the level of impurities, e.g. manganese, europium, and
cobalt, decay times of which are ~ 10 years, ~ 100 years, and ~ 150 years, respectively.
Calcium dominates in concrete for 100 years, and in graphite 14C dominates for 10-12 years.
Therefore radioactive waste in future reactors can be reduced by reducing or eliminating
certain impurities, e.g. replacing nickel with manganese in standard steels, reducing the
amount of cobalt to 0.01 ppm, and reducing niobium and europium content.

After assessing the conditions at the start of decommissioning works, and the measures that
would have to be taken for the reduction of radiation levels in empty pipes from the reactor
vessel and for the replacement of safety and monitoring equipment, it was clear that it would
not be possible to disassemble the primary circuit without robotics in 100 years, although
limited access to structures should be possible in 50 years. Limited access could be gained in
10 years, provided chemical decontamination would be carried out on the primary circuit.

The meetings participants pointed out:

(1) For sodium processing, the goal is to minimize burial disposal and maximize reuse of
sodium, which will be pursued grounded on the following:
To obtain the required purity levels, existing techniques should be refined and,
if needed, new technologies, developed;
Storage options for reusable sodium should be investigated, including the use
of the dump tank, barrels, etc.;

151
 Usage for sodium other than for current and future LMFRs should be
investigated;
Continue to evaluate disposal possibilities, comparing different options with
different technologies. Costs and schedules associated with the deployment of
these options need to be established;
BN-350 has a NaK/mineral oil mixture. Technologies need to be
identified/developed for reprocessing this mixture. Applications for use of the
clean NaK need to be considered;
Identify and/or develop a technology for dealing with liquid waste. The goal in
dealing with liquid waste is waste volume minimization and final waste form
stabilization.

(2) The coolant activity reduction can be achieved by the following means:
For cesium isotopes, by means of purification with the use of carbon-graphite
filters;
For tritium, hydrogen and oxygen, by means of purification with the use of the
cold filter traps;
For corrosion products, by means of purification with the use of the getter
traps.

(3) Two technologies have been developed for disposal of large quantities of both primary
and secondary sodium coolant:
NOAH technology developed and tested by Commissariat lnergie Atomique
(France). In this process, aqueous caustic soda is produced, which can be sent
to the appropriate effluent treatment plants; and
Two-stage technology involving caustic process and carbonate process steps
developed by ANL-West (USA). In the carbonate process, the sodium
hydroxide is reacted with carbon dioxide to form sodium carbonate. A dry
powder, similar in consistency to baking soda, is a waste form acceptable for
burial.

(4) Under certain circumstances (e.g. closed geometries, sodium puddles), the use of
alcohol to clean components or to destroy sodium can be dangerous. This is proven for
ethylcarbitol, but at present it can not be excluded completely for other alcohols. It is
therefore strongly recommended to perform appropriate studies prior to the use of
alcohol together with sodium.

5.2. PRE-SHUT DOWN RADIOLOGICAL CHARACTERISTICS OF THE BN-350

COOLANT AND EQUIPMENT

By the end of life-in-service of the BN-350 reactor the following main radioactive waste will
be formed:

Fuel assemblies (FAs), blanket assemblies (BAs) and control rods;

Reactor;
Primary sodium;
Pipelines and equipment of the primary circuit in the rooms, adjacent to the reactor
vessel cavity;
Pipelines and equipment of the air cooling system of the reactor vessel cavity and
primary circuit cells;
Secondary sodium;
Cover gas system of the primary sodium tanks;

152
 Drainage tanks - primary circuit pressure compensators;
System of lubrication of the main sodium pumps;
Structural elements of the reactor vessel cavity;
Route of the spent fuel movement, spent fuel intermediate storage tank and their
cooling systems; washing systems of the fuel subassemblies, control rods and
equipment.

Below data are presented on the characteristics of the main radioactive waste of the BN-350
reactor facility from the point of view of the problem of its decommissioning.

5.2.1. Primary sodium activity

During reactor operation on power the high activity of the primary coolant is determined by
24
Na (T1/2 = 15 h). By the end of the reactor life after decay of 24Na, the specific activity of the
main radionuclides in sodium, determining radiological conditions for the sodium handling
systems, is as follows:
22
Na (T1/2 = 2.6 years) 30 MBq/kg
137
Cs (T1/2 = 30 years) 250 MBq/kg
134
Cs (T1/2 = 2.2 years) 30 MBq/kg
22
Mn (T1/2 = 2.6 days) 0.3 MBq/kg
65
Zn (T1/2 = 245 days) 0.2 MBq/kg
60
Co (T1/2 = 5.27 years) 0.1 MBq/kg
Plutonium 1 kBq/kg
Tritium (T1/2 = 12.3 years) 10 MBq/kg

5.2.2. Activated equipment of the primary circuit in the area adjacent to the reactor
vessel cavity

The radiological conditions when handling the primary circuit equipment in the rooms
adjacent to the reactor vessel cavity are mainly determined by three components:

Activity of the radionuclides in sodium;

Deposits of radioactive products on internal surfaces of the equipment;
Activation of structural elements of the equipment and cells.

5.2.2.1. Radiological situation in the primary circuit cells (Na remains in the circuit)

In this case the radiological conditions in the cells will be determined by the radionuclides
contained in sodium, first of all by 22Na, 137Cs and 134Cs. For specific activities presented in
Section 8.2.1, the expected calculated dose rate on the surface of thermal insulation 10 days
after the reactor shutdown will be as follows:

For 600 mm pipeline - 3.5 Sv/s, including 0.9 Sv/s caused by 22Na; 1.9 Sv/s
caused by 137Cs and 0.7 Sv/s caused by 134Cs ;
For 500 mm pipeline - 3.1 Sv/s, including 0.8 Sv/s caused by 22Na; 1.7 Sv/s
caused by 137Cs and 0.6 Sv/s caused by 134Cs ;
For the intermediate heat exchanger (IHX) - 1.2 Sv/s, including 0.4 Sv/s caused by
22
Na; 0.6 Sv/s caused by 137Cs nd 0.2 Sv/s caused by 134Cs

153
Mentioned values correlate satisfactorily with the results of measurements made in the cells.

The experimental data on the dose rates of gamma-radiation in the cells of six loops of the
BN-350 are shown in Table 1. Significant dose rates (up to 102 Sv/s) on the primary sodium
pump (PSP) overflow pipeline are caused by the increased deposits of cesium radionuclides
on the surfaces adjacent to the cover gas/sodium interface of the pipelines and components.

TABLE 1. DOSE RATES OF GAMMA-RADIATION IN THE CELLS OF THE BN-350

PRIMARY CIRCUIT, Sv/s (MEASUREMENTS MADE IN 1995)

Measurement points Primary circuit cells

318/1 318/2 318/3 318/4 318/5 318/6
Cell entrance area 0.40 0.50 0.45 - 0.36 0.80
Center of cell, 1 m above the floor 0.58 0.70 0.65 - 0.55 0.90
IHX - PSP 600 mm piping (reference point) 2.80 2.20 2.50 - 1.70 2.90
500 mm cold piping (PSP - reactor) - - - - 1.80 2.80
600 mm hot piping (reactor - IHX) - - - - 1.40 2.50
PSP sodium overflow dump tank - - - 3.80 3.80
PSP sodium overflow dump pipeline - - - - 60-160
PSP - - - - - 3.80
IHX - - - - 0.92 1.00
Note: dose rates were measured on the surface of thermal insulation of pipes and equipment.

In Table 2 data are given showing dynamics of change of gamma-radiation dose rates in the
reference point (IHX-PSP pipeline) for the long period of reactor operation (after the reactor
shutdown and Na22 decay).

TABLE 2. DOSE RATES IN THE REFERENCE POINTS OF THE PRIMARY CIRCUIT

CELLS, Sv/s

Primary circuit cells

Year
318/1 318/2 318/3 318/4 318/5 318/6
1974 - 0.04 0.04 - - -
1976 0.90 - - - 0.67 0.78
1977 - - 3.00* - - -
1978 7.50 10.00 10.00 5.00 6.80 5.00
1979 9.40 7.30 6.00 6.00 5.40 5.50
1981 3.00 3.00 3.20 2.00 1.90 1.80
1993 - - - - 2.90 2.80
1995 2.80 2.20 2.50 - 1.70 2.90
* average value over all boxes

154
5.2.2.2. Deposits of radioactive products on the inner surfaces of the equipment

The radiological situation during dismantling of the equipment, such as heat exchangers, gate
valves and pumps of the primary circuit, located beyond the biological shielding of the
reactor, so that they practically have no induced activity, will be determined by the surface
contamination. After the coolant is drained, different sorbed chemical elements and their
combinations, containing long-lived radionuclides ( 100 g/m2), which determine the
radiation situation during dismantling of the equipment, remain on the surface in addition to
the sodium film.

On the basis of data, obtained as a result of experiments made during the BN-350 reactor
operation, estimations of surface contamination of the primary circuit equipment at the
moment of the reactor shutdown were carried out. The appropriate data are presented
in Table 3.

TABLE 3. SURFACE CONTAMINATION OF THE PRIMARY CIRCUIT EQUIPMENT

Specific surface Total surface

activity, activity of the
Source of activity
MBq/m2 equipment,
TBq
22
Na 3.0 0.03
137
Cs On the surface being in contact with Na 500 5.0
137
Cs On the surface above Na level 5 000 0.9
134
Cs On the surface being in contact with Na 50.0 0.5
134
Cs On the surface above Na level 500 0.09
54
Mn Hot Na area 1 500 13.0
54
Mn Cold Na area 750 1.1
60
Co - 100 1.0

It was supposed, that the surface activity of 137Cs and 134Cs fission products is the same for all
elements of the primary circuit being in contact with sodium. For the structural elements,
located in the gas cavities above the sodium surface, specific surface activity of 137Cs and
134
Cs is about 10 times higher, than that for the surfaces, which are in contact with sodium.

Except for the radionuclides specified in Table 3, 125Sb and 94Nb isotopes may be present in
amounts of up to 5.3 and 0.8 MBq/m2, respectively, on the surface of the equipment of the
primary circuit by the end of the reactor life. Gamma-radiation dose rates on the surface of
thermal insulation of 500 and 600 mm pipelines after the coolant drainage, evaluated for
the values of the surface contamination of the primary circuit equipment given in Table 3, can
make:

For 500 mm pipeline - 0.24 Sv/s;

For 600 mm pipeline - 0.36 Sv/s (the IHX surface is 0.9 Sv/s).

The presented values could be even exceeded because of the increased deposit rate of
radionuclides in gaps, dead-ends (pockets) of sodium piping, and also on some sections of
gas systems.

155
5.2.2.3. Induced activity of structural materials in the primary circuit cells

The in-pile experimental studies have shown that the activation of the main sodium
equipment in the primary circuit cells would be low. It was shown that the activity of stainless
steel caused by 60Co at the time of reactor dismantling would not exceed 1 Bq/g of steel in the
cells, adjacent to the reactor.

The only exception could be the steel of a limited part of sodium pipelines of the primary
circuit, passing through the rooms, which are intermediate between biological shielding of the
reactor vessel cavity and the cells containing the main primary sodium components. The
activation will mainly take place in the area of an outlet of these pipelines from the openings
in the biological shielding of the reactor vessel cavity, and also in the area of the piping inlet
into the shielding. By the end of the reactor life, specific activity of steel caused by 60Co
would be 5 Bq/g for the pipelines sections most remote from the biological
shielding openings. The maximum specific activity of steel caused by the above indicated
radionuclide in the area of the openings may be 50-500 Bq/g. Length of the pipelines in this
sections is 1 m. The dose rate of gamma-radiation on these sections caused only by the
induced activity of 60Co can reach 310-2 Sv/s.

5.2.3. Activity of the radionuclides in the primary cold traps

In the cold traps (CT) of the primary circuit significant amount of long lived radionuclides
will be accumulated by the end of the reactor life. In Table 4 the results of estimations of the
content of radionuclides in all five CTs are given.

TABLE 4. ACTIVITY OF THE MAIN RADIONUCLIDES IN ALL COLD TRAPS OF

THE PRIMARY CIRCUIT

Radionuclides Activity, Bq
137
Cs 2.21013
134
Cs 2.21012
22
Na 3.31011
54
Mn 1.51012
Tritium 1.71015

The estimation was made using results of measurements of gamma-radiation rate patterns in
the cells where CT are located. The dose rate of gamma-radiation on the surface of
thermal insulation of CT after the reactor shutdown and 24Na decay is expected to be at the
level of 4.00 Sv/s.

5.2.4. Activity of air cooling system of reactor vessel cavity and the primary circuit cells

On the BN-350 reactor closed circulation circuit is provided for the cooling of the biological
shielding of the reactor cavity. Air is used as a coolant. During planned shutdowns after decay
of 41Ar dose rates of gamma-radiation in the rooms of this system were analysed, and also
gamma-spectral analysis of wet smears, taken from the external surfaces of heat exchangers
and floor and from the internal surfaces of air ducts was carried out. The dose rates on the
surface of the equipment of this system were:

156
 On air ducts: 510-3 Sv/s;
On the floor and heat exchangers: 610-3-2.510-2 Sv/s.

In Table 5 specific activity values, obtained as a result of gamma-spectrometer measurements

of smears are presented.

TABLE 5. MEASURED VALUES OF SPECIFIC ACTIVITIES OF SURFACE

CONTAMINATION

Radionuclides Activity, MBq/m2

137
Cs 0.5-2.5
134
Cs 0.1
125
Sb 0.06
60
Co 0.03-0.10

On the basis of these specific activities the dose rates of gamma radiation on the
surface of the equipment were calculated. The total calculated dose rate was at the level of
1103-310-3 Sv/s. The difference between evaluated and measured dose rates can be
explained by that using smear technique makes it possible to determine only removable
activity, but not the total amount.

On the basis of presented data it is possible to make a conclusion that the equipment of the
system for air cooling of the reactor vessel cavity will be related to the category of solid
radioactive wastes. On the basis of radiological measurements it will be possible to attribute
also the equipment of air cooling system of the cell of one primary sodium loop (no. 3 loop)
to the category of solid radioactive waste. The dose rate of gamma-radiation on the equipment
of this system is 5.310-3 Sv/s. The radioactive contamination of this system mainly caused
by 137Cs has resulted from earlier repair work for replacement of the drainage valve in the cell
of no. 3 sodium loop.

5.2.5. Activity of the equipment of the secondary circuit

In the secondary circuit of the BN-350 reactor practically only one long lived radionuclide
tritium (T1/2 = 12.3 years) will be present by the end of the reactor life. The specific activity of
tritium in sodium by the time of shutdown will be 0.5 MBq/kg. The activity of tritium in the
total volume of the secondary sodium will be equal to 2.41011 Bq, 1.71014 Bq of tritium
being contained in the cold traps.

5.3. PURIFICATION AND UTILIZATION OF SPENT SODIUM COOLANT OF

PRIMARY AND SECONDARY CIRCUITS

Once decommissioning of the BN-350 reactor has commenced there will be the problem of
disposing of the spent sodium coolant. About 500 tons of sodium will remain in the primary
circuit of the reactor, which after decay of Na24 will be in the category of intermediate activity
radioactive waste (~ 300 MBq/kg). In the secondary circuit ~ 480 tons of sodium remains
with low specific activity (~ 0.5 MBq/kg). The calculated content of the main impurities in
the primary and secondary sodium of the BN-350 reactor are presented in Table 6.

157
TABLE 6. THE CONTENTS OF THE MAIN IMPURITIES IN SPENT SODIUM
COOLANT OF THE BN-350 REACTOR

Impurities concentration
Impurities
I circuit II circuit
22
1. Na, MBq/kg 30 absent
137
2. Cs, MBq/kg 250 absent
134
3. Cs, MBq/kg 30 absent
4. Tritium, MBq/kg 10 0.5
5. -nuclides, kBq/kg 1 absent
6. Manganese-54, MBq/kg 0.3 absent
7. Carbon, ppm 27 26
8. Oxygen, ppm 2 2
9. Nitrogen, ppm 3 3
10. Chlorine, ppm 9 9
11. Potassium, ppm 400 2 000
12. Iron, ppm 3 2
13. Hydrogen, ppm - 0.2

The spent coolant of the BN-350 reactor could be used, first of all, in the nuclear industry for
newly constructed fast reactors with sodium coolant. It could also be used in other industries
as well (e.g. chemistry, metallurgy and mechanical engineering). In the chemical industry
metal sodium is used widely for production of both inorganic (Na2O2, NaNH2, NaCN, NaH)
and organic (C2H5)4Pb, synthetic washing means, dyes, plastics, etc.) compositions. In the
metallurgy and mechanical engineering sodium is used for production of high-strength cast
irons, steels and various alloys. As it can be seen from Table 6, the principal contribution to
the activity of sodium coolant of the primary circuit is made by 137(134)Cs and 22Na nuclides. In
case of secondary circuit use of the spent sodium coolant certain recommendations on its
radioactivity and impurities content will be made. This section is devoted to the analysis of
these recommendations.

5.3.1. Requirements to the coolant quality for its reuse in newly constructed LMFR

5.3.1.1. Radioactive impurities content

The permissible levels of activity of the coolant for its use on the new reactor will be limited,
on the one hand, by the requirements to the transportation of radioactive substances [1], and
on the other hand by the requirement of non-admittance of the NPP staffing overexposure
during operation with the equipment at the stage of sodium receiving, purification and
realization of initial testing activity on the new NPP.

5.3.1.1.1. Transportation

When transporting outside the NPP site, the level of gamma-radiation on a surface of the
transport shielded container and at 1 m distance from the container should not exceed
0.55 Sv/s and 0.028 Sv/s respectively. If transportation is made within the site, these values
are 2.8 and 0.14 Sv/s, respectively. In order to evaluate the dose rates of gamma-radiation
during transportation of spent primary sodium it was assumed, that transportation would be
carried out in the 32 m3 (25 tons of sodium) capacity railway cistern, the activity of the
coolant being defined by three nuclides:

158
22
Na (T1/2 = 2.6 years; a = 3.0107 Bq/kg)
137
Cs (T1/2 = 30 years; a = 2.5108 Bq/kg)
134
Cs (T1/2 = 2.1 years; a = 3.0107 Bq/kg)

The calculations were carried out for two moments of time: just after the reactor shutdown
and 6 years after shutdown [2, 3].

From Table 7 can be seen, that the total dose rates of gamma-radiation on a surface of the
cistern and at 1 m distance exceed permissible levels, stipulated by the rules of radioactive
substances transport even 6 years after the reactor shutdown [1]. Thus the maximum levels of
radiation will be observed on a side surface of thermal insulation at the center of the tank
(4.0 Sv/s on a surface and 2.02 Sv/s at 1 m distance from thermal insulation surface).

TABLE 7. GAMMA-RADIATION DOSE RATES FROM THE TRANSPORT

RAILWAY TANK (length = 10.5 m; diameter = 2.2 m; wall thickness Fe =
12 mm; insulation thickness is = 250 mm; sodium mass = 25 tons)

Dose rate, Sv/s

Points of measurement 22 137 134
Na Cs Cs Total
1 =0 0.97 2.29 1.14 4.40
= 6 years 0.19 1.94 0.18 2.31
2 =0 1.10 2.64 1.41 5.15
= 6 years 0.22 2.29 0.22 2.73
3 =0 0.66 1.67 0.70 3.03
= 6 years 0.13 1.50 0.11 1.74
4 =0 1.76 3.96 2.11 7.83
= 6 years 0.35 3.34 0.31 4.00
5 =0 1.06 1.89 1.23 4.18
= 6 years 0.22 1.63 0.18 2.03
1 filling pipe flange; 2 on the surface of insulation (top of the cistern);
3 at 1 m distance from point 2;
4 on a surface of insulation (sideways on center of the cistern);
5 at 1m distance from point 4.

In order to decrease radiation rate down to permissible level (0.55 and 0.028 Sv/s,
respectively) it is necessary to lower specific activity of the coolant down to the following
values:

on 22Na: 6105 Bq/kg;

on 137Cs: 3.7105 Bq/kg.

The estimations are carried out using the above mentioned technique [2, 3]. In this case 1:0.1
ratio of the contributions in dose rate made by isotopes of 22Na and 137Cs isotopes is assumed
beforehand, taking into account, that the spent coolant can be cleaned from caesium activity
(as it will be shown below), while it is impossible to separate 22Na from the total mass of the
coolant; the decrease of 22Na activity is possible by radioactive decay during storage.

159
5.3.1.1.2. A new NPP start-up

For the sake of estimation of permissible levels of the coolant activity on the preliminary
stage of works the following operations were considered:
Delivery of sodium from the railway cisterns to the sodium dump tanks;
Purification of sodium from impurities in the sodium dump tanks;
Filling main circuits with sodium;
Run-in of the main and auxiliary sodium systems equipment;
Refuelling works on the reactor.
For calculations of restrictions of coolant activity with regard to maintenance and repair of the
equipment in the NPP start-up period it was assumed, that the working time of the
personnel does not exceed two hours (based on the experience of realization of these work),
and that the dose to the worker obtained during this time is equivalent to the weekly dose
(1.010-3 Sv). Thus the radiological situation will be defined mainly by 137Cs. After 6 years
storage and ~ 100 times reduction of caesium nuclides content the radiological situation
will be defined by 22Na (Table 8, row ). As it can be seen from, dose rate in the vicinity
of the surface of the equipment (0.5 m) after 6 years of sodium storage is estimated by
0.15-0.31 Sv/s, which enables work on maintenance and repair of the equipment to be
carried out for 1-2 hours (Table 8).

TABLE 8. GAMMA-RADIATION DOSE RATES FROM THE EQUIPMENT IN SERVICE

Time Dose rate,

Name of Features of Points of
storage, Sv/s
equipment equipment measurement
years 22
Na 137
Cs 137
Cs Sum
Sodium length=10.5 m on the surface =0 1.76 3.96 2.11 7.83
railway diameter=2.2 m of thermal =6 0.35 3.34 0.31 4.00
cistern insulation =6 0.35 0.035 - 0.38
Fe=12 mm at 0.5 m =0 1.40 2.95 1.62 5.97
isol.=25 m distance =6 0.28 2.55 0.24 3.07
=6 0.28 0.028 0.31
Cold trap height=7 m on the trap =0 0.97 2.02 1.10 4.09
of primary diameter=1.2 m surface =6 0.19 1.76 0.17 2.12
circuit =6 0.19 0.018 - 0.19
Fe =25 m at 0.5 m =0 0.69 1.58 0.86 3.13
isol =30 mm distance =6 0.14 1.40 0.13 1.67
=6 0.14 0.014 - 0.15
Lateral length=19 m on the reactor =0 0.18 0.35 0.20 0.73
surface of diameter=7 m surface =6 0.035 0.31 0.035 0.38
the reactor Fe =100 mm =6 0.035 0.003 - 0.04
vessel
Open hole in diameter =300 mm above the =0 0.022 0.06 0.026 0.11
a cover of hole =6 0.004 0.05 0.004 0.06
reactor =6 0.004 0.0005 - 0.005
Dump tank length=14 m on tank =0 1.18 2.77 1.52 5.47
of primary diameter=4 m surface =6 0.24 2.37 0.23 2.84
circuit =6 0.24 0.024 - 0.26
Fe=20 mm At 0.5 m =0 0.79 2.02 1.18 3.99
isol=40 mm distance =6 0.15 1.76 0.17 2.08
=6 0.15 0.018 - 0.17

160
In connection with all stated above, it is possible to consider acceptable variant of
filling of the primary circuit of new FBR with the spent primary sodium of the BN-350
reactor with its preliminary cleaning resulting in 100 times reduction of caesium isotopes
content and ~ 6 years storage. Activity of the coolant thus will be as follows:

On radionuclide 22Na 4.5106 Bq/kg;

On cesium nuclides 2.5106 Bq/kg.

Although the permissible specific activity in case of equipment maintenance is higher than
that for transportation of the coolant, this limiting value is acceptable for the following
reasons:

The transportation of sodium outside the NPP site is a single operation;

It is possible to equip the transport means with biological shielding (~ 8 cm steel layer
is required).

Thus, in order to achieve permissible activity level of the spent coolant of the primary circuit
of the BN-350 (on 22Na 4.5106 Bq/kg and 137Cs 2.5106 Bq/kg) ~ 100 times reduction of
cesium content in sodium and 6 years storage for the decay of Na22 is required. With regard to
the secondary sodium, there are no restrictions for its reuse in new LMFRs from the
standpoint of radioactivity.

5.3.1.2. Non-radioactive impurities content

The necessity of restriction on the content of some impurities in the sodium coolant is
determined by their negative influence on the coolant quality. Among these impurities are
carbon, oxygen, hydrogen, chlorine, nitrogen, potassium and iron. Sources of these impurities
are:

For iron: structural materials;

For carbon: sodium pump lubrication system, structural materials and atmospheric air;
For hydrogen: atmospheric air, corrosion processes in the third circuit, sodium water
interactions;
For oxygen and nitrogen: atmospheric air.

5.3.1.2.1. Carbon concentration

The presence of carbon in sodium results in the embitterment of the chromium-nickel steels of
austenitic grade, loss of their ductility and as a consequence, decrease of their useful life. If
austenitic and pearlitic steels are used in the same sodium system, carbon transfer from
pearlitic steel to austenitic steel takes place. On the basis of the analysis of data on the
changes of mechanical properties of steels caused by the processes of their decarburization
and carburization, it is recommended to establish the following maximum permissible values
of the carbon content for reactor sodium:

For the primary circuit, consisting of only austenitic steel structures, operating at
temperatures up to 700C - 30 ppm;
For the secondary circuit, including structures of both austenitic and pearlitic steels,
operating at temperatures up to 550C - 50 ppm.

161
5.3.1.2.2. Oxygen concentration

Oxygen, contained in sodium as sodium oxide, makes significant effect on the corrosion of
structural materials. For instance, the increase of concentration of oxygen from 5-10 ppm for
the conditions of the BN-350 primary circuit results in ~ 80 kg/year additional corrosion
products yield rate. The concentration of oxygen in sodium is normally maintained at the level
of 2-3 ppm by the operation of a cold trap, however in some periods of work (repair works
carried out on the open circuit, etc.) an increase up to 10 ppm is authorized. On the stage of
sodium delivery, taking into account low temperatures and possibility of the subsequent
additional cleaning by cold traps, the level of the content of oxygen in sodium is
recommended 10 ppm.

5.3.1.2.3. Nitrogen content

The presence of nitrogen containing impurities in sodium makes under certain conditions
negative effect on the properties of structural materials (nitriding, loss of ductility and
reduction of material life). On the basis of the analysis of data on the influence of the impurity
of nitrogen on the characteristics of structural materials, it is recommended the level of the
contents of nitrogen in sodium should be 10 ppm.

5.3.1.2.4. Potassium content

The restriction on the content of the potassium impurity is caused mainly by two factors:

Potassium activation due to (n, p) - reaction with formation 41Ar, which has hard
gamma radiation;
Ability to favour carbon transfer from the source to the austenitic steels.

On the basis of calculations of potassium activation (as compared to the activity of gaseous
fission product 88Kr) it is recommended to establish the level of the potassium contents in the
primary sodium should be 200 ppm and 1 000 ppm in the secondary sodium. On the
BN-350 reactor potassium content in the secondary sodium coolant is higher than the above
presented value (~ 2 000 ppm). However, the influence of potassium on carbon transfer
process is limited because of low temperature of sodium ( 420C). Potassium content in the
primary sodium also exceeds recommended value by a factor of 2.

5.3.1.2.5. Hydrogen content

The restriction on the content of hydrogen is caused by the influence of sodium hydride on the
corrosion process, sensitivity of the monitoring system for inter-circuit leakage, reactivity and
neutrons spectrum. Recommended level of the content of hydrogen in the primary and
secondary sodium is 0.5 ppm.

5.3.2. Requirement for quality of the coolant to be used in industry

Except for the radionuclides with the hard gamma-radiation (22Na, 137Cs, 134Cs), spent sodium
coolant contains impurities, which are dangerous for the human organism by their physical
effect (tritium, plutonium, polonium and strontium). According to the Russian standards of
radiation safety [4] there are levels of activity, below which substance content may not be
regulated. For gama-active long-live isotopes (22Na, 137Cs, 134Cs) this level is 1104 Bq/kg.
Proceeding from these standards, 3000 times reduction of the BN-350 reactor primary sodium
activity caused by 22Na is required (storage for 32 years) in order to use this sodium in

162
industry. By that time 137Cs decays down to the level much lower than permissible activity,
and for 137Cs isotope cleaning is required in order to decrease its content by ~ 2104 times.
Permissible specific values are as follows:

For -active tritium: 1109 Bq/kg;

For -active isotopes:
- 210Po 1104 Bq/kg;
- 239Pu 1103 Bq/kg.

Besides, according to [4] it is required that the total activity does not exceed 1104 Bq for the
listed isotopes, except tritium, for which this value is 1109 Bq. Thus, if the standard
capacities of 100 l volume are used for the sodium transport, minimum levels of isotopes
activity are:

For tritium: 1107 Bq/kg;

For plutonium and polonium: 1102 Bq/kg.

That is, neither primary nor secondary sodium of the BN-350 NPP would require purification
to remove tritium; and 10 times decrease of -nuclides content is required for the primary
sodium. The secondary sodium of the BN-350 reactor can be used in industry without its
purification.

5.3.3. Cleaning of the BN-350 spent sodium coolant from impurities

It follows from the previous section, that the use of the spent sodium coolant of the BN-350
reactor in the nuclear engineering and industries requires the reduction of impurities of
potassium and radionuclides of cesium, 22Na, 39Pu2 and 210Po. The possibility of reducing
these impurities within the necessary limits is now considered.

5.3.3.1. Potassium

Sodium, containing increased concentration of potassium, can be rather effectively cleaned by

a method of unitary distillation [5]. Table 9 gives theoretical value of high purity sodium
output calculated using Relay equation for various maximum contents of the potassium
impurity depending on distillation temperature.

TABLE 9. OUTPUT OF PURE SODIUM IN THE PROCESS OF DISTILLATION OF

COMMERCIAL SODIUM, CONTAINING UP TO 2 000 ppm OF
POTASSIUM

Output in fractions with potassium content not exceeding

Distillation temperature, C
100 ppm 50 ppm 10 ppm 5 ppm
250 94.2% 92.9% 89.9% 88.7%
300 87.6% 84.9% 79.1% 76.7%

In the experiments the distilled metal was kept at temperatures 250-290C, potassium being
mainly sublimated. The output of sodium was at the average above 90%, content of potassium
being at the level of 50-70 ppm. When cleaning spent sodium coolant of the BN-350
secondary circuit to reduce potassium impurity content from 2 000 down to 200 ppm, in case
of its use in the primary circuit, it is not necessary to increase the sodium temperature above
250C, while insignificant loss of sodium (no more than 5%) can be expected. It is expedient

163
to dilute sodium of the BN-350 primary circuit with pure sodium for the reduction of
potassium concentration from 400 ppm down to 200 ppm. It is also reasonable to dilute
sodium of the BN-350 secondary circuit with pure sodium, in case of its use in the secondary
circuits of future LMFRs, in order to reduce potassium concentration from 2 000 down to
1 000 ppm.

5.3.3.2. Radionuclides

Various devices were applied for cleaning sodium coolant from radionuclides of cesium, such
as: cold traps of impurities, traps with the mesh made of steel, nickel, etc. However the traps
with carbon-graphite materials (Russian names MPG-6, PG, VPG, etc.) turned out to be the
most efficient. Alongside with laboratory studies, cleaning of the primary circuit coolant from
cesium nuclides with the help of graphite traps was carried out on the Rapsodie, EBR-II,
BR-10, BOR-60, BN-350 and BN-600 reactors. The efficiency of cleaning varied from
2 times on the BN-600 reactor up to 8 times on the EBR-II reactor. During decommissioning
of Rapsodie reactor, 37 tons of the primary sodium drained from the reactor with total activity
220 GBq ( 6 MBq/L) were cleaned down to the level of 1.5 MBq/L during 12 days [6].
Recently studies on deep cleaning of sodium from caesium radionuclides [7] have been
carried out at the SSC RF IPPE. The idea of the technique consists in step-by-step cleaning of
the coolant, i.e. consecutive pumping (storage) of sodium through each trap till equilibrium
distribution of caesium between sodium and graphite is achieved. The method allows reaching
required level of cleaning of sodium from cesium radionuclides with significant reduction of
amount of the sorbed material as compared to the case of the single loading. The method has
not yet been developed for the use on industrial scale. Improvement of this technique on
experimental rigs and in the reactor will allow carrying out cleaning of sodium coolant up to
the required extent for its reuse in nuclear power engineering and chemical industry.

5.3.3.3. Tritium

Tritium, present in sodium circuits as NaT and NaOT compositions, is easily removed from
sodium by cold traps, the flow rate of tritium in the cold trap depending on concentration of
hydrogen in sodium. The cleaning rate is usually calculated using the following equation
(without taking into account isotope exchange with a crystal phase in a cold trap) [8]:

J = Qs CT (1-CtH/CH),

where:
J intensity of cleaning of tritium in cold trap, kg/s;
Qs flow rate of sodium through cold trap, kg/s;
CH hydrogen concentration in the circuit, ppm;
CtH concentration of hydrogen in a trap, ppm;
CT concentration of tritium in sodium, ppm.

The practice of studies of tritium distribution in the reactor systems shows, that more than
90% of tritium formed in the core is transferred to the coolant and caught by cold traps.
Injection of the adjustable flow rate of hydrogen into the circuit makes it possible to achieve
greater tritium supply to the cold trap. Thus, the cleaning of the spent sodium from tritium
impurity after the BN-350 reactor decommissioning is feasible with the help of cold traps, i.e.
the operation of cold traps can ensure required minimum levels of tritium in sodium, intended
for the reuse.

164
5.3.3.4. Other radioactive isotopes

Metal radioactive impurities, including 54Mn, 60Co and 239Pu, have rather low solubility in
sodium (about 10-3-10-4 ppm), because in principle it is possible to catch them using
methods of settling-out at reduced sodium temperature or by setting up sodium flow through
the cold traps. Feasibility of cold trapping of some radioactive products (54Mn, 60Co, 113Sn,
125
Sb, etc.) has been confirmed experimentally [9].

The radioactive metal impurities were also effectively caught by nickel traps, where nickel
foil was used as a getter. These traps have demonstrated rather high efficiency in catching
corrosion products and fuel. There are also other methods of cleaning, for example, vacuum
gauge separation, which can give high efficiency of cleaning. Applying these methods for
cleaning of spent sodium coolant one can hope reaching required minimum level of
concentration of the radioactive isotopes in sodium, intended for reuse, although today there is
not yet experimental approval of this method.

The only impurity, posing extremely great problems when physical methods of cleaning are
applied, is radioactive isotope of 22Na with half-life period of 2.6 years. In order to decrease
activity of this isotope, methods of storage during the time required for its decay, or dilution
of the active coolant with pure sodium are applied. Taking into account values of required
degree of sodium cleaning indicated above, it is necessary to store the BN-350 primary
sodium during the following period of time: for the safe shipment 14 years, for realization
of commissioning works on newly constructed NPP ~ 6 years, for use in chemical
industry ~ 32 years. Thus, applying various methods of sodium cleaning from impurities
(distillation, pumping through cold traps and absorption traps, settling-out and storage,
dilution with pure sodium), it is possible to lower the content of impurities down to the
required limits. In order to make final decision on the method of the BN-350 spent sodium
coolant management during NPP decommissioning, two options should be compared from the
point of view of cost effectiveness and environmental impact, namely:

Cleaning of sodium coolant for its reuse in the newly constructed fast neutron reactors
or in the industry;
Reprocessing of large amounts of sodium coolant for safe disposal;
Reuse of the secondary circuit sodium on the new LMFRs and in the industry does not
make problems, and no cleaning of this sodium is required.

5.4. REPROCESSING OF RADIOACTIVE SODIUM WASTES INTO SAFE

CONDITION FOR DISPOSAL

Beside the spent sodium coolant, by the time of the BN-350 reactor decommissioning there
will be some radioactive wastes, containing so-called sub-standard sodium, unsuitable for
reuse. This sodium is strongly polluted by the products of its interaction with water or air,
substances used for fire fighting, and sodium, removed from cold traps, caesium traps, etc.
According to the estimations a few tens of tons of sub-standard radioactive sodium can be
accumulated in the BN-350 reactor. Such sodium should be disposed under condition safe for
the environment. That is, should be previously transferred into the solid state, chemically
passive and steady with respect to water leaching.

Methods of radioactive sodium reprocessing by water (Rapsodie reactor [10]) and sodium
hydroxide (EBR-II and DFR reactors [11]), developed by today, cannot be applied for
sub-standard sodium, since it would quickly plug sodium injecting devices used for these

165
methods. In this connection other ways of reprocessing are considered which can be used for
sub-standard sodium.

5.4.1. Burning in air

This method, as well as many others, is based on preliminary transformation of sodium into a
less active chemical condition by its oxidation. It is widely applied for non radioactive sodium
and characterized by sufficient productivity. The rate of reaction is easily adjusted by
controlling supply of sodium or air. The disadvantages of this method are high temperature
and necessity of cleaning of large amounts of damped air from sodium aerosols. In case of
radioactive sodium, cleaning of discharged air from radioactivity is required. When cleaning
air in the water scrubber, alkali and hydrogen will be formed, and the damped air needs in
addition to be dried, since the moisture contains alkali and radioactive substances.

In order to ensure complete sodium burning, occurring in the tray, the proper mixing of
sodium is required. Periodic clearing of the tray from the solid residues is also necessary.
Wastes of the burning process should not be considered as a final product, suitable for the
long storage or disposal. Sodium oxide, filters polluted by sodium oxides and alkali, trays and
other equipment located in the room where sodium will be burnt, the room itself and water
solutions all these will be polluted by the radioactive products, and therefore development
of methods of the their further handling is required. In case of the use of scrubbers attention
should be paid to the hydrogen produced.

5.4.2. Reprocessing using steam and water

Method of steam/gas/water washing was offered earlier for pre-repair washing of the
equipment from sodium, ensuring insignificant corrosion impact on the equipment. This
method is rather simple and it was widely used in national practice for washing the equipment
and circuits from the residues of non-drained sodium. It is sufficient to mention that the
primary circuit of the BR-10 (BR-5) reactor was washed three times using this method.

The steam/gas/water method was used for washing of the standard components of the BR-10,
BN-350 and BN-600 reactors, such as fuel subassemblies, control rods and removable parts of
primary sodium pump (PSP). Modified evaporator modules of the BN-350 steam generators
were also washed by this method. Basically this method can be used for destruction of the
small amounts of sub-standard sodium by the steam or steam/gas mixture. However, this
method is rather insidious, especially if the necessity of destruction of thick layer of sodium
rather than the thin film is considered. The fact is that the interaction of sodium with steam
would result in alkali and hydrogen appearance. The safe method of hydrogen permanent
removal and its cleaning from radioactive impurities is required. In cases when it is not
possible to provide its permanent reliable drainage, the alkali formed can block, as the crust,
access to the inner layers of sodium. Further, in the end of the process, when the equipment is
filled with water, the crust is dissolved, and vigorous, practically uncontrolled process of
interaction of sodium and water becomes probable.

In the recent years modified version of this method has been developed and tested by the
OKBM and IPPE, namely washout of the equipment from sodium by water under vacuum.
This method can be used as well for destruction of sodium. Advantage of this method is its
relative simplicity and insignificant corrosion effect on the structural materials. However, in
case of loss of the vacuum there exists the possibility of the creation of the explosive mixture
[12-15]. The rate of the process is controlled by a combination of temperature and vacuum

166
value and water flow rate. During destruction of sodium, as well as in the previous case,
hydrogen and water solution of alkali are formed.

In 1977 one more modification of the water method, so-called method of water-oil emulsion
was developed and tested on the BN-350 reactor. It was assumed that this technology would
improve safety. However, it was found that under certain conditions water-oil emulsion can
be destroyed. It is necessary to emphasize once again, that in all these cases destruction of
sodium by steam, water under vacuum or water-oil emulsion, development of a safe way of
permanent removal of hydrogen and alkali is required.

Water radioactive solutions of alkali will require additional reprocessing for their
transformation into the form, suitable for long-lived storage. With this aim, the process of
immobilization of alkali into solid mineral-like cements [16] has been developed at the
SSC RF IPPE.

5.4.3. Gas-phase oxidation of sodium

Oxygen, carbon dioxide or their mixtures can be used as oxidizers of sodium [5]. If the
oxygen is used, dry sodium oxide is formed, which is then processed by water, to be turned
into alkali. In case of the use of carbon dioxide, dry powder, consisting of sodium carbonate
and carbon, will be formed. The process of gas-phase oxidation will be carried out
at 300-350C , this being easily controlled by varying gas flow rate.

The absence of hydrogen release in case of completeness of oxidizing reaction, increases

safety of the process. The completeness of sodium oxidation reaction is provided by sodium
blending in the mixture with the mineral additives. Using technology developed at the
SSC RF IPPE, sodium carbonate and carbon powder is further transformed into water-proof
slag-alkaline cement stone, suitable for safe disposal.

5.4.4. Solid state sodium oxidation

Sodium oxidation can be carried out by metal oxides (CuO, Fe2O3, Al2O3, etc.). As a result of
the reaction going at 800C stone-like product will be formed, suitable for disposal. This
high rate reaction cannot be controlled. For the complete sodium oxidizing good sodium
mixing with the oxides of metal and uniformity of reaction process over the entire volume
should be provided [6].

5.4.5. Reprocessing of radioactive alkaline wastes using liquid state oxidation method

The technology of this method assumes the realization of reprocessing of sodium and
sodium-potassium alloy in two main stages [7]:

Stage 1: Dispersion of radioactive waste of the alkaline metal on the solid inert carrier
with the aim of obtaining dry, loose cement-like mixture;

Stage 2: Tempering of the obtained mixture together with solid clay additives by water
solution by metering out the mixture in turn with the water solution, formed
solution-cement mass being mixed.

Composition and ratio of components used in these two operations (alkali metal radwastes,
inert filler, clay additives and tempering water solution) have been determined taking into
account that spontaneous solidification of obtained solution-cement mass would result in the

167
final product of reprocessing, namely high stability slag-alkaline cement stone suitable for
environmentally safe disposal. On the first stage maximum increase of the alkali metal surface
is achieved, allowing to provide required kinetic parameters and completeness of the reaction
on the second stage.

On the tempering stage chemical transformation of alkali metal into alkali with its
simultaneous immobilization into aluminosilicate binder product is reached. Liquid
radioactive waste can be used as tempering water solution, thus enabling the reprocess two
kinds of wastes simultaneously.

The process is controlled by varying input flows of the appropriate components (slag-alkaline
mixture and tempering solution), its monitoring being implemented by the content of
hydrogen in gas cavity. The process temperature is < 150C. The process requires safe
removal and decontamination of hydrogen released.

Final product of reprocessing, namely alkaline aluminosilicate stone contains 10-12 wt% of
alkali metals radioactive waste and has high strength parameters (10-5-10-6 kg/m2 per day
water stability and 15-20 MPa compression strength).

If sodium solid waste, mixed with fire fighting means, are reprocessed by any method, they
should be previously well crushed (pounded). In order to ensure completeness of sodium
reaction permanent mixing of waste should be provided during their reprocessing and
subsequent grouting.

None of the methods of reprocessing of sub-standard sodium, mentioned-above, have been

developed for industrial scale application. Therefore in order to improve technology and
select the optimum variant creation of the experimental facility is required to perform the
necessary scope of experimental studies with subsequent development on their basis of
industrial technology and means for its realization. The fulfilment of this work is supposed to
be carried out during decommissioning of the BR-10 test reactor.

5.5. WASHING THE RESIDUAL SODIUM DEPOSITS AND DECONTAMINATION

OF REACTOR CIRCUITS, EQUIPMENT AND STRUCTURAL MATERIALS

After sodium coolant is drained from the BN-350 reactor systems in the process of its
decommissioning, some sodium will still remain as a film on the inner surfaces of circuits, as
well as in some places where complete drainage is impeded. In connection with its high
chemical and radiation activity sodium should be removed before the secondary use or
disposal of sodium equipment.

The equipment of sodium circuits and systems of LMFRs is made of expensive stainless steel.
Therefore it is expedient from the economical standpoint not to dispose this equipment,
thereby increasing the amount of radioactive waste, but to reuse it wherever possible in
industry or to re-melt it in the metallurgy. For this purpose after washing components from
sodium it is necessary to carry out their decontamination from radioactive deposit, sorbed on
the inner surfaces. Below possible methods of the LMFR pipelines and components
removal-cleaning and washing from sodium, sodium-potassium and decontamination are
discussed.

168
5.5.1. Input data for the choice of methods

5.5.1.1. Primary circuit

The primary circuit of the BN-350 reactor consists of the reactor itself and six identical loops.
According to the estimates made, after draining of 3.5 m3 the sodium remains on the inner
surfaces of the reactor, including those of fuel subassembly mock-ups. About 110-2 m3 of
sodium remain on the inner surfaces of pipelines of the primary circuit systems of impurity
indication and removal after draining, while each one of five cold traps will contain 3.2 m3
of sodium residues. After removal of sodium from 10 primary circuit drain tanks in each of
them would contain on the inner surface and on the bottom 210-2 m3 of sodium.
In total, 22 m3 of sodium remain in the primary circuits of the BN-350 reactor after draining
of the coolant.

5.5.1.2. Secondary circuit

Each one of six independent loops of the secondary circuit of the BN-350 reactor would
contain 0.65 m3 sodium with specific activity 5105 Bq/kg remaining on the inner
surfaces after the coolant draining. Besides 6.7 m3 of sodium would remain in the
non-drained tube bundles on the secondary circuit side of the intermediate heat-exchanger of
each loop. About 110-2 m3 of sodium would remain on the inner surfaces of the pipelines of
the system of impurity indication and removal of each one of six loops of the
secondary circuit after the coolant is drained, while 3.2 m3 of sodium remain in the cold
trap. Each of four drain tanks of the secondary circuit after removal of the coolant would
contain 210-2 m3 of sodium residues. In total, 63 m3 of sodium would still remain in the
secondary circuit of the BN-350 reactor after the coolant draining.

5.5.1.3. System of cold traps cooling by sodium-potassium alloy

After the coolant is drained from the system, 0.4 m3 of non-radioactive sodium-potassium
eutectic would still remain in the pipelines of the system. Each one of two cold traps of the
system would contain 0.16 m3 of sodium-potassium alloy, while 0.35 m3 of
sodium-potassium alloy are kept in the casing and coil of cooling system of each one of
14 cold traps. Five drain tanks of the system would contain 0.12 m3 of sodium-potassium
alloy after the coolant removal. In total, 5.7 m3 of sodium-potassium alloy are kept in the
system of cooling of the cold traps of the BN-350 reactor after coolant draining.

5.5.1.4. Cold traps of spent fuel storage cooling system and coolant processing system

In the cold trap of dismantled cooling system of spent fuel storage, 3.2 m3 of the
sodium-potassium alloy is contained. On the stage of the BN-350 reactor decommissioning,
each one of two cold traps of the coolant processing system would contain 3.2 m3 of
non-radioactive sodium.

5.5.2. Methods for removal of coolant from undrainable parts of reactor and equipment

There are no effective and safe methods for washing and decontamination of the systems,
having elements with the located significant masses of coolant. Therefore coolant removal
from the undrainable equipment is necessary in order to reduce the amount of residual sodium
so that it would be possible from the standpoint of safety to apply some washing technique,
this being necessary condition for the development of particular technologies. One shall

169
consider the possibilities of removal of local mass of sodium from the pipelines and
equipment.

5.5.2.1. Core diagrid and pressure plenum

It is required to remove sodium remaining on the bottom of the core diagrid of the reactor
vessel ( 3 m3) using draining pipe, reaching to the bottom. The requirement of the draining
device is that it should provide < 200 L sodium residual volume after its removal. This
requirement is based on the assumption of instantaneous interaction of sodium mass with the
excess of water, when the pressure in the reactor vessel ( 150 m3), filled with the inert gas at
0.1 MPa, should not exceed permissible value of 0.17 MPa.

5.5.2.2. Cold traps

5.5.2.2.1. Cold traps of secondary circuit and coolant processing system

The contents of cold traps can be released from contents using technology, developed and
applied for the regeneration of cold traps of the secondary circuit of the BN-350 reactor [17].
Firstly, 2/3 of the total amount of sodium is pressed out of the trap from the inlet pipe by the
inert gas, supplied under pressure through the outlet pipe. The transformation of impurities,
accumulated in the trap as Na2O and NaH, into the sodium hydroxide NaOH is based on the
following reaction:

Na2O + NaH NaOH + 2Na (1)

The process is realized by heating of a cold trap up to 470-500C during two days. The
products of regeneration have melting temperature of 390-400C and can be removed from
the trap at the temperature higher than that indicated above. In order to ensure complete
removal of sodium and products of the reaction (1) the trap should be equipped with a special
line providing the possibility of removal of sodium, its residual amount being within
permissible limits to guarantee the safety, i.e. that the internal pressure of the cold trap should
not exceed permissible value (1.0 MPa) in case of instantaneous sodium interaction residues
( 20 kg of sodium) with the excess water.

5.5.2.2.2. Cold traps of the primary circuit

The work with cold traps of the primary circuit is complicated by a high radiation
background, created by the accumulated radionuclides. The feature of the primary cold traps
is in that they mainly accumulate sodium oxide Na2O. Therefore, in order to make hydration,
supply of hydrogen from outside is necessary. The experiments made under special conditions
[18] have shown the possibility of realization of operation of complete draining of a cold trap
of the primary circuit from accumulated impurities. The technology of the draining of cold
traps of the primary circuit should be developed on the basis of existing method developed for
the secondary circuit traps, taking into account high radioactivity.

5.5.2.2.3. Cold traps of Na-K cooling systems of cold traps and spent fuel storage

The technology of draining of the cold traps of these systems should be proved out
experimentally under special conditions using principal approach, described in the previous
sections. For realization of such approach it is necessary to ensure electrical heating of traps,
since electrical heating has not been provided by the design of sodium-potassium alloy
cooling systems.

170
5.5.2.3. Intermediate heat exchangers

Draining, washing and decontamination of the intermediate heat exchanger sections is carried
out using design technology in the rotary decontamination pit of the BN-350 reactor washing
and decontamination system. First, sodium draining (pouring out) is made from the U-shaped
tube bundle by its 93 tilting.

5.5.2.4. Na-K alloy in casings and coils of cold traps cooling systems

It is assumed that sodium-potassium alloy will be removed from casings and coils of the cold
traps cooling systems using method of vacuum distillation with the use of standard systems of
electrical heating of cold traps. The technology of distillation should be tested on the models
under test rig conditions.

5.5.2.5. Drainage tanks

It is expedient to remove sodium residues from drainage tanks using method of vacuum
distillation. This technology should be tested under special conditions on models, taking into
account tanks design and volume.

5.5.3. Methods for washing and decontamination of circuits and equipment

5.5.3.1. Washing the residual sodium deposits

Steam and gas washing method and subsequent decontamination by demineralised water
solutions is considered as the principal method of washing and decontamination of pipelines
and equipment of BN-350. Such conclusion has been proved by the following reasons:

Presence of designed systems for washing and decontamination of the BN-350

equipment;
Experience gained from application of the method under reactor conditions (BR-10,
BOR-60, BN-350, BN-600) [20] and presence of the skilled personnel for its
implementation;
Efficiency, profitability and relative safety of the method;
Presence of designed systems of primary processing of wastes (including radioactive)
and experience gained from its operation;
Successful experience (three times) of application of the method for washing and
decontamination of primary circuit of experimental reactor BR-10 [19].

It is reasonable to apply vacuum distillation method as an auxiliary process (used before

steam/gas washing) of cleaning of circuits and equipment from sodium. This method is
remarkable for the high safety, since it is not accompanied by the release of hydrogen. All
necessary conditions for its implementation are available on the BN-350 reactor, namely the
presence of designed vacuum and heating systems on the most part of the equipment and
pipelines. Results of application of this method will be the reduction of total mass of sodium
to be removed from the pipelines and equipment and improvement of safety of procedures
using water methods.

Washing of the equipment by light spirits cannot be recommended for use on large scale
because of high fire risk, although this method is still used sometimes for washing the
experimental fuel subassemblies from sodium. The method of washing of the equipment and
pipelines from sodium using heavy spirits (butylcellosolve, carbitol) is considered [20].

171
However, these methods have their disadvantages. In this case, as well as in steam/gas/water
washing, there is a danger, connected with the release of hydrogen. Moreover, it is necessary
to take into account hypothetical possibility of decomposition of organic compositions at high
temperatures. It is well known, that washing procedure carried out with the help of carbitol on
the drain tank on the Rapsodie plant, resulted in the burst of the tank wall.

The IWG-FR at its Meeting in May 1999 summarizing results of the discussions took note of
a presentation of the French delegation on the above mentioned accident and made the
following conclusion: under certain circumstances (e.g. closed geometries and sodium
puddles), the use of alcohol to clean components or to destroy sodium can be dangerous. This
is proven for ethylcarbitol, but at present can not be excluded completely for other alcohols. It
was therefore strongly recommended to perform appropriate studies prior to the use of alcohol
together with sodium.

Rather promising technology being developed, is the use of water under vacuum for washing
equipment from sodium [21]. Advantages of this method are simplicity, low level of
temperatures and controllability. Danger caused by the use of this method is related to
possible loss of vacuum and the subsequent filling of the circuit with the atmospheric air that
can lead to formation of explosive mixture.

5.5.3.2. Decontamination

As regards decontamination procedures, application of solutions, tested in fast reactors under

operation [19] is expedient. The general tendency of perfection of decontamination processes
is aimed at softening of their corrosive influence on structural materials, if reuse of the
equipment or the whole circuits is assumed. The main purpose of decontamination of the
equipment and pipelines of the primary circuit on the stage of the BN-350 reactor
decommissioning will probably consist in assurance of removal, to as great as possible extent,
of a superficial layer, polluted by radionuclides of sorption/diffusion nature, with subsequent
utilization of decontaminated structural material. Therefore, at least on the initial stage of
decontamination of the primary circuit equipment it is expedient to use solutions, used
on BR-10, BN-350, and BN-600 reactors [19].

More aggressive solutions could be used for the final decontamination with the purpose of
utilization of metal in the national industry. The possibility of additional cleaning of
superficially polluted metals by their melting has been demonstrated by tests, which could be
taken into account when making choice of final technology of decontamination.

5.5.4. Technology for washing and decontamination of reactor circuits and equipment

5.5.4.1. Primary circuit

Technology of washing and decontamination of the primary circuit of the BN-350 reactor
should be developed on the basis of experience gained on washing and decontamination of the
primary circuit of the BR-10 experimental reactor [19] taking into account the scale factor.
This technology assumes the use of vacuum distillation of sodium with subsequent steam/gas
washing, after sodium is drained from the circuit and undrainable sections (reactor, impurity
indication and removal system and drain tanks) are isolated. In order to implement this
method on the BN-350 reactor, it is necessary to dismantle sections of the intermediate heat
exchangers. The procedures are as follows: the circuit is heated up to 130-150C; the
steam/gas mixture is then supplied to the top of the primary circuit loops at 130C as well;
the steam together with alkali is dumped from the drainage pipelines to the condenser.

172
Permanent hydrogen release is provided from the top of the circuit. The process is monitored
by the hydrogen release into gas phase and by alkali content in the steam condensate.

The next operation i.e. controllable batch filling of the circuit, which is filled with the inert
gas at the bottom, with demineralized water and its subsequent circulation in turn through all
loops by the standard pumps. It is also reasonable to consider the safety requirements for
using method of washing by water under vacuum for this purpose. After washing the loops of
the primary circuit except for the reactor vessel, they are subject to several cycles of
decontamination using the following solutions:

0.5% KMnO4 at 70C;

5% HNO3 + 1% H2C2O4 at 70C;
demineralized water at 60-80C.

Time and the number of decontamination cycles should be determined using model tests,
taking into account the scale factor and measurements of gamma-background carried out
during decontamination process. Final washing of the equipment by demineralized water will
be carried out until the neutral reaction of the washing water is obtained. The washing
procedure is followed by drying of the primary circuit by its heating under vacuum and
blow-down by dry nitrogen.

5.5.4.2. Secondary circuit

Before starting dismantling the secondary circuit, sections of the intermediate heat exchangers
should be replaced by pipelines. Washing of the secondary circuit is made in the order, similar
to that used for the primary circuit (except for decontamination), including the following
operations:

Drainage of sodium;
Isolation of the undrainable sections (systems of impurities control, drainage tanks);
Vacuum distillation of sodium;
Steam/gas washing;
Gradual fillings of the circuit, filled with the inert gas, with demineralized water with
its subsequent circulation in closed circuit of each loop. It is expedient to consider the
possibility of safe washing method using water under vacuum;
Drying of loops.

Application of such approach would allow washing of sodium cavities of the steam generators
before their dismantling. However, it is necessary to consider more carefully the expediency
of preliminary washing of the steam generator, to be made independently of the rest part of
the circuit.

5.5.4.3. Cold trap Na-K cooling system and spent fuel storage coolant cleaning system

Washing of the drainable equipment and pipelines of systems of cooling of the cold traps and
cleaning of the cooling circuit of the spent fuel storage can be carried out using available
technology with step-by-step dismantling with subsequent washing in decontamination pits.
This can be done after the discharge of the drainable sections of the circuit. Washing method
based on cleaning media circulation is inexpedient in view of absence of electrical heating of
the pipelines, valves and equipment, absence of mechanical pumps in the circuit and in views
of necessity of preliminary removal of undrainable amounts of sodium-potassium alloy from
casings and coils of the cold traps.

173
5.5.4.4. Washing of the undrainable equipment (with decontamination of the primary circuit
components)

Washing and decontamination of the undrainable equipment of the primary and secondary
circuits, cooling alloy system of spent fuel storage and cold trap of the cooling circuit of spent
fuel storage can be carried out after removal of the coolant and impurities in order that their
residual amounts would allow safe washing and decontamination procedures to be fulfilled by
the methods, mentioned above. For these purposes special technological system of washing
and decontamination should be created. Those components which cannot be processed using
design means of washing (decontamination pit), should be connected in turn to the special
system. Washing and decontamination of these components have to be made by means of
design technologies, namely: vacuum distillation, washing by steam/gas, demineralized water
or possibly by water under vacuum, and decontamination.

5.5.4.5. Washing of the reactor

The clearing (washing) of the reactor becomes necessary sooner or later to eliminate the
possibility of uncontrollable interaction of the sodium remaining in the reactor vessel and
water with the subsequent probable formation of an explosive mixture. However,
decontamination of the inner surfaces makes no sense because of high of induced activity.
Initial condition of the reactor assumes unloaded core, reactor vessel being isolated from the
circuits, filled with the inert gas and cooled down. The process of cleaning of the reactor will
include the following procedures:

Development and installation for special system of removal of the main amount of
sodium residues;
Heating up of the reactor and drainage of the main amount of sodium;
Vacuum distillation of sodium (if the method of reactor heating works out);
Installation of system for supply of water and steam/gas mixture and removal of water
alkali solution;
Final destruction of the sodium residues by water in the inert atmosphere, water being
supplied in measured portions into the reactor vessel at low temperature.

Instead of three former operations it is necessary to consider:

Possibility of exclusion of the third operation; or

Its replacement by washing of the sodium residues with butylcellosolve, mixed with
indifferent dilute(nt).

It is necessary to emphasize once more that this section gives results of preliminary
development on the system of washing of the equipment and pipelines of the primary,
secondary and auxiliary (sodium-potassium) circuits. This system is based on rather large
industrial experience and results of special experiments, some of which are completed, while
others are still under way. Taking into account the scale factor, hydrogen related danger,
accompanying washing processes, poor development and comparative evaluation of the
alternative methods, the work on substantiation of the optimum and the most safe washing
methods will be continued. As a result of washing and decontamination of the equipment and
pipelines of the primary circuit considerable amount of liquid radioactive wastes will be
accumulated. Evaluation of their amount (without results of washing of the reactor and
undrainable sections) is shown in Table 10, and their total activity is shown in Table 11. As it
can be seen from the tables, the total liquid radioactive waste amount according to

174
 these assumptions is equal from 9 000 to 12 000 tons, with their activity consisting of
4 TBq ( 110 Ci).

TABLE 10. LIQUID RADWASTES AFTER WASHING AND DECONTAMINATION

OF PRIMARY CIRCUIT OF (REACTOR VESSEL AND UNDRAINABLE
ELEMENTS ARE NOT TAKEN INTO ACCOUNT)

BR-10, BN-350, estimation

measurement 1 loop 6 loops Total
2
Surface, m 100 1 800 10 800 -
Volume, m3 2 70 420 -
Steam-gas Steam, t 30 540*/1 050** 3 240*/6 300** 3 660*/6 720**
Distillate, t 2 70 420 -
Decontamination D-solutions, t 30 420 2 520 5 040
Distillate, t 30 420 2 520 -
Note: Estimations are given in the assumption of application of washing and decontamination technology, tested
on the BR-10 reactor. The initial data on the BR-10 reactor are taken from the results of maximum washing;

* calculation was made on the basis of the ratio of the processed areas of surface;
** calculation was made on the basis of ratio of circuit volumes.

TABLE 11. TOTAL ACTIVITY OF LIQUID WASTE AFTER STEAM/WATER

WASHING AND CHEMICAL DECONTAMINATION OF ONE LOOP OF
THE BN-350 REACTOR (COLD TRAP AND EQUIPMENT OF REACTOR
VESSEL CAVITY ARE EXCLUDED) [Hot surface 1220 m2 (50 m2
pipes, 1170 m2 IHX), Cold surface 110 m2 (90 m2 pipes, 20 m2 PSP,
valves)]

Surface Total Total activity of water Total activity of water Total

contami- activity of after steam/water after chemical activity of
nation 1 loop washing, decontamination, water, *
GBq GBq GBq
Nuclides MBq/m2 GBq t=0 5 years t=0 5 years 5 years
22
Na 3 4 4 1 0 0 0
137
Cs 500 660 590 520 70 60 580
134
Cs 50 70 60 12 10 2 14
54
Mn - - - - - - -
hot 1.500 1.820 - - - - -
cold 750 2026 1416 18.5 470 5 23.5
60
Co 100 - 100 50 30 15 65
Total - 2 760 2 170 600 580 82 682
(75 Ci) (58 Ci) (16.2 Ci) (15.5 Ci) (2.2 Ci) (18.4 Ci)
Notes: Total activity of liquid radioactive waste in case of washing of all six loops will be 4 GBq ( 110 Ci);
* after steam/water washing and chemical decontamination

175
5.6. RADIATION CONDITIONS IN THE PRIMARY CIRCUIT CELLS, ACTIVITY OF
REACTOR BLOCK ELEMENTS

Activity values of various coolants and equipment of the reactor installation presented in
Section 8.2 demonstrate the possibility of the selective approach from the point of view of
radiological safety assurance, if some works on the reactor decommissioning are under way.
The following points can have influence on the choice of technical approaches (for example,
remote or usual methods of the equipment handling, etc.):

Wide range of low and high activities of the equipment;

Possibility of realization of technological operations after short or long storage;
Possibility of realization of decontamination of the polluted equipment.

5.6.1. Radiological conditions of dismantling works in the primary circuit cells

Calculated data on the radiological conditions in case of dismantling work carried out in the
primary circuit loops cells are presented in this section. Only the radioactive sources located
in the sodium coolant or on the inner surfaces of the equipment were considered. Radioactive
contamination of outer surfaces of the equipment is absent. Table 12 shows the dose rates of
gamma-radiation in the cells and on the surface of the equipment for both sodium filled and
drained conditions, of the primary circuit after reactor shutdown and decay of 24Na are
shown,. The data are presented for various times after final reactor shutdown.

TABLE 12. DOSE RATES OF GAMMA-RADIATION ON THE EQUIPMENT OF THE

PRIMARY CIRCUIT, Sv/s

Time average Surface of Surface of Surface of Surface of

after gamma- thermal thermal thermal thermal
reactor background insulation insulation insulation insulation
stop, in the cell of 600 mm of PSP of IHX of cold traps
y pipeline

F D F D F D F D F D

0 0.7 0.07 3.5 0.35 3.8 0.38 1.2 0.9 4.0 2.0

5 0.4 0.015 2.0 0.075 2.2 0.08 0.7 0.2 3.0 1.9

10 0.3 0.01 1.6 0.055 1.7 0.06 0.5 0.15 2.0 1.5

50 0.1 0.002 0.5 0.015 0.55 0.015 0.18 0.03 0.6 0.5

100 0.03 0.0005 0.15 0.005 0.15 0.005 0.05 0.009 0.15 0.13

150 0.01 0.0001 0.05 0.001 0.05 0.001 0.013 0.003 0.05 0.04
Note: F - the circuit filled with sodium; D - the circuit drained.

176
The radiological conditions indicated in Table 12 allow estimating the permissible time for
group A8 personnel to stay in the cells of the primary circuit under radiological conditions
indicated in Table 12. In this case, let us take 90 Sv value as permissible dose limit for one
shift. The dose rates of gamma-radiation in the cells of the primary circuit after reactor
shutdown and decay of 24Na are within the range of 0.5-4 Sv/s. Assuming a value of 90 Sv
as permissible dose limit for one shift, it follows that the permissible time for personnel to
stay in the vicinity of the cell entrance is 3 minutes, and 0.3 minutes near the components.
Only visual examination can be made during these short times.

As it can be seen from Table 12, the storage during 510 years results in only 1.5-2 times
increase of permissible operating times for the primary circuit cells without drainage of the
coolant. The work of the personnel of the A category in the primary circuit cells during whole
shift (8 hours) can be permitted only after 150 years. The coolant drainage reduces by about
10 times dose rate of the gamma-radiation from the pipelines and pumps, while the dose rate
from heat exchangers and cold traps is reduced by 1.2-2 times. After coolant drainage the
permissible operating times of personnel in the cells during one shift are estimated as follows:

At the entrance of the cell 30 min;

In the vicinity of pipelines and PSP 7-10 min;
In the vicinity of IHX and cold trap 1-2 min.

As has been already stated, high dose rate of gamma-radiation (50-100 Sv/s) will remain on
the short sections of the PSP sodium overflow pipeline. The storage during 5-10 years reduces
dose rates of gamma-radiation on the drained equipment (e.g. pipelines, PSP and IHX) by
about 57 times, while for the cold traps this reduction is only 30%. Apparently, the work
of the personnel of the A category in the cells of the primary circuit during total shift time
(8 hours) can be permitted only about 100 years after shutdown of the reactor. In order to
increase significantly permissible time of staying of working time of the personnel in the
primary circuit cells, steam/water washing and chemical decontamination of the inner
surfaces of pipelines and equipment are required. For this purpose it is necessary first of all to
isolate reactor from the circuit. The cold traps should be isolated after the drainage of the
coolant before steam/water washing of the sodium loop.
Chemical decontamination, as well as steam/gas/water washing of the sodium loops, should
be carried out after their isolation from the reactor. This is necessary in order to avoid transfer
of radioactive products of corrosion of the in-vessel elements of the reactor, and for the
reasons of washing process safety. In order to reduce contamination of the air in the cells,
caused by the dusting operations (cutting and welding of metal, etc.) during preparatory
stage of decontamination and also during the general dismantling work, the working places
should be equipped with local ventilation, and the personnel should have individual means of
protection of the respiration organs. Table 13 shows the activity reduction ratio of superficial
deposits caused by steam/water washing and chemical decontamination of smooth surfaces of
the primary circuit equipment.

8
Equivalent dose for the personnel of group A dealing directly with industrial sources, in accordance with
Radioactive Safety Standards (NRB-96) is 20 mSv per year.

177
TABLE 13. ACTIVITY REDUCTION RATIO OF SUPERFICIAL DEPOSITS CAUSED
BY STEAM/WATER WASHING AND CHEMICAL
DECONTAMINATION OF SMOOTH SURFACES OF THE PRIMARY
CIRCUIT EQUIPMENT

After
Environment of the After chemical Total
Radionuclides steam/water
primary circuit components decontamination reduction
washing
137
Gas Cs, 134Cs 100 50 5 000
54 60
Gas Mn, Co 2 30 60
137 134
Sodium Cs, Cs 10 300 3 000
54 60
Sodium Mn, Co 4 50 200

Table 14 presents levels of superficial contamination of the primary components, caused by

the main radionuclides, after steam/water washing and chemical decontamination.

TABLE 14. SUPERFICIAL CONTAMINATION OF THE EQUIPMENT OF THE

PRIMARY CIRCUIT AFTER STEAM/WATER WASHING AND
CHEMICAL DECONTAMINATION, kBq/m2

On the surface
Time after reactor On the surface in contact with sodium
in contact with gas
shutdown,
years 137
Cs 134
Cs 60
Co 54
Mn* 137
Cs 134
Cs
0 160 16 500 7.500 1.000 100
5 140 3 260 100 900 20
10 125 0.07 135 1.5 800 5
50 40 0.006 20 - 300 0.05
100 12 - 0.7 - 70 -
150 3 - 0.001 - 20 -
* Hot sodium section

Table 15 shows the calculated dose rates of gamma-radiation on the equipment of the primary
circuit in the sodium loops cells after steam/water (SW) washing and chemical
decontamination (CD). Dose rates at 0.5 m distance from the surface of thermal insulation of
the pipelines and equipment will be about 1.5 times less, than those on the surface. After
steam/water washing and chemical decontamination the permissible operating time of the
operators for one shift inside the cells of the primary circuit is estimated as follows:

178
TABLE 15. DOSE RATES OF GAMMA-RADIATION ON THE SURFACE OF
THERMAL INSULATION OF THE EQUIPMENT OF PRIMARY CIRCUIT
AFTER STEAM/WATER WASHING AND CHEMICAL
2
DECONTAMINATION, kBq/m

Local point
Time
(pipe of
after Average gamma-
600 mm sodium
reactor background in PSP IHX
pipelines overflow
shutdown, the cell
drainage
y
from PSP)

SW CD SW CD SW CD SW CD SW CD

0 0.015 0.001 0.08 0.005 0.09 0.006 0.2 0.013 1.0 0.02
5 0.0025 0.0002 0.012 0.001 0.015 0.001 0.03 0.0015 0.9 0.018
10 0.0015 0.00006 0.0075 0.0003 0.008 0.0003 0.02 0.0008 0.8 0.016
-6 -5 -5 -6
50 0.00025 210 0.0015 110 0.0015 110 0.003 510 0.3 0.005
100 0.0001 - 0.0005 - 0.0005 - 0.0009 - 0.07 0.0015
150 210-5 - 0.0001 - 0.0001 - 0.0002 - 0.02 0.0004

After steam/water washing:

At the cell entrance and in the middle area 1.5-2 hours;

In the area of pipelines and primary sodium pump (PSP) 0.3 hours;
In the vicinity of IHX 0.25 hours.

After chemical decontamination:

At the cell entrance and in the middle area without restriction;

In the area of pipelines and PSP 4 hours;
In the vicinity of 2 hours.

Additional storage of the equipment during 510 years (after the steam/water washing and
chemical decontamination) can practically eliminate the restriction on working time of the
group A in the cells of the primary circuit. Of course, these procedures would result in
formation of certain liquid wastes, requiring additional efforts for their solidification. The
dose rates of gamma-radiation on a surface of the equipment will be about two times higher
than those on the surface of thermal insulation.

5.6.2. Radiological characteristics of activated in-vessel structures, reactor vessel and

reactor cavity

After final reactor shutdown, the project of the BN-350 decommissioning is required for
fulfilment of preparatory work, unloading of all fuel assemblies and control rods from the
reactor, drainage of the coolant and washing of the equipment and pipelines from the residues
of sodium. Further it will be possible to commence the initial stages of dismantling of some
reactor structures. It is necessary to take into account that the unloading of the fuel
subassemblies is accompanied by their replacement with the non-radioactive steel dummy
subassemblies, which should require in the future washing from the sodium residues and
decontamination.

179
The main amount of solid radioactive wastes is caused by the activated structural materials of
the reactor and shielding materials, located in the reactor vessel cavity. Reactor vessel is
located in the concrete cavity and is separated from the concrete cavity walls by the lateral
radiation shielding, consisting of steel plates and the reinforcing cage filled by the iron ore
concentrate. On the foundation concrete beam under the reactor vessel the bottom shielding is
located, which is made of materials, similar to those used for the lateral shielding. In the neck
of the reactor vessel rotating plugs made of steel sheets, graphite and thermal insulation are
located, playing a role of thermal and biological shielding. The upper section of the cavity
beyond the reactor vessel neck is blocked by the upper stationary shielding, consisting of steel
sheets, space between them being filled with the serpentinite concrete.

The activity of this structure was estimated for 2003 as a conditional date of expiration of the
reactor life. Until 1st of December 1997, the real history of the reactor operation was taken
into account, while after this date average capacity of the reactor for the previous period was
used.

5.6.2.1. Calculation codes of spatial-power distributions of neutrons and activation of

materials [2224]

The spatial power distribution of neutrons in the reactor and in the ex-vessel shielding
structures was calculated by the TWODANT code in RZ-geometry with the use of 26-group
data base library ABBN-93. Analytical studies on accumulation of the radionuclides in the
main structures of the reactor, shielding and building materials were carried out by the
FISPACT code, taking into account irradiation history of materials during 30 years of the
reactor operation.

5.6.2.2. Initial contents of elements in structural materials

Impurities were taken into account in the calculations of the radionuclides accumulation in the
materials. The experimental data on the impurities contents in the iron ore filling and Cr18Ni9
steel of the BN-350 reactor are presented in Tables 16 and 17. These tables also give expertise
estimated data on the contents of some impurities, which were not determined by the
experiments, although they can influence the total value of the induced activity.

TABLE 16. THE CONTENTS OF BASIC ELEMENTS AND IMPURITIES IN IRON

ORE FILLING (wt%)

Elements Concen- Elements Concen- Elements Concen- Elements Concen-

trations trations trations trations
O 26.9 Ca 0.079 Cu 0.015 Eu* 0.1E-4
Na 0.096 Sc 0.580E-4 As 0.270E-2 Nb* 0.1E-3
Mg 0.170E-2 Ti 0.10 Zn 0.081 Ni* 0.1E-2
Al 0.036 V 0.650E-2 Mo 0.170E-2 Cs* 0.1E-3
Si 3.3 Cr 0.072 Ag 0.110E-3 Sm* 0.1E-3
P 0.022 Mn 0.210 In 0.013 N* 0.01
S 0.001 Fe 68.6 Sb 0.110E-2 - -
K 0.11 Co 0.380E-2 W 0.800E-2 - -
* Elements which were taken into account in calculation in addition to those measured.

180
TABLE 17. THE CONTENTS OF BASIC ELEMENTS AND IMPURITIES IN
1CR18NI9 STAINLESS STEEL (wt%)

Elements Concen- Elements Concen- Elements Concen- Elements Concen-

trations trations trations trations
C 0.08 Cr 18.0 Mo 0.11E-2 Eu* 0.1E-4
N 0.027 Mn 1.36 Ag 0.71E-4 Cs* 0.1E-3
Mg 0.4E-2 Fe 70.5 Cd 0.11E-2 Sm* 0.1E-3
Al 0.011 Co 0.021 In 0.021 - -
Si 0.43 Ni 9.4 Sn 0.45E-2 - -
P 0.01 Cu 0.064 Sb 0.34E-2 - -
S 0.01 Zn 0.017 W 0.42E-2 - -
Ti 0.013 As 0.21E-2 Pb 0.20E-3 - -
V 0.75E-2 Nb 0.25E-2 - - - -
* Elements which were taken into account in calculation in addition to those measured.

Experimental studies were not made on the contents of the impurities in standard and
serpentinite concrete, thermal insulation of plugs and graphite and 3-grade steel of the
BN-350 reactor. Table 18 shows the most complete data on the contents of the impurities in
concrete were published in the article [26].

TABLE 18. THE CONTENTS OF BASIC ELEMENTS AND IMPURITIES IN

SHIELDING CONCRETE (1.0E-4 wt%)

Element Content Ave. Ele- Content Ave. Ele- Content Ave.

spread cont. ment spread cont. ment spread cont.
H - 0.610 Mn 56.0-990.0 377.0 Ba 20.0-7.060 950
Li - 20.0 Fe* 0.5-24.0 3.9 La 2.9-28.0 13.0
B - 20.0 Co 1.1-31.0 9.8 Ce 6.2-52.0 24.3
C - 1.1 Ni 11.9-87.0 38.0 Nd - 12.6
N - 120.0 Cu 10.0-60.0 25.0 Sm 0.42-4.2 2.0
O - 47.7 Zn 8.4-340.0 75.0 Eu 0.11-1.2 0.55
Na 0.0176-1.89 0.739 Ga 1.05-20.0 8.8 Tb 0.11-.78 0.41
Mg - 0.0200 As 0.89-29.0 7.9 Dy 0.55-4.3 2.3
Al 0.53-6.1 3.1 Se 0.26-2.0 0.92 Ho - 0.9
Si 3.9-32.4 16.8 Br 1.0-5.6 2.5 Yb 0.38-3.0 1.4
P - < 0.5 Rb 2.5-70.0 35.0 Lu 0.15-0.5 0.27
S 0.2-0.46 0.31 Sr 220.0-940.0 438.0 Hf 0.65-5.7 2.2
Cl 11.0-59.0 45.0 Y 3.0-96.0 18.2 Ta 0.09-0.9 0.44
K 0.047-2.5 0.75 Zr 27.0-160.0 71.0 W 0.39-3.9 1.4
Ca 8.8-34.7 18.3 Nb 1.3-9.3 4.3 Ir - < 0.01
Sc 0.73-17.4 6.5 Mo 1.8-36.0 10.3 Pb 5.4-560.0 61.0
Ti 0.023-0.79 0.212 Ag - < 0.2 Th 0.75-12.0 3.5
V 13.3-490.0 103.0 Sb 0.16-13.0 1.8 U 1.4-4.4 2.7
Cr 29.0-540.0 103.0 Cs 0.32-6.2 1.3 - - -

These data on the impurities were taken as the basis for the serpentinite concrete and thermal
insulation of the plugs, the composition of the main elements for these materials being
presented in Table 19.

181
TABLE 19. THE CONTENTS OF BASIC ELEMENTS IN SERPENTINITE CONCRETE
AND IN MINERAL COTTON WOOL (THERMAL INSULATION IN THE
PLUG), wt%

H O Na Mg Al Si S K Ca Fe Dens.,
g/cm3
Mineral
cotton wool - 39.9 2.04 5.04 5.47 19.76 - 2.28 15.8 3.96 0.15
(heat ins.)
Serpentinite
1.27 48.5 - 15.5 1.45 18.6 0.036 - - 5.68 2.2
concrete [6]

Tables 20 and 21 show the expertise estimated contents of the impurity elements in graphite
and 3-grade steel are presented.

TABLE 20. THE CONTENTS OF BASIC AND IMPURITIES ELEMENTS IN

GRAPHITE, wt%

Elements Concentrations Elements Concentrations Elements Concentrations

Li 5.8E-6 Co 5.9E-6 Sm 2.5E-5
C 100 Ni 1.8E-3 Eu 7.6E-7
N 4.7E-4 Nb 7.7E-4 Hg 6.7E-5
Cl 1.3E-3 Ag 9.0E-6 U 2.0E-4
Ca 1.4E-2 Sn 4.9E-5 - -
Fe 2.0E-3 Ba 1.1E-3 - -

TABLE 21. THE CONTENTS OF BASIC ELEMENTS AND IMPURITIES IN 3-GRADE

STEEL, wt%

Elements Concentrations Elements Concentrations Elements Concentrations

Cr 0.25 Co* 0.021 Mo* 0.11E-2
Mn 0.25 Ni 0.25 Eu* 0.1E-4
Fe 99 Nb* 0.25E-2 - -
* Concentrations of elements assumed to be equal to those of Cr18Ni9 steel

5.6.2.3. Results of calculation of materials activity

Values of activity induced in the reactor materials and compositions of radionuclides

produced under irradiation, vary in a wide range. For the purpose of convenience, analysis of
calculation results is given for individual structural elements of the reactor and shielding,
apart from analysis of activity values summed up for all radioactive reactor materials.

5.6.2.3.1. Activity of steel of in-vessel lateral shielding (band), shielding subassemblies and
core diagrid

Tables 22 and 23 represent average specific activity values of nuclides contained in the steel
of the band with the shielding subassemblies and in the steel of the core diagrid. The total
activity values of steels depending on the duration of storage after irradiation are also
presented in these tables.

182
TABLE 22. AVERAGE SPECIFIC ACTIVITIES OF NUCLIDES IN STEEL OF
BAND AND SHIELDING SUBASSEMBLIES (Bq/kg) AND TOTAL
ACTIVITY OF THE STEEL (Bq)

Nuclide Storage duration, years

composition 0 10 50 100 150 10 000
60
Co 2.52E+11 6.77E+10 3.50E+8 4.86E+5 2.39E+3 -
63
Ni 7.21E+10 6.73E+10 5.09E+10 3.60E+10 2.55E+10 -
59
Ni 1.19E+9 1.19E+9 1.19E+9 1.19E+9 1.19E+9 1.09E+9
94
Nb 1.26E+7 1.26E+7 1.26E+7 1.26E+7 1.26E+7 8.93E+6
55
Fe 3.56E+12 2.74E+11 1.30E+7 - - -
66
Mn 6.18E+12 - - - - -
51
Cr 1.23E+12 - - - - -
60m
Co 1.50E+11 - - - - -
59
Fe 7.73E+11 - - - - -
Ni65 5.20E+10 - - - - -
58
Co 2.32E+12 - - - - -
58m
Co 1.15E+12 - - - - -
55
Cr 8.10E+10 - - - - -
52
V 7.37E+10 - - - - -
Sum, Bq/kg 1.79E+13 4.11E+11 5.26E+10 3.72E+10 2.67E+10 1.10E+9
Sum, Bq 6.13E+17 1.41E+16 1.80E+15 1.28E+15 9.15E+14 3.76E+13

TABLE 23. AVERAGE SPECIFIC ACTIVITY OF NUCLIDES IN THE STEEL OF THE

CORE DIAGRID (Bq/kg) AND TOTAL ACTIVITY OF STEEL (Bq)

Nuclide Storage duration, years

composition 0 10 50 100 150 10 000
60
Co 2.33E+11 6.27E+10 3.23E+8 4.50E+5 7.48E+2 -
63
Ni 3.60E+10 3.33E+10 2.53E+10 1.79E+10 1.27E+10 -
59
Ni 2.60E+8 2.60E+8 2.60E+8 2.60E+8 2.60E+8 2.38E+8
94
Nb 2.13E+6 2.13E+6 2.13E+6 2.13E+6 2.13E+6 1.50E+6
55
Fe 6.29E+11 4.84E+10 - - - -
66
Mn 3.55E+12 - - - - -
51
Cr 6.78E+11 - - - - -
60m
Co 1.20E+11 - - - - -
59
Fe 7.63E+10 - - - - -
65
Ni 9.09E+9 - - - - -
95
Nb 1.65E+9 - - - - -
58
Co 7.48E+8 - - - - -
64
Cu 6.05E+8 - - - - -
54
Mn 5.22E+8 - - - - -
Sum, Bq/kg 5.37E+12 1.45E+11 2.58E+10 1.82E+10 1.29E+10 2.40E+8
Sum, Bq 8.97E+16 2.42E+15 4.30E+14 3.04E+14 2.16E+14 4.01E+12

The analysis of results of calculation of induced activity of the bond steel, shielding
subassemblies and core diagrid has shown that the value of activity of steel of these structures
in the whole range of storage duration values exceeds the allowable level, at which no
regulation of the radioactive material by the RSS (radiation safety standards) is required [25].

After several years of storage the activity is determined by 60Co, 63Ni, 59Ni and 94Nb. All these
steel structures should be isolated from the human environment.

183
5.6.2.3.2. Activity of steel of thermal shielding, main and guard reactor vessels and
supporting ring

Tables 24 and 25 present the average specific activities of nuclides in the steel of structures
mentioned above. The total activities of steels depending on time of storage after irradiation
are also given in these tables. The analysis of results of calculation has shown, that the
induced activity of the significant portion of steel mass of the main reactor vessel and guard
vessel (~ 80%) and all steel of the supporting ring exceeds the allowable level, at which the
regulation of the radioactive material by the RSS is not required.

TABLE 24. AVERAGE SPECIFIC ACTIVITIES OF NUCLIDES IN STEEL OF

THERMAL SHIELDING OF THE VESSEL (Bq/kg) AND TOTAL
ACTIVITY OF STEEL (Bq)

Nuclide Storage duration, years

composition 0 10 50 100 150 10 000
60
Co 1.49E+11 3.95E+10 2.06E+8 2.87E+5 4.42E+2 -
63
Ni 2.81E+10 2.62E+10 1.99E+10 1.41E+10 9.98E+9 -
59
Ni 2.41E+8 2.41E+8 2.41E+8 2.41E+8 2.41E+8 2.20E+8
94
Nb 1.28E+6 1.28E+6 1.28E+6 1.28E+6 1.28E+6 9.11E+5
55
Fe 3.89E+11 3.01E+10 1.04E+6 - - -
56
Mn 2.10E+12 - - - - -
51
Cr 4.97E+11 - - - - -
60m
Co 7.75E+10 - - - - -
59
Fe 3.46E+10 - - - - -
65
Ni 5.56E+9 - - - - -
55
Cr 7.79E+9 - - - - -
52
V 2.95E+9 - - - - -
58
Co 4.45E+8 - - - - -
Total, Bq/kg 3.31E+12 9.61E+10 2.03E+10 1.43E+10 1.02E+10 2.21E+8
Total, Bq 1.78E+17 5.17E+15 1.09E+15 7.69E+14 5.47E+14 1.19E+13

TABLE 25. AVERAGE SPECIFIC ACTIVITIES OF NUCLIDES IN STEEL OF THE

MAIN REACTOR VESSEL, GUARD VESSEL AND SUPPORT RING
(Bq/kg) AND TOTAL ACTIVITY OF STEEL (Bq)

Nuclide Storage duration, years

composition 0 10 50 100 150 10 000
60
Co 2.32E+10 6.39E+9 3.31E+7 4.59E+4 6.30E+1 -
63
Ni 3.59E+9 3.13E+9 2.37E+9 1.67E+9 1.19E+9 -
59
Ni 3.52E+7 3.52E+7 3.52E+7 3.52E+7 3.52E+7 3.20E+7
94
Nb 2.13E+5 2.13E+5 2.13E+5 2.13E+5 2.13E+5 1.52E+5
55
Fe 5.12E+10 3.93E+9 1.37E+5 - - -
66
Mn 2.75E+11 - - - - -
51
Cr 6.16E+10 - - - - -
60m
Co 1.32E+10 - - - - -
59
Fe 5.12E+9 - - - - -
55
Cr 9.53E+8 - - - - -
65
Ni 7.34E+8 - - - - -
52
V 1.03E+8 - - - - -
58
Co 5.62E+7 - - - - -
58m
Co 2.82E+7 - - - - -
Total, Bq/kg 4.35E+11 1.35E+10 2.44E+9 1.71E+9 1.22E+9 3.21E+7
Total, Bq 4.42E+16 1.37E+15 2.47E+14 1.74E+14 1.23E+14 3.26E+12

184
The analysis of results of calculation has shown, that the induced activity of the significant
portion of steel mass of the main reactor vessel and guard vessel (~ 80%) and all steel of the
supporting ring exceeds the allowable level, at which the regulation of the radioactive
material by the RSS is not required. After several years of storage the activity is determined
by 60C, 63Ni, 59Ni and 94Nb isotopes.

5.6.2.3.3. Activity of steel of rotating plugs, plugs shielding structure and central column

The central column penetrates through the rotating plugs and shielding structure the activity at
the bottom part making the main contribution to the total activity of steel of the reactor
structures. Some steel sheets located in the bottom part of rotating plugs have the activity
higher than the permissible limits of RSS.

The induced activity of graphite and thermal insulation is lower than the permitted value, i.e.
no RSS regulation of radioactive material is required. Fifty years later the activity of rotating
plugs will be lower than the permitted value, i.e. no RSS regulation of radioactive material is
required. A hundred years later ~ 80% of steel shielding of rotating plugs will have an activity
level, indicated above. Average specific activity values of nuclides in the steel of structures
indicated above are presented in Table 26. In this Table the total activity values of steel
depending on time of storage after irradiation are also presented. Value of the activity after
several years of storage is determined by 60C, 63Ni, 59Ni and 94Nb isotopes.

TABLE 26. AVERAGE SPECIFIC ACTIVITIES OF NUCLIDES IN THE STEEL OF

THE ROTATING PLUGS, SHIELDING STRUCTURE AND CENTRAL
COLUMN (Bq/kg), AND TOTAL ACTIVITY OF THE STEEL (Bq)

Nuclides Storage duration, years

composition 0 10 50 100 150 10 000
60
Co 1.04E+9 2.78E+8 1.43E+6 1.99E+3 - -
63
Ni 6.58E+8 6.14E+8 4.96E+8 3.30E+8 2.33E+8 -
59
Ni 6.20E+6 6.20E+6 6.20E+6 6.20E+6 6.20E+6 5.64E+6
94
Nb 4.0E+3 4.00E+3 4.00E+3 4.00E+3 4.00E+3 2.86E+3
55
Fe 7.04E+9 5.37E+8 - - - -
56
Mn 1.75E+10 - - - - -
51
Cr 1.07E+10 - - - - -
60m
Co 5.78E+8 - - - - -
59
Fe 3.52E+8 - - - - -
55
Cr 1.33E+8 - - - - -
65
Ni 9.67E+7 - - - - -
Total, Bq/kg 3.78E+10 1.44E+9 4.74E+8 3.35E+8 2.39E+8 5.64E+6
Total, Bq 4.73E+15 1.80E+14 5.92E+13 4.19E+13 2.99E+13 7.06E+11

5.6.2.3.4. Activity of steel shielding, iron ore filling and steel lining of concrete

The activity of all shielding steel, located inside the iron ore filling, will exceed the level,
which is critical from the standpoint of the radioactive material regulation by the RSS.
A hundred years from then only steel of the concrete lining and that of the back wall of the
reinforcing cage of the iron ore filling will have activity below the RSS limit value. Average
specific activities of the nuclides in the steel of the structures mentioned above are given in
Table 27. In this table the total activities of steel depending on time of storage after irradiation
are also presented. Value of activity of 3 grade steel after several years of storage is
determined by 60Co, 63Ni, 59Ni, 94Nb, 152Eu, and 154Eu isotopes.

185
TABLE 27. AVERAGE SPECIFIC ACTIVITIES OF THE NUCLIDES IN THE STEEL
OF SHIELDING, IRON ORE FILLING AND CONCRETE LINING (Bq/kg),
AND TOTAL ACTIVITY OF STEEL (Bq)
Nuclide Storage duration, years
composition 0 10 50 100 150 10 000
60
Co 7.65E+9 2.05E+9 1.06E+7 1.46E+4 2.04E+1 -
63
Ni 3.55E+7 3.3E+7 2.50E+7 1.77E+7 1.25E+7 -
59
Ni 3.88E+5 3.88E+5 3.88E+5 3.88E+5 3.88E+5 3.53E+5
94
Nb 8.99E+4 8.99E+4 8.99E+4 8.99E+4 8.99E+4 6.39E+4
152
Eu 7.69E+6 4.58E+6 5.71E+5 4.24E+4 3.15E+3 -
154
Eu 9.38E+6 4.17E+6 1.67E+5 2.97E+3 5.27E+1 -
55
Fe 2.84E+10 2.28E+9 7.52E+4 - - -
56
Mn 1.99E+10 - - - - -
51
Cr 3.46E+8 - - - - -
60m
Co 4.26E+9 - - - - -
59
Fe 2.79E+9 - - - - -
65
Ni 7.69E+7 - - - - -
Total, Bq/kg 6.33E+10 4.26E+9 3.66E+7 1.81E+7 1.29E+7 4.17E+5
Total, Bq 2.18E+16 1.47E+15 1.26E+13 6.23E+12 4.44E+12 1.44E+11

5.6.2.3.5. Activity of the iron ore filling

Iron ore filling serves as concrete shielding against radiation. After 50 years of storage, about
65% of the iron ore filling will have activity above limiting RSS values. A hundred years after
the reactor shutdown the portion of the filling having such activity would not exceed 15% of
its initial volume. Average specific activities of the nuclides in iron ore filling are shown in
Table 28 In this table the total activity values depending on the time of storage after
irradiation are also presented. Value of activity of the iron ore filling after several years of
storage is determined by 60Co, 152Eu, 154Eu, 94Nb, 151Sm, 14C and 41Ca isotopes [27].

TABLE 28. AVERAGE SPECIFIC ACTIVITIES OF THE NUCLIDES IN IRON ORE

FILLING (BQ/KG) AND TOTAL ACTIVITY OF IRON ORE FILLING (BQ)
Nuclide Storage duration, years
composition N 10 50 100 150 10 000
60
Co 2.22E+7 5.92E+6 3.07E+4 4.27E+1 - -
63
Ni 1.94E+3 1.81E+3 1.37E+3 9.73E+2 6.85E+2 -
59
Ni 2.11E+1 2.11E+1 2.11E+1 2.11E+1 2.11E+1 1.92E+1
94
Nb 4.87E+2 4.87E+2 4.87E+2 4.87E+2 4.87E+2 3.47E+2
152
Eu 1.81E+5 1.08E+4 1.35E+3 9.93E+1 7.33E+1 -
154
Eu 1.77E+5 7.06E+3 6.50E+2 5.01E+1 - -
151
Sm 1.59E+4 1.47E+4 1.08E+4 7.33E+3 4.97E+3 -
14
C 7.29E+3 7.25E+3 7.23E+3 7.19E+3 7.06E+3 2.12E+3
41
Ca 2.53E+2 2.52E+2 2.53E+2 2.53E+2 2.53E+2 2.36E+2
39
Ar 3.75E+2 3.66E+2 3.29E+2 2.89E+2 2.53E+2 -
55
Fe 2.66E+8 2.04E+7 7.06E+2 - - -
134
Cs 1.19E+6 4.14E+4 - - - -
59
Fe 1.28E+7 - - - - -
60m
Co 1.24E+7 - - - - -
153
Sm 1.92E+6 - - - - -
42
K 1.13E+6 - - - - -
Total, Bq/kg 3.29E+8 2.64+E7 7.88+E4 1.62+E4 1.34+E4 2.73E+3
Total, Bq 4.28E+14 3.44+E13 1.03+E11 2.12+E10 1.74+E10 3.55E+9

186
5.6.2.3.6. Activity of the concrete
Just after the reactor shutdown the activity of about 381 m3 of concrete exceeds the RSS limit
values. After 100 years of storage the amount of radioactive concrete decreases down to ~ 8%
of the initial value. After 120 years of storage the whole amount of concrete will have activity
not exceeding the limiting RSS values. The average specific activities of nuclides in concrete
are given in Table 29. In this Table the total activities depending on time of storage after
irradiation are also presented. The concrete activity value after several years of storage is
determined by Eu152, Eu154, Co60, H3, Ni63, C14 and Ca41 isotopes.

TABLE 29. AVERAGE SPECIFIC ACTIVITIES OF THE NUCLIDES IN CONCRETE

(BQ/KG) AND TOTAL ACTIVITY OF THE CONCRETE (BQ)
Nuclide Storage duration, years
composition 0 10 50 100 150 10 000
60
Co 3.55E+4 9.53E+3 4.92E+1 - - -
55
Fe 4.04E+5 3.09E+4 - - - -
3
H 2.11E+5 9.91E+4 1.27E+4 7.64E+2 4.61E+1 -
152
Eu 3.52E+4 2.29E+4 2.86E+3 2.12E+2 1.57E+1 -
154
Eu 3.55E+3 1.75E+3 6.87E+1 - - -
41
Ca 1.69E+3 1.69E+3 1.69E+3 1.69E+3 1.69E+3 1.58E+3
63
Ni 2.87E+2 2.68E+2 2.03E+2 1.43E+2 1.02E+2 -
14
C 3.11E+2 3.10E+2 3.09E+2 3.06E+2 3.05E+2 9.30E+1
134
Cs 1.79E+4 1.30E+2 - - - -
59
Fe 1.25E+4 - - - - -
136m
Ba 1.00E+4 - - - - -
139
Ba 7.93E+3 - - - - -
152m
Eu 2.49E+4 - - - - -
60m
Co 2.00E+4 - - - - -
Total, Bq/kg 7.90E+5 1.65E+5 1.78E+4 3.12E+3 2.20E+3 1.69E+3
Total, Bq 6.93E+1 1.45E+11 1.56E+10 2.74E+9 1.93E+9 1.48E+9
5.6.2.3.7. Activity of the graphite
Graphite placed in the plugs, has activity below the limiting RSS value. About 50% of
graphite located in the ionization chambers unit (ICU), also has activity below RSS limit.
After ~ 150 years of storage all graphite will have activity below RSS limit values. Average
specific activities of the nuclides in graphite of ICU are presented in Table 30.

TABLE 30. AVERAGE SPECIFIC ACTIVITIES OF THE NUCLIDES IN GRAPHITE

OF THE ICU (BQ/KG) AND TOTAL ACTIVITY OF GRAPHITE (BQ)
Nuclide Storage duration, years
composition 0 10 50 100 150 10 000
60
Co 1.75E+5 3.01E+4 1.56E+3 2.2E-1 - -
152
Eu 2.88E+5 1.71E+5 2.13E+4 1.58E+3 1.18E+2 -
154
Eu 4.18E+4 1.85E+4 7.39E+2 - - -
94
Nb 4.38E+2 4.38E+2 4.38E+2 4.38E+2 4.38E+2 3.12E+2
63
Ni 6.85E+4 6.37E+4 4.84E+4 3.42E+4 2.42E+4 9.30E+1
14
C 1.05E+4 1.05E+4 1.04E+4 1.03E+4 1.03E+4 3.13E+3
59
Ni 6.27E+2 6.27E+2 6.27E+2 6.27E+2 6.27E+2 5.71E+2
41
Ca 6.32E+2 6.32E+2 6.32E+2 6.32E+2 6.32E+2 5.91E+2
134
Cs 3.67E+6 1.26E+5 - - - -
55
Fe 1.04E+5 7.78E+3 - - - -
Total, Bq/kg 4.36E+6 4.29E+5 8.41E+4 4.78E+4 3.63E+4 4.61E+3
Total, Bq 8.99E+10 8.85E+9 1.73E+9 9.85E+8 7.49E+8 9.49E+7

187
The total activities depending on time of storage after irradiation are also presented in this
table. Graphite activity value after several years of storage is determined by 152Eu, 154Eu, 60Co,
94 63 59 14 41
Nb, Ni, Ni, C and Ca isotopes. Total activities of the nuclides for all reactor materials
are presented in the Table 31.

TABLE 31. TOTAL ACTIVITY OF THE NUCLIDES OF THE REACTOR

MATERIALS (Bq)

Nuclide Storage duration, years

composition 0 10 50 100 150 10 000
60
Co 2.57E+16 6.89E+15 3.57E+13 4.96E+10 1.32E+8 -
63
Ni 5.04E+15 4.68E+15 3.55E+15 2.50E+15 1.78E+15 -
59
Ni 6.26E+13 6.26E+13 6.26E+13 6.26E+13 6.26E+13 5.27E+13
94
Nb 5.90E+11 5.90E+11 5.90E+11 5.90E+11 5.90E+11 4.18E+11
152
Eu 2.62E+12 1.58E+12 1.97E+11 1.46E+9 1.08E+8 -
154
Eu 3.23E+12 1.44E+12 5.75E+10 1.02E+9 1.81E+7 -
55
Fe 1.69E+17 1.30E+16 5.28E+11 - - -
56
Mn 4.21E+17 - - - - -
58
Co 7.96E+16 - - - - -
58m
Co 3.94E+16 - - - - -
51
Cr 8.80E+16 - - - - -
59
Fe 3.12E+16 - - - - -
60m
Co 1.49E+16 - - - - -
55
Cr 3.31E+15 - - - - -
52
V 2.72E+15 - - - - -
65
Ni 2.32E+15 - - - - -
95
Nb 2.75E+13 - - - - -
64
Cu 1.01E+13 - - - - -
54
Mn 8.70E+12 - - - - -
Total, Bq 8.82E+17 2.47E+16 3.65E+15 2.56E+15 1.84E+15 5.77E+13

5.6.2.3.8. Total activity of reactor materials and total volumes of radioactive waste

Table 32 gives total volumes and weights of the radioactive waste, formed as a result of
neutron irradiation of reactor materials, having induced activity above RSS limit values. The
main contribution to the amount of radioactive waste is made by the iron-ore filling and
reactor structures made of stainless steel and 3 grade carbon steel. During 100 years of storage
the amount of the radioactive waste having induced activity higher than the RSS limit value,
will be decreased from ~ 860 m3 (~ 2.900 t) down to ~ 150 m3 (~ 680 t). After storage during
120-150 years the whole mass of concrete and graphite will have activity, at which these
radioactive materials are free from the RSS regulation. The radiological conditions in the
reactor vessel cavity will be determined by the products of activation of Co, Eu and Nb,
contained as the impurities in structural and shielding materials of the reactor.

188
TABLE 32. VOLUMES AND WEIGHTS OF THE REACTOR MATERIALS, HAVING
INDUCED ACTIVITY EXCEEDING RSS LIMITS

Storage duration, years

0 50 100 150 10 000
Stainless m3 42.5 35.6 25.6 25.6 18.8
Steel Cr18Ni9 tons 331.5 277.7 199.7 199.7 146.6
3-grade steel m3 44.2 44.2 30.1 30.1 30.1
tons 344.8 344.8 234.8 234.8 234.8
3
Iron-ore m 376.3 240.3 56.0 56.0 42.0
Filling tons 1303.0 720.9 168.0 168.0 126.0
3
Graphite m 12.9 6.0 3.0 0.0 0.0
tons 20.6 9.6 4.8 0.0 0.0
Concrete m3 381.0 114.3 30.0 0.0 0.0
tons 876.3 263.0 69.0 0.0 0.0
Total for the m3 856.9 440.4 144.7 111.7 90.9
whole reactor tons 2876.2 1616.0 676.3 602.5 507.4

Maximum dose rates of gamma-radiation on the surface of activated structures of the reactor
are presented in Table 33.

TABLE 33. DOSE RATES OF GAMMA-RADIATION ON THE SURFACE OF

ACTIVATED STRUCTURES, SV/S

Elements of reactor Time after reactor shutdown, years

installation
5 10 50 100 150
Elements in Core area 1.5105 7.5104 4103 1.0 0.5
the core Reactor vessel 7.5103 3.7103 1.7101 0.065 0.04
midplane
Iron ore shielding 4103 2103 1.0 0.01 510-3
Concrete 0.08 0.033 510-4 210-5 -
-3
Elements of Reactor vessel 4102 2102 1.0 3.510-3 210
lower reactor Iron ore shielding 4102 2102 0.16 710-4 110-4
structures
Concrete 0.8 0.33 510-3 210-4 10-5
Elements of Bottom part of the 1.5103 7.5102 3.7 7.510-3 2.510-3
upper floating shielding
reactor 610-6
Top part of the 1.5 0.75 410-3 10-5
structures
floating shielding
Rotating plugs shielding 0.25 0.12 610-4 10-6 710-7
Graphite of ICU of the control rods 6.5 2.3 0.08 810-3 2.510-3

189
It was assumed that dose rate of gamma-radiation was formed only due to activation of the
material of the above indicated structure. The possible contribution to the dose rate of gamma-
radiation made by the adjacent structures is not shown. When estimating the radiological
conditions at the reactor top structures, it is necessary to take into account the increased
amount of radioactive products deposits on the reactor structural elements having been in
contact with the primary cover gas. The total values of activity of the materials after several
years of storage are determine by 60Co, 63Ni, 59Ni, 152Eu, 154Eu, 94Nb and 55Feisotopes.

Dose rate of gamma-radiation on the surface of structural elements only due to this
contamination can reach 2 Sv/s value, which is comparable to the effect of induced
activity of structural materials located above the floating shielding. The analysis of data
presented in Tables 29-33 shows that in case of short storage duration it is impossible to carry
out dismantling of the reactor installation without application of robotics. The complete
dismantling of the reactor installation without application of robotics can be carried out only
after 100 years of storage.

5.6.2.4. Protection techniques

It follows from the analysis of the radiological characteristics of the equipment and
technological media of the shutdown reactor, that in order to assure radiological safety of the
personnel during the installation decommissioning it is necessary to meet the following two
requirements:

Maximum use of the operating systems of the reactor installation;

Development of additional systems, allowing safe implementation of dismantling
work.

The ratio between these directions depends on the chosen scenario of the reactor installation
decommissioning. Below are given some specific recommendations, directed to assurance of
the radiological safety of the BN-350 reactor decommissioning.

5.6.2.4.1. Use of operating systems of the reactor

On the stage of the BN-350 reactor decommissioning, the following systems required for the
radiological safety assurance, should be used to the maximum extent:

System of the radiological control;

System of fuel element cladding failure detection and location;
System of special ventilation;
Systems of sodium fire fighting;
System of special drains;
Systems of fuel transport;
Systems of reprocessing and disposal of the liquid and solid radioactive waste;
Hot chambers for works with high activity products;
Protection barriers and sanitary barriers;
System of monitoring of the radiological situation on the NPP site.

The majority of these systems should keep their serviceability, practically at all stages of
decommissioning. After removal of spent fuel and liquid metal coolant of the primary and
secondary circuits and spent fuel storage cooling circuit from the BN-350 NPP, some
systems, assuring radiological safety or their elements, such as the system of fuel element

190
cladding failure detection, the monitoring system of inert radioactive gases and the system of
sodium fire fighting can be dismantled.

5.6.2.4.2. Additional systems for assurance of radiological safety for putting in prolonged
storage and dismantling of the reactor components

In order to assure radiological safety on the stage of putting in prolonged storage and
dismantling of the reactor installation, development of new systems or modification of some
existing systems is required:

Creation of the additional sanitary barriers to prevent spreading of the radioactive

contamination (for example, sanitary lock in the area of cells of the primary circuit
loops, reactor dome, etc.;
Replacement of closed ventilation systems of the primary circuit cells and reactor
vessel cavity by a special ventilation systems having aerosol filters, gas being released
to the atmosphere through the stack (it is necessary to exclude the ingress to the
attended rooms of ozone and nitrogen oxides, caused by gamma-radiation of the high
activity structural elements of the reactor). On the stage of putting reactor cavity
structures in prolonged storage it is expedient to have ventilation system based on the
natural draft provided by the stack;
Creation of local systems of collecting, sorting, reprocessing, storage (putting in
prolonged storage) and disposal of the solid radioactive waste, including high activity
items (control rods and their sleeves, photo neutron sources, cesium absorbers, etc.);
Development of the remote tools, robotics and devices for realization of dismantling
works, including those carried out under water;
Creation of an additional network of stationary instruments, intended for the control of
superficial contamination of personal protection means, equipment, etc. by the
radioactive products;
Organization of system for analysis of radionuclide composition of material samples
from the dismantled equipment (including and control). All these activities should
be taken into account in the project of installation decommissioning and, as a part of
it, in the report on the nuclear and radiological safety on the stage of the BN-350
reactor decommissioning.

REFERENCES

[1] Safety Requirements For transportation of Radioactive Materials, PB RV-73,

Moscow, Russian Federation (1973).
[2] Guidebook on Extended Sources Radiation Protection, Atomizdat, Moscow, Russian
Federation (1965).
[3] Guidebook on Ionizing Radiation Protection, Atomizdat, Moscow, Russian Federation
(1966).
[4] Radiation Safety Standards (NRB-96), Goskomsanepidnadzor of Russian Federation,
Moscow, Russian Federation (1996).
[5] LOKSHIN, E.P., On Production of High Purity Materials by Vacuum Gauge Methods,
Journal of Applied Chemistry, 69, issue 4 (1996) 555561.
[6] MASSE, F., et al., Cleaning and Decontamination: Experimental Feedback from
PHENIX, paper presented in Technical Committee Meeting on Sodium removal and
disposal from LMFRs in normal operation and in the framework if decommissioning,
37 November 1997, Aix-en-Provence, France.

191
[7] LASTOV, A.I., et al., Study on Method of Cesium and Tritium Removal from Sodium
for its Utilization or Recycling, paper presented in Technical Committee Meeting on
Sodium removal and disposal from LMFRs in normal operation and in the framework
if decommissioning, 37 November 1997, Aix-en-Provence, France.
[8] KOZLOV, F.A., et al., Sodium Coolant Purification; Experience and Problems, paper
presented in Technical Committee Meeting on Sodium removal and disposal from
LMFRs in normal operation and in the framework if decommissioning, 37 November
1997, Aix-en-Provence, France.
[9] POLYAKOV, V.I., TCHETCHYOTKIN, YU.V., Efficiency of BOR-600 Reactor
Sodium Coolant Purification From Radioactive Oxides by the Cold Trap, Atomnaya
Energiya, 38 3 (1975).
[10] DE MAGNY, E., BERTE, M., Fast reactors bulk sodium coolant NOAH process
application, paper presented in Technical Committee Meeting on Sodium removal and
disposal from LMFRs in normal operation and in the framework of decommissioning,
37 November 1997, Aix-en-Provence, France.
[11] BROWSER, R., et al., The Design, Construction, Start-up and Operation of Facility at
Dounreay for Sodium Removal from KNK-2, paper presented in Technical Committee
Meeting on Sodium removal and disposal from LMFRs in normal operation and in the
framework of decommissioning, 37 November 1997, Aix-en-Provence, France.
[12] Technical Committee Meeting on Sodium removal and disposal from LMFRs in
normal operation and in the framework of decommissioning, Aix-en-Provence, France,
37 November 1997.
[13] SCOMOROKHOVA, S.N., KOPYLOV, V.S., TRIFANOVA, E.I., Experimental
Studies on Sodium-Water Interaction Under Vacuum, in Report of SSC RF IPPE,
Obninsk, Russian Federation (1995).
[14] Reprocessing of Sodium Coolant Radwaste from Fast Reactors and Immobilization of
Reprocessing Products, paper presented in 2nd Symposium on Radioecology, Obninsk,
1996, published by SSC RF IPPE, Obninsk, Russian Federation (1997).
[15] KONOVALOV, E.E., STARKOV, O.V., MYSHKOVSKY, O.P., Self-propagating
High Temperature Synthesis Of Stable High Radioactive Waste, paper presented in 2nd
Symposium on Radioecology, Obninsk, 1996, published by SSC RF IPPE, Obninsk,
Russian Federation (1997).
[16] SCOMOROKHOVA, S.N., KONOVALOV, E.E., KOPYLOV, V.S., STARKOV,
O.V., KOCHETKOVA, E.A., TRIFANOVA, E.I., Development of Environmentally
Safe Technology of Alkali Metal Radwaste Processing, paper presented in 2nd Russian
Conference on Radiochemistry, Dimitrovgrad, 1997, published by SSC RF NIAR,
Dimitrovgrad (1988).
[17] NALIMOV, YU.P., KOZLOV, F.A., et al, Experience Gained in Operation and
Regeneration of Secondary Cold Traps of the BN-350 reactor, paper presented in
Soviet-French Seminar, 1988, Aix-en-Provence, France.
[18] KOZLOV, F.A., NALIMOV, YU.P., et al., Regeneration of Cold Traps Used in the
Liquid Metal Coolant Systems, Journal Atomnaya Energiya, 44, issue 2 (1978) 104.
[19] KOZLOV, F.A, NALIMOV, YU.P., NIKULIN, M.P., et al., Experience Gained in
Cleaning Components and Piping of Fast Reactor Circuits, paper presented in
UK-USSR Bilateral Seminar, 1991, Shevchenko, Russian Federation.
[20] SCOMOROKHOVA, S.N., KOCHETKOVA, E.., NALIMOV, YU.P., et al,
Washing Components from Alkali Metals Using Butylcellosolve based Compositions,
Journal VANT, Issue: Nuclear Engineering and Technology, No. 6 (1991) 64.

192
[21] KLUKOV, B.P., LEDNEV, A.L., Cleaning of the Equipment from Residual Sodium
by Means of Water-Vacuum Technology, paper presented in Technical Committee
Meeting on Sodium Removal and Disposal from LMFRs in normal operation and in
the framework of decommissioning, 37 November 1997, Aix-en-Provence, France.
[22] FORREST, R.A., ENCLACOTT, D.A., KHURSHEED, J.A., FISPACT Program
Manual, Nuclear Physics Division, Harwell Laboratory, AERE-M3634, Harwell
(1988).
[23] ALCOUFFE, R.E., et al., Users Manual for TWODANT: A Code Package for
Two-Dimensional, Diffusion-Accelerated, Nuclear-Particle Transport, La-10049-M,
Los Alamos National Laboratory (1984).
[24] MANTUROV, G.N., NIKOLAYEV, M.N., TSIBULYA, A.M., Group Constants
System ABBN-93, Verification Report, oscow, Russian Federation TsNIIAI (1995).
[25] Radiation Safety Standards (NRB-96), Hygienic Standards, GN 2.6.1.054-96,
Goskomsanepidnadzor of Russian Federation, Moscow, Russian Federation (1996).
[26] GAVRILOV, S.P., et al., Effect of Long Living Activation Products Of Concrete
Structures On Decommissioning of LWR NPPs, Journal Atomnaya Tekhnika Za
Rubezhom, No. 2 (1990).
[27] Serpentinite Used In Nuclear Reactors Shielding (Ed. Yu.. Egorov, Moscow,
Atomizdat) (1973).

193
6. LEAD-BISMUTH COOLED (LBC) SHIP REACTOR

6.1. INTRODUCTION

The advantages of lead-bismuth and lead reactor cooling are high boiling temperatures and
the relative inertness to water as compared with sodium. The melting and boiling points of
sodium are 98 and 883C, respectively. For lead-bismuth eutectic, these values are 123.5 and
1670C, respectively, and for lead 327 and 1740C at atmospheric pressure. In a lead-bismuth
(Pb-Bi) or lead cooled reactor, the coolant boiling point may increase up to about 2 300C
because of high coolant pressure inside the core. However, the boiling points are well above
cladding failure temperatures. The specific heat per unit volume of lead-bismuth and lead are
similar to that of sodium but the conductivities are lower about a factor of four [1]. Studies of
lead-bismuth cooled (LBC) fast reactors have been carried out in the Russian Federation
organizations SSC RF IPPE (Institute of Physics and Power Engineering and EDO
GIDROPRESS in which a great deal of experience has been accumulated in the course of the
development and operation of submarine reactors cooled with lead-bismuth eutectic [1, 2].
The key results of operating experience of the propulsion nuclear steam supply system
(NSSS) using lead-bismuth coolant, R&D on LBC reactor technology, as well as
SVBR-75/100 reactor design features and parameters are discussed in this section.

6.2. OPERATING EXPERIENCE WITH LBC SHIP REACTOR PLANTS

In the course of operation of NSSS using LBC, accidents have occurred on three units, and
that was the cause of impossibility of their further operation:

The accident in 1968: the core was plugged by the oxides and products of steam/water
interaction causing core partial meltdown [3];
The accident in 1971: the damage of the primary circuit pipelines due to corrosion at
their outer side and radioactive coolant leakage [4];
The accident in 1982: the global corrosion damage of steam generator (SG) tube
bundle caused by poor quality of feedwater; ~ 150 L of radioactive coolant leaked into
the compartment due to personnel errors [5].

6.2.1. The reactor core meltdown

In the early stages of development, the formation of deposits of heavy metal oxide and other
impurities posed problems. An uncontrollable accumulation of significant masses of oxides in
the primary circuit could have formed when, during maintenance and repair operation, the
pipelines of the primary circuit gas system were depressurized, and thus air penetrated into the
primary circuit. Besides that, the primary circuit was contaminated by products of oil
pyrolysis, which was used as lubricant for the pump shaft seal designed to prevent gas leaks
from the primary circuit. The lubricant entered the primary circuit because of faulty seals.
When the rate of SG leakage increased suddenly (it had started some time before the
accident), the oxides accumulated and other impurities filled the core, which was the cause of
the violent deterioration of heat removal. A negative temperature reactivity effect was the
cause of transfer of the automatic power control rod to the upper switch terminal and
spontaneous power reduction to 7% of rated value. This was the first symptom of the accident
happened in May 1968.

However, operational documentation did not include any necessary instructions for the
operator how to act when that kind of situation arose. Instead of resetting the emergency
protection (EP) at the left side reactor, the operator followed the commander's directions (it

194
occurred in the course of navy training) and tried to maintain the given power level by
continuous removal of shim control rods (SHRs) out of the core. All reactivity reserve for
12 SHRs was released in about 30 minutes, though it was intended to provide for the power
reserve generation for about 4000 efficient hours. When SHRs were pulled out, the fuel in the
core area, where heat removal was deteriorated, melted and left the core together with the
coolant flow. Signals of radiation hazard in the compartment that called for shutting down the
reactor and removing the crew out being distantly removed from the reactor were not taken
into account. After this accident, the work on the coolant technology problem has been
launched. For many years, this work has been carried out at the number of organizations
under the scientific supervision of SSC RF IPPE. As a result, the problem has been solved
successfully and it was corroborated by the many-years experience of further NSSS operation.
The following main technical measures have been developed for eliminating the causes of
such accidents:

In order to eliminate the accumulation of oxides, excess inert gas pressure was
maintained in the primary gas system in case of equipment repair and reactor
refueling.
In order to eliminate the possibility of air penetration into the primary circuit and the
radioactivity release to the environment, the most possible tightness of the primary
circuit has been provided. For this purpose, special repair and refueling equipment
have been developed;
The sensors of thermodynamic oxygen activity enabling control of the content of
oxygen dissolved in LBC and detection of alloy oxidation at the very early stages have
been designed and introduced;
Rejecting the use of the oil seals of the pump shafts and adoption of water seals or
gas-tight electric drives of the primary pumps. This eliminates oil penetration into the
primary circuit and contamination of LBC by the products of oil pyrolysis;
Using the ejection system for high-temperature hydrogen regeneration built-in into the
NSSS in order to ensure chemical recovery of lead oxides by hydrogen (the explosion
proof compound of helium and hydrogen is used) and enable, if necessary, purification
of even strongly contaminated circuit from lead oxides;
Sing continuously operated system of coolant purification from irreducible impurities
on the glass fabric filters;
Using automatic system of coolant quality control equipped with sensors of
continuous control of coolant and cover gas quality ensures preservation of oxide
films on the surface of the primary circuit structural materials contacting with the
coolant, eliminates their corrosion deterioration and ensures the early diagnostics of
abnormal conditions.

6.2.2. Radioactive coolant leaks caused by damage to the primary pipelines

Since the beginning of the NSSS tests in 1970 and its further operations in 1971 and 1972, its
operation has been accompanied by a higher content of moisture in the air of the tight
compartment (TC) where the NSSS was installed. The tests have shown that the causes of
moisture content increase were poor air-tightness of the seal of one SG cover because of the
flaw in the nickel gasket (which was later replaced) and steam leak through the steam heating
system welds, which have been made unsoundly, and there was no possibility to eliminate this
leakage because of tight assembly.

As a result of cold surface sweating inside TC, water drops were the cause of wetting the
thermal insulation and dry protection materials containing chlorides. The drops of water

195
saturated with chlorides touched the primary circuit hot auxiliary pipelines made of austenitic
steel and gave rise to their corrosion cracking on the outer surface that has been fully verified
by the results of the NSSS inspection performed. Corrosion damages of the primary circuit
auxiliary pipelines at two of three loops and impossibility of their repair because of compact
assembly caused the decision of removing this unit out of the Navy and carrying out the
NSSS inspection. Below are described engineering measures eliminating the causes of such
accidents.

In the designed advanced NSSS, pool type primary circuit arrangement has been used. This
fully eliminates any primary circuit pipelines extending from unit vessel including
comparatively thin-wall auxiliary pipelines of small diameter, and no valves. Therefore,
ramified steam heating system was eliminated. Manufacturing the NSSS monoblock unit
under the plant conditions ensures high quality and delivery of the reactor unit available for
operation. The integral (pool) arrangement almost completely eliminated the possibility of
coolant leakage.

6.2.3. Radioactive coolant leaks owing to corrosion damage of the steam generator

Global corrosion damage of the SG evaporator sections tubes made of perlitic steel occurred
as a result of disregard of requirements to water-chemical regime (WCR) for SG feed water. It
was the result of that under real operating conditions, the way of reducing oxygen content in
the feed water by electron-ion-exchanging filter with copper-containing charge provided in
the design, caused copper release to the secondary circuit that was the cause of severe
electric-chemical corrosion of the piping system of the SG evaporator sections.

As a result of through piping damage, steam from the secondary circuit began to penetrate
into the primary circuit, where after separation from coolant it condensed in the emergency
condenser (EC) specially provided in gas system for the case of leakage in the SG. As EC
internal cavity had been filled with the condensate step by step, a signal was produced to the
operator who repeatedly drained the EC by removing the condensate accumulated to the
appropriate reservoir, and thus he eliminated the essential pressure increase in the primary
circuit gas system. However, the EC drainage was stopped because of unclear reasons. Heat
transfer surface of the EC was completely flooded by water and the condensation of
penetrating steam stopped.

Pressure increase began in the primary gas system. The strength of the gas system and
primary circuit could bear full working pressure of the secondary circuit. That is why in that
case there could not be any loss of integrity of the primary circuit.

Nevertheless, loss of integrity occurred, and it was caused by the events described below. The
gas pocket in the leakage re-injection pump (LRP) located in the pump tank below the
LBC level had adjusting manometer with ultimate pressure of 4 kg/cm2. According to the
instruction, if the RI was in operation, this manometer had to be shut off by the valve. The
instruction requirement was violated and the valve occurred to be open. Due to this fact, when
steam pressure in the gas pocket of the LRP tank reached ~ 6 kg/cm2 value and the LBC level
in the internal pocket of the LRP increased with the increase of gas pressure, the sensitive
manometer element was destroyed, gas escaped from the pump pocket, that was the cause of
filling up gas pocket of the LRP with lead-bismuth alloy and its further leakage through the
damaged manometer into the inhabited section of the reactor compartment (Fig. 1).

196
 P

Emergency
condenser

H
Excessive
gas
pressure

Condensate
draining line
steam

Steam-coolant
mixture
in SG-leak
accident

Primary
pump

Pump tank

FIG. 1. Primary pump with auxiliary system: flow diagram.

Air contamination by 210Po aerosols reached 10 maximum permissible concentration (MPC).

Due to the following proper actions, the crew irradiation and radioactive contamination were
within the permissible limits. The analyses of crew bio-samples, which had been performed
by medical service, demonstrated that none of the crew had the content of 210Po more than
10% of maximum permissible value.

Reactor installation (RI) examination showed that it could be reconditioned. However,

another decision was accepted. It was decided to replace the whole reactor compartment of
this unit with a new one manufactured earlier. The motive for this decision was as follows: in
the course of this RI fabrication at the machine-building plant, there was faulty replacement of
the relevant piping made of high-nickel corrosion-resistant steel by the same size piping made
of common stainless steel. This error was found out after the NSSS was fabricated and it was
impossible to change the piping. Since service lifetime of stainless steel piping was restricted
by corrosion, a decision was made to limit service lifetime of the reactor unit by 25 000 hours
and fabricate a reserve NSSS unit in order to use it for the replacement of off-spec one in the
course of the submarine overhaul period. In 1982, service life of stainless steel piping had to
be expired, and that was the motive for replacing reactor the compartment. The analysis
performed showed that the cause of this accident discussed in the previous section did not
relate to the use of LBC. The following technical measures ensure the elimination of such
accidents at the new NSSS generation:

Elimination of use of copper-containing materials in the water-steam circuit;

Preferable use of corrosion-resistant steel for SG water-steam tubes instead of
low-alloyed perlitic steel;
Providing passive drainage of the emergency condenser (EC) when filling it up by the
condensate up to the given level;

197
 Uniting gas volume inside the pump electric motor with the total one above the free
coolant surface;
Providing great extent of control-fitness and repair-fitness of the NSSS.

6.3. R&D ON LBC REACTOR TECHNOLOGY

Among the key problems, the problem which have been solved in the course of design and
operation of this type installations was that of LBC technology, i.e., development of systems
and devices ensuring measurement and maintenance of the LBC quality required during its
long-time operation both under normal conditions of leak-proof circuit and in the case of
partial loss of integrity of the circuit during repair and reactor refueling. Functioning of those
systems and devices are necessary for eliminating structural material corrosion and circuit
slagging by the lead oxides [6]. It should be pointed out that in the early days of mastering
LBC, when the necessity of developing and implementing measures on the coolant
technology had not been realized, there were cases of reducing the coolant flow cross sections
down to the full blockage of coolant flow because of depositing lead oxides and other
impurities and all the resulted consequences.

The basic kind of corrosion damage, which is the most dangerous for structural materials both
in Pb-Bi and in Pb coolants, is local corrosion of materials appearing as the separate
corrosion-erosion centres (pittings). Local through corrosion damages of structural elements
may appear at temperatures over 550C after holding for some hundred hours under the
following conditions: unbalance of alloying elements and impurities in steel, poor quality of
metal, absence of coolant quality control and non-optimal coolant flow regimes. The typical
corrosion rate in such cases is estimated as 2.55 mm/year.

The principle solutions ensuring high corrosion resistance of structural materials in heavy
liquid metal coolant were found using oxygen dissolved in the coolant. It has been shown as a
result of long-term studies that this corrosion resistance essentially depends on concentration
of dissolved oxygen.

Upon reaching certain level of concentration of dissolved oxygen corrosion processes is

stopped due to protective oxide film formed on the steel surface. At high temperatures an
indispensable condition of corrosion inhibition is presence of silicon in steel as additional
alloying element. The silicon content in steels is varied within 1-3.5% range depending on
steel type. Oxide films formed on the steel surface prevent it from interaction with liquid lead.
Since breakdown of oxide films is possible during operation, precautions must be taken for
resuming and maintaining their thickness and density.

Thus, steel corrosion in molten lead can be significantly slowed down by the oxide film
formed on the steel surface. The main technological problem is maintaining such oxygen
content in the coolant which, on the one hand, would provide stability of oxide film (Fe3O4)
on the steel surfaces, but, on the other hand, would preclude generation of lead oxide (PbO) in
the coolant, that could result in the circuit slagging. There are some ranges of content of
oxygen dissolved in lead meeting these two conditions, for instance (~ 510-6-10-3 wt %)
range. Oxygen content in lead can be controlled either by injecting gaseous oxygen or by
dissolving solid PbO.

Required oxygen content in lead can be maintained in two ways: (a) bubbling of argon,
hydrogen and water vapour mixture or gaseous oxygen through molten lead; (b) lead oxide
filling through which molten lead is pumped. In order to change oxygen content and remove

198
surplus PbO, reactions with water vapour or hydrogen can be used. To determine oxygen
content in molten lead (similarly to Pb-Bi technology development) galvanic cell can be used.

The problem of hyperthermal corrosion resistance of structural materials was got over by
development of preliminary protective coatings for the working steel surfaces. In particular,
the most important structural units of circuit, e.g. fuel rod claddings and steam generator
tubes, are covered by these coatings at the final stage of their manufacture. Additional barriers
are also formed directly on the inner surfaces of liquid metal circuit under effect of the
coolant in the early stage of the reactor operation.

The best results of using technology of preliminary oxidation of circuit components were
achieved by application of media with low partial pressure of oxygen, namely, Pb-Bi-O,
H2O + H2 and CO2. These methods, first of all, make it possible to avoid critical kinetic stage
of preliminary passivation of uncoated surfaces of steel structures of the circuit. Moreover,
they prove to extend the range of permissible decrease of oxygen concentration in the coolant.
Therefore, the basic factors ensuring high corrosion-erosion resistance of structural materials
in heavy liquid metal coolant (Pb, Pb-Bi) are as follows:

Application of silicon alloyed steels;

Passivation by oxygen using special regime of coolant;
Using additional corrosion barriers such as oxide films formed on working surfaces of
circuit components under reactor start-up conditions.

In reaction of PbO reduction, water vapours are efficiently removed from the circuit. Small
amount of moisture acts as diluted oxidizer preventing from achieving reduction conditions
for oxide films on the steel surface. Parameters of all these processes have to be developed
with necessary control of hydrogen content in cover gas and oxygen activity in liquid lead.

In conclusion, it appears that corrosion resistance of structural materials have been ensured by
using special steel alloys, applying protective films to them in advance and maintaining
necessary concentration of corrosion inhibitor - dissolved oxygen - in LBC. The importance
of these measures has been corroborated by the fact that when the necessary coolant quality
was maintained, for several thousands of hours at 650C temperature, there was no corrosion
of the fuel element steel cladding. However, when dissolved oxygen concentration was
inadequate (under specially provided conditions), it took about 20 hours for through corrosion
damage of the 5 mm thick pipe under the same temperature [7, 8].

Melting point of LBC is about 124C. Maintaining liquid state of LBC in all NSSS operation
regimes is ensured by using SG with multiple circulation over the secondary circuit, besides
that, water temperature at the SG inlet is higher than LBC melting point. For initial heating-up
and maintaining the primary circuit under hot condition at a low power level in the core, a
system of steam heating or electrical heating may be used.

The substantiation of the possibility of multiple coolant freezing-unfreezing in the NSSS

was an important practical problem. Low shrinkage of LBC during solidification and rather
high plasticity with low strength in solid state facilitate the elimination of reactor plant
damage when alloy is transiting from liquid to solid form and its further cooling down to the
ambient temperature. A special order of the temperature-time heating regime has been
developed for safe reactor plant unfreezing. This problem is dealt with in paper [9].

199
The following characteristics have been obtained in the course of NSSS test and operation:

The specific feature of LBC is the formation of -active 210Po radionuclide with a
half-life of ~ 138 days when bismuth is irradiated with neutrons.
The major reason for its radiation danger is the formation of radioactive polonium
aerosols when hot LBC contacts with air. It could happen under conditions of
emergency integrity loss of the primary circuit and coolant spillage. In this case, as the
NSSS operating experience in the nuclear plant has shown, the yield of polonium
aerosols and air radioactivity (according to thermodynamics laws) reduces rapidly
with temperature decrease and spilled alloy solidifying. Rapid solidification of spilled
LBC restricts the area of radioactive contamination and simplifies its removal in the
form of solid radioactive waste.
Low polonium concentration in the coolant (about 10-8 at.%) and formation of
thermodynamically proof chemical polonium-lead compound additionally reduces the
polonium pressure by 1000 times.
Personnel individual and collective protection facilities, methods for equipment
decontamination and recording activity on the surfaces, as well as, performing repair
procedures (also in case of LBC penetrating into the secondary circuit as a result of
unit crew errors) have been developed.

There have been no cases of personnel overdose by this radionuclide above the permissible
limits. This positive practical result is in a good agreement with conclusions of foreign
experts who have investigated polonium hazard problem if LBC is used for nuclear reactor
cooling [10, 11]. Reference [12] has focused on the analysis of this important issue.

In summary, the following characteristics have been obtained in the course of NSSS tests and
operation: power and parameters of installation, core lifetime, reactivity margin, reactivity
coefficients, poisoning effects, temperature distribution, dynamic parameters, coolant
radioactivity, and dose rates of caused by neutron and -radiation behind the shield. They
were in sufficiently good agreement with calculation results.

Among the positive properties of the reactor plant using LBC, which have been discovered in
the course of operation, one can point out the following: the simplicity of control, high
maneuverability and short time of reaching the power regime out of subcritical reactor state,
the possibility of NSSS operation if there is small leakage in the SG pipe system, high
repair-fitness of the SG by plugging the depressurized pipes, the possibility of reactor plant
stable operation at any low power levels, the possibility of quick changing the circulation
regime of coolant with essential change of its flow rate, and almost complete generation of
designed power by cores under normal and acceptable conditions of tightness of fuel rod
claddings.

6.4. CONCLUSIONS

In the early stages of development, the formation of deposits of heavy metal oxide and other
impurities posed problems. A careful control of the purity of the coolant is required to avoid
the formation of such deposits. It was necessary to improve corrosion resistant steels and
pre-treat the surface of components, and also to use special inhibitors in the lead-bismuth
coolant.

Lead-bismuth alloy is inflammable in air or water. In principle, lead-bismuth cooled reactor

would not have to have an intermediate circuit separating the primary coolant and
water/steam. However, as it has been mentioned above and emphasised in Ref. [3], there have

200
been incidents with lead-bismuth cooled reactor. Some areas of the reactor core were plugged
by lead oxides and other impurities as well as by the products of water and lead-bismuth
interaction due to SG leaks, causing meltdown of the core. Therefore, the elimination of the
intermediate circuit needs additional R&D efforts.

It was found [13] that specific -activity of the typical lead - bismuth coolant is defined by
210m
Bi (half-life = 3.6106 years), generated in reaction 209Bi (n,) 210mBi. The long-lived
-activity of 208Bi (half-life = 3.65105 y) is produced in reaction 209Bi (n, 2n) 208Bi. Thus, the
residual activity of lead-bismuth coolant is expected to be as high as millions of years. As it is
pointed out in [13], that purification of lead-bismuth from the long-lived radionuclides would
be too expensive9.

Eight submarines with lead-bismuth reactors have been constructed in the former USSR [14].
In the course of operation of NSSS with lead-bismuth reactor, there were accidents on three
units that have resulted in premature closing down [4].

The production of -radioactive 210Po having 138 days half-life undergoes -decay, some
problems are caused by bismuth because of its migration from the coolant to the cover gas
and formation of aerosols. 210Po is volatile, so that the leakage from the cover gas poses some
hazard to the plant operators [1].

REFERENCES

[1] INTERNATIONAL ATOMIC ENERGY AGENCY, Comparative Assessment of

Thermophysical and Thermophydraulic Characteristics of Lead, Lead-Bismuth and
Sodium Coolants for Fast Reactors, IAEA-TECDOC-1289, IAEA, Vienna (2002).
[2] GROMOV, B.F., SUBBOTIN, V.I., TOSHINSKY, G.I., Application of Lead-
Bismuth Eutectic and Lead Melts as Nuclear Power Plant Coolant, Atomnaya
Energiya, Vol. 73, issue 1 (1992) 19 (in Russian).
[3] OSIPENKO, L., ZHILTSOV, L., MORMOUL, N., Nuclear submarine epopee.
Heroic deeds, failures, catastrophes, A/O Borges, Moscow, Russian Federation
(1994) 208-209 (in Russian).
[4] ZRODNIKOV, A.V., ET AL, "Use of Russian technology of ship reactors with lead-
bismuth coolant in nuclear power", Small Power and Heat Generation systems on the
basis of Propulsion and Innovative Reactor Technologies, IAEA-TECDOC-1172,
IAEA, Vienna (2000) 127156.
[5] GROMOV, B.F., GRIGORIEV, O.G., DEDOUL, A.V., TOSHINSKY, G.I.,
STEPANOV, V.S., NIKITIN, L.B., The Analysis of Experience of Operating
Reactor Installations Using Lead-Bismuth Coolant and Accidents Happened, paper
presented in Heavy Liquid Metal Conference (HLMC), 58 October 1998, Obninsk,
Russian Federation.

9
After expiring the reactor installation (RI) lifetime, the radioactive LBC can be many times recycled in the
new RIs after conditioning. In 1000 years of irradiation, slight residual long-lived radioactivity of LBC caused
by 208Bi and 210mBi radionuclides will be lower than natural radioactivity of the uranium ore (in terms of U3O8).
It will be only important at the final stage of nuclear power functioning. In this connection, lead-bismuth coolant
in the form of solid radioactive waste being disposed in the deep geological formations will not disturb the
natural radioactivity equilibrium. Low chemical activity of lead and bismuth rules out radioactivity release into
the biosphere. Therefore, the radio-ecological consequences of this disposal will be of no risk for the population
of the next generations. There is a similar problem for the LWRs as long-lived radionuclide 93Zr is forming in
the fuel elements' zirconium claddings and channels [16].

201
[6] GROMOV, B.F, ORLOV, YU.I., MARTYNOV, P.N., GOULEVSKY, V.A., The
Problem of Heavy Liquid Coolants Technology (Lead-Bismuth, Lead), paper
presented in Heavy Liquid Metal Conference (HLMC), 58 October 1998, Obninsk,
Russian Federation.
[7] YACHMENEV, G.S., ROUSANOV, A.E., GROMOV, B.F, BELOMYTSEV, V.S,
SKVORTSOV, N.S., DEMISHONKOV, A.P., The Problems of Structure Materials
Corrosion in Heavy Liquid Metal Coolant paper presented in Heavy Liquid Metal
Conference (HLMC), 58 October 1998, Obninsk, Russian Federation.
[8] ROUSSANOV, A., et al., Corrosion resistance of structure materials in lead coolant
with reference to reactor installation, paper presented in Int. Seminar Cost,
Competitive, Proliferation Resistant Inherently and Ecologically Safe Fast Reactor
and Fuel Cycle for Large Scale Power, Moscow, 29 May1 June 2000.
[9] PYLCHENKOV, E.H., The Problems of Maintaining the RIs Operation Ability
under the Regimes of Coolant (Lead-Bismuth Eutectic Alloy) Freezing-Unfreezing,
paper presented in Heavy Liquid Metal Conference (HLMC), 58 October 1998,
Obninsk, Russian Federation.
[10] TUPPER, R.B., MIMISHKIN, B., PETERS, F.E., KARDES, Z.L., Polonium
Hazards Associated with Lead Bismuth Used as a Reactor Coolant, paper presented
in Conf. Kyoto, Japan, 28 October1 November 1991.
[11] WIGGS, L.D., COX-DE VORE, C.A., VOELZ, G.L., Mortality Among a cohort of
Works Monitored for Po-210 Exposure: 1944-1972, Health Physics, Vol. 61, No. 1
(1991).
[12] PANCRATOV, D.V., YEFIMOV, E.I., BOLHOVITINOV, V.N., BOUGREEV,
M.I., KOURANOV, V.D., Polonium Problems for Nuclear Power Installations Using
Lead-Bismuth Coolant, paper presented in Heavy Liquid Metal Conference (HLMC),
58 October 1998, Obninsk, Russian Federation.
[13] OUSANOV, V.I., PANKRATOV, D.V. et al., Long-lived Radionuclides of Sodium,
Lead-bismuth and Lead Coolants at Fast Reactors. Atomnaya Energia, Vol. 87, No.
9 204210 (1999) (in Russian).
[14] ASSOCIATION BELLUNA REPORT, Radioactive Contamination Causes in
Murmansk and Arkhangelsk Regions (Books 8874, Youngstorget N-0028, Oslo,
Norway, Version 1) (1994).
[15] ZRODNIKOV, A.V., TOSHINSKY, G.I., DRAGUNOV YU.G., et. al.,
Lead-bismuth reactor technology conversion: from NS reactors to power reactors and
ways of increasing the investment attractiveness of nuclear power based on fast
reactors, paper presented in Int. Conf. Fifty Years of Nuclear Power The Next Fifty
Years, Moscow/Obninsk, Russian Federation, 27 June2 July 2004.
[16] TOSHINSKY, G.I., IPPE, ROSATOM, Private communication, e-mail of
2005.09.16.

202
7. LMFR CHARACTERISTICS AND THEIR DEMONSTRATION DURING THE
FINAL STAGE OF OPERATION

7.1. PASSIVE CHARACTERISTICS WITH REGARD TO REACTOR SHUTDOWN

The history of industrial development suggests that complex engineering problems are often
solved in simple ways if the subjects intrinsically inherent properties of self-regulation are
grasped and brought into play in a timely manner. If any permissible limits of the operating
parameters - having appeared at first glance to be critical from the standpoint of normal
conditions - are expanded, the difficulties are overcome by virtue of natural feedbacks. As a
rule, this has led to refined and technologically simple solutions.

In this context, the range of temperatures permitted under emergency conditions for the fast
reactor as an important issue. The experience which has been acquired through the operation
of experimental reactors and rigs demonstrates that the temperature of the coolant and
structural elements of a reactor, which are made of high-temperature stainless steels, can be
increased appreciably for a short time (a period of a few hours) without negatively affecting
either strength characteristics or serviceability. For a fast reactor, this is possible due to the
thermal properties of the sodium coolant: a high boiling margin (300C), the constancy of the
efficient cooling of fuel element cladding in the case of a dramatic coolant velocity decrease
in the core, and the low level of the corrosive and chemical activity that exists between
sodium, the structural material (stainless steel), and almost all fuel materials at temperatures
that appreciably exceed the operating temperature.

As calculations and experiments demonstrate, allowance for an appreciable short-term

temperature increase within a reactor is of decisive importance to the realization of maximum
values for radial and axial core expansion, the centrifugal bending of the subassemblies and
the elongation of the absorber rod transfer bars, and the corresponding movement of the
absorber rods into the core. The formation of negative reactivity in the core gives rise to the
development of self-regulation factors. When the design process is target-oriented, this in turn
results in reducing reactor power to the decay heat level, even in the case of failure of the
safety rods to operate and reactor heating-up as consequence of pump shutdown and ceasing
of feed water supply to the steam generators. In other words, allowance for possible peak
temperature conditions within a reactor and the concomitant appearance of self-regulation
effects can, in principle, make the reactor sub-critical without scram.

Increased temperatures of the coolant and the external structural elements, as exemplified
by the main and guard vessels, go a long way toward laying the foundation for heat removal
based on passive principles. In such situations, heat can be discharged into the atmosphere as
a result of natural air circulation in the space between the vessel's hot wall and the well in
which the reactor or steam generator is located, as well as into special sodium-air heat
exchangers that also operate on the principle of natural circulation.

The core of a fast reactor is characterized by negative power reactivity coefficients that result
from both axial core expansion and a negative Doppler coefficient. The latter is especially
noticeable in the presence of considerable burnup, when the fuel adheres to the cladding and
expands with it. An increase of core power and coolant temperature rise in the core brings a
reactivity effect into play that result from the S-shaped centrifugal bending of the
subassemblies in the core region. It is suggested that when the assembly is spaced in a height-
wise fashion at three points (an articulated joint where the subassembly joins the housing of
the core diagrid, a spacer plate at the upper edge of the core, and a spacer plate at the upper
end of the subassembly), its bending leads to a negative reactivity effect (Fig. 1).

203
1-core former; 2-duct load pads; 3-S-shaped configuration due to contact of the load pads;
4-articulated joint of subassembly foot with diagrid; 5-fuel; 6-core barrel; and 7-core support structure

FIG. 1. Sketch of the restraint and spacing of the fuel assemblies for enhancing the negative feedback
reactivity during elevation of the temperature.

As with other measures, reactor design optimization for the purpose of intensifying negative
reactivity feedback also affects the selection of method of the vessel support in a pool type
reactor. The conclusion is reached, that a vessel bottom support and the co-location of the
rotating plugs and the control rod drive column on the cover plate could be the optimum. In
this case, the effect of the elongation of the safety system absorber rod transfer bars is utilized
to the fullest extent possible so as to facilitate the easy insertion of these rods into the core.

The self-regulated reactor concept makes allowance for the possible introduction of positive
reactivity due to the erroneous withdrawal of the control and/or safety rods.

Self-actuating absorber rods are being developed for the purpose of preventing serious
residual effects in advanced LMFR reactor designs.

Such absorber rod designs make use of electromagnetic clamps and transfer bars that are
located in the coolant. The electromagnet's windings are powered by a heat-sensitive
transformer that is mounted on the rod guide tube above the core. A stream of hot sodium
from the core flows around the transformer. When the temperature of the transformer's core
reaches the threshold value (Curie point), it loses its magnetic properties, the connection
between the primary and secondary windings is broken, and the supply of electric power to
the electromagnet's winding is terminated, which causes the rod to detach from the transfer
bar and drop into the core. Alternatives involving hydraulically suspended safety rods are
being examined for the purpose of resolving similar problems with respect to pump failure.

Other engineering measures that are aimed at reducing the positive component of reactivity
feedback include the use of a neutron absorber that is dissolved in the primary circuit coolant.
This alternative is similar to the concept employed in thermal reactors where boric acid is

204
dissolved in water. In such designs, the rapid absorber elimination and reactor runaway
become physically impossible. Indium is being considered possible absorbing solute. The
melting point of indium is 156C, i.e. about 60C higher than that of sodium. The requisite
indium concentration in the sodium can be maintained over a wide range of temperatures via
the crystallization of sodium-indium alloy in the cold traps. At the suggested indium
concentration, the melting point of the sodium-indium alloy does not exceed 204C, as a
result of which the solidification of the coolant in the loop is prevented. The indium dissolved
in the sodium is gradually retained in the cold traps over the course of reactor operation,
which offsets the reactivity decrease associated with fuel burnup. The neutron capture
cross-section of indium is equal to 0.45 barns for the neutron spectrum of an LMFR reactor,
or 20% of the capture cross-section of 10B. For example, the indium content needed in the
sodium in order to offset 10 dollars excess reactivity in the 400 MW (e) reactor is 6.5 atomic
percent. The interaction occurring between a liquid sodium-indium alloy and austenitic steel
was investigated on a test bed represented by a closed loop made of the 304 steel. The
welding joints present in this loop made it possible to study the influence exerted by indium
compounds on the welding material as a function of microstructure. Tests performed for a
period of 2 000 hours at a steel temperature of 540C demonstrated the absence of any
noticeable structural material penetration by the indium. The welds and adjoining sections
were not excluded from the overall pattern. Indium-nickel-manganese compound crystals
were detected in the cold sections of the loop. These crystals precipitated onto the walls of the
piping. Based on this finding, the conclusion can be reached that the only type of interaction
engaged in by indium that has been precipitated from the coolant is the reaction occurring
between indium and nickel that might be present.

7.2. PASSIVE CHARACTERISTICS WITH REGARD TO EMERGENCY HEAT

REMOVAL

The process of advancing nuclear-power engineering is accompanied by the inevitable

complication of reactor systems and equipment. As a result, the structure and complement of
components of the nuclear power plant (NPP) are more complex than those of a fossil fuel
power plant. For example, the NPP equipped with a light-water reactor having an electrical
capacity of 1000 MW contains about 30 thousand valves (gate and other types) as compared
to 4 thousand valves in a fossil fuel power plant having identical capacity.

NPP development has resulted in two challenges:

(i) Heat removal systems that are unrelated to nuclear processes must be designed and
manufactured in compliance with strenuous engineering standards, and
(ii) The complexity and multifunctional nature of the systems involved have increased the
likelihood of failures, thereby a high degree of quality control is needed. This has
resulted in a concomitant cost increase. In principle, the above mentioned
complexicity is also likely to increase the possibility of human error and of automatic
system malfunctions.

Among the specialists of various countries, the race is on to create emergency heat removal
systems that are based entirely on passive features and do not require intervention on the part
of an operator or an automatic system. Being based on precise physical laws, such systems
would function continuously. However, their capability would be intensified if a menacing
deviation in limiting parameters were to happen; thus, the parameters of the reactor and any
associated equipment would be maintained within predetermined limits by virtue of
self-regulation effects.

205
The nuclear power plant operator must concentrate attention on observing the course of
normal processes and equipment performance, as well as on preventing minor malfunctions
from turning into major accidents. The provision of emergency shutdown cooling in fast
reactors based on the use of passive features alone without bringing the main heat removal
systems into play i.e. the elimination of secondary circuit and tertiary circuit systems from
this process does away with a number of the strenuous and costly requirements that would
otherwise be imposed on the design, manufacture, and operation of such systems.
As previously mentioned, this makes the reactor in particular and the power block as a whole
less expensive. The essence of passive features with respect to fast reactors consists of making
use of the natural circulation of the coolant and the dissipation of the heat into the
environment via thermal radiation, thus removing the heat from the core and the reactor
vessel.

It should be noted that the possibility of employing passive features in systems that are
essential to safety has already been given consideration in early reactor designs. However, full
use was not made of these features. As previously mentioned, special active safety systems
that are switched on by an operator or by an automatic system when the threat of an
emergency situation is impending have performed and continue to perform protective
functions. Special systems are also employed. In all the designs that have been developed up
to date, passive factors of are represented by:

An extended pump coast down time (due to the use of special flywheels), with the
transition being made to coolant natural circulation, and the incorporation of special
emergency shutdown-cooling heat exchangers into the vessel of pool type reactors.
These heat exchangers are connected to the air heat exchangers by means of external
sodium lines. In all three loops, heat removal is accomplished by means of natural
circulation owing to the height-wise separation of the emergency shutdown-cooling
heat exchangers and the air heat exchangers (Fig. 2).
The removal of heat from the guard vessel via thermal radiation and convection on
the part of the air circulating through the annular space between the vessel and the
reactor well lining (Fig. 3). In the presence of a sodium volume increase that results
from heating up in the upper chamber, the sodium overflows into the annular cavity
and streams downward along it, transmitting heat through the vessel wall, the gas
space, and the guard vessel's wall to the outside air (Fig. 4). Design measures provide
the requisite natural sodium circulation flow rate within the reactor. In order to
ensure natural air circulation, the annular well space is connected to a short
ventilation pipe. Power dissipation in this case does not exceed 0.4% of the reactor's
thermal capacity.

A heat removal system based on the external air cooling of the reactor's backup vessel is
examined. The fining of either the guard vessel or the well lining is being considered for the
purpose of intensifying heat removal by the air. It has been proposed that a copper with a
linear expansion coefficient identical to that of steel, a high degree of thermal conductivity,
and a high level of blackness (as a result of surface oxidation) should be used as the fin
material. These fins would be fashioned on the well lining. One potential problem might be
the fouling and clogging of the space between the fins by impurities present in the air.
A positive natural circulation propelling head is created by means of heating the air present in
the space between the guard vessel and the finned well lining.

206
1-air heat exchanger; 2-heat exchanger for emergency cooling; 3-core; 4-insulation;
5-finned shell; 6-reactor vessel

FIG. 2. The system for emergency cooling of sodium advanced fast reactor (SAFR)
based on passive features.

207
1-core; 2, 3-inlet, and outlet respectively; 4-guard vessel; 5-finned shell; 6-fins; 7-concrete; 8-vessel
FIG. 3. Passive heat rejection system from the vesse1 of the power reactor
innovative small module (PRISM) reactor.

FIG. 4. Primary sodium flow path and passive emergency heat rejection
from the vesse1 of PRISM reactor.

208
Under emergency conditions, when the temperature of the sodium and structural elements are
elevated, a rapid increase in the heat transferred into the air take place proportional to a
cubic of the wall temperature value. Calculations have shown that such a system is capable of
removing a thermal output of 1% of the rated value. Here, the coolant temperature in the
reactor does not exceed 650C. Thus, the system at hand is entirely passive and is self-
regulated. The efficiency of the subject system increases with an increase in the temperature
value in place in the reactor. This system is activated without the participation of the operator
and can cool the reactor for an extended period of time. Calculation results are presented
further on in this report which demonstrate that the safety of the power reactor innovative
small module (PRISM) reactor over the course of an accident involving the total failure of the
heat removal system without scram is ensured exclusively by an air-cooling system.
Of course, the possibility of employing such a simple emergency shutdown-cooling system -
one based on passive principles alone - is a result of the low capacity of the PRISM reactor;
i.e. of the existence of beneficial relationship between the surface of the vessel and the
thermal capacity of the reactor. This constitutes one of the principal advantages of modular
reactors. The aforementioned relationship is diminishes as reactor output increases.
In addition to an external heat removal system, the sodium advanced fast reactor (SAFR),
having an electrical capacity of 350 MW, also employs a heat removal system of the internal
type, the latter taking the form of emergency shutdown-cooling heat exchangers that are built
into the vessel and are connected to the external sodium-air exchangers. The thermal
processes taking place in the SAFR reactor in the presence of the emergency conditions under
discussion are depicted in Fig. 5.

1-boiling temp. of sodium; 2-reactor cooling due to the heat capacity of sodium and steel; 3-cooling by air only
through the guard vessel; 4-heat rejection by the emergency heat exchanger and by air cooling of the guard
vessel; 5-normal shutdown of the reactor.

FIG. 5. Temperature of the coolant in the upper chamber of SAFR reactor during an accident
situation with loss of heat sink without scram.

In this reactor, the external natural air circulation vessel cooling system removes thermal
outputs of 1.5 MW under normal conditions and 5 MW under emergency conditions. It should

209
be noted that the internal emergency shutdown-cooling system does not function when the
reactor is operating normally. The flow of air through the air heat exchanger is closed off by a
damper and the emergency shutdown cooling heat exchanger is filled with helium. The fact
that the aforementioned passive systems directly abut on the reactor is important, since the
equipment supporting the heat removal loops and the steam-water circuit is thus eliminated
from the makeup of the systems that are essential to safety. This reduces the volume of vital
and complex equipment and systems, curtails the complement of operating personnel,
shortens equipment manufacture time frames, and curbs expenditures for equipment repair. In
addition, it becomes possible to enhance safety while simultaneously simplifying and
reducing the cost of a nuclear steam supply system. The chief advantages of the utilization of
passive features in the reactor protection system design are as follows:
The assessment of the risk of serious accidents is simplified due to a decrease in both
the number of systems that ensure safety and the number of probable operator errors;
and
It becomes possible to experimentally verify reactor safety, which is important both to
specialists and to the public, since the latter doesn't often understand the complex
probabilistic criteria that characterize NPP safety at all.
7.3. DEMONSTRATION OF SAFETY CHARACTERISTICS WITH EXPERIMENTAL
LMFRs
Following the accidents at Three Mile Island and particularly at the Chernobyl NPP, those
countries which were engaged in the development of fast reactors began to focus a great deal
of attention on reactor protection systems that function in accordance with passive principles
without the participation of automatic systems or the operator. Experiments conducted in
direct relation to the safety of three fast reactors - the Rapsodie, the EBR-II, and the FFTF -
gave great impetus to the evolution of this research. During these experiments, the most
serious emergency situations were simulated, as exemplified by the failure of the regular heat
removal systems with simultaneous failure of the shutdown system. On the whole, the subject
experiments demonstrated the existence of a direct dependence between core power rate and
power removal. For example, when the absorber rods of the EBR-II reactor jammed, reactor
output was correspondingly reduced in proportion to the feed water flow rate decrease that
resulted from reactivity feedback effects. During the shutdown of the primary pumps
accompanied by the safety rods failure reactor output was reduced to the level of decay heat
that could be removed from the core by means of natural circulation. Thus, the marked self-
regulation properties of fast reactors were clearly confirmed by these experiments. Reactor
experiments aimed at simulating the most serious emergency situations were begun in
Europe. Such experiments were first authorized and set up in France.

7.3.1. Rapsodie reactor

7.3.1.1. Main features
The Rapsodie experimental sodium cooled reactor, the first French fast neutron reactor
located at the Cadarache Research Centre and was operated by the CEA. The construction
was started in 1962 within an association of CEA and EURATOM. The reactor went critical
on 28 January 1967, reaching 20 MW(th) power on 17 March 1967. The core and equipment
were modified in 1970 to increase the thermal power level to 40 MW (FORTISSIMO) with a
peak neutron flux of 3.21015 ncm-2s-1. In operation, the primary sodium entered the core at
400C and flowed out at a mean temperature of 550C. The operating parameters were similar
to those in large commercial size reactors.

210
 FIG. 6. Rapsodie primary and inert gas system.

211
The reactor core was cooled by two identical loops each comprising a primary sodium circuit
from which thermal power is transferred to a secondary sodium circuit through an
intermediate (sodium/sodium) heat exchanger (IHX) by means of a primary pump (Fig. 6).
The system lines are enclosed in concrete cells inside a double containment barrier (Fig. 7).

FIG. 7. Rapsodie reactor cross section.

212
The principal geometric specifications of the primary pipes system are the following:

Core to intermediate heat exchanger (IHX): D = 300/314 mm, 16 m long;

IHX to pump: D = 300/314 mm, 8.5 m long;
Pump to Y junction: D = 200/208 mm, 18 m long;
IHX vessel dimensions: D = 884 mm, 5.2 m high;
Pump vessel dimensions: D = 850 mm, 4.5 m high;
Expansion tank: 36 m2 surface area.
7.3.1.2. The installation
The installation consists of six main buildings, access to three of them is restricted. These are:

(1) The reactor building or secondary containment including the reactor vessel and its
upper closures as well as the two primary loops, each of which was equipped with a
mechanical pump and an intermediate heat exchanger. All these components were
enclosed in concrete cells to provide radiation shielding. The secondary, non-
radioactive sodium, is piped to a conventional building containing the components of
the two secondary loops including a sodium/air heat exchanger in each;
(2) The active building comprising interim storage facilities for both fresh and used fuel,
and various other facilities such as the washing cell for decontaminating components
polluted with primary sodium, and a dismantling hot cell used for conditioning used
irradiated equipment for long term storage as waste;
(3) The fuel assembly dismantling building comprising hot cells for non-destructive
examination of fuel pins, and the assembly of experimental sub-assemblies.
All the circuits and components are made of austenitic stainless steel, the main pipes and
vessels have a double wall. The reactor vessel is immediately surrounded by special high
density concrete containing rare earth oxides, called Sercoter. This was protected externally
by a steel liner which was considered to constitute the second barrier.
Rapsodie was designed, built and operated to obtain data on the physical behaviour of a fast
neutron reactor under static and dynamic conditions, to offer information of direct use for the
design of future LMFRs, and to supply a fast neutron flux for irradiation tests of fuels and
materials. Mixed oxides were used as reactor fuel. During its 15 years of operation, more than
30 000 fuel pins of the driver core were irradiated, of which about 10 000 reached a burnup
beyond 10%, and 300 irradiation experiments and more than 1000 tests were performed.
In 1971, the irradiations performed in the core revealed a phenomenon of irradiation swelling
in the stainless steel of the wrapper and the fuel cladding in the high neutron flax. The
Rapsodie results have been extrapolated in the Phnix reactor. The operating history of
Rapsodie has been reviewed in journals and conference papers, and have regularly been
published by the TWG-FR. The decision to stop running the reactor was taken after two
successive defects were detected in the primary system containment.
The first defect, which appeared in 1978, consisted of a sodium micro leak: radioactive
sodium aerosols were found in the double wall reactor vessel. Investigations did not find any
liquid sodium in the gap nor locate the defect. The reactor was subsequently operated at a
reduced power level (~ 0.6 PN), which was high enough for irradiation needs but did not cause
the leak to reappear. The second defect appeared in 1982 and consisted of a small leak from
the nitrogen blanket surrounding the primary system. Before the final shutdown of the reactor,
a series of end-of-life tests were conducted in April 1983. The LMFR Rapsodie was shut
down in April 1983. Pre-decommissioning operations were then conducted until 1986. They

213
consisted essentially in unloading the fuel and fertile assemblies, and in draining the sodium
from the primary and secondary circuits. Decommissioning operations started in 1987.
7.3.1.3. Experiments with the Rapsodie reactor
Two series of tests performed on the Rapsodie reactor, the purpose of which was to
investigate the serviceability of this reactor's core and of the reactor as a whole under extreme
conditions that were characterized by an exceedingly high temperature. The first series of
tests to be performed called for an experimental inquiry into the behavior of fuel elements
(FEs) during fuel melting. Over the course of these tests, the fuel pin linear power observed
on two test subassemblies reached 1000-1060 W/cm; i.e. two times greater than that normally
used in commercial reactors. Subsequent materialogical investigations of FEs with different
fuel-cladding gaps (air or helium) revealed that 54% of the fuel present was melted in the first
instance, while 29% was melted in the second instance (Figs 8 and 9).

FIG. 8. Rapsodie fuel subassembly and fuel pin.

214
 1-empty space; 2-molten fuel
FIG. 9. Fuel pin containing molten fuel of Rapsodie reactor.
Profilometric and metallographic research demonstrated the absence of deformation and of
integrity loss of the FE claddings. These experiments provided the careful monitoring of the
temperature status by the thermocouples that were incorporated into the cladding and the
fuel. The FEs were held at the aforementioned peak linear power for 10 minutes. The
second series of experiments simulated the most serious accident, which consisted of the
shutdown of the primary-circuit and secondary-circuit pumps, as well as the tertiary-circuit
fans, and the non-operation of the safety rods. Here, reactor output reached 21.2 MW (more
than 50% of the rated value), while the mean coolant temperatures at the reactor inlet and
outlet came to 402 and 507C, respectively. The principal characteristics of these tests are
depicted in Fig. 10.

FIG. 10. Rapsodie end-of-life test: characteristics during a LOFWS transient10.

10
In the related Rapsodie tests the pump stop time was about 400 s; the halving time appears to be
about 45 s.

215
Yet another important fact should be mentioned with respect to the results emanating from the
experiments conducted on the Rapsodie reactor. Since the Doppler effect of the Rapsodie core
is equal to zero, it was possible to define, with a high degree of precision, the influence
exerted by axial fuel expansion. Here, it was essential to understand the manner in which the
fuel expanded, independent of the cladding or in conjunction with the cladding. A comparison
of calculation results and experimental data demonstrated that the fuel residing in the core
shared a state of coalescence with the FE cladding and expanded with the cladding upon
heating-up. It is in such instances precisely that good agreement is reached between the
calculation results and the experimental data concerning the coolant temperature at the
subassembly outlet.

FIG. 11. EBR-II equipment and pipes.

216
 FIG. 12. EBR-II reactor cut off.

The primary sodium contained within this tank represents the primary cooling system for
removal of the heat from the reactor core. Liquid sodium, with a boiling point of
approximately 900C, has excellent thermal properties and is thus an optimum coolant. The
primary system contains about 330 m3 of sodium, and transfers heat to the secondary sodium
system (about 50 m3) through a sodium-to-sodium intermediate heat exchanger.

The secondary sodium was circulated in a closed loop through superheaters and steam
generators outside of the reactor containment (Fig. 13). The high pressure steam produced in
the steam generator drove a turbine generator to produce electricity. EBR-II fuel subassembly
and fuel element are shown in Fig. 14.

217
218
FIG. 13. EBR-II schematic heat transfer and flow diagram.
 FIG. 14. EBR-II fuel subassembly and fuel element.

7.3.2. EBR-II reactor

7.3.2.1. Design features
The Experimental Breeder Reactor-II (EBR-II) was a sodium cooled metal-fuelled liquid
metal research reactor located in the South-Eastern portion of the Idaho National Engineering
and Environmental Laboratory (INEEL). The EBR-II was a 62.5 MW(th) reactor that began
operation in July 1964, and when fully operational, was capable of producing ~ 19.5 MW(e).
Plant capacity factors typically exceeded 70% and approached 80%, even though the reactor
supported an extensive testing program while simultaneously producing electricity as a
complete LMR power plant. The following key characteristics common to LMRs made

219
reliable, safe operation feasible: low pressure sodium coolant, limited thermal stress, limited
corrosion of components, and simplicity of layout in both the primary and secondary sodium
systems.

The EBR-II reactor building was connected to the fuel conditioning facility, a large inert
atmosphere hot cell facility. The EBR-II reactor building, a cylindrical structure with a
hemispherical domed top, had a steel containment shell with an inner diameter of 24.4 m and
a height of 42.4 m. The bottom and sides were 2.54 cm thick steel plates and the dome was
1.27 cm thick, lined with a 10.2 cm concrete missile shield. The 1.70 m diameter reactor
vessel and its shield were immersed in a sodium pool within the 7.9 m diameter by 7.9 m
height primary tank (Figs 11 and 12). The reactor served as a test facility for fuels
development, materials irradiation, system and control theory tests, and hardware
development. The EBR-II core and blanket subassemblies were contained within the reactor
vessel prior to defuelling (Fig. 12).

7.3.2.2. History of operation

The EBR-II had been operated for 30 years. Given the scope of what has been developed and
demonstrated over those years, it is arguably the most successful test reactor operation ever.
Tests have been carried out on virtually every fast reactor fuel type, and the reactor itself has
been extensively characterized. The most dramatic safety tests were conducted on 3 April
1986, when it was demonstrated that an LMR with metallic fuel could safely accommodate
anticipated transients: loss of flow or loss of heat-sink without scram.
The EBR-II before closed was operated as the integral fast reactor (IFR) prototype,
demonstrating important innovations in safety, plant design, fuel design, and actinide recycle.
The ability to passively accommodate anticipated transients without scram has resulted in
significant benefits related to simplification of the reactor plant, primarily through less
reliance on emergency power and by virtue of not requiring the secondary sodium or steam
systems to be safety-grade. These advantages have been quantified in a probabilistic risk
assessment (PRA) conducted for the EBR-II that demonstrated considerable safety advantages
over other reactor concepts. The uranium-plutonium-zirconium alloy fuel is fundamental to
the superior safety and operating characteristics of the reactor. The results of assessments,
analyses, and tests, indicated that the reasonable expected technical lifetime estimate for
EBR-II was well beyond 30 years and possibly 50 years or more before approaching any
aging limits. The plant-life extension program was refocused to build on the original results of
the plant engineering and operational assessment.
In January of 1994, the Department of Energy mandated the termination of the Integral Fast
Reactor (IFR) Program, effective as of 1 October 1994. To comply with this decision,
Argonne National Laboratory-West (ANL-W) prepared a plan providing detailed
requirements to place the EBR-II in a radiologically and industrially safe condition, including
removal of all irradiated fuel assemblies from the reactor plant, and removal and stabilization
of the primary and secondary sodium used to transfer heat within the reactor plant.

The EBR-IIs long, successful operating history provided an important source of

informationon for the long-term reliability of LMRs. Major programs conducted in the EBR-
II included metal fuel irradiation testing and demonstration of the inherently safe response of
a metal-fuelled, pool-type LMR. The EBR-II also served as an important test bed for key
features of innovative LMR designs, such as flexible pipe joints and materials and
improvements in instrument and control systems. Other major tests conducted included those

220
to determine the efforts of running beyond cladding breach, and the response of oxide fuel to
operational transients in the US-Japanese joint program.

7.3.2.3. Experiments with the EBR-II reactor

A number of devices were fabricated in order to conduct a series of tests. Among them,
instrumented fuel subassemblies were designed, and one is shown in Fig. 15.

FIG. 15. Sketch of EBR-II instrumented fuel subassembly.

There are two flowmeters located in the lower shield below the active core region. The
spacer-wire thermocouples were contained in a metal sheath of 316 stainless steel and were
made of type K chromel/alumel with magnesium oxide insulation. There are a total of
28 thermocouples: two located in the flowmeters near the inlet (TC), five at the active core
midplane (MTC), thirteen near the top-of core (TTC), several above the core (TC), two at the
subassembly outlet (OTC) and two in the thimble region of the subassembly (ATC).

Several series of tests were performed on the EBR-II reactor for the purpose of simulating the
most serious of accidents: for example, the loss of the supply of electric power to the main
pumps and the safety rod failure at full reactor power, as well as the failure of the secondary
and tertiary circuits, together with the simultaneous failure of the safety rods.

221
The characteristics of this accident are depicted in Fig. 16.

1, 2-the maximum fuel cladding temperature and the coolant outlet temperature, respectively;
thermocouple reading; 3-power/flow; 4-reactivity; 5-sodium temperature rise; 6-power; 7-flow rate

FIG. 16. Characteristics of the EBR-II reactor during LOFWS transient.

222
Experiments simulating the loss of flow without scram (LOFWS) were conducted at full
power in the presence of different pump coast-down times: active rundown, by controlling the
pump speed, pump stop time 300 seconds (Fig. 16) and 100 seconds, passive coast-down,
accompanied by the shutdown of the auxiliary electromagnetic sodium pump (Fig. 17).

1, 2-the maximum fuel cladding temperature and coolant outlet temperature, respectively;
thermo-couple reading; 3-power/flow; 4-reactivity; 5-coolant temperature rise; 6-power; 7-flow rate

FIG. 17. Characteristics of the EBR-II reactor during LOFWS transient.

223
The EBR-II pump system has small inertia, leading to a fast coastdown. Therefore it was
decided to use the stored energy in both the pumps and the motor-generator set, and
controlling the coastdown with the magnetic clutch which couples the motor and generator.
A comparison of peak temperatures demonstrated the decisive influence exerted by the
rundown time of the primary pumps on fuel element cladding and reactor coolant temperature
values. Experiments relating to the loss of heat sink without scram (LOHSWS) were also
conducted at full power. The reactor outlet temperature was reduced, while the reactor inlet
temperature was increased.
Owing to the experimental confirmation of the fact that safety properties are intrinsically
inherent in fast reactors, design work has been undertaken, the purpose of which was to
ensure the creation of conditions which favour, to the greatest extent possible, the utilization
of inherent factors. It is generally known that an increase in the reactor inlet temperature and
the reactor coolant temperature rise present therein is accompanied by the thermal expansion
of the core diagrid, subassembly bending directed outward from the center, as well as by the
elongation of the control rod transfer bars, thereby resulting in the introduction of negative
reactivity. In order to reduce power, it is essential that the positive reactivity effects produced
by the coolant density decrease associated with an increase in both the coolant temperature
and the Doppler effect, should be overcome by the negative effects generated by heating-up
and the thermal expansion of the reactor inner structures, since the fuel temperature value is
also necessarily reduced.
As previously mentioned, the most impressive results with respect to self-regulation were
obtained on the EBR-II reactor when using metal fuel. During experiments involving pump
shutdown in circuits I, II, and III together with the safety rod failure, this reactor's output was
spontaneously reduced to the decay heat level and the coolant temperature value at the outlet
of the hottest core subassembly was increased from 520 to 720C (a boiling margin of
150C).
In summary, the EBR-II experiments, which were done in April 1986, simulated both the two
major heat imbalance accidents that happened (pumps failure which stopped coolant flow
through the reactor - essentially the Chernobyl accident, on the one side, and failure to
transfer heat from the reactor coolant to the steam system - essentially the TMI-2 accident11,
on the other side). In neither case was any operator or safety system action taken, and in
neither case was the reactor or its fuel harmed in any way.
7.3.3. The FFTF reactor

7.3.3.1. Design features

The concept of a Fast Flux Test Facility (FFTF) dates back to the late 1950s. In April 1965,
the USAEC authorized Battelle Pacific Northwest Laboratories to perform a conceptual
design and cost study of FFTF. The objective was to establish a unique and flexible reactor
plant capable of intense fast flux irradiation of fuels and structural materials in view of the US
LMFBR development programme. The components, fuels, materials, and core environment of
FFTF were to be as similar as practicable to those required for larger power stations. The
FFTF was a 400 MW(th) sodium-cooled fast reactor specifically designed for development
and testing of fast breeder reactor fuels, materials, and components (Fig. 18).

11
Not included in this Report.

224
 FIG. 18. FFTF reactor elevation.

The reactor was a loop-type plant with three parallel heat transport system loops. The plant
has neither steam generators nor blanket assemblies for fissile breeding, consistent with its
role as a test reactor. The outer three rows of core assemblies were stainless steel radial
reflector assemblies which serve to enhance the neutron flux in the core interior.

The FFTF was equipped with a great deal of instrumentation. Each core assembly was
provided with instruments for measurement of sodium flow rates and sodium outlet
temperature. Three instrument trees, one of which serves each of the three core sectors,
provide outlet instrumentation for all fuel assemblies, control and safety assemblies, and
selected reflector assemblies. In addition, 8 of the 73 core positions were equipped for full
in-core instrumentation. Two of these eight positions were available for closed-loop facilities.

In these closed test-loops components inserted in the reactor core, the coolant, instrumentation
and heat transfer systems were completely separated from the main FFTF core, permitting the
testing of fuels and materials over a wide range of temperatures in a controlled environment
independent of the main reactor coolant system. The open loop test positions and integral
components of the reactor core for testing large quantities of candidate fuel pins and
assemblies were cooled by the reactor primary coolant system.

225
7.3.3.2. Operating experience

The operating histogram of FFTF is given in Fig. 19.

The FFTF began its power ascent in November 1980. In December 1980 full power of
400 MW(th) was reached. A series of natural circulation tests proved that the FFTF loop-type
system could be operated safely under conditions of long-term decay heat removal by natural
convection without any sodium pumps working.

The FFTF has completed basically flawless operation over 10 years confirming the design
assumptions and material performance of mixed oxide fuel, sodium reactor systems, and the
overall safety and robustness of the modern LMR.

Advanced core materials have also been developed and tested at the FFTF. The most notable
material among them has been the ferritic steel alloy HT9, which has been irradiated to
ultra-high neutron fluence levels with little or no neutron-induced swelling and has been
selected for use in the U.S. ALMR. The FFTF also conducted materials experiments for the
development of the fusion reactor. A special test zone in the FFTF core allowed continuous
monitoring of tritium production occurring in fusion reactor blanket materials under a variety
of irradiation and thermal conditions.

The FFTF completed Cycle 12 in March 1992, accumulating a total of 2278 effective full
power days (EFPD) since the beginning of operation. During this time only one fuel pin from
all of the standard driver fuel assemblies had developed a leak; this assembly was well beyond
its design exposure. Irradiation of the reconstituted fusion materials open test assembly
experiment was initiated in May 1991 and functioned as designed until reactor shutdown in
March 1992. It has been removed from the reactor for irradiation in other facilities.

Nine core demonstration experiment for fuel assemblies, including lead tests, continued
irradiation until the reactor was shut down in March 1992. A lead test assembly reached a
world's best fuel assembly burnup of 238 MWd/kg. The highest burnup assembly reached a
burnup of 221 MWd/kg. It was agreed to process a hot channel lead test for post irradiation
examination by PNC. All nine assemblies have achieved their current exposures without
operational difficulty.

The possibility of future DOE missions along with collaborative international programs for
the FFTF had been evaluated. The plant was in steady state hot standby conditions for a long
time and was finally shutdown in 2000.

226
 1992 1993

FIG. 19. FFTF operating histogram.

227
7.3.3.3. The FFTF reactor experiments

The experiments conducted with respect to this reactor were carried out at 50% power over a
fairly low range of temperatures (the main pumps functioned at a 100% flow rate prior to
shutdown). The results emanating from these investigations are presented in Fig. 20.
Experiments involving pump shutdown without scram demonstrated that reactors based on
oxide fuels also possess intrinsically inherent safety properties (self-controllability); however,
in such reactors, these properties become readily apparent at higher temperatures, sometimes
exceeding the permissible value.

Supplemental measures were considered for the passive introduction of negative reactivity in
medium power and high power reactors that use oxide fuels. During the aforementioned FFTF
reactor experiments, such measures included so called gas expansion modules, GEMs (hollow
pipes, sealed at the top, open at the bottom, and set into the internal row of the radial blanket).

When the pumps were in operation, the coolant was pumped into the aforementioned GEMs
and gas compression occurred. Following pump shutdown, the core negative reactivity was
introduced simultaneously with the displacement of the sodium. This effect on the part of the
FFTF is essential to the subject reactor's self-shutdown.

The experiments conducted in the United States especially those performed on reactors
that use metal fuels are considered to be both outstanding and promising. Based on the
results emanating from these experiments, the largest U.S. reactor manufacturers, with the
support of government agencies (US DOE) have launched efforts aimed at the development of
absolutely safe reactors, in which both the shutdown system and the heat removal system
function exclusively on the basis of passive principles; i.e. their activation does not require the
participation of automatic systems or an operator.

In the most serious situations, such as those exemplified by LOFWS12, the spontaneous
reduction of power to a few percentage points of the rated value was essential in facilitating
the achievement of emergency heat removal from the reactor based entirely on passive
principles. As previously mentioned, the simplest of the subject techniques is the cooling of
the reactor vessel (guard vessel) by natural air flow. Here, the positive properties of reactors
that use a liquid-metal coolant also come into play. Owing to the high boiling point of the
coolant (~ 850C), as well as to the fact that the structural materials used are stainless steels,
the temperature of the structural elements (the reactor vessel, the guard vessel, etc.) can be
increased to 600650C when low-probability one-time emergency situations arise. Relatively
low vessel wall thickness values (3050 mm), together with the great difference between the
temperature of the surface being cooled and that of the free air, ensure efficient heat removal,
even when the air present in the annular space around the non-insulated guard vessel in the
reactor well (with respect to small reactors) or in the special sodium-air exchanger (with
respect to medium and large size reactors) is circulated naturally.

12
The FFTF primary pump was found in acceptance testing to stop in 100 to 115 s and have a halving
time of about 6 s.

228
FIG. 20. Characteristics of FFTF during LOFWS.

229
 BIBLIOGRAPHY

ASOMOTO, R., et al., Economics of the PRISM modular nuclear power plant, presented at
the international conference on fast breeder systems, 1317 September 1987, Richland,
Washington, USA.
DONALDS, R., et al., The Westinghouse approach to an inherently safe liquid metal reactor
design, paper presented in ANS International Topical Meeting on Reactor Safety, 1985,
Knoxville, USA.
ESSIG, C., Dynamic behaviour of Rapsodie in exceptional transient experiments, paper
presented in ANS International Topical Meeting on Reactor Safety, 1985, Knoxville, USA.
FOX, J., Safety features of small modular liquid metal reactor. Trans. ANSV. 52 (1985)
336-340.
GESA, L., et al., Safety and licensing of the PRISM liquid metal cooled reactor, paper
presented in 10th Int. Conf. Small and medium-sized nuclear reactors, SMIRT, 14-18 August
1989, Anaheim, California, USA.
GLUECLER, E., et al., Safety characteristics of a small modular reactor. Trans. ANSV. 52
(1985) 5058.
GRIFFITH. J.D., HORTON, K.E., Status of liquid metal reactor development in the United
States of America, Status of National Programmes on Fast Reactors, IAEA-TECDOC-741,
IAEA, Vienna (1994).
INTERNATIONAL ATOMIC ENERGY AGENCY, Status of Liquid Metal Cooled Fast
Breeder Reactors, Technical Reports Series No. 246, IAEA, Vienna (1985).
INTERNATIONAL ATOMIC ENERGY AGENCY, Status of Liquid Metal Fast Reactor
Development, IAEA-TECDOC-791, IAEA, Vienna (1993).
INTERNATIONAL ATOMIC ENERGY AGENCY, Progress in Liquid Metal Fast Reactor
Technology, IAEA-TECDOC-876, IAEA, Vienna (1995).
INTERNATIONAL ATOMIC ENERGY AGENCY, Status of Liquid Metal Cooled Fast
Reactor Technology, IAEA-TECDOC-1083, IAEA, Vienna, (1999).
INTERNATIONAL ATOMIC ENERGY AGENCY, Conceptual Designs of Advanced Fast
Reactors, IAEA- TECDOC-907, IAEA, Vienna (1996).
LANCET, R., et al, Inherent safety of the SAFR plant, paper presented in ANS International
Topical Meeting on Reactor Safety, 1985, Knoxville, USA.
LUCOFF, D.M., Passive safety testing at the FFTF, Proc. Int. Conf. Experience Gained and
Path to Economical Power Generation FBR, 1317 September 1987, Richland, USA, ANS,
Vol. 2, 11.8-1 - 11.8-7.
PLANCHON, H., et al., Implications of the EBR-II inherent safety demonstration test, paper
presented in Conf. Fast Breeder Reactors, 1986, Chalk River, USA.
TILL, C.E., Technical note: Advanced reactor development, Ann. Nucl. Energy, Vol. 16,
No. 6 (1989) 301305.

230
8. DOUNREAY FAST REACTOR: DESIGN FEATURES, EXPERIMENTS
DURING THE FINAL STAGE OF OPERATION

8.1. DESIGN FEATURES AND MAIN OPERATION RESULTS

As an initial experimental stage of a programme that began in the early 1950s to ensure
long-term security of the nuclear component of the UK's electricity supplies, a decision was
made in 1954 to build the 60 MW(th), 15 MW(e) Dounreay Fast Reactor (DFR), which was
subsequently operated from 1959 to 1977. The main design features of DFR were as follows
(Figs 1-3):

Upwards coolant flow through the core to avoid gas entrainment;

The driver core with a vented metallic enriched uranium fuel elements in the form of
hollow cylinders;
An electromagnetic pumps in primary and secondary circuit;
A low-rated, double-walled matrix design steam generators.

The philosophy of DFR at that time was to have the experimental part of the system only
inside the reactor vessel, and in the outside zone every effort was to be made to minimize the
risk of breakdown of the cooling system. This explains the unusual feature of 24 coolant
loops, which results in a size of pumps and heat exchanger where experience had been
accumulated in previous experimental work.

The DFR was designed primarily to confirm the feasibility of the fast reactor concept, but
quickly assumed a more enduring role as a test bed for candidate fuel, clad and structural
materials. After several years of successful operation, the DFR was shut down in 1967/68 for
one year to locate and repair a small leak in one of the coolant outlet pipes inside the reactor
vessel. Perniciously, the leak disappeared every time the reactor was shut down, making it
very difficult to locate and assess. The DFR continued to operate until March 1977, when it
was finally shut down. At its closure, mixed oxide fuel experiments had reached a peak
burnup of over 20%. Fuel pins with leaking cladding were irradiated following failure to a
further 3% burnup with little deterioration.

Until 1967 the major problem of damage to cladding materials was embrittlement. However,
in 1967, evidence was firstly announced of considerable void swelling taking place in
austenitic stainless steels irradiated to high fluences in the DFR.

During the final stages of normal power operation of the DFR, a series of experiments was
performed with the objective of exposing bundles of typical mixed oxide fuel pins to coolant
boiling for prolonged periods. The series, known as the DFR special experiments programme,
were comprised of eight separate experiments; they utilized both unirradiated and previously
irradiated fuel pins; and, in three experiments, included a thin steel plate simulating a local
blockage in the heated section.

231
232
FIG. 1. DFR plant cross section.
 FIG. 2. DFR reactor primary circuit.

233
234
FIG. 3. DFR reactor cross section.
8.2. INCORE COOLANT BOILING EXPERIMENTS [14]

8.2.1. Introduction

This section comprises the results of DFR in-core coolant boiling experiments, including a
brief description of the different rigs. It summarizes the operating experience and conclusions
from post irradiation examination (PIE) for each rig and includes conclusions concerning the
relevance of the programme and its results to ongoing LMFR safety studies. The work was
involved a large number of people at DNE and other centers, e.g. the intrinsic thermocouple
was supplied by the Nuclear Research Centre, Karlsruhe, Germany.

In each experiment, coolant boiling was initiated by reducing the coolant flow through the rig
using an inlet flow valve built into the experimental vehicle and controlled from the reactor
control room. The coolant flow rate was reduced to a level which was 25% below that needed
to induce boiling. Each experiment contained an array of thermal and acoustic sensors, the
precise complement of devices differing from rig to rig. The installed instrumentation served
two purposes; firstly it provided a means of monitoring conditions within the experiment;
secondly it enabled a comparison to be made between the efficacy of different types of
technique for detecting the initiation and development of a subassembly incident.

In particular, the complementary and contrasting merits of the thermal and acoustic noise
detection techniques were assessed at the Risley Nuclear Laboratories (RNL). Interpretation
of the experiments was carried out using the subchannel boiling code CLAYMORE which
was specially developed for the DFR programme. As a further aid to interpretation, a replica
of one of the rigs was studied, using electrical heaters, in a water rig at BNL. The range of
tests carried out included both single and two-phase behaviour, the influence of flow
direction, effect of different blockage porosities, and the consequences of gas injection. The
experiments were not designed to study dry-out mechanisms, so the investigation did not
attribute the encouraging behaviour to these mechanisms.

The experiments produced minimal interference with normal reactor operation; they were
loaded as part of the normal reactor inventory and, with one exception, not unloaded until the
scheduled end of the run, some 65 days later; therefore in most instances the fuel was
subjected to significant periods of irradiation following exposure to boiling conditions. PIE
of the fuel pins provided insight into how fuel might behave in incident situations analogous
to those modeled in this programme. Despite exposure to severe boiling conditions, the
majority of the fuel pins examined to date showed either no damage or minor damage such as
clad splitting; in one blockage experiment where conditions were severe enough to lead to
clad melting there was no evidence of fuel melting.

8.2.2. Objectives of the experiment

In broad terms the experimental objectives were to investigate the consequences of local and
inlet blockages in fuel pin bundles under reactor conditions, hopefully to demonstrate the
pessimism inherent in safety analyses. More specifically, the results were expected to
contribute to a more realistic understanding of the following aspects of LMFBR safety
arguments:

Fuel failure characteristics and propagation under abnormal service conditions;

The conditions for the initiation and stabilization of local boiling;
The possibility of blockage growth through instability of the pin cluster under
conditions of high temperature and high temperature gradient;

235
 Heat removal under bulk boiling conditions in a multipin geometry;
The effectiveness of different instruments in detecting and distinguishing specific fault
conditions;
The timescales available for detection and for remedial action.

In addition, it may be possible to make qualitative comments upon:

(i) Fuel redistribution under escalating local incident conditions;

(ii) The dynamics and extent of fuel-coolant chemical reactions.

The special experiments programme involved eight separate in-pile experiments using a total
of ninety-nine fuel pins. The coolant flow conditions were different in magnitude and
direction from those in current LMFRs such as the prototype fast reactor (PFR) but it is
believed that these atypical conditions make the DFR experiments pessimistic and that the
conditions experienced by the fuel were more severe than for the corresponding stages in a
similar incident in a commercial LMFR. These experiments explored fresh ground by
reducing the areas of speculation and by providing quantitative data, both supported the
development of theoretical methods used in safety analysis and provided pointers to the most
appropriate areas for further development.

8.2.3. Design of the instrumented vehicles

The experiments were mounted in two different designs of instrumented vehicle; five
experiments used a vehicle called a mini-subassembly containing 18 fuel pins and an
unheated instrument probe; three experiments used a smaller vehicle involving only three fuel
pins and called a trefoil. The coolant flow through the eight experiments was determined by a
valve sited upstream of the fuel bundle in each vehicle and operated remotely from the reactor
control room. A typical mini-subassembly is shown schematically in Fig. 4.

It contained 18 fuel pins within a pair of concentric cluster walls sitting inside a massive steel
carrier. The cluster walls confined an argon gas layer which insulates the coolant flow in the
rig from the comparatively cold bulk reactor flow. The five mini-subassemblies are divisible
into two categories, the 523 series for bulk boiling and the 539 series for local blockage
studies.

The fuel was mixed oxide, either pelleted or vibrocompacted, with burnups ranging from zero
to 10%, and clad in cold worked M316 stainless steel. The nominal pin outside diameter (OD)
was 5.3 mm and the pitch spacing similar to the PFR; the pins were spaced by honeycomb
grids axially separated by approximately 100 mm. The fuelled length was comparatively
short, being 505 mm in the 523 series and between 286 and 375 mm, respectively, in the
539 experiments. Because of the low inlet temperature of 230C, and short heated length, the
axial temperature gradient under boiling conditions was between three and four times the
gradient for a comparable incident in current LMFR designs. Typical peak linear rating for the
fuel pins was 32 kW/m (280 W/g).

The local blockage in the 539 series was a thin steel plate mounted eccentrically and covering
approximately 70% of the available flow area (Fig. 5).

236
FIG. 4. Simplified sketch of mini-subassembly.

237
 FIG. 5. Simplified sketch of trefoil.

238
The plate was welded to the downstream edge of a honeycomb grid and sited approximately
midway along the heated length. There was a narrow, internally shaped, annular gap between
the blockage plate and each of the fuel pins which passed through the plate; the effect of the
resulting small leakage flow upon the wake has been examined in out-of-pile water tests.

The trefoil vehicle in shared many features with the mini-subassemblies, e.g. the flow control
valve design; the most distinct difference, apart from the number of fuel pins, was the
introduction of instrumentation directly into and below the fuel cluster without the use of a
separate instrument probe; this was made possible by the use of unirradiated fuel allowing the
rigs to be assembled in one piece in the laboratory (Fig. 5). Thus trefoil experiments possess
thermocouple hot junctions sited in the main coolant flow, i.e. proud of any supporting
structure, and, in the final trefoil experiment, an acoustic sensor positioned close to the boiling
region. All three trefoils contained unirradiated mixed oxide fuel; for two trefoil experiments,
522 and 540, the specification was similar to that used for the 523 series mini-subassemblies;
for the remaining trefoil, 536, the fuel pins were 6.68 mm outside diameter, spirally wrapped
and contained vibrocompacted fuel at 80% theoretical density (TD) over a height of 569 mm.

In general, acoustic sensors were situated at upstream of the fuel bundle and thermal sensors
were positioned at inlet, midplane, outlet, and, when appropriate, in the wake region
downstream of the blockage. The instrument probe carrying the thermal sensors into the
mini-subassembly fuel bundles passed centrally in the 523 series but was displaced to an off
central position in the local blockage experiments so as to provide a better guide to the wake
temperatures (Fig. 4).

Three types of thermocouple were used in the experiments; mineral insulated stainless steel
sheath bifilar chromel/alumel 1 mm diameter conventional thermocouples; fast-response
thermocouples consisting of filaments of chromel and alumel butt welded together
inside mineral insulated stainless steel sheaths to form a single core coaxial cable, OD from
0.5 to 1 mm [2]; and, in the last of the 528 mini-subassemblies only, an 'intrinsic
thermocouple' based upon the principle described in [3] and allowing a very high frequency
response. Experience with all three types of sensor was excellent throughout the programme
and of the many thermocouples used only one gave any sign of failure.

Three types of acoustic sensor were employed in the experiments, a conventional

waveguide/accelerometer, a RNL high temperature microphone, and a miniature Capacitance
microphone. Apart from the final trefoil experiment all the acoustic sensors were sited
approximately 300 mm upstream from the start of the fuel bundle. The RNL microphone was
based around a lithium niobate crystal and was reduced in size especially for the
DFR programme to an outside diameter of 11 mm. The DNE miniature microphone was built
from 3 mm MISS single core cable and has a flattened 'sensitive' end of approximately
35 mm5 mm2 mm; because of its small size it can be inserted into otherwise inaccessible
locations. In the final trefoil experiment a miniature microphone survived a temperature of
nearly 1 000C for many hours. In addition to the thermal and acoustic sensors described
above selected rigs also incorporated void detectors and electromagnetic devices designed to
detect small rapid changes in local coolant resistivity caused by temperature fluctuations or
voids.

During the whole of each experiment's in-pile life the signals from the installed sensors were
recorded visually on paper charts and, as a permanent record, on magnetic tape; on-line
processing of thermal and acoustic noise data was provided by RNL during the majority of
boiling runs. All these data have now been edited and a series of reference 'archive' magnetic
tapes produced. The raw analogue data from the magnetic tapes has been normalized, to

239
account for the amplification in the recording system, and digitized to provide a computer
data bank of temperatures for each boiling run of each rig.

8.2.4. Operation

The experiments were irradiated during the last three power runs of the DFR, i.e. in the period
December 1975 to March 1977; the experiments were loaded and discharged, with one
exception, as part of the normal reactor loading schedule and were subject to the same
stringent safety criteria as other rigs, both with regard to the safety of the core and to possible
interference with other experiments.

To complement out-of-pile hydraulic tests, in-pile calibrations of the throttled rigs, i.e. curves
of valve position against corresponding coolant temperature rise across the rig, were carried
out during the period of low-power operation at the beginning of each reactor run.

Additional calibrations were also carried out, if sufficient time was available, at the end of a
reactor run and following any reduction in reactor power in an attempt to identify changes in
the flow characteristics of the rigs as a result of boiling. Low power calibrations could be
scaled to higher reactor powers and deviations between such scaled calibrations and data
measured at the higher power attributed, in principle, to the presence of coolant boiling.

The simple intention behind the operation of each rig was to expose the fuel pins to periods of
prolonged boiling; the mode of operation was straightforward. Because there was little
previous experience to act as a guide, coolant boiling was approached slowly in the early
experiments and the periods of boiling kept relatively short; conditions were made more
rigorous as confidence was gained. Thus in the lead experiment, DFR 522, the outlet
temperature was raised from 620 to 900C, at which sub-cooled boiling was first detected in
slightly more than 3 hours and boiling conditions were maintained for less than one hour. For
the final boiling run of an exactly similar rig, the DFR 540, at the end of the programme,
outlet temperatures in excess of 900C were reached in less than an hour and boiling
maintained for approximately 3.5 hours.

The coolant flow rate was reduced to a level which was 25% below that needed to induce
boiling. A typical boiling run proceeded by reducing flow through the rig using the control
valve until coolant boiling was detected; the flow was reduced by a further 10 to ~ 30% and
the rig left for a predetermined period of time; finally the valve was opened and the rig left
until the next approach to boiling or the end of the reactor run. During their in-pile lifetime
the rigs were subjected to several boiling runs, e.g. in the local blockage experiments DFR
539/2 and 539/3 they boiled 5 and 11 times, respectively. A summary of the overall
programme is provided in Table 1 and represents a total of 600 pin boiling hours; the longest
total boiling time, 24 hours, was experienced by 13 pins in DFR 539/3.

Three different indicators of boiling were recognized during the programme. Bulk boiling at
the rig outlet was readily detected by the measured outlet temperature reaching the saturation
temperature and by noting the invariance of the outlet temperature against a further decrease
in coolant flow, as measured by the increase in midplane temperature.

Analysis of the lead experiment, DFR 522, showed that the inception of sub-cooled boiling
within the bundle was accompanied by an increase in the thermal noise signal at the outlet
thermocouples; the signal increased with further decrease in flow (i.e. increasing severity of
boiling) until it reached a peak and thereafter decreased to a low value as bulk boiling was
approached. For all subsequent rigs thermal noise was monitored on line by a team of experts

240
from RNL; details of both this and the acoustic technique referred to in the next paragraph
will be published elsewhere and it is sufficient to say here that consistent and apparently
reliable indications of the onset and development of boiling were made. Water tests in an
electrically heated replica of the local blockage experiment confirmed that the thermal noise
signal at the outlet thermocouple increased following the inception of fairly extensive
nucleate boiling and increased further as the boiling developed.

In these experiments it was found that on-line examination of the RMS level and spectral
content of the signals from the acoustic sensors sited upstream of the bundle was unable to
provide any reliable indication of boiling inception.

A pulse technique developed by RNL and based around two separate detectors provided the
only reliable acoustic indications of boiling and for the majority of rigs gave results consistent
with those from the thermal noise analysis. In one rig an acoustic detector was sited close to
the boiling region and provided signals which have been interpreted as indicating sub-cooled
nucleate boiling well before the outlet thermocouple showed any significant rise in thermal
noise signal; this was consistent with signals from an electromagnetic sensor at the rig outlet
and with observations from the water rig.

8.2.5. Analysis of the results and conclusions

The most important feature of the experiments is the behaviour of the fuel pins and the first
objective of the supporting analytical programme was to assist and complement the Post
Irradiation Examination (PIE) of the fuel pins by specifying the thermalhydraulic conditions
experienced by each set of pins. A subsidiary objective is the determination of timescales at
which significant points- in the rigs' histories were reached, such as fuel failure, from changes
in the measured thermalhydraulic data.

The CLAYMORE [4] code was developed to help reach these objectives. CLAYMORE is a
thermal hydraulics sub-channel code capable of describing local and bulk boiling, it includes
a homogeneous two-phase flow model with slip, and reverse flow, and i.e. it can represent the
wake downstream of a local blockage. It is necessary to substantiate predictions from a code
such as CLAYMORE and to this end comparisons were made with:

Measurements made in a water-cooled replica of the local blockage rigs using

electrically heated pins; and
With data from single pin and multi pin out of pile sodium experiments.

The results of such comparisons showed there to be a broad measure of agreement between
prediction and measurement, in particular the description of the development of local boiling
appears to be qualitatively correct.

When applied to the DFR experiments the code has adequately predicted the inception of
boiling and such gross effects as the two-phase pressure drop/flow curve but detailed
comparisons show that there are aspects of the rigs' behaviour which remained unexplained.

8.2.5.1. Post irradiation examination (PIE)

The results with regard to PIE are summarized in Table 1.

241
 TABLE 1. SUMMARY OF THE DFR INCORE COOLANT BOILING EXPERIMENTS [1]

242
Rig Brief rig Boiling Boiling Comments
description history duration

T522 3 unirradiated prepressurized One boiling run at end of run 79. Approx. 1 Boiling was clearly detected at the time by the constant
mixed oxide fuel pins. Max. This constituted a lead hour boiling temperature limiting at saturation temperature and,
rating, 315 W/cm per pin. experiment for the series. retrospectively, by thermal noise analysis; examination of fuel
pins showed no sign of damage.

M(b)539/l 18 reference design fuel pins, One boiling run, but experiment Despite the severity of the conditions to which the rig was
burnups in range 0-10%, vibro terminated because of severe gas subjected (failed pins exposed to hot coolant for the order of
and pelleted fuel, different clad entrainment problems leading to 10 days) there was no sign of fuel melting; nor was there any
types. Blockage plate covering extensive clad failure in the wake sign of rapid escalation of the fault condition.
approx. 70% of flow area. region and self induced boiling.

T536 3 fat unirradiated spiral wrap Three boiling runs of approx. one Approx. 4 Boiling detected by thermal noise and acoustics on line.
mixed oxide pins. Max. rating hour duration each. On two hours Examination of fuel pins showed no sign of damage.
416 W/cm per pin. occasions a self limiting temp.,
excursion was observed. First and
second boiling runs approx. 1
week apart.
M528/1 18 reference design fuel pins at Two boiling runs, each run Approx. Boiling detected by both thermal and acoustic noise
0.3% and 6% B.U. One pin showed similar, steady and 2.5 hours techniques on line at similar rig temperatures. Examination of
contains an artificial defect at its repeatable behaviour. bundle showed bowing of pins at the hot end and one pin had
hot end. failed - no signs of clad melting or of any fuel loss.

M(b)539/213 As for 539/1. Five boiling runs spread over 4 Approx. 4 Boiling detected by thermal noise measurements on line.
weeks. In the intervals between hours Examination of rig showed marked bowing of pins at hot end.
the first four boiling runs the rig Six pins had failed with numerous longitudinal cracks. One
was left at approx. 100C below also had a small zone of molten clad. No observable fuel loss.
saturation at hottest point.

13
M represents mini subassembly, M(b) represents mini-subassembly with plate blockage, T represents trefoil
 TABLE 1 - SUMMARY OF THE DFR INCORE COOLANT BOILING EXPERIMENTS [1]
Cont'd
Brief rig Boiling Boiling
Rig Comments
description history duration

M(b)539/214 As for 539/1. Five boiling runs spread over 4 weeks. In the Approx. 4 Boiling detected by thermal noise measurements on line.
intervals between the first four boiling runs the hours Examination of rig showed marked bowing of pins at hot
rig was left at approx. 100C below saturation end. Six pins had failed with numerous longitudinal cracks.
at hottest point. One also had a small zone of molten clad. No observable fuel
loss.
M528/2 As for 523/1, Two boiling runs of approx. 3 hours duration Approx. 6 Boiling clearly detected with thermal noise measurements
proportion of pins ten days apart. hours and consistent results from acoustic noise. Evidence of
at differing coolant chugging for outlet temp, exceeding 900C.
burnups changed. Nondestructive examination showed the pin array to be
regular with no obvious pin failures. Swelling and distortion
at hot end.
M(b)539/3 As for 539/1. 11 approaches to boiling spread over period of Approx. 24 Increase in thermal noise signal with boiling less pronounced
a month. For most runs the period of boiling hours although qualitatively the same as in previous rigs.
lay between 2 and 3 hours. A total of more than Radiographs of the rig showed obvious signs of pin failures
7 hours was spent with the rig outlet temp with possible fuel loss. During some of the later boiling runs
exceeding 900C with a significant fraction flows were reduced to 75% of their value at boiling
above 960C. inception.

T540 Similar to 522, but Two boiling runs of approx. 1.5 and 3.5 hours Approx. 5 Thermal noise and acoustic noise both indicated boiling.
pins unpressurised duration. Saturation temp., was maintained hours Examination of pins showed distortion but no significant
and fuel in steady at outlet and reached at the midplane. diameter change - no sign of failure of any description.
pelletted form.

14
M represents mini subassembly, M(b) represents mini-subassembly with plate blockage, T represents trefoil

243
8.2.5.2. Trefoil behavior

PIE shows no signs of failure in any of the trefoil (previously unirradiated) fuel pins.
Measured diameter changes are small showing no apparent effect caused by prolonged high
temperature operation: the DFR 540 coolant outlet temperature was above 800C for more
than 6 hours, above 900C for more than 5.6 hours, and at a steady 960C for just over
4 hours.

8.2.5.3. Mini-subassemblies without local blockage

The first of the DFR 528 series has been radiographed and examined visually, the
second experiment has only been radiographed. Both bundles show significant bowing of the
pins, and some swelling over the hottest region, otherwise the pin arrays are regular.
DFR 523/1 contained only one failure; there was either sign neither of fuel loss nor of clad or
fuel melting.

8.2.5.4. Mini-subassemblies with local blockade

The first local blockage experiment inadvertently entrained gas into the wake region causing
fuel failure, including clad melting, soon after the first rise to full power. Boiling of the rig
occurred after a further 10 days irradiation and led to a further region of fuel failure
downstream of the wake. Detailed Post Irradiation Examination has been performed on the
rig; there is no evidence of any fuel melting despite the severity of the incident; there are
several examples of sintered fuel remaining intact despite removal of clad from the failed
region; there was no sign of any secondary blockage formation at the downstream grids
despite the loss of complete sections of some fuel pins. The mechanism of gas entrainment
into the wake and subsequent fuel failure has been demonstrated using the heated water rig.

The second blockage experiment has been examined visually. Six pin failures were found,
including small regions of clad melting. There was some small observable fuel loss but no
sign of steady blockage growth. The final blockage experiment has only been examined
radiographically but there were obvious signs that at least three pins have failed extensively;
however the rig was boiled vigorously for nearly 24 hours. In the experiments 600 pin-hours
of detectable boiling were accumulated. No failures occurred in the pins with 3% burnup,
seven failures arose amongst the seventeen pins with greater than 6% burnup. From the results
obtained it can be concluded that:

DFR fuel pins were extremely durable, even when subjected to severe incident
conditions;
There was no sign that fuel failure, even when subjected to continuing irradiation and
boiling, leads to a rapidly developing subassembly incident;
The time available for the detection of the initial incident is long compared to the
response times of all potential reactor monitoring systems;
Thermal noise was an extremely effective and reliable method for detecting coolant
boiling.
8.2.5.5. Relevance to safety studies
Because of the low, downward flow in these experiments the effect of buoyancy upon boiling
and entrained gas behaviour ensured that the experiments represent conditions which are more
severe than the corresponding situation in a modern LMFR with high, upward coolant flow.
In particular the results have shown that:

244
 Boiling was detected at an early stage;
Prolonged boiling was stable;
Boiling did not necessarily lead to fuel failure;
Fuel failure did not necessarily lead to fuel release;
Fuel release did not lead to the formation of a local blockage;
A local blockage did not lead to fuel melting;
The combination of gas entrainment and a local blockage with consequent hold-up of
gas in the downstream wake did not lead to fuel melting.

Experimental studies on heat removal from the core by the boiling coolant carried out at the
IPPE (fuel element simulators were tested in sodium-potassium coolant) shown possibility of
the long term heat removal from the fuel elements by the boiling coolant under conditions of
natural flow and heat flux up to 270 kW/m2 (without fuel element simulator damage).
8.2.5.6. Relevance to sodium voiding reactivity

As pointed out in Ref. [5], historically, the desire to lower the amplitude of the positive
sodium void worth in LMFR was driven by a concern over extremely high rates of sodium
voiding. This concern had several origins:

First, an observation in laboratory experiments showed that the onset of boiling in

liquid sodium can occur at temperatures significantly higher than the boiling point
before any vapor was formed. This would imply that once vapor began forming, the
entire superheated coolant inventory in the core could void almost instantaneously
without further heat addition;
Second, there was the potential for vapor explosions wherein extremely high
temperature fuel suddenly breached through cladding and mixed intimately with liquid
sodium [6]. Finally, any case of rapid voiding could result in high positive reactivity
ramp rates which might act faster than the negative feedbacks could respond and
would take the core into the super prompt critical range.

However, experimental data on boiling in tube bundles shown that the in-core sodium boiling
process in fact does not reach high superheat, but rather comprises a series of local
pressurization and flow reversals which voids part of an assembly for a short period of time.
Detailed analyses have shown substantial spatial and temporal incoherence in the boiling
process, with incoherent chugging and a few assemblies leading the rest of core [6].

8.2.5.7. Relevance to reactor core acoustic instrumentation

Acoustic devices have been used successfully to detect boiling at DFR. Therefore, an acoustic
detection of boiling can play an important part in the plant protective system, because it
affords a means of detecting local overheating and allows protective action to be taken before
damage spreads to otherwise unaffected parts of the structure.

The objective of the protective system is to intervene by shutting down the reactor, preferably
soon enough to prevent fuel melting, certainly soon enough to prevent the propagation of
damage beyond the affected subassembly. An acoustic instrumentation has an important role
in providing a signal, indicating that something is amiss, to initiate the protective action. Its
importance lies in the fact that it may afford the only way of knowing promptly that the
coolant is boiling.

245
As an example of the type of local fault against which acoustic boiling noise detection
(ABND) provides protection, it is common to consider a complete instantaneous blockage of
the coolant flow at inlet to a core subassembly while the reactor is operating at full power.

The performance of ABND systems was the subject of an IAEA Coordinated Research
Programme [7], aimed at determining whether boiling could be detected reliably, in the
presence of background noise, with an acceptably low spurious detection rate. Background
noise from the dummy cores of PFR and Super-Phnix, and from the operating cores of PFR
and KNK-2, was recorded, and from various out-of-pile test rigs, was recorded.

Various software methods for detecting boiling noise in the presence of background have been
investigated [8]. The use of filtration, pattern recognition techniques, or a combination of
both, allows boiling to be detected in the presence of a background which is noisier than the
boiling source.

The IAEA programme showed that with a signal-to-noise ration of -12dB [i.e. with the root
mean square method (RMS) boiling signal power about a factor of 16 lower than the RMS
background signal power], boiling can be detected with a reliability of less than one error in
106, and with less than one spurious indication in 106years.

REFERENCES

[1] SMITH, D.C.G., BAGLEY, K.Q., GREGORY, C.V., LEET, G.O., TAIT, D.,
DFR special experiments, paper presented in Symp. Design, construction, and operating
experience of demonstration LMFRs, 1014 April 1978, Bologna, Italy.
[2] THOMSON, A., FENTON, S.W., High-integrity, small-diameter, mineral-insulated
thermocouples, Temperature Measurement, paper presented in Conf., Teddington, 1975,
Inst. Phys. London (1975).
[3] BOYER, J., et al., Probeless temperature measurements at the wall of a liquid metal
flow, Heat Transfer in Liquid Metals, published in Progress in Heat and Mass Transfer,
Vol. 7, Pergamon, Elmsford, NY, USA (1973).
[4] GREGORY, C.V., LORD, D.J., The CLAYMORE code and its use for describing
coolant boiling in LMFBR sub-assemblies, Nucl. Energy, 17 (1978) 107114.
[5] WADE, D.C., Recent innovations in IFR safety research. paper presented in Int. Topical
Meeting on Advanced reactor safety, 1721 April 1994, Pittsburgh, Pennsylvania.
[6] HICKS, E.P., MENZIES, D.C., Theoretical studies on the fast reactor maximum
accident, Proc. Conference on safety, fuels and core design in large fast power reactors,
ANL-7120, October 1965.
[7] INTERNATIONAL ATOMIC ENERGY AGENCY, Acoustic Signal Processing for the
Detection of Sodium Boiling or Sodium-Water Reaction in LMFRs, IAEA-TECDOC-
946, IAEA, Vienna (1997).
[8] Signal processing techniques for boiling noise detection, paper presented in the
International Working Grpup on Fast Reactors, Vienna, 1989.

246
9. SODIUM COOLED FAST REACTORS DEVELOPMENT AND OPERATING
EXPERIENCE: LESSONS LEARNED IN THE PAST AND CHALLENGES
FOR THE FUTURE

9.1. MASTERING OF SODIUM COOLANT TECHNOLOGY

Excellent thermophysical properties of sodium were demonstrated at the first test facilities as
early as in the 1950s, and promised its large-scale application in the power area.

During the period 1958-1963, design studies were made on the use of sodium as a coolant of
traditional boilers: sodium heated by burning fossil fuel in the furnace was used to transport
heat to the steam generator (SG), located in the vicinity of the turbine. This resulted in
considerable decrease the length of steam/water pipelines. However, this option was
abandoned because of complicated sodium handling for traditional power engineering.

At the same period, the sodium-graphite thermal neutron power reactor Hallam
[240 W(th), 75 W(e)] was constructed and put into operation in the United States in 1962.
Sodium was used to cool the reactor core and to transport heat to the double wall SG; graphite
blocks with steel lining being used as moderator. In case of lining integrity loss, sodium-
graphite interaction occurred in the blocks resulted in graphite swelling and lining rupture.
This circumstance along with the complexity of sodium technology in comparison with water
caused premature closing down of the reactor in 1966, and abandonment of this concept.

A large size solar facility [500 kW(e)] at Almeria in Spain with 70 tons of sodium used to
transfer the solar heat energy to the boiler, had faced technological and safety problems. In
order to work on the sodium circuit, the operators should have cooled the piping to freeze the
sodium at ambient temperature, then they could cut the circuit. But a problem occurred with
the formation of the solid sodium plug; pressurised sodium spewed through the cut made,
splashing off the nearby steel structures and causing a fire in the hall. Thus, non-observation
of special requirements to maintenance of sodium components resulted in pouring out large
amounts of sodium (14 tons at 225C). The sodium burnt in the atmosphere and the
temperature reached an estimated 1200C during approximately 15 minutes. It caused
considerable damage (the metallic beams distorted and a hole was blown in the roof), and the
facility was closed down in 1986 [1].

Ten experimental and six prototype and commercial size sodium cooled fast reactors have
been constructed and operated. The worldwide investment already made in the development
of this unique technology exceeds US$ 50 billion [2, 3]. Robustness of the sodium systems
(active zone cooling and heat transfer to SG or sodium/air heat exchanger) as a whole was
mastered by operation of fast reactors in various countries.

The sodium production meeting reactor standard requirements was developed and mastered at
plants; and measuring instruments and devises were developed for controlling sodium quality
and parameters required for reactors systems. For the most vulnerable element of the nuclear
power plant, i.e. the SG heated by sodium, special systems were developed for detection of
water-to-sodium leaks and protection of the SG against consequences of sodium-water
reaction. Problems related to the design of pumps with bearings lubricated by the pumped
liquid (sodium) were successfully solved.

Appropriate steels (mainly standard stainless steels) were chosen out of the available range of
structural materials on the basis of their irradiation in research and then in demonstration
reactors. The reactors themselves and, more frequently, components have showed remarkable

247
performance well in excess of design expectations. The overall experience with sodium
cooled fast reactors in many cases has been extremely good, in spite of some setbacks.

Incidents occurred in the process of removal of sodium residues from the drain tanks using
heavy alcohol resulted in the extensive local damage of the auxiliary building structures at
two test facilities in Europe. Appropriate studies to the use of alcohol together with sodium
had been performed. It was concluded that under certain circumstances (e.g. closed
geometries, sodium puddles) the use of alcohol to clean components or to destroy sodium can
be dangerous.

The Japanese prototype fast reactor MONJU of 280 W(e) power was shut down in
December 1995 during the 40% power pre-operational testing phase due to a leak and fire in
the non-radioactive secondary sodium heat transport system. The cause of the failure of the
well tube is considered to be high-cycle fatigue due to flow-induced vibration.

These events gave rise not only to perform design modifications aiming at the confinement of
possible leaks and the protection of structures against sodium fire, but also to accelerate R&D
and code development with respect to sodium spray fires.

The above mentioned issues, as well as fast reactor development declined, have forced
researchers to complete the knowledge on fast reactor technology with alternative coolants:
gas (He and CO2), steam and heavy liquid metals15 (lead-bismuth and lead). Less imminent
requirements to the breeding capability have made it possible to expand the range of coolants
under study, since one of advantages achieved by sodium, i.e. high power densities of cores
assuring effective fuel breeding was no longer so important. Of course, some sodium cooled
fast reactors had faced operating and technological issues, but it should be noted that this
would be true for any other reactor line at the initial stage of development, as this type of
reactors is. Full industrial development of fast reactors has not been completed yet. It is
simply too early at the prototype stage of development to more general view of sodium cooled
fast reactor technology. Other reactor technologies, including water cooled reactors, achieved
high reliability when their respective large scale introduction had taken place. We cannot say
that this will not happen in the case of LMFR. In this context, sodiums attractive properties
should be recalled, namely: compatibility with traditional structural materials and all fuel
compounds up to high temperatures owing to its corrosion inertness and pressure close to
atmospheric value, excellent thermohydraulic characteristics, thus assuring effective heat
removal under conditions of either nominal flow rate or natural circulation flow rate at the
reactor. Therefore, on condition of proper design and manufacture of components, there are
no physical factors provoking failures. Consequently, in most countries involved in fast
reactors R&D activities, sodium cooled reactor line is considered showing promise.

Sodium is relatively cheap compared to other liquid coolants, and its resource is vast. In the
nineties of the last century, the technology of sodium as fast reactor coolant was mastered and
brought to a demonstration commercial level experts believe, that sodium could even more
consolidate its stand as fast reactor coolant, if the designers used vast experience gained in
this area in order to develop advanced approach as to the design of nuclear steam supply
systems, for its simplification and economics improvement.

15
Although, on the basis of experience, it is fair to assert that problems and risks from recognized shortcomings
(e.g. sodium fire) are less worrisome than those due to hidden effects chronically present in the system (e.g.
corrosion effects by heavy coolant).

248
Three fast reactors, which are now under construction: the experimental CEFR (China), the
prototype PFBR (India), and the commercial BN-800 (Russian Federation) use sodium as
coolant. Out of the selected reactor systems for Generation IV, the sodium cooled fast reactor
received international interest and support as reported at the Policy Group Meeting of the
Generation IV International Forum (GIF), Cape Town, South Africa, March 2003.

There is no doubt that experience gained in the world could serve as a basis for the
development of sodium cooled fast reactors using so called traditional approaches. However,
there is a great need to continue design studies to decrease the cost of construction, operating
and decommissioning of the reactor, and to improve the robustness of the sodium cooled fast
reactors technology in general. The points causing higher costs compared to that of LWR are
the large mass of structural material (mainly steel) in the reactor system as well as the
comparatively complicated reactor and balance of power design. It is an important
achievement in the fast reactor area of nuclear power that two demonstration commercial
size pool type fast reactors: BN-600 and Super-Phnix have been constructed and
operated (Fig. 1).

FIG. 1. Cross-section of the primary circuits: Super-Phnix - on the left,

BN-600 - on the right in the same scale.

249
9.2. DEMONSTRATION COMMERCIAL FAST REACTORS OPERATING
EXPERIENCE: ACHIEVEMENTS AND SETBACKS

9.2.1. BN-600

The power plant BN-600 was connected to the grid in April 1980. Since to reach the steady
state of the core three refuelling cycles were needed, full power was reached in October 1981.
Reactor operation is stable and the average load factor () is about 75% (Fig. 2).

80 83.53
80.29 79.89
78.19 77.35
76.60 75.89 76.32 76.43
74.11
70 71.76 72.75 72.48 73.46 72.97 73.23
69.83 70.31

65.91
60

56.00
50
50.10
47.93

40

30
27.50

20

10

0
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

FIG. 2. BN-600: annual load factors16,%.

Routine events, including four small (0.1-3.0 L) leaks in the primary auxiliary sodium and gas
systems and more leaks in the secondary auxiliary system occurred mainly during the
pre-operational testing phase (SGs valve sealings 4 leaks; drain and blow-of lines 11
leaks; sodium reception and storage 7 leaks). However, almost no events exceeding normal
operating conditions have occurred except one: the large (800 kg) sodium leak in October
1993 from the external primary sodium purification system rated at level 0-1 of the
International Nuclear Events Scale (INES) [4]. Due to the fire fighting and protective systems
operated in a proper way, the local damage was insignificant and repair was affected quickly.
Of the twelve cases of steam into sodium leak occurred mainly in the first year of operation in
the superheater and reheater modules owing to develop of latent manufacturing defects; in
only two cases was it necessary to shut down the corresponding steam generator (SG), and in
one case the reactor. In the other cases, since there is some excess of heat transfer surface of
the modular SG and the presence of valves on steam-water and sodium side, the appropriate
section was disconnected and the SG as a whole continued to operate without reducing the
reactor power. That was the outstanding operating advantage of the modular SG, but not the
cost advantage of the complicated and steel-intensive design.

The overall experience with BN-600 power plant has been rather successful, reactor
themselves as well as, components with sodium coolant: intermediate heat exchangers,
pumps, reactor refuelling systems, reactor cooling and heat transfer systems showing
remarkable performances well in accordance with the design expectations. The majority of the
BN-600 reactor components were tested at the sodium facilities simulating the design
operating conditions; larger R&D work programs ($ 12 billion) [4] led to success.
16
Reactor is being shutdown two times per year for refuelling: in 1998 - for the rotating plug repair.

250
9.2.2. Super-Phnix (SPX)

The largest LMFR in the world SPX was successfully constructed and connected to the grid in
January 1986. The SPX plant reached its full power was in December 1986. However, these
have been a difficulty in operation: the operation was interrupted by tenth events and
incidents nine of which rated at level 1-2 of INES [5, 6]. Figure 3 displays an operating and
administrative history of the SPX power plant.

53 months normal operation (production and maintenance): 4.5 years

25 months recovery after incidence: 2 years

administrative delay: 4.5 years

54 months
(including 24 months of public hearing)
11 years in total

January May March April July March July August September December
86 87 88 89 90 91 92 94 95 96

1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997
Total

17 months (1 st cycle) 15 months (2nd cycle) 5 months (3rd cycle) 16 months (4th cycle) 53 months
Na purification Ar leak in exchanger
Storage drum
10 months 8 months 7 months 25 months
Administrative delay Administrative delay Public hearing Administrative delay

13 months 16 months 24 months 1 month 54 months

FIG. 3. The operating and administrative history of SPX.

During 11 years of the SPX power plant existence (1986-1997, Fig. 3), it was in operation for
4.5 years. Repair procedures for elimination of consequences of incidents took two years;
administrative and social procedures concerning license issue for reactor start-up took more
than 4 years. During 4 cycles (53 months, ~ 4.5 years, Fig. 3), the SPX produced 7.9 billion
kWh [6]. The outstanding success of the SPX operation has undoubtedly been the
demonstration of reliable (without leaks) operating of SGs with high self power
[750 MW(th)]. Three incidents marred the commissioning procedure and caused lengthy
delays:

Cracks and a leak in the space between the storage drum vessels on 8 March 1987
were caused by inappropriate use fuel storage drum vessel (FSD) of carbon steel
instead of stainless steel. The repair and other procedures (a 23-month reactor
shutdown) required dismantling the FSD and conversion of 700 tones of the residual
sodium into sodium carbonate through a controlled additional of water vapor and
carbon dioxide gas.
Lack of instruments for impurities control in cover gas of the SPX reactor and
practically atmospheric cover gas pressure caused uncontrolled air inflow into the
reactor in July during 3 weeks 1992 through a failed compressor membrane of the gas

251
 radiometer that caused oxidation of a large sodium fraction (Fig. 4) [7]. Purification of
reactor coolant from oxides (~ 400 kg) required installation of additional cold traps
into the reactor. The result of the air ingress into the reactor was a 24-month reactor
shutdown.
Gas (argon) leakage to sodium through the crack in the weld of piping of the argon
supply to the sealing bell surrounding the IHX in January 1995 (8-month reactor
shutdown).

1-compressor bode, 2-main rod, 3-intermediate ring, 4-top of connecting rod, 5-head, 6-closures,
7-compression chamber, 8-neoprene working diaphragm, 9-neoprene safety diaphragm,
10-safety chamber, 11-control unit

FIG. 4. Scheme of the radiometer compression.

These events as well as a very strong anti-nuclear social movement in the post Chernobyl
period, and a sodium fire at the Almeria facility that coincided with SPX reactor start-up tests
and power increase, had considerable political repercussions, caused long reactor outage
(Fig. 3) and provided one of the reasons for which the SPX reactor was prematurely closed
down by the left French Government. It should be noted, that the experience gained in the
development, construction and operation of the BN-600, SPX and other demonstration and
prototype reactors are invaluable for the development of fast reactor technology, in particular,
for the design of the commercial size advanced reactor projects, e.g. EFR, BN-800 and other,
having much better technical characteristics and indicating that the goal of robustness fast
rectors is within reach.

9.3. CHALLENGES FOR THE FUTURE

9.3.1. Increase of safety margins

There is a widespread opinion that the liquid metal cooled fast reactor advanced projects can
meet present safety standards quite adequately, and are expected to do so for the foreseeable
future. Nevertheless it is essential to continue the search for greater safety because:

252
 No opportunity to improve technology should be ignored;
Safety standards may be raised by the regulatory authorities;
Reactors suitable for countries with less developed infrastructure will be needed; and
More cost-effective ways of meeting existing safety standards may be found.

A fast reactor with the internal breeding ratio ~ 1.0 and sodium coolant, on principle, provides
a potential to design a nuclear system in which the power would eventually passively adjust
itself due to self-regulated heat production and heat removal [8] owing to:

Near zero burnup reactivity swing;

Strong negative temperature reactivity effects;
Sufficiently small hot to cold reactivity swing;
Possibility of core coolant temperature rise changing in wide range,
Impossibility of local criticality formation under conditions of any perturbation of the
neutron field;
Absence of poisoning effect (such as xenon poisoning) followed by positive reactivity
insertion;
A considerable portion of radioactive fission products is held by sodium and caught by
the cold traps in case of their release from the fuel elements to the coolant;
High vaporization heat of sodium;
System operated at low pressure and having large margins to boiling temperature;
Negligible thermal energy stored in the coolant available for release in the event of a
leak.

Figures 5 and 6 depict variations in the principal parameters of the PRISM (power reactor
innovative small module) reactor during two types of emergency situations [9]:

Pump shutdown in circuits I, II, and III, accompanied by the non operation of the
safety rods an unprotected loss of flow (ULOF), Fig. 5; and
The erroneous withdrawal of the control rods accompanied by the non operation of the
safety rods an unprotected transient overpower (UTOP), Fig. 6.

It follows from the parameters presented in Fig. 5 that, despite the use of low-inertia
electromagnetic pumps, a quite gradual flow rate decrease is achieved in the primary circuit of
the PRISM reactor apparently as a result of the presence in this reactor's power supply
system of synchronous electric motors with large moments of inertia. The subject reactor
possesses a low pressure drop core, which ultimately ensures a high natural circulation flow
rate (7-8%). The first reaction to the flow rate decrease is a fuel temperature increase,
followed by a coolant temperature increase. A positive sodium density reactivity effect is
achieved during the first few seconds of this process. However, negative components the
Doppler effect, the axial expansion of the fuel, and the radial expansion of the core diagrid
combine to form a resultant negative reactivity that also provides a power reduction. During
the ensuing reduction of power, the fuel is cooled and the Doppler effect takes on a positive
value at a particular moment in time. However, due to the use of a metal fuel with sodium
filled gap that has a high degree of thermal conductivity, the difference in temperatures
between the fuel and the coolant is small, which diminishes the positive component of this
effect. The effect that results from the axial expansion of the fuel behaves in a similar fashion.

253
 1-fuel temperature, 2-central fuel subassembly 1-central FSA outlet temperature, 2-reactor
(FSA) coolant outlet temperature, 3-reactor power, 3-core outlet temperature, 4-sodium
power, 4-coolant flow rate, 5-sodium density, density, 5-total reactivity, 6-control rod drive
6-Doppler, 7-axial fuel expansion, 8-control rod lines expansion, 7-axial fuel expansion, 8-radial
drive expansion, 9-total reactivity, 10-radial grid grid expansion, 9-Doppler
expansion
FIG. 5. PRISM parameters: ULOF. FIG. 6. PRISM parameters: UTOP.

It is apparent from that these two effects - the positive sodium density effect and the negative
effect owing to the radial thermal expansions of the core diagrid are antagonistic toward one
another (Fig. 5). The favourable behaviuor of the Doppler effect and the high sensitivity of
reactivity to core geometry are dictated by the high energy neutron spectrum of a reactor
which has relatively small dimensions and a metal fuel heterogeneous core. The
temperature of the coolant reached a peak of 680C at the 10-s mark, and then dropped to a
constant value of ~ 600C (Fig. 5).

Figure 6 depicts variations in the principal parameters of the PRISM reactor during the most
dangerous reactor incident - the erroneous withdrawal of the control rods accompanied by the
non operation of the safety rods an unprotected transient overpower (UTOP). It was
assumed over the course of the calculations performed that all six control and safety rods
began to be withdrawn from the core, thereby introducing positive reactivity at a rate of
0.02 /s. The complete withdrawal of these rods introduced a positive reactivity of 0.35 . The
power increase occurring at this moment in time was moderated by the negative reactivity
effects: the Doppler effect -0.14 ; the radial expansion of the core diagrid 0.13 , and; the
axial expansion of the fuel element, 0.06 . The elongation of the control rod drive lines set in
after a slight delay, with the reactivity introduced thereby coming to -0.04 at the 80-second
mark. As a result, total reactivity began to decline and the effect produced by the withdrawal
of the control rods had been neutralized entirely at the 200-second mark, at which point
reactivity came to ~ 0. Despite the considerable increase in power by 1.7 times the

254
temperatures in the core and throughout the reactor did not exceed the maximum permissible
values [9]:
The coolant temperature at the outlet of the most hot subassembly rose from 500 to
620C;
The mean coolant temperature at the core outlet was increased from 465 to 575C;
The temperature of the fuel reached a peak at 920C, which is 200C below the
melting point of a metal (U-Pu-Zr) fuel.

It should be noted that the relatively favorable course of the emergency process under
consideration is due to the diminished nominal temperature values used in this PRISM's
design (300/465C). Thus, in the right design, the reactor with above mentioned features
would be passively protected against nearly all control system failures and operator errors.
Implementation of a fast reactor design with safety based on self-regulation principles
(physical laws) requires the core and its support and surrounding structures to be designed in
such a way that negative reactivity insertion is provided in case of their thermal expansion and
bowing in the right direction, that gives net reactivity sufficient for reactor power decrease
down to the level at which heat removal is possible by the natural circulation coolant flow.

The importance of these approaches has confirmed by the results of studies, development of
PRISM/ALMR reactors and during the final stage of operation of the test fast reactors
Rapsodie, EBR-II and FFTF [10-12]. In these tests, the reactor pumps were switched off and
reactor safety system failure was simulated with absorber rods prevented from falling down
by the operator, core power decreased abruptly owing to negative reactivity effects, and heat
was removed by sodium natural circulation flow. It should be pointed out, that the LMFR
could become more attractive for the investors and utilities, if the designers managed to make
the most of the physical and thermohydraulic advantages of this fast reactor. The complete
development of the self-control and self-protection design of the sodium cooled fast breeder
reactor against probable accidents including the most severe accidents, could compensate for
the complication of the LMFR technology. Designers are making an effort to assure LMFR
safety in the system of Generation IV reactors exclusively on a self-regulation basis. It is just
because of its inherent safety features that LMFR is included in the Generation IV program.
Some institutes and design organizations having know-how on this issue, are capable of
solving this problem by giving a self-control capability to the sodium cooled fast reactor,
which is not a characteristic for any other type of reactor, and this would reliably compensate
for some drawbacks of sodium coolant.

9.3.2. Simplification of NSSS structure and operation & maintenance (O&M)

In existing nuclear power plants, the reactor and heat removal/transport systems are
enmeshed by over half a dozen external auxiliary systems with their elements spread over
various rooms belonging mainly to the nuclear island. The main systems and elements are
as follows:

Fuel storage drum (FSD) with related auxiliary sodium systems;

External system for oxides removal from radioactive sodium;
Emergency systems for decay heat removal (including heat exchangers, pumps, etc.)
connected to the reactor or secondary circuit;
Sodium heated steam reheater;
Lubrication systems of the main sodium pumps (for the upper bearing and sealing
units).

255
The potential of pool type fast reactors has not been exhausted regarding the integration of the
main radioactive circuit and the rejection of external auxiliary sodium systems, e.g. sodium
purification and emergency reactor cooling. There are some elements of sodium pipelines
with T-joints in this system, where sodium flows having different temperatures are mixed,
causing temperature fluctuations and, hence thermal stresses that resulted in crack formation.
As was pointed out before, a crack due to the temperature fluctuations in the coolant close to a
structure appeared in the T-joints of the BN-600 piping of the primary sodium purification.
A design option locating cold traps and their related systems inside the reactor vessel has been
realized in the SPX reactor and studies have been made of to apply advanced design option to
the BN-1600 reactor project (Fig. 7). A refuelling option has been developed for the BN-350
reactor without use of ex-vessel fuel storage drum (FSD) following its failure. Similarly to
SPX, designer and operator proposed a solution on the reactor refuelling without use of the
FSD (there was also an incident causing the FSD failure).The latter demonstrates that this
component can be avoided by means of holding spent fuel subassemblies (FSAs) in the
in-vessel storage during reactor operation in order to decrease their decay heat. An additional
reduction in the residual power of unloaded FSAs could be reached by incorporating one or
two rows of B4C-filled assemblies between the reactor core and the store (Fig. 7).

1-reactor core; 2-main reactor vessel; 3-guard vessel; 4-submerged heat exchanger for decay heat removal; 5-
cold trap; 6-rotating plug; 7-core diagrid; 8-core catcher; 9-stationary shield; 10-hot sodium collector; 11-baffle;
12-well liner; 13-containment; 14-refuelling mechanism; 15-elevator (provides removing FSAs from in-vessel
storage); 16-refuelling cell; 17-fuel transfer mechanism.

FIG. 7. BN-1600 pool type reactor (project) with in-vessel fuel storage and cold trap.

256
The complicated and labor-consuming operation and maintenance (O&M) technology for the
repair of sodium components and the absence of a guaranty that components will function
after removal of sodium residues (washing) and decontamination necessitate carrying make
out tests of forerunners or representative models of reactor mechanisms and components on
sodium facilities under designed operating conditions. All components manufactured without
defects and tested, have been operated in LMFRs for decades owing to the corrosion inertness
of sodium and the low pressure.
Serious incidents occurred during the cleaning of components from sodium residues because
complete sodium draining has not been enabled at the design stage. This approach was
justified on the initial stage of mastering fast reactors in order to eliminate the probability of
coolant leakage from the drain tanks and components vessels. The presence of non-drained
cavities in the components can be explained by insufficient understanding of the problem at
that stage. Accumulated technological experience provides the basis for studying the
possibility of draining the tanks and the vessels of advanced reactor facilities with
deterministic elimination of unauthorized draining by design.
Oil ingress into the primary circuits of an LMFR is undesirable because of the potential
release of methane gas through the core. This could cause positive reactivity effects and
possible blockage of the fuel subassemblies by solid carbon debris. In future project oil
bearings are probably best avoided. Some advanced pump design was changed following the
PFR and other reactors oil ingress incident by introducing innovative magnetic bearings and
ferro-fluid seals [13] to eliminate oil completely and remove the potential hazard of its ingress
into sodium.

9.3.3. Steels stabilized by titanium

The hot leg pipes around the steam generator up to including the buffer tanks, steam generator
sodium headers and modules were made of 321 steel stabilized by titanium to improve the
mechanical strength at high temperature.

The cracks in the secondary sodium piping elements with this steel (type 321) were observed
at the Phnix reactor. The French expert evaluations showed that the cause of the delayed
reheat crack in the thermally affected (during the welding) zones is the hardening of the steel
due to tine precipitations of titanium carbide inside the crystalline structure resulted the plastic
deformation capacity the periphery of the grains that provoked the crack. This phenomenon
particularly has been observed in the complex geometry parts (T-junction welds, welds
connections with different thickness and walls deviation) [15]. This discussed events were not
observed/examined at other reactors (PFR, BN-600, BN-350), possibly, owing to lower
temperature and thicker wall of the piping (Table 1) as well as using steel 18Cr 9Ni type
without adding Ti (BN-350, BN-600).

TABLE 1. THE LMFRS HOT LEG SECONDARY SODIUM PIPING DATA [14, 15]

Reactor Steel type Temperature, C Pipe diameter, mm Pipe wall thick, mm

Phnix 321 (Ti) 55017 510 6.0
PFR 321 (Ti) 530 2360 10.0
BN-350 18Cr 9Ni 415 529 12.0
BN-600 18Cr 9Ni 515 630 13.0

17
The higher the operating temperatures, the more cracks were found [15]

257
In the future sodium cooled reactor equipment designs the using of the 321 steel stabilized by
titanium are probably best avoided.

It should be pointed out that an invaluable knowledge for future LMFR designs has been
gained on the structural material behavior and component options thanks to Phnixs long
time operation with a high level of temperatures and NPP thermal efficiency.

9.3.4. Steam generators principal design and structure materials

A total of 37 leaks were experienced in the UK prototype fast reactor (PFR) steam generator
units in the period 1974 to 1984. All leaks originated at the butt/fillet welds between the
tubes and the tubeplates (Fig. 8c) [15].

For the first ten years (1974-1983), electrical output of the PFR plant was limited, mainly
because of a series of leaks in the steam generator units, and the highest load factor in any
year was 12% [15].

(a) 'stub/butt (b) butt/fillet (c) PFR SG: butt/fillet weld

FIG. 8. Steam generator tube/tubeplate weld joints.

The stub/butt weld (Figs 8 a and 9), permitted heat treating after manufacture, radiographic
and other inspection of an individual tube to tubeplate joints, should be envisaged to avoid the
risk of stress corrosion cracking and reheat cracking. Austenitic steels are in question for
LMFR steam generators because of the high risk of caustic stress corrosion damage following
even small leaks [15].

Two types of tube/tubeplate weld for the UK PFR SG: typical stub/butt and new butt/fillet
welds were selected for detail consideration (Fig. 8 a-b). The 'butt/fillet' weld was used for the
design (Fig. 8b). It was found that the machining costs for the stub/fillet weld type of joint
were considerably cheaper than for the stub/butt' and thus additionally there was no possible
weakness due to the cross-grain of the machined stub.

The tube to tubeplate weld has been recognised as the most critical item in the robustness of
LMFR steam generators. It was concluded, that the type of direct tube-to-tubeplate weld (the
butt/fillet weld, Fig. 8c) adopted at PFR, which could not be heat treated after manufacture,
should be avoided in future SG [16].

258
 FIG. 9. A typical straight-tube steam generator tube/tubeplate weld joints (EFR) [17].

The design of sodium heated steam generators has been largely changed during the
development of fast reactor technology. Studies were aimed at the creation of a reliable,
low-cost design, which is easily inspected (diagnosed) in the process of operation after a
steam generator unit switch-off. The experience gained during development and operation has
shown that neither micro modular nor macro modular, neither double wall modular steam
generator meets completely these criteria; they are complicated and metal-intensive.
Consideration of capital cost suggests that the units should be few in number and as large as
possible.

Upon experimental confirmation of the long-term resistance of steels such as mono-metallic

modified 9Cr-lMo (Grade 91) steel in sodium, water and steam including wet steam, a
possibility of assuring once-through process in the tube (water heating, boiling, evaporation,
and steam superheating) has appeared. This result, along with sodium replacement with steam
in the reheater, has made it possible to use a single-vessel steam generator, which was first
applied at the Super-Phnix reactor. A steam generator unit of 750 W(th) power, having
welded coil tubes, operated reliably at Super-Phnix. However, it turned out to be
comparatively expensive and complicated as far as the tube coil diagnosis under operating
conditions (after a steam generator unit switch-off) is concerned. It was possible to eliminate
these drawbacks in the single-vessel steam generator [600 W(th)] with straight long tubes
made of 9Cr-lMo (Grade 91) steel, which was designed for the EFR reactor.

REFERENCES

[1] Technical Committee Meeting on Evaluation of radioactive materials release and

sodium fires in fast reactors, 1114 November 1996, O-arai, Japan.
[2] MOUROGOV, V.M., et al., Energy, Vol. 23, No. 7/8 (1998) 637648.
[3] MITENKOV, F.M., Prospect for the development of fast breeder reactors, Atomnaia
Energia, Vol. 92, No. 6 (2002) 423432.
[4] SARAEV, O.M., Operating experience with Beloyarsk fast reactor BN-600,
Unusual Occurrences During LMFR Operation, IAEA-TECDOC-1180, IAEA, Vienna
(2000) 101116.
[5] RAHMANI, L., DECHELETTE, S., BANDINI, C., SPX significant events and
whether it would have happened on EFR, Unusual Occurrences During LMFR
Operation, IAEA-TECDOC-1180, IAEA, Vienna (2000) 57110.

259
[6] INTERNATIONAL ATOMIC ENERGY AGENCY, Status of Liquid Metal Cooled
Fast Reactor Technology, IAEA-TECDOC-1083, IAEA, Vienna (1999).
[7] MARTH, W., The story of the European fast reactor cooperation, KFK 5225,
Karlsruhe, Germany, 1993.
[8] WADE, D.C., Recent innovations in IFR safety research, paper presented in
International Topical Meeting on Advanced Reactor Safety (ARS 94), 1721 April
1994, Pittsburgh, USA.
[9] SLOVIC, G.C., Evaluating Advanced LMR Reactivity Feedbacks Using SSC, paper
presented in Int. Top. Mtg. Safety of Next Generation Power Reactors, 15 May 1988,
Seattle, Washington, USA.
[10] ESSIG, C., Dynamic Behaviour of Rapsodie in Exceptional Transient Experiments,
paper presented in Int. Top. Mtg. Reactor Safety, 1985, Knoxville, USA.
[11] PLANCHON, H., et al., Implications of the EBR-II Inherent Safety Demonstration
Test, paper presented in Conf. Fast Breeder Reactors, 1986, Chalk River, USA.
[12] LUCOFF, D. M., Passive Safety Testing at the FFTF. Proc. Int. Conf. Experience
Gained and Path to Economical Power Generation FBR, 1317 September 1987,
Richland, Vol. 2, ANS (1987) 11.8-1 - 11.8-7.
[13] Project Rapide 1500 MW, published by EdF (1984).
[14] INTERNATIONAL ATOMIC ENERGY AGENCY, Fast Reactor Database,
IAEA-TECDOC-866, IAEA, Vienna (1996).
[15] SAVAGE, J.- F., Phenix 30 years of history: the heart of reactor, CEA/EDF, CEA
Valrno-BP 17171-30207 Bagnols-sur-Ceze cedex, 2004.
[16] CRUICKSHANK, A., JUDD, A.M., Problems experienced during operation of the
prototype fast reactor, Dounrey, 19741994, Unusual Occurrences During LMFR
Operation, IAEA-TECDOC-1180, IAEA, Vienna (2000) 942.
[17] DUMN, K., EFR-600 MW straight tube steam generator, the strategy towards the
definition of a DBA, paper presented in Specialists Mtg. Steam Generator Failure and
Failure Propagation Experience, 2628 September 1990, Aix-en-Provence, France.

260
 ABBREVIATIONS

AHX Air heat exchanger

ALMR Advanced liquid metal reactor
BN-350 Bystrie neytrony (Fast neutrons)
BN-600 Bystrie neytrony (Fast neutrons)
CDA Core disruptive accident
CEFR China experimental fast reactor
CR Conversion ratio
CRC Central rotating column
CRP Central rotating plug
DFBR Demonstration fast breeder reactor
DFR Dounreay fast reactor
EBR-II Experimental breeder reactor-II
EC Emergency condenser
EFPD Effective power day(s)
EFR Experimental fast reactor
FA Fuel assembly
FE Fuel element
FEM Finite elements method
FFTF Fast flux test facility
FSD Fuel storage drum
IFC Irradiated fuel cave
IHX Intermediate heat exchanger
IFR Integral fast reactor
JSFR Japan sodium-cooled fast reactor
KAEC Kazakh atomic energy committee
KALIMER Korean advanced liquid metal reactor
KNK-II Kompakte Natriumgekuhlte Kernreaktoranlage
LMFR Liquid metal cooled fast reactor
LRP Leakage re-injection pump
MINATOM Ministry of Atomic Energy
MPC Maximum permissible concentration
NIV Neutron-induced voidage
OD Outside diameter
PIW Post irradiation examination

261
 PFR Prototype fast reactor
RI Reactor installation
RNL Risley Nuclear Laboratories
SBB Sodium boiler building
SG Steam generator
PSP Primary sodium pump
TC Tight compartment
TD Theoretical density
ULOF Unprotected loss-of-flow
WCR Water-chemical regime

262
 CONTRIBUTORS TO DRAFTING AND REVIEW

Ashurko, Y. Institute of Physics and Power Engineering, Russian Federation

Astegiano, J.- C. Centre dEtude de Cadarache, France

Badulescu, A. International Atomic Energy Agency, Austria

Bhoje, S. B. Indira Gandhi Center for Atomic Research, India

Chetal, S. C. Indira Gandhi Center for Atomic Research, India

Daogang, L. China Institute of Atomic Energy, China

Hahn, D. Korea Atomic Energy Research Institute, Republic of Korea

Haihong, X. China Institute of Atomic Energy, China

Kim, Y. I. Korea Atomic Energy Research Institute, Republic of Korea

Kosilov, A. International Atomic Energy Agency, Austria

Krechetov, S., National Atomic Company Kazatomprom,

Republic of Kazakhstan

Mandl, W. International Atomic Energy Agency, Austria

Nakai, R. Japan Nuclear Cycle Development Institute, Japan

Rineiskii, A. International Atomic Energy Agency, Austria (consultant)

Stanculescu, A. International Atomic Energy Agency, Austria

Xu, Mi China Institute of Atomic Energy, China

Yamaguchi, K. Japan Nuclear Cycle Development Institute, Japan

Yanev, Y. International Atomic Energy Agency, Austria

Yarovitsin, V. Institute of Physics and Power Engineering, Russian Federation

263