Libro Verde

EPRI Underground Transmission Systems
Reference Book
2006 Edition
Place Image Here

EPRI Underground Transmission
Systems Reference Book
2006 Edition
1014840
Final Report, March 2007
EPRI Project Manager

S. Eckroad
ELECTRIC POWER RESEARCH INSTITUTE

3420 Hillview Avenue, Palo Alto, California 94304-1338 ▪ PO Box 10412, Palo Alto, California 94303-0813 ▪ USA
800.313.3774 ▪ 650.855.2121 ▪ askepri@epri.com ▪ www.epri.com
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES
THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN
ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH
INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE
ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM:
(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I)

WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR
SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS
FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR
INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL
PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S
CIRCUMSTANCE; OR
(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER

(INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE
HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR
SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD,
PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.
ORGANIZATION(S) THAT PREPARED THIS DOCUMENT
Electric Power Research Institute
NOTE
For further information about EPRI, call the EPRI Customer Assistance Center at 800.313.3774 or
e-mail askepri@epri.com.
Electric Power Research Institute, EPRI, and TOGETHER…SHAPING THE FUTURE OF ELECTRICITY
are registered service marks of the Electric Power Research Institute, Inc.
Copyright © 2007 Electric Power Research Institute, Inc. All rights reserved.
CITATIONS
This report was prepared by

3420 Hillview Avenue
Palo Alto, CA 94304
Principal Investigator
S. Eckroad
The authors of each chapter of this book are listed with the chapters. This document describes
research sponsored by the Electric Power Research Institute (EPRI).
This publication is a corporate document that should be cited in the literature in the following
manner:
EPRI Underground Transmission Systems Reference Book: 2006 Edition. EPRI, Palo Alto, CA:
2007. 1014840.
Photographs on cover are reprinted courtesy of (clockwise from upper left): South Carolina
Electric & Gas, ITCTransmission, BC Hydro, and Southwire. Center: Prysmian Cables and
Systems.
iii
PRODUCT DESCRIPTION
This report is an updated edition of the Underground Transmission Systems Reference Book,
which was originally published in 1992. Published in the first edition with a green cover, the
book has become commonly known throughout the industry as “the Green Book.” The book
provides a desk and field compendium on the general principles involved in the planning, design,
manufacture, installation design, installation, testing, operation, and maintenance of underground
cable systems.
Results & Findings

This book is a state-of-the-art guide to underground cable systems. The content of the earlier
edition has been significantly expanded and updated. The new edition provides information to
better reflect the latest technology, new materials and methods, recently issued standards and
regulations, and current utility needs and practices. Separate chapters have been developed to
provide detailed information on extruded-dielectric, pipe-type, and self-contained fluid-filled
cables. Special application cables—gas-insulated lines, dc cables and long-length submarine, and
superconducting cables—are given coverage. New chapters have been added on hydraulic design
and grounding and cathodic protection.
In all, information is provided on the history of underground cables, different cable types, basic
design considerations, the construction of cables, cable accessories, hydraulic design and
equipment, grounding and cathodic protection, ampacity, installation design, cable system
construction, testing, operation and maintenance, cable system considerations, and the future of
underground transmission.
Efforts have also been taken to make the book more user friendly, including reorganizing
chapters to make information easier to find and expanding the glossary and index.
Challenges & Objectives

Today companies that own and operate underground transmission systems face a number of
challenges. They must reduce costs, improve system reliability, increase power throughput, and
ensure health and safety. At the same time, many of these companies are losing their core of
experienced cable engineers. Since its publication, the Green Book has provided underground
transmission companies with a valuable tool for meeting these challenges. However, despite the
value and popularity of the Green Book, some elements of the book’s content have become
increasingly outdated over the years since its publication and do not reflect the latest
developments in technology and utility practices. For example, in recent years, there has been an
increased use of extruded dielectric cable at higher voltages, a trend not reflected in the 1992
edition. In addition, there is a need today for the book to incorporate more of an international,
rather than an exclusively U.S., perspective on underground transmission design and use.
v
Applications, Values & Use
The book provides a desk and field compendium on underground cable systems. The new edition
will enable energy companies to simplify the design and reduce the cost of new extruded and
traditional pipe-type cable systems. Although the Green Book compiles a wealth of information,
it is not intended to provide an exhaustive treatment of all subjects associated with the
underground cable industry—more detail is available in the references cited in the book. Instead
the book serves as a handy reference guide, compiling a large amount of information in one
place. For example, design engineers can use the book as a checklist to determine the relevant
design phase issues that must be addressed in the design phase of a project; the tradeoffs are
identified but not necessarily solved. The book can also serve as basis for training new cable
engineers. When used in conjunction with EPRI’s Underground Transmission Workstation, it
enables new cable engineers to come up to speed more quickly, thereby saving time and money.
EPRI Perspective
The Green Book is one of a series of landmark EPRI transmission reference books, first
published in the 1980s and 1990s and currently being updated. These books—on overhead line
transmission, underground transmission, wind-induced conductor motion, and compact line
design—brought together leading experts to compile state-of-science information on the design,
construction, and maintenance of these systems. This much needed revision to a popular and
well used standard reference work brings together the combined wisdom and practical
experience of today’s acknowledged experts in the field and does so in a convenient format that
was designed to meet the needs of utility company underground transmission engineers and
planners, and the consultants who serve them.
Approach
To capture the current state-of-the-art in the field of underground cables, EPRI initiated a three-
year research project. EPRI assembled a team of international experts to review the existing
edition of the Green Book, propose a plan for revision and expansion, and write an updated
edition. The new edition was peer-reviewed by members of the team, EPRI member advisors,
and other experts in the field.
Keywords
Extruded dielectric cables
Pipe-type cables
Self-contained fluid-filled cables
Transmission
Underground transmission
vi
EPRI Underground Transmission Systems Reference Book
Foreword
The Electric Power Research Institute is pleased to AUDIENCE AND SCOPE

present the EPRI Underground Transmission Systems In planning for and producing the 2006 Edition of the
Reference Book–2006 Edition. This book covers all Green Book, EPRI and its utility advisors decided to
stages of cable system design and operation, from ini- generally follow the approach of the earlier edition with
tial planning studies to failure analysis. It contains con- respect to audience and scope. The intended users of
tributions from many of the industry’s experts, and the Green Book are primarily utility UT design engi-
represents practices both in North America and inter- neers and consultants. Addressing issues related to the
nationally. aging workforce and continuing loss of utility expertise,
the Green Book strives to preserve the institutional
Energy companies must achieve a number of objectives knowledge that is being eroded in today’s downsized
to satisfy performance-driven expectations from inves- and restructured industry. Engineers in utility organiza-
tors and customers: cut costs, improve transmission tions with even modest or minimal preparation or
system reliability, increase power throughput, and background in the specifics of underground transmis-
ensure health and safety. EPRI’s Underground Trans- sion systems will benefit from the “tutorial” nature of
mission (UT) Program has historically provided its the book, while experienced professionals will find the
members with valuable research results and guidance technical depth they need to carry out their assign-
for the planning, design, construction, operation, and ments. To make it manageable in size, the book does
maintenance of UT systems. The EPRI Underground not include fundamental derivations and tables, as
Transmission Systems Reference Book–2006 Edition these are widely available in standard textbooks and
continues a valued tradition of providing transmission other references. The minimal technical preparation
engineers and planners with a ready reference to under- needed to fully utilize the book would be third-year col-
ground cable design, as well as theory and practical lege engineering or physics. Finally, recognizing that a
application knowledge for construction, operation, and significant portion of today’s design and manufacturing
maintenance. of transmission cable systems occurs outside of North
America, attempts were made to expand the focus of
This 2006 Edition is a thorough revision and updating the Green Book to include international practice and
of the widely used and popular 1992 Edition of the terminology. To that end, a number of chapters have
same title, commonly known as the “EPRI Green international co-authors or reviewers.
Book” because of its green case binding. In turn, the
1992 Green Book was an outgrowth of the oft-
PROCESS
referenced 1957 Edison Electric Institute Underground
Systems Reference Book. In publishing this revised Production of the 2006 Edition began with a thorough
edition, we recognize that, although many of the basic review of the 1992 Edition. Thirteen UT professionals,
principles and practices described in the 1992 Edition many of them contributors to the earlier edition, per-
are unchanged, there is a need to accurately and more formed individual reviews of each of the 1992 Green
fully reflect the innovations and recent changes in the Book chapters. These reviews focused on identifying
transmission cable industry, as well as the changing outdated or inaccurate text and/or figures, proposing
needs of utility companies that face a business environ- new material to reflect more recent developments, and
ment very different from that of the late 1980s and suggesting overall enhancements. Additionally, some
early 1990s. reviewers provided detailed outlines of entirely new
chapters. This stage was completed at the end of 2004,
and the results were published in the EPRI report
vii
Foreword EPRI Underground Transmission Systems Reference Book
Review of the Underground Transmission Systems Refer- NEW MATERIAL AND ORGANIZATION
ence Book (1008718). An extended three-day workshop A number of important changes from the structure and
held in March 2005 led to consolidation of the com- content of the 1992 Edition were incorporated in the
ments for each chapter and consensus on the required 2006 Edition, some of which are highlighted below.
changes. Results included a detailed description of and Foremost among these was significantly expanded
specification for major revisions required to existing coverage of extruded dielectric cable systems. These
chapters, content and scope of new chapters, new infor- cable systems, at transmission voltages, were in their
mation to be provided across chapters, chapters to be infancy when the 1992 Edition was written. In order to
removed, etc. enhance the coverage of extruded dielectrics as well as
make the content more manageable, the chapter on
In the second phase of the project, expert teams of one cable construction, which in the 1992 Edition covered
to four experts (depending upon chapter complexity all cable types, was expanded into four distinct chapters
and resources) were formed to rewrite each chapter or in the present work: extruded dielectrics, pipe-type, self-
compose new chapters. The text was then indepen- contained fluid-filled, and special applications. The
dently reviewed by other experts as well as utility engi- latter includes gas-insulated line (GIL), dc and subma-
neer members of the EPRI Underground Transmission rine, as well as superconducting cables.
Task Force. Contributors to the book were drawn from
among those who conducted the chapter reviews; how- Another significant enhancement in the 2006 Edition is
ever, others were selected as well, often upon the rec- the addition of two new chapters: Hydraulic Design
ommendation of the reviewers. An effort was made to and Equipment (Chapter 9), and Grounding and
represent all major stakeholders, both domestic and Cathodic Protection (Chapter 10). Although some of
international, on the author/reviewer teams, including the material in these two chapters was covered in other
consultants, engineering and construction companies, chapters in the 1992 Edition, much of it is new.
cable system equipment and services suppliers, and
utilities. An important change in the chapter on Ampacity
(Chapter 11) was the decision to use an approach
A second three-day workshop was held in early 2006 focused on the IEC standard methods—generally rec-
with utility advisors and all authors and reviewers. The ognized as providing better accuracy for ac losses with
structure of the book and the content of each chapter a larger variety of cable systems and being easier to fol-
were finalized, and areas requiring coordination across low—while accounting for cyclic loading by using the
chapters were defined. A publications editor was “loss factor” approach from Neher-McGrath, for any
assigned to the project to coordinate writing/reviewing arbitrary load shape. The IEC standards are also sub-
activities and assist authors with the drafting and layout ject to regular maintenance and improvements.
of each chapter. The editor also secured copyright per-
missions for the many photos and graphics provided by A new Appendix provides an extensive, annotated list
third parties to enhance the book’s effectiveness and of EPRI reports to aid in further exploration of specific
interest. areas of interest. Research conducted by EPRI over the
past two decades has resulted in a significant body of
Writing began in April 2006, and a draft of the book literature on underground transmission systems,
was completed by December 2006. An independent including their design, construction, operation and
“whole book” review by an engineer not associated maintenance, and condition assessment. The results of
with the initial writing and review process was con- much of this work are still valid today, and the reader
ducted in early 2007. This review focused on ensuring will find in these resources additional in-depth informa-
the “unity” of technical depth and content across the tion on many of the topics covered in the Green Book.
17 chapters, particularly where similar or identical Additionally, a Glossary has been added to the 2006
material was discussed in more than one place. The Edition, and the Index has been expanded.
final review helped eliminate many inconsistencies and
overlaps in the book and corrected errors that
remained. A further round of corrections and edits was
undertaken before submitting the text for final layout.
In this way, every effort was made to ensure that the
material is accurate and complete.
viii
EPRI Underground Transmission Systems Reference Book Foreword
ACKNOWLEDGMENTS significantly as both an author and a reviewer. Finally, I

During the three-year course of this project I had the want to express my appreciation for the support and
very distinct pleasure of meeting and interacting with assistance, especially in the early stages of the book’s
many world-renowned transmission cable engineering development, of Walter Zenger, who was the Program
professionals. These men and women sacrificed much Manager of EPRI’s UT Program until May 2006.
of their professional and personal time and put much Last but not least, this book would never have been
of their hearts into the development of this reference published without the incredible efforts of Jonas Weisel
work. They can be rightfully proud of the results. Per- of Jonas Weisel and Associates. Besides his excellent
sonally, I am proud to have had the rich opportunity to work of technical editing and assistance to authors with
work with them and learn from them. I would like to formatting and other textual and graphics problems,
make special note of the contribution of Rachel Mosier Jonas very persistently goaded all of us to get our mate-
of Power Delivery Consultants, who performed the rials in on time and faithfully followed up resolutions
“whole book” review under a very demanding schedule to the many obstacles that we encountered along the
and working at times with difficult text files. The way. Jonas was ably assisted by Lee Lehrman, who laid
authors and reviewers of this book are further acknowl- out the entire book and solved many issues related to
edged by name and organization in the Acknowledg- placement of figures and tables. Brian Fies, an associate
ments section. of Jonas, also helped during peak periods of editorial
Without the financial and technical support of the work. I am most grateful to each of these individuals
EPRI Underground Transmission Task Force (UTTF), and their unnamed team members who ensured that
the project to revise the Green Book would not have this book was published on time.
succeeded. I am grateful for the confidence that the It is my aspiration that this 2006 Edition of the EPRI
members of the UTTF placed in both EPRI as an orga- Underground Transmission Systems Reference Book will
nization and me personally to bring this project to com- serve well the electric utility industry and the men and
pletion, in spite of schedule and budget constraints. In women who energize it for many years, and will come
particular, I would like to thank Mohammad Pasha of to be regarded with at least the same high esteem as its
United Illuminating Company and Ron Knapwurst of predecessor, if not greater.
American Transmission Company for their persistence
in reviewing all the chapters and their assistance in
obtaining both the quality and timeliness desired for Steven W. Eckroad
this product. I also want to acknowledge the chairman Electric Power Research Institute
of the Task Force, Nirmal Singh of Detroit Edison, who Charlotte, North Carolina
supported this project from the outset and contributed USA
ix
Contents EPRI Underground Transmission Systems Reference Book
x
Acknowledgments
This book reflects the efforts of many cable industry Adriano Santini, Consolidated Edison Company
engineers and experts over a period of three years. of New York, Inc.
Thirty expert authors and reviewers took time from Axel Schlumberger, Southwire HV Solutions
their very busy schedules to prepare text; review the
contributions of others; resolve issues of content, com- Nirmal Singh, Detroit Edison Company*
pleteness, and accuracy; and finalize text and figures to Stephen E. Turner, Underground Systems, Inc. (USi)*
produce the highest quality results. These authors and Milan Uzelac, G&W Electric Co.
reviewers are listed below (an asterisk indicates that the
individual also participated in the review of the 1992 Steven P. Walldorf, Underground Systems, Inc. (USi)*
Edition). Brief profiles of the authors also appear at the Thomas B. Welles, Advanced Testing Systems Inc.
start of each chapter. Jay A. Williams, Power Delivery Consultants, Inc.
Pierre Argaut, Silec Cables (PDC)*
Phil Bolin, Mitsubishi Electric Power Products, Inc. Ernesto Zaccone, Prysmian Cables and Systems*
Earle C. (Rusty) Bascom, III, Power Delivery Joseph T. Zimnoch, Consultant
Consultants, Inc. (PDC)*
John H. Cooper, Power Delivery Consultants, Inc. Also assisting in reviewing the 1992 Edition were: Paul
(PDC)* Cinquemani of Prysmian Cables and Systems; Robert
B. Gear and John S. Engelhardt of Underground Sys-
Steven W. Eckroad, Electric Power Research Institute tems, Inc. (USi); and David Notman of Cable Consult-
(EPRI) ing International Limited.
Anthony Ernst, Underground Systems, Inc. (USi)
Stephen J. Galloway, Cable Consulting International Members of the EPRI’s Underground Transmission
Limited Task Force (UTTF), chaired by Nirmal Singh of
Detroit Edison Company, took an active role in guiding
Reza Ghafurian, Consolidated Edison Company and supporting the development of this book. Many of
of New York, Inc.* the UTTF members served on the Green Book Work-
Brian Gregory, Cable Consulting International Limited ing Group that met with authors and reviewers in
Frank L. Kuchta, Prysmian Cables and Systems extended, three-day planning sessions on two occasions
and participated in numerous teleconferences during
Dennis E. Johnson, POWER Engineers, Inc.* the writing of the book. The members of the UTTF
W. Graham Lawson, Energy Cable Consultants, Inc. Green Book Working Group over the three-year course
David Lindsay, Southwire, Inc. of the project were (listed alphabetically by company):
Allen MacPhail, BC Hydro Howard Jones, American Electric Power (AEP)
Rachel I. Mosier, Power Delivery Consultants, Inc. Ron Knapwurst, American Transmission Company
(PDC) (ATC)
Frank Musick, WA Chester, LLC (WAC) Takashi Kojima, BC Hydro
Deepak Parmar, Geotherm, Inc. Richard Raymond, CenterPoint Energy
Mohammad A. Pasha, The United Illuminating Erich F. Schoennagel, CenterPoint Energy

Company Reza Ghafurian, Consolidated Edison Company of
Dave Purnhagen, Underground Systems, Inc. (USi)* New York, Inc.
John S. Rector, Black & Veatch, Inc. (B&V) Nirmal Singh, Detroit Edison Company
Don Koonce, Dominion–Virginia Power
xi
Acknowledgments EPRI Underground Transmission Systems Reference Book
Mark R. Smith, Georgia Power Company Rachel Mosier, Northeast Utilities

Staci Beckwith, Georgia Transmission Company Peter L. Tirinzoni, Northeast Utilities
Vince Curci, Los Angeles Department of Water and Jerry Ruschkofski, South Carolina Electric and Gas
Power (LADWP) Mohammad Pasha, United Illuminating Company
Ed Hahn, New York Power Authority (NYPA)
xii
Preface
This EPRI Underground Transmission Systems Refer- liography for each topic was neither planned nor pre-
ence Book–2006 Edition is intended to serve as a desk sented. For the reader’s convenience, the symbols used
and field compendium of the general principles in the equations in Chapters 11 and 16 are listed and
involved in the planning, design, manufacture, installa- defined at the end of those chapters.
tion design, installation, operation, and maintenance of
underground transmission cable systems. It is not With regard to the units of measure used within, in
intended to provide an exhaustive treatment of all sub- most cases, the units most commonly used by cable
jects associated with the underground transmission engineers are provided first, followed by their equiva-
industry but, rather, a basic understanding of the issues. lents—i.e., English units are followed by SI (metric)
More detailed information is available in the many ref- units in parentheses. Where SI units are more com-
erences cited throughout the text and elsewhere. For monly used, such as in forced-cooling calculations, they
example, the design engineer could use the book as a are presented first, followed by their equivalent English
checklist to determine the relevant design issues that units. Because international engineering approaches
must be addressed in the design phase of a project; the can vary substantially from U.S. practice, “hard metri-
tradeoffs are indicated, but not necessarily solved. The cation” was not utilized. When unit conversions are
many possible variations and contingencies involved in presented, the exactness of the conversion is related to
each individual cable project preclude a more detailed its engineering significance. For example, 10 feet might
approach. be converted to 3 meters in one instance, but 10.00 feet
would be converted to 3.05 meters in another.
The book is organized by chapters, sections, and sub-
sections. Although a common overall structure was To provide users with a handy reference, Section 2.11
attempted, each chapter reflects the unique style and provides a summary of industry specifications and stan-
presentation of its authors. Chapters are self-contained dards related to underground cables, updated to the
and may be read independently, for the most part, but a present time.
conscious effort was made to avoid repetition of con-
tent. During the chapter content planning phase, a Every effort has been made to locate, identify, and give
number of common, or “core,” topics were identified. It credit to the source for previously published material
was recognized that these topics would be germane to (especially figures and tables) used within, even for
more than one chapter but, rather than provide a com- public-domain material. If any sources were not identi-
plete treatment in each chapter, the authors decided fied or credited, it was by inadvertent error. We are very
which chapter should contain the more detailed treat- grateful for the use of all previously published material
ment of a particular core topic. Thus, where needed, and apologize for any omissions.
the book employs cross-referencing to direct the reader
to additional information. Likewise, within a chapter, Comments and suggestions for improving the presenta-
material common to that chapter’s major subsections is tion or content of the book should be addressed to Mr.
treated at the beginning of the chapter rather than Steven W. Eckroad, Electric Power Research Institute,
being repeated throughout. 1300 West WT Harris Boulevard, Charlotte, NC 28262.
These contributions will be addressed prior to the pub-
Figures and tables are numbered sequentially within lication of any revised editions, although none is
each chapter. References are listed immediately follow- planned at present.
ing the text of each chapter and are intended to be as
current and inclusive as possible, but an exhaustive bib- Jonas Weisel
xiii
xiv
Contents
Chapter 1 Historical Perspective 2.10 SELECTING CABLE TYPES . . . . . . . . . . . . . . . . . . . . 2-17

1.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2 System Considerations. . . . . . . . . . . . . . . . . . . . . . . . . 2-18
Installation Considerations . . . . . . . . . . . . . . . . . . . . . . 2-19
1.2 EARLY CABLE SYSTEMS IN EUROPE . . . . . . . . . . . . . 1-2 Considerations for Cable Operation . . . . . . . . . . . . . . . 2-20
Environmental Considerations . . . . . . . . . . . . . . . . . . . 2-21
1.3 BEGINNINGS OF THE U. S. UNDERGROUND
Cost Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-21
CABLE INDUSTRY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-3
2.11 INDUSTRY SPECIFICATIONS AND STANDARDS . . . 2-22
1.4 PROGRESS THROUGH THE 1950s . . . . . . . . . . . . . . . 1-6
Utility Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-22
1.5 PROGRESS THROUGH THE 1960s . . . . . . . . . . . . . . . 1-7 Specification Listing . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-23
Contact Information . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-24
1.6 PROGRESS THROUGH THE 1970s . . . . . . . . . . . . . 1-8
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-26
1.7 PROGRESS THROUGH THE 1980s . . . . . . . . . . . . . . . 1-8
1.8 PROGRESS THROUGH THE 1990s . . . . . . . . . . . . . . 1-10
1.9 PROGRESS THROUGH THE 2000s . . . . . . . . . . . . . . 1-11 Chapter 3 Basic Design Considerations
3.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-2
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-13
3.2 CABLE SYSTEM DESIGN . . . . . . . . . . . . . . . . . . . . . . . 3-2
Circuit Layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-2
Chapter 2 Cable System Types Circuit Cross Sections . . . . . . . . . . . . . . . . . . . . . . . . . . 3-2
Cable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
2.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-3
Accessories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
Features of Underground Cable Systems . . . . . . . . . . . .2-3 Ancillary Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
2.2 CONSIDERATIONS FOR SELECTING AN Hydraulic and Pneumatic Scheme . . . . . . . . . . . . . . . . . 3-5
UNDERGROUND CABLE SYSTEM . . . . . . . . . . . . . . . .2-3 Bonding Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
Thermal Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6
2.3 ELECTRICAL CHARACTERISTIC . . . . . . . . . . . . . . . . .2-4 Thermomechanical Design. . . . . . . . . . . . . . . . . . . . . . . 3-6
Installation Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6
2.4 CABLE COMPONENTS . . . . . . . . . . . . . . . . . . . . . . . . .2-4
Operation, Maintenance, and Testing. . . . . . . . . . . . . . . 3-6
Conductor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-4
Semiconducting Shield . . . . . . . . . . . . . . . . . . . . . . . . . .2-5 3.3 ELECTROMAGNETIC FIELD . . . . . . . . . . . . . . . . . . . . 3-6
Insulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-5 Electrostatic Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
Outer Layers, Shielding, and Sheath . . . . . . . . . . . . . . .2-6 Magnetic Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
2.5 EXTRUDED-DIELECTRIC CABLES. . . . . . . . . . . . . . . .2-7 3.4 POWER RATING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11
2.6 PIPE-TYPE CABLES . . . . . . . . . . . . . . . . . . . . . . . . . . .2-8 Direct Current (DC) Cable Systems . . . . . . . . . . . . . . . 3-12
Alternating Current (AC) Cable Systems . . . . . . . . . . . 3-13
2.7 SELF-CONTAINED FLUID-FILLED CABLE SYSTEMS .2-9
3.5 CURRENT RATING . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-15
2.8 SPECIAL APPLICATION CABLES . . . . . . . . . . . . . . . . 2-11 Continuous Steady-State Rating . . . . . . . . . . . . . . . . . 3-15
Gas-insulated Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-11 Short-Circuit Rating . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-16
Submarine Cables and Dc Cables . . . . . . . . . . . . . . . . 2-11 Emergency Rating . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-17
High-temperature Superconducting Cables . . . . . . . . .2-13
3.6 CABLE DESIGN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-18
2.9 CABLE ACCESSORIES . . . . . . . . . . . . . . . . . . . . . . .2-14 Conductor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-18
Splices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-14 Conductor Shield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-23
Terminations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-15 Insulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-23
Link Boxes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-16 Insulation Shield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-34
Fluid-handling Equipment . . . . . . . . . . . . . . . . . . . . . . .2-16 Ground Return Conductor . . . . . . . . . . . . . . . . . . . . . . 3-34
Pipe-type Cathodic Protection and Grounding . . . . . . .2-17 Jacket . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-36
Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-17 Cable Length . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-37
xv
3.7 CABLE ELECTRICAL PARAMETERS . . . . . . . . . . . . .3-37 4.15 CONSTRUCTION: CONDUCTOR AND INSULATION
Cable as a Two-Port Network . . . . . . . . . . . . . . . . . . . .3-38 SHIELDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-34
Inductance L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-39
4.16 CONSTRUCTION: INSULATION . . . . . . . . . . . . . . . . . 4-36
Conductor Resistance, R . . . . . . . . . . . . . . . . . . . . . . .3-43
Positive, Negative, and Zero Phase Impedances . . . . .3-46 Insulation Thickness . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-36
Characteristic Impedance . . . . . . . . . . . . . . . . . . . . . . .3-46 Calculation Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-36
Selection Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-37
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-47
4.17 CABLE CONSTRUCTION: OUTER LAYERS. . . . . . . . 4-39
Bedding Tapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-39
Water-blocking Tapes . . . . . . . . . . . . . . . . . . . . . . . . . . 4-40
Chapter 4 Cable Construction: Extruded Dielectrics Metallic Shields and Sheaths . . . . . . . . . . . . . . . . . . . . 4-41
4.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-3 Shield Conductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-42
Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-3 Wet Designs of Cable with Metallic Shields . . . . . . . . . 4-42
Technological Position of Extruded Cables . . . . . . . . . . .4-3 Metallic Sheaths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-43
Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-4 Other Sheath Considerations for Extruded
Dielectric Cables . . . . . . . . . . . . . . . . . . . . . . . . . 4-46
4.2 PE CABLE: BENEFITS AND CHALLENGES . . . . . . . . .4-5
4.18 CONSTRUCTION: JACKETS. . . . . . . . . . . . . . . . . . . . 4-47
Benefits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-5
Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-5 4.19 MANUFACTURE OF COMPOUNDS . . . . . . . . . . . . . . 4-49
4.3 GENERAL DESCRIPTION OF POLYETHYLENE Manufacture of XLPE Insulating Compound. . . . . . . . . 4-49
INSULATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-5 Manufacture of Semiconducting Compound. . . . . . . . . 4-50
4.4 LANDMARKS IN PE CABLE TECHNOLOGY. . . . . . . . .4-9 4.20 MANUFACTURE OF EXTRUDED CABLE CORE . . . . 4-51
Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-9 Extruders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-51
Water Trees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-12 Types of CV Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-54
VCV (Vertical Continuous Vulcanization) . . . . . . . . . . . 4-55
4.5 XLPE CABLE: BENEFITS AND CHALLENGES . . . . . .4-13 CCV (Catenary Continuous Vulcanization)-Nitrogen . . 4-57
Benefits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-13 CCV (Catenary Continuous Vulcanization)-
Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-13 Silicone Oil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-59
MDCV (Mitsubishi Dainichi Continuous Vulcanization). 4-59
4.6 GENERAL DESCRIPTION OF XLPE INSULATION . . .4-13 Degassing Process. . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-60
Chemical Crosslinking Process. . . . . . . . . . . . . . . . . . .4-15
4.21 QUALITY CONTROL . . . . . . . . . . . . . . . . . . . . . . . . . . 4-60
4.7 LANDMARK DEVELOPMENTS IN XLPE CABLE Materials Inspection (XLPE and EPR) . . . . . . . . . . . . . 4-61
TECHNOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-17 Materials Inspection Specific to EPR . . . . . . . . . . . . . . 4-62
Worldwide Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . .4-17 Quality Control of CV line . . . . . . . . . . . . . . . . . . . . . . . 4-63
Japanese Landmarks in XLPE System Technology . . .4-18 Inspection of Cable Insulation . . . . . . . . . . . . . . . . . . . 4-64
Commercial Applications of XLPE Cable Systems . . . .4-19 Tests on Extruded Core . . . . . . . . . . . . . . . . . . . . . . . . 4-67
Electrical Tests on Main Insulation . . . . . . . . . . . . . . . . 4-68
4.8 ELECTRICAL DESIGN STRESSES FOR XLPE
SYSTEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-21 4.22 SHIPPING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-69
4.9 PROPERTIES OF XLPE CABLE INSULATION . . . . . .4-23 4.23 DESIGN STRESS AND SERVICE LIFE. . . . . . . . . . . . 4-69
Electrical Breakdown Mechanism . . . . . . . . . . . . . . . . .4-23 Factors Affecting Design Stress . . . . . . . . . . . . . . . . . . 4-70
Electrical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . .4-24 Comparison with Other Manufacturers’ Experience . . . 4-71
Physical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-25 Mechanistic Approach to Design . . . . . . . . . . . . . . . . . 4-71
Thermal Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-25 Statistical Breakdown Testing . . . . . . . . . . . . . . . . . . . 4-72
Endurance Tests on Cables . . . . . . . . . . . . . . . . . . . . . 4-73
4.10 EPR CABLE: BENEFITS AND CHALLENGES. . . . . . .4-26
Endurance Estimation on Full-Sized Cable Systems,
Benefits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-26 CSEM Method . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-73
Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-27
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-75
4.11 GENERAL DESCRIPTION OF EPR CABLE . . . . . . . .4-27
Formulation of EPR. . . . . . . . . . . . . . . . . . . . . . . . . . . .4-28
Mixing of High-Voltage EPR Insulation Compounds . . .4-29
Cable Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . .4-29
Chapter 5 Cable Construction: Pipe-Type
5.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3
4.12 LANDMARK DEVELOPMENTS IN EPR CABLE . . . . .4-30
General Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3
4.13 PROPERTIES OF EPR INSULATION. . . . . . . . . . . . . .4-31
5.2 CONDUCTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3
EPR Electrical Breakdown Mechanism . . . . . . . . . . . .4-31
EPR Electrical Properties . . . . . . . . . . . . . . . . . . . . . . .4-31 5.3 SHIELDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-5
EPR Physical Properties . . . . . . . . . . . . . . . . . . . . . . .4-31 Conductor Shield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-5
EPR Thermal Properties . . . . . . . . . . . . . . . . . . . . . . . .4-32 Insulation Shield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-5
4.14 CONSTRUCTION: CONDUCTOR . . . . . . . . . . . . . . . .4-33
xvi
EPRI Underground Transmission Systems Reference Book Contents
5.4 INSULATION AND IMPREGNANT . . . . . . . . . . . . . . . . .5-5 6.16 CABLE FLUID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-16

Background—Paper and Laminated Paper Polypropylene
6.17 SPECIAL CONSTRUCTIONS FOR
Insulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-5
SUBMARINE USE . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-17
Impregnated Paper . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-7
Laminated Paper Polypropylene . . . . . . . . . . . . . . . . . .5-10 6.18 SELF-CONTAINED FLUID-FILLED CABLES FOR
Cable Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-10 OPERATION ON DC . . . . . . . . . . . . . . . . . . . . . . . . . . 6-18
Insulation Electrical Properties . . . . . . . . . . . . . . . . . . .5-12
Physical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-16 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-21
Breakdown Mechanisms . . . . . . . . . . . . . . . . . . . . . . . .5-17
Impregnating Liquid . . . . . . . . . . . . . . . . . . . . . . . . . . .5-17
Chapter 7 Cable Construction: Special Applications
5.5 CABLE OUTER LAYERS AND FILLING MEDIUM . . . .5-18
Outer Layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-18 7.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-3
Pipe-Filling Medium. . . . . . . . . . . . . . . . . . . . . . . . . . . .5-20 7.2 GAS-INSULATED LINES (GIL) . . . . . . . . . . . . . . . . . . . 7-3
5.6 QUALITY CONTROL . . . . . . . . . . . . . . . . . . . . . . . . . .5-29 SF6 Gas and SF6-N2 Mixtures for High-Voltage
Cable Manufacturing Quality Control . . . . . . . . . . . . . .5-29 Insulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-3
Quality Control of Dielectric Liquids . . . . . . . . . . . . . . .5-30 Epoxy Spacers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-5
Enclosures, Enclosure Joints, Mechanical Design, and
5.7 SPECIFICATIONS AND STANDARDS RELEVANT TO Corrosion Protection . . . . . . . . . . . . . . . . . . . . . . . . 7-6
PIPE-TYPE CABLES . . . . . . . . . . . . . . . . . . . . . . . . . .5-32 High-Voltage Conductor and Conductor Joints. . . . . . . . 7-7
AEIC Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . .5-32 Factory-assembled and -Tested GIL. . . . . . . . . . . . . . . . 7-7
ASTM Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-32 Field-assembled and -Tested GIL. . . . . . . . . . . . . . . . . . 7-8
IEEE Standards. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-32 Field Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-8
Operation and Maintenance . . . . . . . . . . . . . . . . . . . . . 7-8
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-33
7.3 DC CABLES AND LONG-LENGTH SUBMARINE . . . . . 7-9
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-9
Chapter 6 Cable Construction: A Review of Worldwide Submarine Cable Experience . 7-11
Self-Contained Fluid-filled A Review of HVDC Underground Cable Experience . . 7-13
HV Submarine and Dc Cable System Design . . . . . . . 7-14
6.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-2 HVDC and Submarine Cable System Tests . . . . . . . . . 7-20
6.2 SYSTEM DEVELOPMENT. . . . . . . . . . . . . . . . . . . . . . .6-2 Reliability in Service . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-22
SCFF Cable Landmarks . . . . . . . . . . . . . . . . . . . . . . . . .6-2 7.4 SUPERCONDUCTING CABLES . . . . . . . . . . . . . . . . . 7-24
Description of Design . . . . . . . . . . . . . . . . . . . . . . . . . . .6-4 Superconducting Materials and Cable Systems . . . . . . 7-24
6.3 CONDUCTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-5 HTS Cables—Design Basics . . . . . . . . . . . . . . . . . . . . 7-26
Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-29
6.4 CONDUCTOR SHIELD. . . . . . . . . . . . . . . . . . . . . . . . . .6-6 Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-32
Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-33
6.5 INSULATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-6
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-34
6.6 INSULATION SHIELD. . . . . . . . . . . . . . . . . . . . . . . . . . .6-8
6.7 LAYING UP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-9
6.8 DRYING AND IMPREGNATION . . . . . . . . . . . . . . . . . .6-10 Chapter 8 Cable Systems Accessories

6.9 METAL SHEATHING . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-11 8.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-3
6.10 REINFORCEMENT. . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-11 8.2 JOINTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-3

Theory and General Description . . . . . . . . . . . . . . . . . . 8-3
6.11 JACKETING OR OVERSHEATHING . . . . . . . . . . . . . .6-12 Joints for Extruded-Dielectric Cables . . . . . . . . . . . . . 8-11
6.12 ROUTINE TESTING . . . . . . . . . . . . . . . . . . . . . . . . . . .6-12 Pipe-Type Joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-20
Self-Contained Fluid-Filled Joints . . . . . . . . . . . . . . . . 8-34
6.13 SHIPPING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-12 Industry Specifications and Standards . . . . . . . . . . . . 8-38
6.14 ELECTRICAL TESTS APPLICABLE TO 8.3 TERMINATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-41
FLUID-FILLED CABLES . . . . . . . . . . . . . . . . . . . . . . . .6-13 Theory and General Description . . . . . . . . . . . . . . . . . 8-41
High-Voltage Withstand Test . . . . . . . . . . . . . . . . . . . . .6-13 Terminations for Extruded-dielectric Cable . . . . . . . . . 8-45
Measurement of Dielectric Loss Angle . . . . . . . . . . . . .6-13 Pipe-type Cable Terminations . . . . . . . . . . . . . . . . . . . 8-51
Self-Contained Fluid-Filled Terminations . . . . . . . . . . 8-63
6.15 HYDRAULIC AND FLUID QUALITY TESTS. . . . . . . . .6-13
Industry Specifications and Standards . . . . . . . . . . . . 8-66
Expulsion Test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-13
Impregnation Coefficient Test . . . . . . . . . . . . . . . . . . . .6-14 8.4 OTHER ACCESSORIES . . . . . . . . . . . . . . . . . . . . . . 8-66
Flow Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-14 Sheath Bonding, Sectionalizing, Link Boxes, and Sheath
Residual Gas Pressure Test . . . . . . . . . . . . . . . . . . . . .6-15 Voltage Limiters . . . . . . . . . . . . . . . . . . . . . . . . . . 8-66
Fluid Cleanliness Test . . . . . . . . . . . . . . . . . . . . . . . . . .6-15
Fluid Electrical Tests . . . . . . . . . . . . . . . . . . . . . . . . . . .6-16 REFERENCES 8-70
xvii
Chapter 9 Hydraulic Design and Equipment Chapter 11 Ampacity

9.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-2 11.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2
9.2 ENGINEERING DESIGN PRINCIPLES . . . . . . . . . . . . .9-2 Ampacity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2

Cable System Pressure Limits. . . . . . . . . . . . . . . . . . . . .9-2 11.2 STEADY-STATE AMPACITY CALCULATIONS. . . . . . 11-2
Basic Hydraulic Equations . . . . . . . . . . . . . . . . . . . . . . .9-3 Calculation Principles . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2
Thermal and Hydraulic Driving Forces . . . . . . . . . . . . . .9-4 Utility System Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-7
9.3 HYDRAULIC CALCULATIONS . . . . . . . . . . . . . . . . . . . .9-5 Cable Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-9
Installation Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-11
Expulsion Calculation . . . . . . . . . . . . . . . . . . . . . . . . . . .9-5
Calculation Procedure . . . . . . . . . . . . . . . . . . . . . . . . 11-14
Demand Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . .9-5
Cables in Air and Tunnel. . . . . . . . . . . . . . . . . . . . . . . 11-40
Pressure Drop Due to Flow . . . . . . . . . . . . . . . . . . . . . . .9-8 Ampacity for Trenchless Installations . . . . . . . . . . . . . 11-43
9.4 PRESSURIZATION EQUIPMENT. . . . . . . . . . . . . . . . . .9-8 Ampacity Implications for Crossing External
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-8 Heat Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-45
Fault Current Calculations . . . . . . . . . . . . . . . . . . . . . 11-46
Pressurization Plants (HPFF /SCFF Cable Systems) . . .9-9
Reservoir Systems (SCFF Cables) . . . . . . . . . . . . . . . .9-15 11.3 TRANSIENT RATING AND TEMPERATURE
Gas Cabinets (HPGF cables) . . . . . . . . . . . . . . . . . . . .9-17 CALCULATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-48
Computing Reservoir Volume . . . . . . . . . . . . . . . . . . . .9-17
Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-48
9.5 FORCED-COOLING EQUIPMENT. . . . . . . . . . . . . . . .9-18 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-48
Considerations for Distributed vs. Lumped Thermal
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-18
System Components . . . . . . . . . . . . . . . . . . . . . . . . . . .9-18 Capacitance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-48
Selection of Solution Method . . . . . . . . . . . . . . . . . . . 11-48
Overall Cooling Plant Design . . . . . . . . . . . . . . . . . . . .9-21
Multisection R-C Network Method . . . . . . . . . . . . . . . 11-49
9.6 SYSTEM DESIGN CONSIDERATIONS . . . . . . . . . . . .9-23 Composite R-C/Exponential Integral Method . . . . . . . 11-50
Pressurization Plants . . . . . . . . . . . . . . . . . . . . . . . . . .9-23 Special Topics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-56
Forced-Cooled Systems . . . . . . . . . . . . . . . . . . . . . . . .9-26 11.4 FORCED COOLING . . . . . . . . . . . . . . . . . . . . . . . . . . 11-56
9.7 WORKED EXAMPLE . . . . . . . . . . . . . . . . . . . . . . . . . .9-28 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-56
Units, Nomenclature, and Conventions . . . . . . . . . . . 11-57
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9-29
Calculating Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-57
Thermal-Hydraulic Characterization . . . . . . . . . . . . . . 11-58
Axial Thermal Calculations . . . . . . . . . . . . . . . . . . . . . 11-61
Chapter 10 Grounding and Cathodic Protection Pressure-Loss Calculations . . . . . . . . . . . . . . . . . . . . 11-63
Alternative Calculation Approach . . . . . . . . . . . . . . . . 11-64
10.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-3
11.5 AMPACITY VERIFICATION AND UPRATING . . . . . . 11-65
10.2 GROUNDING OF SINGLE-CORE CABLES . . . . . . . . .10-3
Ampacity Verification . . . . . . . . . . . . . . . . . . . . . . . . . 11-65
Sheath Bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-3
Uprating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-66
Structure Grounding . . . . . . . . . . . . . . . . . . . . . . . . . . .10-8
Design Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . .10-9 11.6 WORKED EXAMPLES . . . . . . . . . . . . . . . . . . . . . . . . 11-69
10.3 CORROSION AND CATHODIC PROTECTION . . . . . 10-12 Example 1 – 230-kV Crosslinked Polyethylene Cable
Normal Rating . . . . . . . . . . . . . . . . . . . . . . . . . . 11-69
Corrosion Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-12
Example 2 – 230-kV Crosslinked Polyethylene Cable
Corrosion Coordinating Committees . . . . . . . . . . . . . . 10-15
Emergency Rating . . . . . . . . . . . . . . . . . . . . . . . 11-78
Corrosion Prevention . . . . . . . . . . . . . . . . . . . . . . . . . . 10-15
Example 3 – Calculation of Shield/Sheath Fault Current
Cathodic Protection Criteria . . . . . . . . . . . . . . . . . . . . 10-16
Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-81
Applying Cathodic Protection to Pipe-type Cables . . . 10-16
Example 4 – 345-kV Pipe-Type Cable Normal Rating 11-82
Stray Current Mitigation. . . . . . . . . . . . . . . . . . . . . . . .10-23
Pipe to Soil Potential Surveys . . . . . . . . . . . . . . . . . . .10-26 11.7 NOMENCLATURE . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-89
Coatings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-27
Disbonded Coating . . . . . . . . . . . . . . . . . . . . . . . . . . .10-28 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-95
Atmospheric Corrosion . . . . . . . . . . . . . . . . . . . . . . . .10-29
Casings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-30
Lead-Sheathed Cable . . . . . . . . . . . . . . . . . . . . . . . . .10-31
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10-33
xviii
Chapter 12 Installation Design 13.5 DIRECT-BURIED CABLE INSTALLATION . . . . . . . . . 13-16

12.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12-3 Trench Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-16
Cable Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-16
12.2 COMMON DESIGN CONSIDERATIONS . . . . . . . . . . .12-3 Direct-Buried Cable Joints . . . . . . . . . . . . . . . . . . . . . 13-17
Cable-Type Selection . . . . . . . . . . . . . . . . . . . . . . . . . .12-3
13.6 HIGH-PRESSURE FLUID-FILLED (HPFF) PIPE
Route Considerations . . . . . . . . . . . . . . . . . . . . . . . . . .12-3
INSTALLATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-17
Surveying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12-6
Rerouting Existing Lines . . . . . . . . . . . . . . . . . . . . . . . .12-7 Pipe Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-17
Overhead-to-Underground Transition . . . . . . . . . . . . . .12-7 Corrosion Coatings and Repairs. . . . . . . . . . . . . . . . . 13-20
Soil Thermal Properties and Special Backfill . . . . . . . .12-9 Electrical Tests on Pipe Coatings . . . . . . . . . . . . . . . . 13-20
Mechanical Considerations . . . . . . . . . . . . . . . . . . . . 12-17 Typical Work Area. . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-21
Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-19 Field Bends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-21
Grounding/Cathodic Protection . . . . . . . . . . . . . . . . . .12-21 Quality Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-22
Installing Pipe in Tunnels or Casings and on Bridges. 13-23
12.3 INSTALLATION MODES . . . . . . . . . . . . . . . . . . . . . . .12-21 Installing Pipe in Manholes. . . . . . . . . . . . . . . . . . . . . 13-23
Splices: Directly Buried vs. in Manholes . . . . . . . . . .12-21 Installing Termination Assemblies . . . . . . . . . . . . . . . 13-24
Trenching. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12-22 Cleaning and Testing the Installed Pipe System. . . . . 13-24
Trenchless Installations . . . . . . . . . . . . . . . . . . . . . . . .12-25
13.7 BACKFILLING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-25
Bridges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12-31
Underwater Installations . . . . . . . . . . . . . . . . . . . . . . .12-32 Placing of Backfill Material . . . . . . . . . . . . . . . . . . . . . 13-25
Surface Restoration . . . . . . . . . . . . . . . . . . . . . . . . . . 13-26
12.4 CABLE SYSTEM DESIGN CALCULATIONS . . . . . . .12-38
13.8 CABLE INSTALLATION . . . . . . . . . . . . . . . . . . . . . . . 13-26
Steady-State, Transient, and Emergency Ampacity . .12-38
Cable-pulling Tension and Sidewall Pressure . . . . . . .12-39 Storage of Cable. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-27
Extruded-Dielectric Cable Installation. . . . . . . . . . . . . 13-27
12.5 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12-46 Pipe-type Cable Installation . . . . . . . . . . . . . . . . . . . . 13-28
Self-Contained Fluid-Filled Cable Installation. . . . . . . 13-31
13.9 SPLICING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-31
Chapter 13 Cable System Construction Extruded-Dielectric Cables . . . . . . . . . . . . . . . . . . . . . 13-32
13.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13-3 High-Pressure Pipe-Type Cable Splice. . . . . . . . . . . . 13-32
General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13-3 Self-Contained Fluid-Filled Cables . . . . . . . . . . . . . . . 13-33
Responsibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13-3 Link/Grounding Box . . . . . . . . . . . . . . . . . . . . . . . . . . 13-33
Project Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13-3
13.10 TERMINATIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-33
Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13-3
Quality Control/Inspection of Work . . . . . . . . . . . . . . . .13-4 Extruded-Dielectric Cables . . . . . . . . . . . . . . . . . . . . . 13-34
Contractor Licensing . . . . . . . . . . . . . . . . . . . . . . . . . . .13-4 Pipe-Type Termination . . . . . . . . . . . . . . . . . . . . . . . . 13-34
Permitting/Security . . . . . . . . . . . . . . . . . . . . . . . . . . . .13-4 Self-Contained Fluid-Filled Termination . . . . . . . . . . . 13-35
Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13-4 13.11 DIELECTRIC FLUID INSTALLATIONS. . . . . . . . . . . . 13-35
13.2 TRENCHING AND MANHOLES . . . . . . . . . . . . . . . . . .13-4 HPFF Cables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-35
Survey of Route . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13-5 SCFF Cables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-35
Excavating the Trench . . . . . . . . . . . . . . . . . . . . . . . . . .13-6 13.12 NITROGEN GAS INSTALLATION . . . . . . . . . . . . . . . 13-35
Water Removal/Well Pointing . . . . . . . . . . . . . . . . . . . .13-6
Shoring/Sheeting . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13-7 13.13 INSTALLATION OF PRESSURIZATION PLANTS . . . 13-36
Traffic Control/Protective Barriers and Signs . . . . . . . . .13-7 High-Pressure Fluid-Filled Pressurization Plants . . . . 13-36
Bed of Trench . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13-8 Gas Cabinets for High-Pressure Gas-Filled Systems . 13-37
Backfilling the Trench . . . . . . . . . . . . . . . . . . . . . . . . . .13-8 Self-Contained Fluid-Filled Pressurization Cabinet . . 13-37
Street and Highway Crossings . . . . . . . . . . . . . . . . . . .13-9
Manhole Installation . . . . . . . . . . . . . . . . . . . . . . . . . . 13-10 13.14 AFTER INSTALLATION TESTS . . . . . . . . . . . . . . . . . 13-37
Extruded-Dielectric Cable Tests . . . . . . . . . . . . . . . . . 13-37
13.3 TUNNELING, JACK-AND-BORE, AND DIRECTIONAL High-Pressure Pipe-Type and SCFF Tests . . . . . . . . . 13-38
DRILLING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-11
Tunneling and Microtunneling . . . . . . . . . . . . . . . . . . . 13-11 13.15 INSTALLATION OF PIPE FREEZES . . . . . . . . . . . . . 13-39
Jack-and-Bore Installations . . . . . . . . . . . . . . . . . . . . . 13-12 13.16 INSTALLATION OF FIBER-OPTIC CABLES . . . . . . . 13-39
Horizontal Directional Drill Installations. . . . . . . . . . . . 13-13
13.17 SUBMARINE CABLE INSTALLATION . . . . . . . . . . . . 13-39
13.4 CONDUIT INSTALLATION FOR SCFF AND EXTRUDED-
Installing Extruded-Dielectric and Self-Contained
DIELECTRIC CABLES . . . . . . . . . . . . . . . . . . . . . . . . 13-14
Fluid-Filled Cables . . . . . . . . . . . . . . . . . . . . . . . 13-40
Joining Sections of Conduit. . . . . . . . . . . . . . . . . . . . . 13-14 Installing Pipe-Type Cables . . . . . . . . . . . . . . . . . . . . 13-42
Field Bends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-15 Repair of Damaged Cables . . . . . . . . . . . . . . . . . . . . 13-42
Proof Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-15
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-43
xix
Chapter 14 Testing Cable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-5

Dielectric Fluid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-6
14.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-2
Joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-7
General Testing Categories. . . . . . . . . . . . . . . . . . . . . .14-2 Terminations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-7
Test Standards and Guides. . . . . . . . . . . . . . . . . . . . . .14-4 Pressurization System . . . . . . . . . . . . . . . . . . . . . . . . . 15-8
Standards Organizations. . . . . . . . . . . . . . . . . . . . . . . .14-4
15.4 ELECTRICAL FAULT LOCATION . . . . . . . . . . . . . . . . 15-10
14.2 HIGH-VOLTAGE TESTS AND MEASUREMENT
Electrical Faults on Underground
PROCEDURES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-5
Transmission Circuits . . . . . . . . . . . . . . . . . . . . . 15-10
AC Voltage Withstand and Step-Breakdown Tests . . . .14-5 General Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-10
Dc Hi-Potential Tests . . . . . . . . . . . . . . . . . . . . . . . . . . .14-7 Pinpointing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-11
Impulse and Switching Surge Tests . . . . . . . . . . . . . . .14-7 Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-11
Insulation Dissipation Factor Test . . . . . . . . . . . . . . . . .14-8
Factory Partial-discharge Measurements . . . . . . . . . . .14-9 15.5 ELECTRICAL FAILURE REPAIR . . . . . . . . . . . . . . . . 15-12
Field Partial-discharge Measurements . . . . . . . . . . . .14-12 Extruded-Dielectric . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-12
Pipe-Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-13
14.3 TYPE/QUALIFICATION TESTS AND FACTORY
Self-Contained Fluid-Filled . . . . . . . . . . . . . . . . . . . . . 15-16
PRODUCTION TESTS . . . . . . . . . . . . . . . . . . . . . . . .14-16
Qualification/Type Tests for Extruded-Dielectric Cables 15.6 FAULT REPAIR MATERIAL . . . . . . . . . . . . . . . . . . . . 15-17
(ICEA S-108-720-2004) . . . . . . . . . . . . . . . . . . .14-16 Extruded-Dielectric Cable. . . . . . . . . . . . . . . . . . . . . . 15-17
Production Tests for Extruded-Dielectric Cables Pipe-type Cable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-17
(AEIC CS9-06 and ICEA S-108-720-2004). . . . .14-17 Self-Contained Fluid-Filled Cable . . . . . . . . . . . . . . . . 15-18
Routine, Special, Type, and Prequalification Tests for
Extruded Cable Systems 15.7 LEAK MANAGEMENT TECHNIQUES FOR
(IEC 60840 and IEC 62067) . . . . . . . . . . . . . . . .14-18 FLUID-FILLED FEEDERS . . . . . . . . . . . . . . . . . . . . . 15-18
Qualification Tests for Pipe-type Cables Background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-18
(AEIC CS2-97) . . . . . . . . . . . . . . . . . . . . . . . . . .14-20 Causes of Leaks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-19
Production Tests for Pipe-type Cables Leak Detection Techniques . . . . . . . . . . . . . . . . . . . . 15-19
(AEIC CS2-97) . . . . . . . . . . . . . . . . . . . . . . . . . .14-21 Leak Location Techniques . . . . . . . . . . . . . . . . . . . . . 15-20
Qualification Tests for Compressed-Gas-Insulated Pinpointing Methods . . . . . . . . . . . . . . . . . . . . . . . . . . 15-21
Cables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-22 PFT Leak Location Method . . . . . . . . . . . . . . . . . . . . 15-21
Tests for Cable Joints . . . . . . . . . . . . . . . . . . . . . . . . .14-22 Field Trial Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-22
Tests for Cable Terminations . . . . . . . . . . . . . . . . . . . .14-23 Leak Prevention Methods . . . . . . . . . . . . . . . . . . . . . . 15-23
14.4 AFTER-INSTALLATION TESTS . . . . . . . . . . . . . . . . .14-24 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-24

Extruded-dielectric Cable After-installation Tests . . . .14-25
HPFF and HPGF Cable After-installation Tests. . . . . .14-26
SCFF Cable System After-installation Tests . . . . . . . .14-26 Chapter 16 Cable System Considerations
Sheath-bonding Verification Tests. . . . . . . . . . . . . . . .14-27
16.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-2
Series Impedance Measurements . . . . . . . . . . . . . . .14-27
16.2 IMPEDANCE CALCULATIONS . . . . . . . . . . . . . . . . . . 16-2
14.5 FIELD MAINTENANCE/PROOF TESTS . . . . . . . . . . .14-28
Shunt Impedance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-3
Electrical Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-28
Series Symmetrical Component Impedances . . . . . . . 16-4
Other Maintenance Tests . . . . . . . . . . . . . . . . . . . . . .14-29
Single-Conductor Cable Symmetrical Component
14.6 SPECIAL-PURPOSE TESTS . . . . . . . . . . . . . . . . . . .14-29 Series Impedance Formulas . . . . . . . . . . . . . . . . 16-4
Long-Term Accelerated Life Test . . . . . . . . . . . . . . . .14-29 Generalized Calculation Method for Symmetric Component
Mechanical Bending Tests . . . . . . . . . . . . . . . . . . . . .14-30 Series Impedance . . . . . . . . . . . . . . . . . . . . . . . . 16-7
Thermomechanical Bending Tests . . . . . . . . . . . . . . .14-30 Worked Examples—Single-Conductor Cable
Longitudinal Water Penetration Tests Sequence Impedances . . . . . . . . . . . . . . . . . . . . 16-9
(ICEA T-34-664) . . . . . . . . . . . . . . . . . . . . . . . . .14-30 Symmetrical Component Series Impedances for
Pipe-Type Cables . . . . . . . . . . . . . . . . . . . . . . . . 16-13
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14-32
16.3 INTEGRATION OF CABLES INTO THE
UTILITY SYSTEM. . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-16
Capacitance and Reactance Effects. . . . . . . . . . . . . . 16-16
Chapter 15 Operation and Maintenance
Reactive Compensation . . . . . . . . . . . . . . . . . . . . . . . 16-17
15.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15-2 Load Sharing and Power Flows . . . . . . . . . . . . . . . . . 16-20
Hybrid Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-22
15.2 SYSTEM OPERATIONS . . . . . . . . . . . . . . . . . . . . . . . .15-2
System Restoration . . . . . . . . . . . . . . . . . . . . . . . . . . 16-23
15.3 ROUTINE MAINTENANCE. . . . . . . . . . . . . . . . . . . . . .15-2 System Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-23
Cable Installed in Pipe or Duct . . . . . . . . . . . . . . . . . . .15-3 Harmonic Resonance. . . . . . . . . . . . . . . . . . . . . . . . . 16-24
Pipe Coating Damage . . . . . . . . . . . . . . . . . . . . . . . . . .15-3 Insulation Coordination. . . . . . . . . . . . . . . . . . . . . . . . 16-25
Manhole and Joint Casings. . . . . . . . . . . . . . . . . . . . . .15-4 System Protection. . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-29
xx
16.4 MAGNETIC FIELDS . . . . . . . . . . . . . . . . . . . . . . . . . .16-31 17.4 RESEARCH IN CABLE INSTALLATION DESIGN . . . . 17-6
Cables without Ferromagnetic Components . . . . . . . .16-31 Construction Cost Reduction . . . . . . . . . . . . . . . . . . . . 17-6
Cables with Ferromagnetic Components . . . . . . . . . .16-34 Subsurface Structures Detection . . . . . . . . . . . . . . . . . 17-6
Magnetic Field Management Methods . . . . . . . . . . . .16-36
17.5 FUTURE CABLE SYSTEMS—HV/EHV DC CABLES . 17-6
16.5 INDUCTIVE COORDINATION . . . . . . . . . . . . . . . . . .16-41 The SuperGrid Vision . . . . . . . . . . . . . . . . . . . . . . . . . . 17-6
Industry Standards and Guides for Inductive
Coordination . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-42 17.6 SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-7
Calculation of Induced Voltages and Currents . . . . . .16-43 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-8
Inductive Coordination with Pipelines . . . . . . . . . . . . .16-44
Inductive Coordination with Fences . . . . . . . . . . . . . .16-48
Inductive Coordination with Communications Lines . .16-49
Railway System Inductive Coordination . . . . . . . . . . .16-51 Appendix 1 EPRI Reports Related to Underground
Transmission Systems . . . . . . . . . . . . A1-1
16.6 NOMENCLATURE . . . . . . . . . . . . . . . . . . . . . . . . . . .16-53
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16-55
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .G-1
Chapter 17 The Future of Underground Transmission

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .I-1
17.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-3
17.2 CURRENT RESEARCH IN CABLE TECHNOLOGY . . 17-3
Dielectrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-3
Dry-type Terminations and High-Stress Splices . . . . . . 17-4
Superconducting Cables . . . . . . . . . . . . . . . . . . . . . . . . 17-4
17.3 IMPROVEMENTS IN DESIGN AND OPERATION . . . . 17-5

Educational Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-5
Thermomechanical Bending in Pipe-type Cables . . . . . 17-5
Recabling Upgrades . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-6
xxi
xxii
CHAPTER 1 Historical Perspective

Author: Nirmal Singh, Detroit Edison Company
Reviewer: Reza Ghafurian, Consolidated Edison Company of New York, Inc.
This chapter presents an overview of the history of underground transmission—from the

early development of underground insulated communication and electric power distribu-
tion and transmission lines in the early 1800s to the present.
1-1
Chapter 1: Historical Perspective EPRI Underground Transmission Systems Reference Book
1.1 INTRODUCTION cation. In 1812, Baron Schilling, a Russian, fired a mine

This chapter reviews the history of underground electric under the River Nava in St. Petersburg with an electric
transmission systems. Section 1.2 describes early cable impulse sent through a rubber-insulated cable. This
systems in Europe, including the development of new application represents the first record of a continuously
materials, constructions and landmark installations. insulated conductor. Following the successful efforts of
Section 1.3 outlines the start of the cable industry in the Schilling, in 1886, Francis Ronalds in Hammersmith,
United States, including the first use of paper-insulated England, used bare copper wires drawn into butted
cable and the development of oil-filled and gas-filled glass tubes and joined by sleeves, which were placed in
cables from low to high pressures. Sections 1.4 through wooden troughs—a kind of wooden pipe—for a tele-
1.9 summarize developments in the underground trans- graphic system. This design was intended to allow ready
mission industry decade by decade from the 1950s access to the wire and the localization of faults.
through the first years of the twenty-first century.
Gutta percha or gutta taban, a natural latex extracted
Section 1.4 introduces the users to the potential of poly- from trees, was the predominant form of insulation in
ethylene and crosslinked polyethylene as an alternative cables from the early 1840s to the late 1860s. Dr. Will-
insulation to oil-paper systems and other natural and iam Montgomerie, a physician to the British Governor
synthetic rubbers. Section 1.5 shows the acceptance of in Malaya, introduced gutta percha to the West in 1843,
high-pressure fluid-filled (HPFF) cable systems up to and Sir Michael Faraday suggested its use as an insulant
the 345-kV level, largely as a result of the successful to Williams Siemens, whose brother was involved in mil-
completion of field trials at Cornell University. itary underground cables in Germany.
Section 1.6 covers the dominance of pipe-type cable sys-
tems and lack of enthusiasm of U.S. utilities for In 1847, Werner Siemens invented a machine for apply-
extruded cable systems, including the commencement of ing seamless gutta percha to wire because his previous
active involvement of the newly formed EPRI in under- effort to cover the wire with strips of this material expe-
ground transmission. Section 1.7 highlights the develop- rienced moisture ingress at the strip-interface. In 1847,
ment and installation of laminated paper polypropylene Siemens produced telegraphic wire insulated with gutta
insulated HPFF cables in the United States, continued percha, and in 1850, similar telegraphic cable was
confidence in HPFF cable systems, and strides being placed across the English Channel. In 1848, a process
made in extruded cable systems in Europe and Japan, was developed in England covering copper wire with
including the discovery of high temperature supercon- vulcanized (or crosslinked) natural rubber strips, and in
ductivity (HTS). Section 1.8 includes the growing inter- 1859, it was used for telegraph cables in England. In
est and acceptance of extruded cables, increasing focus 1879, Francois Borel of Switzerland invented a lead
on condition and life assessment of HPFF cables and press for extruding a lead sheath directly over an insu-
means for better use of their capacity, as well as evalua- lated core, and this made a product satisfactory for use
tions of alternative HTS cable designs. Section 1.9 in power cables. In 1884, Borel also made two 2-kV
marks the introduction of long-length extruded cable cables for Vienna, which were insulated by a combina-
systems with a large number of splices at 230-345 kV, tion of paraffin-impregnated jute. By 1890, single-core,
and the launching of several commercial demonstration underground dc power cables, insulated with this mate-
projects of HTS cable systems, including a few extruded rial and covered with lead, rated for 3-kV service, were
HVdc cable installations. installed in Paris. Siemens also installed similar cables,
rated at 2 kV, in 1886 in Vienna, Rome, and Milan.
1.2 EARLY CABLE SYSTEMS IN EUROPE
The successful installation by Sebastian Ferranti of
A history of cables has to start with the technique of England, of a 30-mile, 10-kV high-voltage paper-insu-
making wires by drawing metal through dies, which lated cable with about 7500 joints in 1890 is an epochal
dates back to the fourteenth century. Wires were gauged event in the history of high-voltage cables. This cable
by numbers in the early eighteenth century, laying the consisted of 20-ft lengths of concentric copper tubes,
foundation of the geometric system, which forms the insulated with ordinary rag paper soaked or impreg-
basis of current practices. nated in ozokerite, which is a naturally occurring min-
eral for m of paraffin wax. This cable marks the
According to early references, cables were utilized in introduction of paper to high-voltage cable. An overall
mining and telegraphy, with the latter providing the lead sheath was provided. This installation, connecting
impetus for the development of insulated cables in the the Deptford Generating Station to Grosvenor Gallery
early-to-mid 1800s for this emerging mode of communi-
1-2
EPRI Underground Transmission Systems Reference Book Chapter 1: Historical Perspective
Generating Station, was commissioned on February 15, The first use of paper tape, applied helically as a conduc-
1891, and the last length was only removed in 1933—the tor insulation, is attributed in 1872 to John Wortendyke,
42-year service offered by the Ferranti cable might have a Virginia paper manufacturer, for a bell system in his
incidentally led to the general acceptance of a 40-year home. In 1884, Edwin McCracken improved on Wor-
cable life. tendyke’s crude taping machine and patented the con-
cept of paper-insulated wire. The following year, Jacques
Following the introduction of impregnated paper cable patented a process combining vacuum-drying and hot
for high-voltage cables by Dr. Ferranti, paper-insulated, oil impregnation under pressure to provide the manufac-
lead-covered cable was used extensively in Europe for turing process, which was later extensively used for
underground transmission at increasingly higher volt- paper-insulated power cable suitable for high voltages; it
ages. By 1911, paper-insulated cable was installed in is still an integral part of paper-cable manufacturing.
Germany for operation at 60 kV and in Spain at 50 kV. The first oil-impregnated paper cable made with helically
In 1914, Martin Hochstadter of Germany invented the applied paper tapes, and covered with a lead sheath was
use of a metallic shield over the insulation; this was a a telephone cable designed by Barrett of the American
major advance in the reliability of solid-type, paper- Telephone and Telegraph Company in the late 1880s.
insulated cable. In fact, this concept is still the basis of The Norwich Wire Company, founded by McCracken,
all high-voltage cable design today. commenced manufacturing impregnated-paper cables in
1892. This Norwich company later became the National
In 1920, Luigi Emanueli of Italy introduced a hollow- Conduit and Cable Company.
conductor cable in which low-viscosity mineral oil could
be fed from external, pressurized reservoirs. The oil per- The impetus for underground transmission was supplied
meated the paper insulation and filled voids; this design by Thomas Alva Edison with the invention of the incan-
soon extended the operating range of paper-insulated descent lamp in 1879. Prior to Edison, telegraph, tele-
cable up to 138 kV. In 1931, gas-pressurized, oil-impreg- phone, and arc-lighting companies installed their cables
nated paper cable was introduced by M. Hochstadter above ground on poles. After a demonstration at Menlo
and W. Vogel in Germany and by Bowden in Great Brit- Park, New Jersey in 1880, the Edison Tube Works began
ain for use at 66 to 132 kV. A thin lead sheath separated the installation of a rigid-tube dc system emanating
the cable insulation from the external gas pressure, from the Pearl Street Station, New York City. The sys-
which deformed the lead and prevented voids in the tem was energized in 1882 and consisted of 10.5 miles
insulation during contraction. The external gas pressure (16.8 km) of underground mains. This was followed by
of 200 psi (1380 kPa) was contained by a steel pipe or by the first three-wire, 240-V, underground, Edison-tubed
a second reinforced lead sheath. European practice was system in Brockton, Massachusetts, which was placed in
to bury power cables directly in the ground, which is still service in 1883. The tubes and the necessary splice boxes
favored today. were placed underground, and the system operated at
+/-120 V dc. By 1907, New York City had 54 miles
1.3 BEGINNINGS OF THE U. S. (86.4 km) of such two- and three-wire feeders, and
UNDERGROUND CABLE INDUSTRY approximately 180 miles (288 km) had been installed
in various cities in the United States. Edison-tubed,
The transmission of electrical energy by means of
dc cable use then declined as flexible, duct-type cable
underground cable in the United States generally paral-
and ac transmission became more popular, due to
leled the European evolution, beginning with the tele-
the pioneering efforts of Nikola Tesla and George West-
graph industry. In 1841, Samuel B. Morse designed a
inghouse in ac transmission and distribution in the
cable insulated with rubber, hemp, tar, and pitch for use
United States.
in a submarine telegraph line between the Battery and
Governors Island, New York, and it was placed success-
The Edison low-voltage dc, underground cable system
fully in operation in 1842. Vulcanized rubber-insulated
was supplied from a central generating station and had
cable came into general use by 1860 for the telegraphy
two insurmountable disadvantages: the voltage could
industry, thanks to the pioneering efforts of Charles
not be regulated within reasonable limits, and heavy
Goodyear in rubber processing, and by 1890 for street
transmission losses made the system uneconomical. The
and other lighting cables. Subsequently, rubber-insu-
ac system devised in Europe by Lucien Gaulard and
lated cables were successfully developed for high-voltage
John Gibbs overcame both of these disadvantages. This
underground feeder cables up to 138 kV.
system, using overhead cable and operating at 1 kV ac,
was initiated by Westinghouse and installed in Buffalo
in 1886.
1-3
Paper-insulated cable for lighting and power took the on part of the Minneapolis and St. Paul 25-kV system.
form initiated by John Barrett for telephone use, and However, after this installation, power utilities in general
was later adapted to the power industry by Edwin favored paper-insulated cable for service over 5 kV. For
McCracken and his associates of the Norwich Wire secondary networks, rated 120–280 V, rubber-insulated
Company. It was composed of narrow paper tapes cable has been the preferred cable.
applied helically around the conductor and impregnated
wi t h o i l u n d e r p r e s s u r e. I t i s n o t ewo r t hy t h at In 1902, General Electric (GE) Company introduced
McCracken’s patent on wire being insulated with paper varnished cambric insulated cables for general power
tapes was already six years old at the time of the cele- purposes, up to 15–27 kV. The varnished cloth consisted
brated Ferranti rigid cable installation in England. Oil- of cotton cloth, coated with baked films of varnish, both
impregnated paper insulation quickly became the most applied helically around the conductor. These cables
common for voltages above 5 kV, and even today there are still found on utility systems, particularly in generat-
are enormous quantities of paper cables, ranging from ing plants.
5 to 525 kV. By 1897, the Chicago Electric Company
(now Commonwealth Edison and part of Exelon) had Figures 1-1 and 1-2 show early cable trucks and mecha-
installed three-conductor, belted-paper-insulated cable nized equipment from 1916.
rated 4.5 kV, Brooklyn Edison Company had installed
similar cable rated 6.6 kV, and the first 13-kV paper- The inherent suitability of paper cables was such that
insulated cable, made by the National Conduit and three-conductor cables were introduced within about a
Cable Company, was installed in Minneapolis and St. decade of the introduction of such one-conductor cable.
Paul. In later decades, belted cables up to 24 kV were Around 1910, the first sector-shaped conductor, paper-
made, with some still in use. insulated cable was installed in the United States by
New York Edison. By 1920, three-conductor, shielded,
In 1900, a 25-kV system using paper-insulated cable was paper-insulated cable (Type H, so called after Dr. Hoch-
installed in Minneapolis and St. Paul. The impregnating stadter), began to replace three-conductor, belted cable
oil used in these early installations was a mixture of rosin as the preferred design. The first type H cable used com-
oil and wood rosin, which was compatible with paper mercially in the United States was made by Standard
and sufficiently viscous that it would not drain exces- Underground Cable Company and installed at Fall
sively from the insulation structure during normal load River, Massachusetts for 25-kV service in 1920. During
cycle. Thus rosin, which is derived from trees, served as a the 1920s, the use of Type H cable rapidly expanded to
viscosity modifier. Compared to Europe, rosin found less many cities in the United States at voltages up to 35 kV,
favor in the United States, where the use of Petrolatum and even higher. The U.S. efforts culminated in 1923
was more prevalent. All Petrolatums are derived from with the installation of the world’s first 66-kV impreg-
paraffinic oils. Rubber-insulated cable was also installed nated-paper cable (solid-type) with oil reservoirs at each
Figure 1-1 Mechanized cable pulling into manholes, 1916 Figure 1-2 Cable truck with incorporated winch and
(courtesy of Detroit Edison Company). removed cable pieces, 1916 (courtesy of Detroit
Edison Company).
1-4
joint at the Cleveland Electric Illuminating Company New York placed a similar 132-kV oil-filled cable in ser-
(now FirstEnergy) (see Figure 1-3). vice. The cables operated at a nominal pressure of 15 psi
(105 kPa).
In the early 1920s, the phenomenon of “wax” formation
in oil-paper cables was discovered by Charles Hirschfeld Work in the 1920s through 1940s resulted in reducing
of Detroit Edison Company (Dunsheath 1929). This dielectric losses in the insulation, which permitted using
unusual substance, which is formed as a result of the paper insulation for higher voltages. In 1927, H. W.
polymerization of the impregnant due to ionization Fisher and Ralph Atkinson demonstrated that oil-
activity, drew a lot of attention in the industry at that impregnated paper insulation could operate at higher
time. Because of the material’s insulating properties, it voltage stresses if the impregnant is maintained at pres-
was concluded that this material does not lead to cable sures in the range of 150–225 psi (1035–1550 kPa). In
failure, a conclusion that still holds today. The applica- 1933, the concept of the radial evaluation of the electri-
tion of pressure eliminated wax formation in oil-paper cal and mechanical properties of paper tapes removed
cables altogether. from manufactured and in-service cables introduced by
Detroit Edison Company readily allowed examination
The paper-insulated cable manufactured during most of of such cables. This approach facilitated a proper under-
the 1920s had disadvantages in that, during normal load standing of the aging of oil-paper cables, leading to
cycle in service, the impregnant had a tendency to drain product improvements.
from the insulation, and the lead sheath expanded when
the cable was hot and did not contract to its original size This discovery led the way to extend the voltage range of
due to a marked differential between the lead and the oil-filled cable above 132 kV. But, more importantly, for
impregnant. This problem was overcome by the afore- the United States, it led to the introduction of pipe-type
mentioned oil-filled cable invented by Luigi Emanueli. cable by Charles Bennett of Okonite. The first installa-
Instead of the viscous oil, the “oil-filled” cable used an tion of high-pressure, fluid-filled (HPFF), pipe-type
oil of low viscosity, which could move in and out of the cable in the United States was in 1932 at the Plymouth
insulation as the cable was exposed to daily thermal Meeting, Pennsylvania substation of Philadelphia Elec-
cycling. In 1927, Commonwealth Edison of Chicago tric Company by Okonite. This was an experimental
placed in service the first low-pressure, oil-filled (LPOF) installation, which operated at 69 kV, with an internal
132-kV underground cable in the world. The LPOF pressure of 200 psi (1380 kPa). Okonite then supplied
cable is now referred to as self-contained liquid filled or the first commercial installation in 1935 at a higher
self-contained fluid filled (SCLF or SCFF) cable. This 138 kV operating voltage to the Pennsylvania Railroad
was an oil-filled cable of the Emanueli design, and later in Baltimore, Maryland. Three paper-insulated conduc-
in the same year, Consolidated Edison Company of tors were pulled into a steel pipe having a corrosion pro-
tective covering. This design of HPFF pipe-type cable
soon became the preferred cable for transmission volt-
age of 115–345 kV in the United States. The designation
of this cable was eventually changed to high-pressure,
fluid-filled in the 1970s, recognizing the widespread use
of synthetic insulating fluids. The HPFF terminology
will be used in this book.
The first full-size installation of low-pressure, gas-filled

(LPGF) cable was placed in service in 1938 in Yonkers
and New York City. This three-conductor, lead-covered
cable was designed to operate at 13–35 kV and 8–15 psi
(55–103 kPa). Dry nitrogen was used to prevent mois-
ture ingress in the event of a sheath leak. Later, medium-
pressure, gas-filled cable (MPGF) was placed in service
for 35–40 kV and 69 kV operation, using pressures from
24–40 psi (165-245 kPa). Such nitrogen-assisted cables
are still in use today, particularly at voltages up to 40 kV
Figure 1-3 The world’s first 66-kV solid-type cable, at several U.S. utilities. However, because of age, nitro-
Cleveland Electric Illuminating Company, 1923 (courtesy of gen leaks, and lack of maintenance, these cables are gen-
General Cable Catalog, 1949). erally replaced with extruded cables.
1-5
For operation at 69–138 kV, the principle of high- cables were in operation. Although most of these cables
pressure, gas-filled cable (HPGF) was applied to pipe- were of the solid-type, there also were LPGF and SCFF
type cable. In this design, the pressurizing medium is cables in service.
dry nitrogen. The first HPGF pipe-type cable was
placed in service in 1941 at the Detroit Edison Com- The solid-type cable proved to be somewhat marginal in
pany for operation at 120 kV (see Figure 1-4). This operation above 27 kV and unsatisfactory at 69 kV
cable belongs to the 138-kV class. However, it was because of its large diameter, ionization, and lead sheath
placed in service at 120 kV, the prevalent voltage level at fatigue problems. The LPGF cables were satisfactory for
Detroit Edison Company. After 20 years of operation, a operation up to 27 kV, but for 40 kV, the nitrogen pres-
major portion of this cable was changed from nitrogen- sure had to be increased from 8-15 psi (55–105 kPa) to
filled to oil-filled, permitting 230-kV operation and the 24–40 psi (165–275 kPa), and even then the cable
resultant increase in power transfer. This cable is still in (MPGF) was not generally accepted because of its low
service today, a testimony to longevity of the pipe-type margin of safety at operating voltage. There was limited
cable system, irrespective of the nature of the pressuriz- use of gas-compression, paper-insulated cable, where
ing medium. nitrogen gas was applied outside an oval-shaped poly-
ethylene sheath. This design, used in the voltage range
In January 1943, the wartime concern arose regarding of 69–138 kV, was not widely accepted, although
obtaining maximum use of materials and facilities con- 161-kV and 230-kV designs using oil pressurization
sistent with tolerable rates of failures, and new conduc- remain in service.
tor temperature guidelines were issued. During this
time, many cables found repeated applications of higher Mineral-base and oil-base, natural-rubber-insulated
current loadings, particularly at lower voltage levels. cables continued to be used occasionally up to 25 kV. In
the 1950s, GR-S (Buna-S) and GR-I (butyl) became
1.4 PROGRESS THROUGH THE 1950s commercially available and were suitable for under-
ground cable insulation. Because they were lower in
Until the late 1940s, underground transmission and dis-
cost, easier to process in the factory, and formulated to
tribution cables were similar to those of the late 1930s.
provide better performance, they quickly began to
The highest voltage in service was 132 kV. At this volt-
replace natural rubber insulation. By 1960, butyl-
age, two types of cable were in operation: the single-con-
insulated cable, which has good heat, water, and ozone
ductor SCFF cable, and the HPFF pipe-type cable with
resistance, was used extensively rather than oil-base rub-
three conductors in a common pipe. The latter type ulti-
ber and paper-insulated cables for voltages up to 15 kV
mately became the most widely accepted in the United
and occasionally up to 35 kV. However, this type of
States. Both types were installed with oil-impregnated
cable was not suitable for voltages at 69 kV and higher.
paper, as were practically all other cables in service
above 15 kV. In the voltage range of 5–69 kV, both
The use of polyethylene, developed in England in 1933
single-conductor and three-conductor, paper-insulated
by ICI (Imperial Chemical Industries), was first
employed in British tanks during World War II. Its use
in the United States began in 1942 for cables up to 5 kV
in industrial plants, residential distribution-type circuits,
and street lighting. Polyethylene was being used for cir-
cuits rated 15 kV by 1947, and by 1955 at 35 kV. In
1957, Consolidated Edison Company of New York
installed its first polyethylene-insulated, 15-kV cable as
a replacement for the sector-conductor, paper-insulated
cable. However, it experienced early service failures, as
did similar 24-kV cable that was installed shortly after.
Other utilities installed other forms of polyethylene-
insulated cable for operation up to 35 kV, without a pro-
tective metallic covering. Most of the cable installed
underground during the 1950s did not have a metallic
covering; some, however, had polyvinyl chloride, poly-
ethylene, or extruded lead as the outer protective cover-
Figure 1-4 The world’s first 138-kV class high-pressure
ing. The successful crosslinking of polyethylene by GE
gas-filled pipe-type cable, Detroit Edison Company, 1941
(courtesy of Detroit Edison Company). (Al Gilbert and Frank Precopio), a landmark develop-
1-6
ment in the late 1950s, was indispensable in furthering HPFF cables for 230-kV and 345-kV operation, all pro-
the cause of extruded cables that gained increasing viding satisfactory service.
acceptability in the following decades.
In the early 1960s, the development of crosslinked poly-
The 1950s saw great strides in underground cables and ethylene (XLPE) insulation made possible by the earlier
tremendous growth in underground transmission. GE work, greatly increased the use of extruded cable for
Extensive research in cable ratings, soil thermal proper- underground transmission circuits rated up to 69 kV. In
ties, and cable characteristics resulted in a large number 1963, GE extended its earlier chemical crosslinking
of technical papers, many of which are classics today. In techniques to include fillers in polyethylene, and trade-
1957, the Underground Systems Reference Book was named the material Vulkene. Although some of the min-
published, and there were many innovations in cable eral-filled XLPE cables did not perform well, a 69-kV
design and installation. It was also in this decade that mineral-filled crosslinked polyethylene of the first gener-
the plans for the field testing of full-size 345-kV paper ation cable has been operating satisfactorily at Ameri-
cables at Cornell University were launched by AEIC can Electric Power Company (AEP) since 1968.
(Association of Edison Illuminating Companies) in
cooperation with the then four leading U. S. cable man- The crosslinked variety closely resembles polyethylene
ufacturers, of which only one is still active in transmis- in its excellent electrical characteristics, with the addi-
sion paper cables today. tional advantage that it is thermosetting, thus permit-
ting higher-temperature operation. The earliest
1.5 PROGRESS THROUGH THE 1960s commercial use of chemically crosslinked polyethylene
(XLPE) insulation for cable was in 1961. Its use up
Principally because of lower installed cost, pipe-type
to 1963 was limited to 15 kV; however, by 1965, many
cable generally superseded self-contained, fluid-filled
utilities were using XLPE cable for voltages up to 46 kV.
cable, which had been introduced about 10 years earlier
A few utilities used 69-kV XLPE cable, in many cases
than pipe type. The early installations of fluid-filled
replacing solid-type, paper-insulated cable. For the most
cable, beginning in 1932, were of HPFF pipe-type in the
part, there was no outer covering such as a lead sheath,
voltage range of 69-138 kV. By 1954, the voltage was
as is used with paper-insulated cables. It was felt that the
extended to 230 kV, and by 1964, the voltage was
extra cost of the lead sheath would negate economic
extended to 345 kV. Both SCFF and HPFF 345-kV
advantages of extruded cables over paper cables, all the
cables and their accessories were proven acceptable in
more at distribution voltage levels.
an extensive test program at Cornell University. As a
result of the Cornell University program completed in
Some use of polyethylene-insulated (PE) cable at trans-
1964, utilities began to increasingly accept 345-kV
mission voltages continued during the 1960s, but it
HPFF cable systems.
began to lose favor when compared to XLPE cable. The
first 115-kV voltage-stabilized polyethylene cable
Fluid circulation in HPFF cable, to smooth hot spots in
(VSP), by Simplex Wire and Cable, was installed at the
the transmission system or to increase the circuit capac-
Baltimore Gas & Electric Company in 1965. Subse-
ity above the self-cooled level, was initiated in 1948 and
quently, numerous 115 to 138-kV cables were installed
became used on an increasing scale during the 1950s.
by utilities during this period, encompassing VSP,
Slow fluid circulation generally was used for tempera-
XLPE, high-molecular-weight polyethylene, and EPR
ture averaging, where the fluid is circulated between two
(ethylene-propylene-rubber).
parallel feeder circuits. Forced cooling was used for
increasing capacity as well as temperature averaging. In
Prior to 1969, there were several installations of XLPE-
this case, the pipe fluid was circulated through an exter-
insulated transmission cable in the voltage range of
nal heat exchanger and returned in a separate pipe.
69–138 kV at utilities such as Tampa Electric, Common-
wealth Edison Company, Sacramento Municipal Utility
The first 345-kV, HPFF pipe-type cable was installed by
Department, City of Lakeland, and Florida Power &
Consolidated Edison Company in New York City and
Light. For new circuits, there was a trend to use XLPE
was put into service in 1964. Since that time this com-
cable in place of LPGF, MPGF, and SCFF cable in the
pany has added many 345-kV circuits, including some
voltage range of 15–69 kV. Some rubber-insulated cable
under the East and Hudson Rivers. Several other utili-
continued to be used, particularly for industrial applica-
ties—such as NSTAR, Commonwealth Edison, Ari-
tions. In addition, ethylene-propylene-rubber (EPR)
zona Public Service Company, and Public Service
began to be produced commercially in the late 1960s,
Electric & Gas Company—have installed forced-cooled
offering a marked improvement over butyl rubber and
1-7
oil-based rubber as a cable insulant. The properties of placed in service in 1970 for 138-kV operation between
EPR were deemed suitable for cables rated up to 138 kV. the Connecticut Light and Power Company and the
Long Island Lighting Company (now LIPA). In addi-
The Department of Energy (DOE) initiated efforts to tion, two self-contained cables at 525 kV, still the high-
develop low-temperature superconducting cables in the est cable voltage in the United States, were installed at
mid-1960s, laying the foundation of the technology that the Grand Coulee Dam.
came to fruition in the following decades at the high-
temperature superconducting level. Helium and hydro- The first domestic, underground, gas-insulated trans-
gen were utilized in these efforts, along with several mission line (GITL) was installed by Consolidated Edi-
metal alloys based on niobium. son Company in 1972 in Westchester County, and by
1975, an underground GITL cable system was installed
1.6 PROGRESS THROUGH THE 1970s by Bonneville Power Administration for operation at
550 kV. The prevalent design was three isolated-phase
During the 1970s, HPFF pipe-type, paper-insulated
aluminum pipes, each containing a coaxial aluminum
cable continued to dominate the voltage range of
conductor, supported by epoxy insulating spacers, and
115–345 kV. Aluminum, rather than copper, was used in
pressurized with sulfur hexafluoride (SF6) gas, usually
some cases as the conductor, largely for economic rea-
at 50 psi (345 kPa). These aluminum pipes were placed
sons. The potential of sodium as a conductor was
in a common trench and were protected with a suitable
addressed and soon given up. Polybutene, a synthetic
anticorrosion coating.
fluid, was employed increasingly as the impregnant or
as the pipe-filling fluid. In the early 1970s, a two-year
Late in the 1960s, the Edison Electric Institute’s Electric
acceptance test program was performed at the Waltz
Research Council, the predecessor of EPRI, set up a
Mill Test Site on HPFF cable systems. This program
development program for 138-kV extruded-dielectric
demonstrated that conventional HPFF cable systems
cable, and by 1971, five XLPE cables and one EPR cable
were satisfactory for operation at 550 kV, further
from six manufacturers were installed at the Waltz Mill
strengthening the position of HPFF cable systems. Nev-
Test Site for field trial. The field trial later was extended
ertheless, a few significant extruded cable (XPLE)
to include 230-kV XLPE cable. These cables did not
installations at 138 kV level, the highest voltage that
perform satisfactorily, and reduced U.S. utility enthusi-
U.S. utilities would consider, were made at Cincinnati
asm for extruded-dielectric transmission cables.
Gas and Electric (now Duke) and Cleveland Electric
(now FirstEnergy); both are operating satisfactorily,
The formation of EPRI greatly accelerated the transmis-
although the latter experienced a few termination prob-
sion cable research and development in the country,
lems that were taken care of by replacements of the
covering a very broad range of cable issues. Some
same termination design. It is noteworthy that the
projects of note included Elastimold’s 138-kV splice for
FirstEnergy extruded cable has no metallic sheath.
extruded cables, forced-cooling studies, LPP investiga-
Unlike the Duke cable, the FirstEnergy cable has no
tions, extruded dc cable development, and cable oil
splices and is exposed to water due to close proximity to
studies prompted by the oil embargo by the Organiza-
a river.
tion of Petroleum Exporting Countries (OPEC) in 1973.
DOE cooperated with EPRI on several projects at Waltz
During the 1970s, investigations were conducted on
Mill and elsewhere.
replacing paper insulation with polymeric tape insula-
tion, but did not result in any commercial cable produc-
tion in this decade. The research was conducted under 1.7 PROGRESS THROUGH THE 1980s
the sponsorship of the Electric Power Research Institute The dominance of HPFF pipe-type cable for under-
(EPRI), which was established in 1973 to serve as the ground transmission continued in the 115–345 kV volt-
research and development arm of member utilities. This age range. However, Sun Oil Company, the major
work culminated in the development of laminated paper mineral oil supplier to the U.S. cable industry, discontin-
polypropylene (LPP or PPP) tapes for insulation appli- ued this product line in 1982, resulting in the nonavail-
cations in the 1980s. ability of mineral oil from U.S. sources. The U.S. utilities
commenced the use of alkylbenzene fluid, which was
Although HPFF pipe-type cable remained the preferred first introduced by the Italians for SCFF cables in the
transmission cable during the 1970s, one of the longest late 1960s, as a pipe-filling fluid for circulating and static
ac fluid-filled submarine cable crossings in the world HPFF cables. For the latter cable types, alkylbenzene
(Long Island Sound Crossing) was an SCFF cable was blended with polybutene fluid to raise its viscosity.
1-8
The highest voltage for pipe-type cable at this time was ethylene and installed several 225-kV extruded systems
345 kV, and—except for special SCFF and GITL cable and one 400-kV system. The Japanese relied on XLPE
circuits rated 525 kV—it continues to be the highest and reported short connections without joints at 275
voltage used for underground transmission in the and 500 kV.
United States. Changes during the 1980s in paper-
insulated cable designs were in components and in the Most installations at this time employed HPFF pipe-type
modification of some pipe-type cable circuits. Cable cable, which continued to provide virtually trouble-free
manufacturers and material suppliers also undertook service. Nevertheless, the inherent advantages of
efforts to develop laminated paper propylene film insu- extruded systems were gradually being appreciated by
lation, and by the early 1980s, such HPFF pipe-type U.S. utilities. As a result of worldwide efforts, extruded
cable for service up to 765 kV was under test at Waltz dielectric cables underwent improvements in material
Mill with EPRI support. The first LPP (or PPP) cable at cleanliness, manufacturing, testing, and quality control.
345 kV was installed at Boston Edison in 1987, and it The most notable improvements involved super-clean
was followed by LPP cable installations at Consolidated and extrudable insulation and shielding compounds, tri-
Edison Company of New York and Public Service Elec- ple extrusion, dry curing, increased acceptance test levels,
tric & Gas Company. Work to increase the capacity of dramatically increased partial discharge detection sensi-
existing circuits continued in the 1980s. As expected, the tivity, use of jacketed constructions for XLPE and EPR
Europeans focused on SCFF cables insulated with LPP cables, and, in some cases, sealed-strand and/or laminate
in the 400–500 kV range, resulting in the first commer- radial water barriers. In 1986, Iowa Power & Light Com-
cial 500-kV installation in 1986 in the Far East. The pany installed underground XLPE cable for service at
majority of these LPP installations are in the United 161 kV in duct, the highest voltage rating of XLPE cable
Kingdom, Japan, Hong Kong, and Singapore. in the United States at that time. The cable did not have
an outer moisture barrier other than the polyvinyl chlo-
In 1982, Detroit Edison Company reported the conver- ride jacket, and the installation included several field-
sion of 138-kV HPGF pipe-type cable to 230-kV HPFF molded joints with prefabricated terminations.
by replacing nitrogen gas with dielectric liquid, based on
experimental work that had started in the mid-1970s. In The changing utility business brought about by compe-
1983, Florida Power & Light Company reported recon- tition forced utilities to address operating and mainte-
ductoring an existing 69-kV HPFF pipe-type cable to nance issues in an effective manner and run the aging
operate at 138 kV by using a smaller conductor and equipment harder. As a result, great progress was made
reduced insulation thickness. In 1987, United Illuminat- in condition assessment techniques, particularly for
ing Company reported an increase of 35% in ampacity H P F F a n d S C F F c abl e s y s t e m s t h ro u g h D G A
by converting an existing 115-kV HPGF pipe-type cable (dissolved-gas analysis). Under the leadership of Con-
to an HPFF cable by replacing the nitrogen gas with solidated Edison and EPRI, plans were formulated to
dielectric liquid to smooth a serious thermal bottleneck. assess the condition and remaining life of the HPFF
The use of GITL transmission cable systems continued cable systems, about 25% of which were approaching or
for short applications through the 1980s. Typical instal- had exceeded a 40-year life. HPGF cable systems were
lations have been for station get-aways, risers, links receiving increasing attention due to the absence of a
inside substations, and systems inside tunnels and verti- dielectric liquid and the recognized ruggedness and reli-
cal shafts. Considerable development effort was ability of pipe-type cable systems. However, EPRI-spon-
expended to improve components, increase corrosion sored efforts to extend the 138-kV operating voltage of
protection, provide flexibility, and confirm the feasibil- HPGF cable systems to the 230-kV level were not suc-
ity of raising the voltage range to 1100 kV. cessful. EPRI research did address the problem of TMB
(thermomechanical bending) in HPFF cables and iden-
The use of XLPE and EPR cable for transmission cable tified and implemented mitigating measures. Under
rated 115–138 kV was very limited during the 1980s, EPRI sponsorship, significant progress was made in the
probably due to the unsatisfactory outcome of extruded generation of accurate information on the thermal
cable testing at Waltz Mill. This understandably led to properties of the soil along a cable’s route and an effec-
reduced extruded cable manufacturing activity, with tive Soil Thermal Property Analyzer developed. EPRI
some suppliers even discontinuing the extruded cable also issued a Designer’s Handbook on forced-cooling of
line, particularly at transmission voltage level. However, HPFF cable systems.
the European (particularly the French) and Japanese
manufacturers did not slacken, making great strides. This decade saw the installation of the world’s first
The French focused on low-density uncrosslinked poly- 3 4 5 -k V L P P H P F F c ab l e a t B o s t o n E d i s o n
1-9
(now NSTAR). Several major submarine installations number of splices. The Japanese extruded system consti-
were made, including the 525-kV ac connection between tuted the world’s first 500-kV XLPE cable system, with-
Vancouver Island and the Canadian mainland and the out any splice. This installation is far above the then
400-kV dc link between Sweden and Finland. existing extruded voltage level of about 300 kV. Consid-
ering the growing importance of extruded cable systems,
The discovery of high-temperature superconducting EPRI continued its focus on such cable systems, and
(HTS) cable in 1986 by IBM scientists in Switzerland participated in the testing of three 345-kV extruded cable
renewed worldwide interest in superconducting cables, systems at IREQ (Institut de recherche d’Hydro-
and both EPRI and DOE made significant contribu- Québec), and diagnostic testing of 138-kV extruded
tions, laying the foundation for future work in HTS cable systems at NEETRAC (the National Electric
cable systems. Energy Testing Research & Applications Center). The
results of the latter project were not conclusive.
1.8 PROGRESS THROUGH THE 1990s
The HPFF cables continued to receive favor at 230 kV
This decade saw growing interest and gradual accep-
and above, and the renewed interest in HPGF cables at
tance of extruded cable systems. The absence of a dielec-
the 138-kV level persisted. Utilities that made major
tric fluid and the inherent simplicity of extruded cable
HPGF installations included Colorado Springs, AEP,
systems had a great deal of appeal. Exelon embarked on
and South Carolina Electric & Gas. Under EPRI spon-
the replacement of its aging 138-kV SCFF system with
sorship, an effective HPFF cable leak location technique
extruded cables that had built-in fiber-optic capability
using perfluorocarbon tracers was developed and uti-
and prefabricated splices. Although 138-kV extruded
lized at Consolidated Edison Company. To make opti-
cables offered enough confidence, there was concern at
mum use of HPFF cable systems, EPRI sponsored the
the 230-kV level for the accessories. In fact, problems of
development of the Dynamic Feeder Rating technique
extruded accessories have been reported from time to
in cooperation with Consolidated Edison. This system,
time at 138 kV, although the number of such problems
which emanated from an earlier EPRI project called
has been decreasing significantly in recent years.
DRUMS (Dynamic Rating and Underground Monitor-
ing System), has been effectively employed by Consoli-
A 230-kV extruded cable installation, with splice-free
dated Edison, and several other utilities such as Public
extruded XLPE, was undertaken in 1992 at the Orlando
Service Electric & Gas, NSTAR, and Arizona Public
Utilities Commission for a substation connection. Such
Service. To ameliorate the erosion of cable expertise at
a cautious approach to evaluate new types of high-volt-
utilities, EPRI continued the development of design
age equipment is not uncommon with both North
tools and software for underground transmission.
American and overseas utilities. The experience at
Orlando Utilities on their 230-kV extruded cable system
The advancing age of pipe-type cables, coupled with
was not entirely satisfactory. This was followed in 1994
the need to make the most of cable assets brought about
by another splice-free 230-kV extruded cable substation
by deregulation, placed increasing focus on condition
connection at Colorado Public Service (now Xcel
and life assessment of such cables. EPRI sponsored a
Energy), without any problems.
six-year HPFF Cable Life Management study at Waltz
Mill. The project drew significantly from a previous
The most significant 230-kV extruded cable system,
comprehensive EPRI study sponsored at Detroit Edison
with a 1.75 mile circuit-length, was installed at Arizona
relating to condition assessment of fluid-filled cables
Public Service in 1999. It was installed in duct and had
through DGA and related paper testing, including the
12 joints with 12 outdoor terminations; the moisture
development of EPOSS (EPRI Pressurized Oil Sam-
barrier consisted of an aluminum foil radial moisture
pling System) and its modification EDOSS (EPRI Dis-
barrier. No problems have been reported on this Ari-
posable Oil Sampling System)—both developed
zona installation.
specifically for HPFF cable systems. As the owner of the
largest HPFF cable system in the United States, Consol-
Of the many extruded cable installations made in Europe
idated Edison played a significant role in all the key
and Japan at transmission-class voltage levels, two are
EPRI projects relating to pipe-type cable systems.
noteworthy: a Danish and a Japanese installation put
into service in 1997 and 1998, respectively. The 400-kV
HTS cables continued to receive increasing attention,
Danish system consisted of 66 km of cable, 72 joints, and
and EPRI sponsored feasibility studies at Philadelphia
12 terminations. This was the first direct-buried 400-kV
Electric and South Carolina Electric & Gas, including
XLPE cable system in the world containing a significant
1-10
the testing of prototype HTS cable designs at the constraints to improve overall reliability. The incorpora-
138-kV level. These efforts led to an HTS cable project tion of low levels of properly chosen, treated, and pro-
at Detroit Edison. cessed inorganic fillers into XLPE and EPR has shown
outstanding improvements in critical properties relating
1.9 PROGRESS THROUGH THE 2000s to design stresses, defect tolerance, and reduced losses.
This EPRI-sponsored effort has the potential to revolu-
This decade marks the extensive installation of extruded
tionize extruded materials for cable applications.
cable systems at 230 kV and their introduction at
345 kV in the United States, including some practical
Mention should be made of a major 500-kV extruded
HTS cable systems and a few extruded dc cable installa-
tunnel system with 240 splices commissioned in 2000
tions. Despite this long-awaited acceptance of extruded
by the Japanese. This decade also includes the develop-
cable systems by U.S. utilities, this decade also saw the
ment of extruded cable manufacturing capability by
installation of major HPFF cable systems at NSTAR,
a domestic U.S. manufacturer at the 230-kV level,
Northeast Utilities, Dominion Resources, and Potomac
resulting in some 230-kV installations, a change from
Electric. To reduce environmental concerns posed by
the earlier decades when even 138-kV domestic fabrica-
cable fluid spills, EPRI addressed the characteristics of
tion was abandoned.
cable fluids, including risk-based cleanup goals in the
early 2000s, the first such study of its kind. In addition,
Following various paper and experimental studies car-
a 115-kV HPGF cable system was installed by South
ried out by EPRI and DOE in conjunction with HTS
Carolina Electric and Gas including a 7200-ft (2195 m)
material suppliers, cable manufacturers, utilities, and
successful directional boring underwater in the Charles-
refrigerant suppliers, several designs were identified and
ton area.
refined, and their relative merits fully addressed. These
sustained efforts led to several HTS cable system dem-
In this decade, U.S. utilities launched in earnest evalua-
onstration projects at Detroit Edison, AEP, Niagara
tions of the first, long-length XLPE extruded cable sys-
Mohawk (now part of National Grid), and Long Island
tems containing many joints at 230 kV and above. In
Power Authority (LIPA), with promising results. Chap-
2001, the Los Angeles Department of Water & Power
ter 17 provides brief descriptions of the latter three
(LADWP) successfully installed a 5.5 circuit-mile,
projects. Such cables have also been gaining broader
230-kV extruded cable system, with 78 prefabricated
acceptance worldwide due to the international scope of
joints. EPRI also played a role in the commissioning
HTS cable efforts, including China and Mexico.
tests of this LADWP cable system, deemed to be the
first significant 230-kV extruded installation in the
HTS cables, capable of carrying three to five times more
United States. This has been followed by several
power than conventional cables, can meet increasing
extruded 230- and 345-kV cable systems at many utili-
power demands in urban areas though retrofit applica-
ties such as Mystic Station, Reliant Energy, Calpine
tions and improvements in capacity, flexibility, and reli-
Corporation, Public Service of Colorado (now Xcel
ability. The Detroit Edison project consisted of the
Energy), Northeast Utilities, United Illuminating, and
design, manufacturing, installation, and operation of a
ITC Transmission Company. PGE installed the longest
24-kV, three-phase warm dielectric EPR cable replacing
230-kV extruded cable system (24 circuit-miles) in the
existing paper-insulated lead-covered (PILC) cables at
United States—a system with 2500 kcmil copper and
its Frisbie Station, which fed a large industrial customer.
234 prefabricated splices. Although such cables started
This pioneering commercial effort met most of its goals.
as short-length get-aways or substation connections, as
It included the HTS cable with a splice being success-
confidence increased, cables with splices and sufficient
fully pulled through demanding ducts having six right-
lengths were introduced. The significant extruded cable
angled bends. Although the system was never put in
installations that are providing satisfactory service
operation because of problems with leaks in the cry-
across the world in the 220- through 500-kV range are
ostat, its positive outcome and the lessons learned paved
too numerous to mention, demonstrating the strong
the way for HTS cable projects that soon followed.
confidence being placed in these easy-to-install cable
systems, which require minimum maintenance.
Although HVdc cables, based on mass-impregnated
paper and other types of oil- or gas-filled cables, have
EPRI-sponsored work should lead to the minimization
served the electric power industry well for several
of risk of thermomechanical damage to extruded cable
decades, the interest in harnessing the advantages of
systems installed in pipe or duct-manhole systems. The
extruded cables for this application has been gaining
benefits of this significant project include cost savings
strength. The modification of polyethylene or another
through proper selection of cables, joints, and manhole
1-11
polymer through additives to control resistivity holds ket for cable insulating materials is so limited, these
great promise as regards such cables. Several extruded materials need to have other uses. The HTS cable is an
HVdc cables have been installed in the world. The sole exception to some extent in that it is the revolutionary
U.S. installation is a 150-kV, 330-MW submarine con- conductor that makes such a cable possible. If the past is
nection across Long Island Sound. It employs an alter- any guide, refinements in materials, processing, and
native HVdc system based on extruded-dielectric cable, cable manufacturing techniques coupled with the never-
which has also been used in Sweden, Norway, and Aus- ending search for revolutionary materials would ensure
tralia. The Japanese have also been active in extruded an exciting future for cable systems. The need for reli-
HVdc cables and have put several HVdc extruded cables able and cost-effective underground electric transmis-
in service. A recent EPRI report (EPRI 2004) covers the sion, dictated by the escalating cost of overhead line
application of HVdc to extruded cables. rights-of-way and the unwanted visual intrusion, is only
going to sharpen the focus on cables in the years to
The history of cables is intertwined with the evolution of come. The industry seems to be well placed to address
insulating materials. As various insulations became underground transmission with available and emerging
available, related cables became available. Since the mar- cable system options and technologies.
1-12
REFERENCES EPRI. 1992. Underground Transmission Systems Refer-

Barnes, C. C. 1953. Power Cables: Their Design and ence Book. TR-101670. Palo Alto, CA.
Installation. Chapman & Hall. London.
EPRI. 2004. DC Cable Systems with Extruded Dielec-
Barnes, C. C. 1964. Electric Cables. Sir Isaac Pitman & trics. 1008720. Palo Alto, CA.
Sons, Ltd.
Gilbertson, O. I. 2000. Electrical Cables for Power and
Barnes, C. C. 1966. Power Cables: Their Design and Signal Transmission. John Wiley & Sons.
Installation. Chapman & Hall. London.
Moore, G. F. 1997. Electric Cables Handbook. BICC
Bartnikas, R. and K. D. Srivastava. 2000. Power and Cables Limited.
Communication Cables: Theory and Applications. IEEE
Press Series on Power Engineering. National Electric Light Association. 1931. Under-
ground Systems Reference Book. Prepared by an Edito-
Del Mar, W. A. 1924. Electric Cables: Their Design, rial Staff of the Underground Systems Committee,
Manufacture and Use. McGraw-Hill Book Company. Engineering National Section.
Dunsheath, P. 1929. High Voltage Cables, Theory and U. S. Department of Interior and Electric Research
Practice of their Design and Operation. London. Sir Council. 1971. Underground Power Transmission. Pre-
Isaac Pitman & Sons, Ltd. pared by Arthur D. Little for Project No. RP78-8.
Edison Electric Institute. 1957. Underground Systems Winspear, G. G. 1968. The Vanderbilt Rubber Hand-
Reference Book. Prepared by the Editorial Staff of the book. R.T. Vanderbilt Company, Inc.
Transmission and Distribution Committee.
1-13
CHAPTER 2 Cable System Types

Authors: Jay A. Williams, Power Delivery Consultants, Inc.
Allen MacPhail, BC Hydro
Reviewer: Brian Gregory, Cable Consulting International
Chapter 2 summarizes basic components of ac transmission cable systems, describes

cable construction for the major types of transmission cable used in North America and
overseas, and discusses system implications of installing underground cables. The chapter
addresses typical conditions when a cable would be installed rather than an overhead line
and provides a summary of design, installation, and operating considerations for the
cable system itself.
This chapter is an overview of underground transmission cable systems; subsequent

chapters provide additional details on various topics. Chapter 3 discusses basic design
considerations, and Chapters 4, 5, and 6 discuss extruded-dielectric cables, pipe-type
cables, and self-contained fluid-filled cables, respectively. Chapter 7 addresses special
cable applications (submarine cables, dc cables, gas-insulated transmission lines, and
superconducting cables). Sections 2.5–2.8 of the present chapter describe each cable type.
Jay Williams has degrees from Brown University and New York Univer-
sity. He worked in the cable section at Con Edison in the 1960s and
early 1970s when the utility was installing major lengths of 345-kV
cable. He was part of the design team for the Waltz Mill Test Facility,
and was in charge of the utility advisory committee for early EPRI
superconducting cable projects. Mr. Williams spent 20 years at Power
Technologies, Inc., where he headed the cable engineering group. He co-
founded Power Delivery Consultants, Inc. in 1992, and has continued
to specialize in underground transmission cable systems. He was technical editor and
principal author of the 1992 edition of the Underground Transmission Reference Book,
and was project engineer for several other major EPRI projects including the original
Underground Transmission Workstation. Mr. Williams is a Fellow of the IEEE and a reg-
istered Professional Engineer in New York and Ohio.
G. Allen MacPhail graduated from the University of British Columbia,

Vancouver, with a BApSc (EE) in 1972. He then joined BC Hydro, spe-
cializing in underground and submarine cable engineering for most of
his career. Mr. MacPhail’s experience includes most phases of a typical
project cycle from feasibility studies through to commissioning tests
and maintenance, for cable systems ranging from 15 to 525 kVac and up
to +/-450 kVdc. It also includes HV and EHV cable condition assess-
ment, life expectancy and extension; up-rating feasibility studies; real-
time dynamic rating and fiber optic distributed temperature sensing systems. In addition
to in-house responsibilities, Allen has been active with BC Hydro International subma-
rine cable projects in Central America, Alaska, Ireland, Egypt, and Jordan, at voltage lev-
els from 35 to 400 kV. Activities also include acting as an advisor to EPRI’s Underground
Transmission Task Force. He is a contributor to the IEEE ICC as past Vice Chair of the
2-1
Chapter 2: Cable System Types EPRI Underground Transmission Systems Reference Book
Accessories Subcommittee, past Chair of the Cable Insulation Power Cables and Their Accessories Rated
Construction and Design Subcommittee, and partici- above 46 kV through 345 kVac." Since retiring from BC
pant in many discussion and working groups. He is also Hydro in 2006, Mr. MacPhail has formed a small con-
a CIGRE member and past Chair of AEIC’s Task sulting company, Cabletricity Connections, specializing
Group for developing CS9 "Specification for Extruded in power cable application engineering.
2-2
EPRI Underground Transmission Systems Reference Book Chapter 2: Cable System Types
2.1 INTRODUCTION by laying concrete cover slabs directly above buried

This chapter discusses cables in general and their acces- cables, or by installation in tunnels. Pipe-type cables
sories. A cable system consists of the installed cable, enjoy the protection of steel enclosure pipe. In spite of
which includes pipe for a pipe-type cable, splices, termi- this protection, dig-in is the largest single cause of
nations, grounding/bonding, and fluid-pressurizing forced outages for transmission cables.
systems if needed. The term “cable system” implies that High Energy Efficiency
all components—including civil works—have been Electrical losses are a very small proportion of power
evaluated and designed to meet the requirements for transmitted. A conductor with low electrical resistivity
normal operation as well as emergency and transient must be selected to limit the temperature rise when heat
voltages and currents. Typically, the cable system is dissipated through the insulation and ground. This is
consists of all components from the top of the achieved by selecting a large conductor area—for exam-
termination at one end of the cable to the top of the ple, 2000 to 5000 kcmil (1000 to 2500mm2)—of high
termination at the other end of the cable. Many North purity copper or aluminum. Additionally, modern
American utilities include the high-voltage (HV) surge designs of transmission class cables employ low-loss
arrester as a cable system component. insulations such as extruded crosslinked polyethylene
(XLPE) or laminated paper-polypropylene tapes.
Chapter 2 provides an overview on underground cable Extruded and SCFF circuits are specially bonded to
systems. Section 2.2 summarizes a number of consider- achieve high current ratings by eliminating the sheath
ations for selecting an underground cable system. Sec- circulating current loss in solidly bonded systems.
tion 2.3 reviews electrical characteristics of underground
cables, and Section 2.4 discusses cable components and High Short-Time Rating
their functions. Sections 2.5 through 2.8 briefly discuss High transient and emergency ratings result from the
four basic types of cables—extruded dielectric, pipe- high mass and specific heat of the conductor, insulation,
type, self-contained fluid-filled, and special application duct, or pipe enclosures—and the earth itself for emer-
(gas insulated, dc and long-length submarine, and gency durations longer than several hours. These pro-
superconducting). Section 2.9 covers cable accessories. duce a slow rate of temperature rise and extended
Section 2.10 provides information on a few factors to be emergency operation time before the limiting tempera-
considered when choosing among the different types of ture is reached. In addition, industry specifications per-
cable. Section 2.11 offers a summary of industry specifi- mit a higher-than-normal temperature for certain
cations and standards related to underground cables. durations during emergencies.
Low Electro-Magnetic Field (EMF)
2.1.1 Features of Underground Cable Systems
The electric field outside a cable is essentially zero
Underground cable systems have the following features because of the grounded cable shield/sheath and the
relative to other power system components: grounded steel pipe for pipe-type cables.
Nonobtrusive
Cables are installed out of sight underground, in tun- The magnetic field from underground extruded and
nels, or under water. Terminal ends are often the only SCFF cables can be higher than that for overhead lines,
visible evidence of an underground cable’s presence. directly over the cable itself. However, the field attenu-
ates much more quickly than that for overhead lines.
Immune to Weather Magnetic fields for pipe-type cables are very low
Buried cable systems are unaffected by weather because of the closeness of the three phases and the
extremes, except for aboveground terminal ends. attenuation due to the steel pipe (CIGRE 2001, 2005).
High Reliability
Cable systems must be highly reliable because most 2.2 CONSIDERATIONS FOR SELECTING AN
components are inaccessible for visual inspection and UNDERGROUND CABLE SYSTEM
maintenance; they are installed in a hostile environ- There are a great many considerations required for
ment, usually underground; and they require a long out- selecting an underground cable system versus an over-
age repair time following electrical failure. head line and for selecting among cable system types.
This chapter and other chapters describe considerations
Vulnerability to Third-Party Damage
in more detail, but a few important considerations com-
Cables are vulnerable to third-party dig-ins. Protection
mon to all cable system types are summarized below:
for extruded-dielectric and self-contained fluid-filled
(SCFF) cables is provided by encasing ducts in concrete, • Power transfer levels are lower than those for over-
head lines because heat from buried cables must pass
2-3
through the earth before reaching air—which is the of widely spaced conductors, provides capacitive reac-
ultimate heat sink. Therefore, the cable might require tive compensation and increased stability (stiffness)
two cables per phase (essentially two cable systems) under transient operation. Disadvantages of cable
to match the capacity of the overhead line. capacitance are additional inductive compensation may
• The cable’s coaxial electrodes and outer shielding cre- be required in a long all-cable ac network, an overvolt-
ate a capacitance that affects power transfer, equip- age may result at the end of an overhead line feeding an
ment requirements, and system operation. off-load cable, and the magnitude of harmonic overvolt-
ages present in the network may be increased.
• Dielectric losses in cable insulation are present any
time the cable is energized and reduce the amount of Low Inductance
power transfer. The close spacing of phase conductors significantly
• Cable installation will affect every part of the route, reduces external magnetic flux linkages and yields a low
inductance. Inductance is further reduced under tran-
whereas overhead lines require foundations every
sient operation as an impulse of voltage and current
600-1200 ft (183-366 m) or farther apart.
travels in co-axial mode principally between the conduc-
• Even though the cable itself may be buried and out of tor and the annular shield/sheath conductor. Coaxial
sight, termination structures are required at each end geometry gives the lowest possible inductance.
to connect to overhead lines or substation equipment.
In addition, HPFF cable systems require pumping Low Impedance
plants and/or fluid tanks at one or both ends. The ac cable exhibits low impedance in power frequency
operation and is predominantly capacitive. This results
• Cables have a lower forced outage rate than over-
from low-inductive reactance in series with a higher-
head lines, but the outage durations are typically
capacitive reactance in parallel. Because the capacitive
much longer.
and inductive reactances are nearly equal and in
• Cables may be the only feasible approach for some opposite time phase in a gas-insulated line, they nearly
installations such as downtown urban areas or long cancel, producing in an ac system a near-ideal
underwater crossings. resistive impedance.
• Cable installations are almost always more expensive Low Characteristic Impedance
than equivalent power-transfer overhead lines—and Characteristic impedance is proportional to the ratio of
are sometimes significantly more expensive. inductance to capacitance. Characteristic impedance is
low, in the range of 10-40Ω, compared to 400Ω for an
Four cable systems can be considered for underground overhead line, a ratio of 10-40:1.
installations. These systems are described in greater
detail in other chapters, and are summarized as follows: Chapter 16 provides a great deal of additional informa-
• Extruded-dielectric with XLPE or EPR insulation tion on cable system electrical characteristics.
• Self-contained, either fluid-filled (SCFF), gas-filled
(SCGF), or mass-impregnated (MI) 2.4 CABLE COMPONENTS
• Pipe-type, either fluid-filled (HPFF) or gas-filled All cable types have the same basic components since
(HPGF) they all are required to carry current, safely support
steady-state and transient voltages, carry occasional
• Special (Gas-Insulated Lines [GIL], High Tempera-
fault currents, withstand the rigors of installation, and
ture Superconducting [HTS] and Submarine)
operate satisfactorily for many decades under thermal
and mechanical stresses. Chapter 4 on extruded dielec-
2.3 ELECTRICAL CHARACTERISTICS trics, Chapter 5 on pipe-type cables, Chapter 6 on self-
Electrical characteristics of underground cables differ contained fluid-filled cables, and Chapter 7 on special
from overhead lines—and these differences should be cable constructions provide details specific to each cable
taken into account when evaluating underground versus type. The present section discusses component functions
overhead lines and when comparing types of under- in general and provides comparative information among
ground lines. cable types.
High Capacitance 2.4.1 Conductor
High capacitance is a consequence of thin insulation
and higher dielectric permittivity than air—for example, The purpose of the conductor is to carry the required
2.3-3.5 or higher. A cable circuit installed in an ac over- current for the cable system. Conductor material, size,
head line network, which is naturally inductive because and construction are selected to achieve the required
2-4
current for installation and operating conditions with- ing installation or in the event of damage during opera-
out exceeding allowable temperatures and with consid- tion. Chapter 4 provides more details.
eration of effects on cable system operation and cost.
Insulating individual conductor strands reduces the ac
Copper and aluminum are used as conductor materials. resistance of copper conductors (aluminum oxide
Copper has lower electrical resistivity but also has quickly forms on stranded aluminum conductors, giving
higher cost and weight. Generally, copper conductors the same effect.) Enamel coatings may be applied to the
are employed for high-power transfer such as tie lines strands, or the strands may be intentionally oxidized.
and major generator leads and form a large percentage Chapter 11 quantifies the benefits on cable ratings.
of new installations. Aluminum conductors are used for
lower power-transfer applications and where their Conductor binders, often consisting of stainless steel
lighter weight gives an advantage. Almost all conduc- tapes, are tightly wound around many conductors to
tors are stranded to reduce ac losses and give greater ensure that the conductor stays round during the
flexibility, but solid aluminum conductors have been manufacturing process. Shifting of strands or segments
used for sizes as large as 4000 kcmil (2000 mm2). could lead to a non-round conductor and intensify
electrical stresses.
Conductor constructions are optimized for cable types
and specific applications. The most common types for 2.4.2 Semiconducting Shield
flexible cables are described below. Chapter 7 describes Semiconducting materials are applied over the conduc-
the tubular aluminum conductor used for gas- tor (including binder, if present) and over the cable insu-
insulated lines. lation to form an equipotential surface that smoothes
• Compact segmental: four to six segments are used for irregularities and protrusions from the conductor or
extruded-dielectric and pipe-type cables. A layer of outer cable shielding. This is necessary to avoid areas of
insulating or semiconducting paper is placed between high electrical stress in the insulation.
segments to reduce ac losses due to skin and proximity
effects in the conductor. Compact segmental conduc- As described in Chapter 4, extruded-dielectric cables use
tors are used in high-current capacity applications a semiconducting compound. Paper-insulated cables use
with large conductor sizes. carbon black or metallized paper tapes, as described in
Chapters 5 and 6.
• Compact round has the same number of strands as
concentric conductors, but the strands are compacted
2.4.3 Insulation
to 90% the size of the equivalent concentric conduc-
tor. This conductor is used for smaller extruded- The insulation forms the heart of a cable and is the
dielectric and pipe-type cables since ac loss effects are primary differentiation among cable types. Insulation is
lower for these smaller conductors. required to reliably withstand normal and transient
voltages over the lifetime of the cable in the presence of
• Compressed conductors, which are concentric round thermo-mechanical stresses imposed by cable loading.
conductors passed through a die to flatten the outer Subsequent chapters describe in detail insulation
strands, are specified for many extruded-dielectric systems for each cable type. A brief summary is
cables. These conductors are about 97% the size of provided below.
the equivalent concentric conductor.
• Concentric conductors, which are essentially compact Extruded-dielectric
round conductors without the compaction, are used The insulation material is extruded over the conductor
principally at distribution voltages, but are sometimes shield and then crosslinked for crosslinked polyethylene
used for transmission-voltage conductors. or ethylene-propylene rubber. Chapter 4 describes the
material and manufacturing procedures. Note that a
• Hollow-core (annular) conductors have a central true triple extrusion is universally employed for extra-
hollow core to allow fluid passage for self-contained high-voltage (EHV) extruded-dielectric cables (conduc-
fluid-filled cables. Several constructions are used. See tor shield, insulation, and insulation shield are applied
Chapter 6. in a common set of extrusion heads) and that dry curing
is standard (steam heating is no longer used). Some
Conductors for extruded-dielectric cables sometimes are manufacturers also employ dry cooling.
specified to have a strand-block material, often applied
to distribution-voltage conductors. A water-swellable Impregnated Paper
powder, gummy material, or water-swellable tape is Individual paper or laminated paper polypropylene
applied to prevent longitudinal water penetration dur- (LPP) tapes are helically wrapped around the conduc-
2-5
tor, then the cable core (conductor plus insulation and • A sheath provides a pressurizing enclosure for SCFF
semiconducting tapes) is impregnated with dielectric liq- cables.
uid. This process is described in Chapter 5 for pipe-type
cables and in Chapter 6 for self-contained fluid- The metallic shielding/sheath has traditionally been
filled cables. extruded lead-alloy or aluminum for XLPE extruded-
dielectric and SCFF cables. Alternative shield/sheath
Gas-spacer
constructions are becoming more common with copper
GIL systems use rigid epoxy spacers and gaseous (SF6
drain wires for fault currents and a thin copper or alu-
or nitrogen/SF6 mixtures) to provide the insulation
minum tape water barrier bonded to the underside of
between tubular conductor and tubular enclosure.
the jacket, or longitudinally seam-welded and corru-
gated copper or aluminum sheath. Chapter 4 provides
2.4.4 Outer Layers, Shielding, and Sheath
more details for extruded-dielectric cables, and Chapter
As indicated in Section 2.4.2, a semiconducting shield is 6 describes sheaths for SCFF cables.
placed over the insulation for both extruded-dielectric
and paper-insulated cables to avoid electric field Metallic shielding tapes for a pipe-type cable are typi-
enhancements that can lead to cable failure. cally stainless steel, copper, or zinc, intercalated with a
polyester tape. This assembly has a high shield resis-
Outer layers are provided over the semiconducting tance designed to reduce shield losses as described in
shielding. These outer layers can serve several functions: Chapter 5. The steel pipe carries the majority of any
Moisture Absorption Layer—Extruded-dielectric fault current.
Cables Jacket and Other Layers—Extruded-dielectric and
Many XLPE extruded-dielectric cables have water- SCFF
swellable tapes or powder that restrict the passage of Extruded-dielectric and SCFF cables have an extruded
moisture in the event of damage to the cable jacket and polymeric jacket (called oversheath outside North
moisture barrier. Chapter 4 provides more detail on the America) to isolate the shield/sheath from ground and to
moisture-absorbing layer. provide mechanical and corrosion protection. See Chap-
Moisture Barrier—Pipe-type Cables ters 4 and 6. Submarine cables are typically supplied
This layer also is called the moisture seal or outer with armoring to withstand tensions during cable laying
shielding assembly. It is a semi-permeable seal to limit and retrieval and contain other protective layers, as
impregnant drainage and reduce moisture penetration described in Chapter 7. Each of these cable types also
during storing, shipping, and installing pipe-type cables. can contain an integral or external optical fiber to per-
Typically, a metallized polyester tape or polyester tape mit using a distributed fiber optic temperature sensing
intercalated with a metal tape is used for the moisture (DFOTS) system to determine temperature along the
barrier; see Chapter 5. cable line. Chapter 16 describes the DFOTS system and
its application to cable operation.
Cushioning Tapes—Extruded-dielectric Cables
XLPE extruded-dielectric cable insulations have high Skid Wires, Armor Wires, Temporary Lead-alloy
coefficients of thermal expansion, and operation at Sheath—Pipe-type
emergency temperatures could damage the outer metal- Pipe-type cables have a skid wire applied over the outer
lic shielding. Cushioning tapes—which are often also shielding of each core to protect it from damage and to
water-swellable tapes—are provided to absorb much of reduce the coefficient of friction during cable pulling.
this expansion as described in Chapter 4. Two D-shaped skid wires, made of stainless steel, brass,
or zinc, are applied helically along the cable. For long
Metallic Shielding/Sheath—Extruded-dielectric and vertical installations, pipe-type cables also can be sup-
Paper-insulated Cables plied with long-lay stainless steel armor wires to share
The metallic layer serves the following purposes: the loading caused by cable weight. A temporary,
• Provides a return path for capacitive charging current. removable lead-alloy sheath may be applied to a pipe-
type cable section that is to be kept in storage for long
• Provides adequate short-circuit current-carrying periods of time. See Chapter 5 for a further discussion
capacity for extruded-dielectric and SCFF cables. of these items.
• Gives some protection against external mechanical
Dielectric Fluid
damage.
Pipe-type and SCFF cables require a pressurized
• A sheath prevents moisture ingress in addition to the dielectric fluid. A low-viscosity fluid is used in an SCFF
three items listed above. cable so that gaseous voids cannot form during thermal
2-6
contraction. In a pipe-type cable, a more viscous fluid is The cables must make a transition from three cables in
used to transmit a high pressure of 200 psi (1380 kPa) or one pipe for the majority of the length to individual
greater to any voids that may form to prevent them from cables in individual pipes to reach the cable termina-
discharging. Historically, mineral oils were used for tions—which may be 8 to 15 ft (2.4 to 4.6 m) apart.
liquid-filled systems, but synthetic liquids—alkylben- Non-magnetic stainless steel pipes are used for this
zenes and polybutenes or a mixture of the two—have application, although some older systems used copper
predominated since the 1970s. Nitrogen gas can be used pipes. This transition takes place in a special joint—
for pipe-type cables up through 138 kV for a high- known as a trifurcating joint—or in a jointless assembly
pressure gas-filled (HPGF) cable system. Nitrogen gas known as a trifurcator or spreader head.
has been used for self-contained gas filled (SCGF) cable
systems up to 150 kV. Chapter 5 provides more detail on the pipes, and
Chapter 13 describes the steps that are needed for
Typical operating pressures are 180-250 psig (1240- pipe installation.
1725 kPa) for pipe-type cables, and 15-40 psig (103-
276 kPa) for SCFF cables. However, pressures can be For extruded-dielectric and SCFF cables, the duct is not
substantially higher due to elevation changes or pres- considered an integral part of the cable system (both
sure drop due to fluid circulation for cooling a high- cable types can be installed without ducts—directly bur-
pressure fluid-filled cable system pressurized with a ied or in tunnels). Ducts are addressed in Chapter 13.
dielectric liquid. (Note that HPGF cable systems do not
have significant elevation heads, and gas circulation is Sections 2.5 through 2.8 describe how these components
not feasible.) Special designs are needed for the cable are applied to the different cable types.
systems to allow them to tolerate these high pressures.
Chapters 5 and 6 describe dielectric fluid properties and 2.5 EXTRUDED-DIELECTRIC CABLES
operating pressures in more detail. Expansion and con-
Extruded-dielectric cables, also known as solid-
traction of dielectric liquid with load changes or ambi-
dielectric cables, use either crosslinked polyethylene
ent earth temperature changes require a pressure control
(XLPE) or ethylene-propylene rubber (EPR) insulation.
system and reservoirs. Section 2.9.4 of this chapter pro-
Low- and high-density polyethylene (LDPE or HDPE)
vides a summary of pressure control systems, and Chap-
insulations also have been used but are very uncommon
ter 9 provides more detail.
for cables installed since the 1980s.
Steel Pipe—Pipe-type Cables
The three insulated conductors are installed in a carbon- XLPE and LDPE/HDPE cables are characterized by
steel line pipe. The pipe’s functions are to very small dielectric losses relative to paper and EPR
cables. These losses are essentially constant up to
• contain the fluid operating pressure, and
emergency operating temperatures in contrast to
• provide a path for fault currents. impregnated paper-insulated cables where dielectric
losses can increase substantially at high temperatures.
However, the pipe also provides the following added At transmission voltages, the low dielectric loss of
benefits: XLPE cables results in important improvements in
• a conduit for installing the cables, ampacity and offers an additional safety factor against
thermal runaway. (Thermal runaway is the condition
• mechanical protection, and where a cable temperature increases, causing an increase
• a hermetic seal to keep moisture from the cables. in electrical losses that further raises the temperature,
causing additional losses until the cable overheats and
The pipe is coated internally to prevent rust before fails.) The extruded-dielectric cable’s lower capacitance
installation, and it is coated externally to prevent corro- extends the critical line length (the length at which all
sion during operation. The most commonly used pipe current-carrying capacity is required to supply charging
size is a nominal 8 in. (20 cm), although pipes as small current) significantly and does not contribute as severely
as nominal 5 inches (13 cm) have been used for 69-kV to overvoltage problems on weak transmission systems.
and 138-kV cables. Pipes as large as nominal 12 in.
(30.5 cm) have been installed for some 345-kV installa- Transmission-voltage extruded-dielectric cables have
tions. As described in Chapter 13, pipe ends are flared been used extensively in North America at voltages
and a backing ring is inserted between adjacent sections 138 kV and higher since the mid-1980s, and they have
to ensure a smooth internal profile and avoid cable dam- seen extensive use overseas since the 1970s. XLPE-
age during installation and operation. insulated cables have been used at 400 kV and higher
2-7
throughout the world since the late 1990s (Sakuma and ular for underwater installations and within substations
Shinagawa 2000). Two 25+ mile (40+ km) 500-kV requiring tight bending radii.
installations have been in service in Japan since 2001
(Ohki and Yasufuku 2002). Installation for several Chapter 4 provides a great deal more information on
major 345-kV XLPE installations, totaling more than extruded-dielectric cable construction and manufactur-
100 circuit miles (160 circuit km), began in North ing, and Section 2.10 of this chapter provides compari-
America in the early 2000s. sons of extruded-dielectric cables to other cable types.
Extruded-dielectric cable installations in North 2.6 PIPE-TYPE CABLES

America are commonly in duct banks, one cable per
Pipe-type cables have the three cable phases insulated
duct, because direct burial and tunnel installations have
with tapes of kraft paper or laminated paper polypropy-
not proven feasible in city streets. Overseas, most
lene, installed in a common steel pipe. The pipe is pres-
transmission-voltage extruded-dielectric cables are
surized, usually with a dielectric liquid, and is called a
installed directly buried, with some in utility tunnels.
high-pressure fluid-filled (HPFF) or sometimes a high-
E x t r u d e d - di e l e c t ri c c abl e s a re c o n s i d e re d l e s s
pressure liquid-filled (HPLF) cable system. This cable
complicated in terms of installation and accessories
system was called high-pressure oil-filled (HPOF) until
than SCFF or pipe-type cables, but the sheath bond-
the 1970s when synthetic dielectric liquids began to
i n g / g ro u n di n g s y s t e m c a n b e c o m p l ex , a s wi t h
replace the mineral oils that had been used earlier. Pipe-
SCFF cables.
type cables rated up to 138 kV (and possibly 230 kV in
the future) that are pressurized with nitrogen are called
Figure 2-1 shows a typical 345-kV XLPE-insulated
high-pressure gas-filled (HPGF) cables.
cable, and Figure 2-2 shows a typical trench cross-
section for a duct-bank installation.
North America is the major user of pipe-type cables,
with the earliest installation in the 1930s and significant
EPR-insulated cables are installed at voltages through
installations at 345 kV beginning in the early 1960s.
138 kV. They are not employed at higher voltages, prin-
There also are installations in Ireland and the Middle
cipally because of their relatively high dielectric losses
East, and Japan has a few pipe-type installations.
compared to XLPE-insulated cables. However, their
Although 345 kV is the highest commonly used voltage
greater bending flexibility, and greater resistance to
moisture degradation and ionization, makes them pop-
Figure 2-1 345-kV XLPE-insulated cable (courtesy of

Figure 2-2 Typical duct bank for city-street installation
Silec Cable).
(courtesy Northeast Utilities).
2-8
for pipe-type cables (and the highest voltage in wide- erations do not exist with a high-pressure gas-filled
spread use in North America), there has been at least cable system.
one commercial installation at 500 kV overseas. Exten-
sive testing at the EPRI Waltz Mill facility in Pennsylva- Figure 2-3 shows a 345-kV HPFF cable, and Figure 2-4
nia in the 1980s proved its suitability for voltages shows a typical two-pipe trench installation.
to 765 kV.
Chapter 5 provides additional information on pipe-
Pipe-type cables have been the most commonly used type cable systems, and Chapter 13 describes cable
transmission cable type in North America through the system installation.
early twenty-first century for several major reasons:
• The pipe provides rugged protection against third- 2.7 SELF-CONTAINED FLUID-FILLED CABLE
party damage. SYSTEMS
• The system is reliable. The first transmission cables used in North America
were self-contained fluid-filled in the mid-1920s in
• The pipe can be installed in relatively short street Chicago and New York. The self-contained cable is
openings, minimizing traffic disruption in crowded insulated with kraft paper or laminated paper-polypro-
urban areas, and it can generally be installed more pylene tapes and internally pressurized with a dielectric
quickly than a concrete-encased duct bank. liquid, sometimes called self-contained liquid-filled
• The dielectric liquid provides several options for (SCLF) cables in North America, but more commonly
cooling to improve the circuit rating. SCFF. This cable type was called low-pressure oil- filled
(LPOF) until the 1970s when synthetic dielectric liquids
A major innovation in pipe-type cable systems was the began to replace the mineral oils that had been used
EPRI-funded development of laminated paper- earlier. Early cables were generally pressurized to 5-15
p o ly p ro py l e n e ( L P P ) i n s u l at i o n , wi t h t h e fi r s t psig (35-105 kPa) while newer designs with aluminum or
commercial installation in Boston in 1987 (Allam 1988). reinforced lead-alloy sheaths commonly operate at 5-
This insulation has higher electrical strength and lower 75 psig (35-520 kPa). Commercial installations have
dielectric losses than kraft paper—its higher power pressures as high as 400 psig (2760 kPa) using reinforced
transfer has permitted 345-kV cables to be installed in a sheaths (Ray et al. 1974). Small liquid reservoirs are
nominal 8-in. (20-cm) pipe rather than a nominal 10-in. placed along the route, although long submarine lines
(25-cm) pipe used for 345-kV kraft paper insulated have a pressurizing plant at one or both ends. The
systems. The smaller pipe has many advantages, presence of the dielectric liquid and reservoirs, plus the
including cost, bending radii, trench size, and amount of complexity of sheath bonding, are disadvantages of the
dielectric liquid. Most newly installed 345-kV HPFF SCFF cable system.
cable is now insulated with LPP, and it is used for many
230-kV and some 138-kV installations.
The three phases making up a line are pulled together

into a previously installed coated and cathodically
protected steel pipe, with distances commonly 3000 ft
(980 m) between splices for 345-kV cables. After the
entire line is installed, including splices and termina-
tions, and the pipe welding is completed, the line is
evacuated, then filled with a dielectric liquid (or
possibly nitrogen gas for lines up through 138 kV). The
liquid is pressurized to a nominal 200 psig (1380 kPa)
using a pressurizing plant that also has a reservoir tank
to accept volume changes due to system thermal
expansion and contraction, as well as pumps, controls,
and alarm systems. The presence of the dielectric liquid Figure 2-3 345-kV HPFF cable (courtesy of the Okonite
and need for a pressurizing plant must be taken into Company).
account for the HPFF cable system, but those consid-
2-9
The highest voltage SCFF application in North Ameri- cables are now beginning to displace them. Chapter 6
can commercial service is 525 kV, for the double-circuit provides detailed descriptions of the SCFF cable system.
Vancouver Island submarine cable interconnection
(Crowley 1983). SCFF cables have been tested to 800 kV The self-contained cable system consists of three individ-
in Canada (Couderc 1996) and 765 kV in Europe. ual phases, each contained within a hermetically sealed
Designs at 1100 kV have undergone preliminary tests metallic sheath that is typically extruded lead-alloy or
(Occhini et al. 1978). SCFF cables may be directly bur- aluminum. The cables are insulated with a high-quality
ied, or they may be installed in ducts or in tunnels. They taped insulation. Both kraft paper and laminated paper-
have been the principal cable type for EHV installations polypropylene insulations are employed. The fluid pres-
outside North America, with conductor sizes up to sure necessary to prevent ionization is maintained
6000 kcmil (3000 mm2), carrying the highest power levels through a hollow core in the center of the conductor.
for a cable in the world. However, extruded-dielectric Figure 2-5 shows a SCFF cable. The cross-section of a
Figure 2-4 Two-pipe trench installation.
Figure 2-5 Self-contained fluid-filled cable (courtesy of Prysmian Cables and Systems).
2-10
SCFF cable duct bank would be very similar to that cable is supplied in rigid 40- to 60-ft. (12- to 18-m) sec-
shown in Figure 2-2 for extruded-dielectric cables. tions. Field joints are made between each section. There
is minimal service experience in buried installations. The
An advantage of SCLF LPP insulation cables compared GIL system is normally installed aboveground or in
to extruded-dielectric cables is the insulation is thinner underground troughs or tunnels.
for the same conductor size and voltage class. This
allows longer cable lengths on the same reel size and, in Terminations consist of a cast resin insulator interfacing
some instances, can provide the opportunity to install with SF 6 -insulated switchgear bus bar or a porcelain
230- to 275-kV-class cables into existing 5-in. (125-mm) shell similar in appearance to a conventional pothead.
ID ducts. Internal stress control generally requires only a horn-
shaped electrical grading shield in the shell. The same
2.8 SPECIAL APPLICATION CABLES gas that pressurizes the bus also pressurizes the porce-
lain. Insulators for GIL cables are subject to distortion
Extruded-dielectric, pipe-type, and self-contained fluid-
of the stress distribution through accumulation of space
filled cable systems account for the great majority of
charge and particulate contamination, although manu-
installed lengths of transmission cable systems through-
facturers have developed particle traps to control con-
out the world. However, other novel cable types for spe-
taminants. Since SF6 is considered a “greenhouse gas,”
cial applications exist. These installations can be high
care must be taken to minimize any leakage from the
cost and require a great deal of engineering analysis and
GIL system, hence the development of the 90% N 2 –
design. These special application cables are summarized
10% SF 6 mixture. Figure 2-6 shows a section of GIL,
below and are described in greater detail in Chapter 7.
including an elbow for making major direction changes.
2.8.1 Gas-insulated Lines
2.8.2 Submarine Cables and Dc Cables
Gas-insulated lines (GIL) have been used in a few appli-
The core (conductor, insulation, outer semiconducting
cations where high-power capabilities are required, such
layer, and fluid supply, if present) of a submarine cable
as in substation ties and through tunnels. The first com-
is essentially the same as that for SCFF and extruded-
mercial underground line was a 345-kV, 600-ft (183-m)
dielectric land-based cables described earlier in this
dip installed in 1971, which had a 3300-ampere
chapter. One major difference is that the cable is
(2000-MVA) capacity (Williams 1972). Since GIL are
mechanically designed to be either coiled or wound
more costly than other cable types, they have not yet
onto a turntable to permit very long splice-free lengths
been applied to long transmission lines.
and to be installed—and perhaps to be retrieved—at
significant water depths.
GIL is similar to the SF6-insulated metal-clad bus bar
connected to switchgear. GIL has been internationally Alternating-current Cables
defined as having a circuit length of greater than 1640 ft As described in Chapter 7, most long submarine lines
(500 m) and being installed outside utility owned land— use dc cables, although there are important ac subma-
on roads or rights-of-way (CIGRE 2003).
Several names have been used to describe this cable: gas-

insulated line, electronegative gas/spacer, SF 6 cable,
compressed-gas-insulated cable, and gas-insulated
transmission line.
GIL cable was used for a 525-kV underground applica-

tion in the Bonneville Power Administration system—
and it has been used aboveground for a 765-kV applica-
tion approximately 100-ft (30.5-m) long on the Ameri-
can Electric Power system.
The GIL system is inherently simple. The conductor

and enclosure are made of aluminum tubing, cast-
epoxy-resin spacers hold the conductor in place, and the
gas gap insulation is typically SF 6 gas pressurized to
30-90 psig (200-600 kPa)—or more recently a mixture of
90% nitrogen and 10% SF 6 at a higher pressure. The Figure 2-6 GIL, showing elbow (courtesy of CGIT).
2-11
rine cable links (Crowley 1983; Grzan et al. 1993). The several 138-kV extruded-dielectric submarine cables
choice of a dc power transmission cable system over an appeared in North America beginning in the early
ac cable would generally be made in cases where the 1990s. Both extruded-dielectric and SCFF cable types
power transfer requirement is greater than 250-300 MW may be supplied in very long lengths splice-free or with
and the length is typically greater than 25 miles factory splices. Recent installations have occurred in
(40 km)—or between two large ac transmission net- North America with 138-kV three-conductor XLPE-
works where power transfer control is a potentially seri- insulated submarine cables. At the time of writing, a
ous problem. A dc cable provides an asynchronous or contract has been awarded for three 11-mile (19-km)
flexible transmission interconnection. 138-kV 3-core XLPE, 150-MW submarine cable
circuits across Long Island Sound to replace the seven
Figure 2-7 shows the 345-kV self-contained fluid-filled existing 30-year-old SCFF cables. A 1.9-mile (3.2-km),
cable used for a crossing of the Long Island Sound. 660-ft (200-m) deep 420-kV 1-core XLPE submarine
cable connection is also pending to Gossen Island,
In the HV and EHV voltage ranges, from 115 kV associated with the offshore Ormen Lange production
upwards, existing ac submarine cable lines longer than platform in the North Sea. This will be the highest
about 6 miles (10 km) are dominated by the SCFF type. voltage XLPE submarine cable system in the world.
The cables differ from the SCFF land cables described
in Section 2.7 in several regards: the hollow core is often Pipe-type submarine cables would require at-sea splices
larger to accommodate fluid expansion and contraction every mile or so and are, therefore, limited to shorter
over very long lengths, the dielectric liquid may be a lines for practical reasons.
lower viscosity to reduce pressure drop during expan-
sion and contraction, and the cable has several addi- EPR-insulated cables are employed for 138-kV and
tional layers external to the core to account for rigors of lower submarine cable lines with lengths less than a
installation and operation in a marine environment. In mile. EPR insulation does not deteriorate at the same
addition, many long SCFF submarine cables have active rate in water, and some installations use a “wet design”
pressurizing plants similar to those for pipe-type cables with no moisture barrier.
to accommodate the large fluid volumes and large pres-
Direct-current Cables
sure changes inherent to a long submarine cable. See
As described in Chapter 7, almost all direct-current
Chapter 7.
cables are installed in long submarine crossings. With
few exceptions, land dc cables are limited to the shore
Extruded-dielectric cables are beginning to be applied
ends of submarine crossings. Pipe-type cable has been
at the HV and EHV range for ac submarine cables—
tested for 600-kVdc operation, but pipe-type cable is not
used for long submarine crossings, and there is no pipe-
type dc installation on land.
Three types of submarine power cable can be considered

for high-voltage, direct-current (HVDC) operation,
namely
• self-contained, fluid-filled cables;
• mass-impregnated cables; and
• extruded-dielectric cables.
Self-contained, fluid-filled cables are very similar to the

ac cables previously described. Mass-impregnated cables
are paper-insulated cables that employ a very high viscos-
Figure 2-7 Self-contained fluid-filled cable (courtesy of ity compound to prevent drainage during operation.
Prysmian Cables and Systems). Mass-impregnated cables have no pressurizing source.
They account for the largest percentage of dc submarine
cables by far. The longest submarine cable interconnec-
2-12
tion in the world, the 363-mile (580-km), 1345-ft (410-m)- tion projects were under way in 2006, with voltages to
deep, 450-kV dc, 700-M W Norway-Netherlands 115 kV and lengths to 1140 ft (350 m). These short lines
‘NorNed’ link uses mass-impregnated cables. are installed both for laboratory testing and operating
in a commercial utility grid.
Figure 2.8 shows two mass-impregnated submarine
cables. Commercial applications for HTS cables should be able
to use the unique features of superconducting materials
Extruded-dielectric cables had not been used for high- to deliver more power in a smaller space with reduced
voltage dc power transmission until the early part of the impact to the surrounding environments. Thermal and
twenty-first century because of space-charge buildup electromagnetic emissions to the surrounding soil and
due to the unidirectional field and subsequent distortion adjacent equipment are virtually eliminated with HTS
of the electrical stress distribution in the cable insulation cables. Environmental impacts are further reduced ver-
that reduces the dc and impulse strength of the cable. sus other cable types due to the environmentally friendly
However, recent progress in developing modified XLPE nature of liquid nitrogen as the cooling medium.
has apparently been successful in reducing the space-
Cable Design Options
charge problem. As a result, the first modified XLPE
The electrical insulation materials for HTS cables may
HVDC submarine cable link (the 150-kVdc, 330-MW
be inside or outside the cryostat, and single or multiple
Cross Sound Cable Link) was installed in Long Island
electrical phases may be placed inside a single cryostat.
Sound in 2002. Another modified XLPE HVDC sub-
All HTS cables start at the inner most layer with a
marine cable installation was commissioned in 2006.
former. This former may be a stranded copper member
This link between Finland and Estonia, a distance of
or a hollow tube. The function of the former includes
some 47 miles (75 km), operates at 150 kVdc and can
serving as a structural member to strand HTS wires on
transmit 350 MW of power.
top of and, in some cases, a flow path for liquid nitro-
gen. In the case of stranded copper cores, the copper
2.8.3 High-temperature Superconducting Cables
serves as fault current stabilization to protect HTS wires
Beginning in the early 1990s, utilities and manufacturers in the event of a fault current on the cable.
began developing high-temperature superconducting
(HTS) cables—cables with conductors that become On top of the former, one or more layers of HTS wires
superconducting at temperatures near liquid nitrogen are helically wound to make up the phase conductor of
temperature (around -320°F, or 77 K). Nine demonstra- the cable. It is these HTS wires that replace the copper
or aluminum conductor in a conventional cable.
Options for the placement of the dielectric materials dic-
tate if this cable core becomes a warm dielectric or cold
dielectric HTS cable.
Warm Dielectric
The HTS core may be placed inside the cryostat directly.
Conventional dielectric layers are then extruded on the
outside of the cryostat to operate at ambient (warm)
temperatures. Dielectric materials for this application in
previous projects have included EPR and polyethylene.
Conventional concentric wire screen may be applied to
the outside of the dielectrics.
Figure 2.9 shows a warm-dielectric HTS cable produced

by nkt cables in Denmark.
Cold Dielectric
In cold dielectric cable designs, the stranded HTS core is
wrapped with layers of dielectric materials that may be
made of paper or polymer tapes. These tapes operate
with the liquid nitrogen coolant saturating the material,
Figure 2-8 450-kVdc Baltic Cable (left) and 100-kVdc
Gotland Island Cable (right) (courtesy of ABB HV Cables). filling butt-gaps and serving as part of the dielectric sys-
tem. The liquid nitrogen in these cables serves much the
same purpose as impregnating liquids in conventional
2-13
paper cables. The fully insulated conductor may have a Chapter 7 provides a more complete description of high-
second layer of HTS wires helically wrapped to serve as temperature superconducting cable systems and their
the concentric screen. The superconducting screen in potential benefits to utilities.
these cables may operate at 100% of the load current,
resulting in a fully shielded cable design with no mag- 2.9 CABLE ACCESSORIES
netic field outside the cable structure. This lack of an
Since accessories for cable systems are installed in the
electromagnetic field (EMF) eliminates eddy current
field, they are often the weak point in terms of cable
heating in surrounding equipment or steel duct and
operation and reliability. Great care must be taken in
allows cold dielectric HTS cables to operate at higher
designing, qualifying, specifying, and installing accesso-
load densities than warm dielectric cables.
ries. Chapter 8 provides details of accessories for each
cable type. The listing below provides a summary of
Cold dielectric cable designs afford the opportunity to
accessory requirements. Splices and terminations are
optimize the three-phase structure to reduce material
required for all cable types; other accessories may be
usage and reduce physical spacing. Options for cold
required for specific cable systems.
dielectric cables include single-phase per cryostat, three
phases twisted together inside a single cryostat, and
2.9.1 Splices
three concentric phases on a single core all inside a com-
mon cryostat. Splices—also called joints when all components such as
casings and valves are included—are required to con-
nect two sections of cable. The distance between splices
is generally determined by allowable cable lengths,
which may be limited by weight or diameter of ship-
ping reels, pulling tensions, or steady-state sheath volt-
ages in the case of high-current extruded-dielectric and
SCFF cables.
All splices consist of a connector to join adjacent

conductors, a semiconducting layer to smooth the
electric field, restoration of the insulation, outer semi-
conducting layers, and outer metallic shielding/sheath.
Additional components are required depending on cable
type and application. Proper evaluation of thermo-
mechanical forces is important for satisfactory long-
Figure 2-9 Warm dielectric HTS cable (courtesy of nkt). ter m splice operation. Figure 2-10 shows basic
components required for a splice.
Figure 2-10 Splice components.
2-14
Considerations for specific cable types are summarized ations where high differential thermo-mechanical or
below: gravitational forces exist, such as on inclines.
• Extruded-dielectric and SCFF cables require sealing
Splices for extruded-dielectric and SCFF duct systems,
against moisture ingress. Both types generally require
and for pipe-type cables, are typically placed in concrete
insulating breaks in the outer semiconducting layers
manholes. Splices for directly buried cables can be made
and shielding/sheath and means to connect bonding
on concrete pads with sidewalls called joint bays that are
leads to permit applying special bonding methods to
then covered with thermal backfill and have concrete
reduce circulating currents and maintain acceptable
slabs placed over them for mechanical protection. Typi-
sheath voltages. Manufacturers have different meth-
cal splicing durations for three-phase 345-kV cables are
ods of accommodating forces and movement due to
6 to 10 eight-to-twelve-hour days for extruded-dielectric
thermal expansion and contraction ranging from rig-
cables, and 6 to 8 days around-the-clock for pipe-type
idly clamping the splice to prevent movement, to pro-
and SCFF cables.
viding expansion loops in the manhole to
accommodate expansion/contraction.
Chapter 13 describes splice installation in more detail.
• Pipe-type cable splices require supports to prevent
excessive movement in the joint casing, and they 2.9.2 Terminations
require casings to restore the continuity of the steel
Terminations—also called potheads or sealing ends—
pipe. As described in Chapter 8, special-purpose
are required to bring the cables from three individual
pipe-type splices include features to isolate the dielec-
conductors in close proximity to three separate units
tric liquid, trifurcate from three cables in a single car-
that may be 8–15 ft (2.4–4.6 m) apart for connection to
bon steel pipe to individual cables in stainless steel
substation bus, overhead lines, or apparatus such as
pipes, or support the cable for large elevation
transformers. Terminations provide electrical stress con-
changes.
trol, insulation between the conductor and ground, a
• Self-contained fluid-filled splices are either “through seal against the external environment, anchor conductor
splices” that have an open channel in the conductor movement, and withstand positive pressure for pressur-
for passage of the dielectric liquid, stop joint splices ized cable systems.
to isolate dielectric liquid sections, or “feeding”
splices to provide fluid passage to the reservoirs. The Termination housings for extruded-dielectric cables may
latter two-splice types are commonly combined— be either porcelain or composite; termination housings
with a central stop feature and a feeding feature on for pipe-type, SCFF, and GIL systems are almost
both sides always porcelain.
• GIL splices just restore the continuity of the conduc-
tor and the enclosure; the SF6 gas or nitrogen / SF6 There are three types of terminations:
mixture maintains the electrical insulation. Gas seg- • Outdoor/indoor terminations
regation splices are provided at regular intervals to
• Metal-enclosed GIS terminations
limit the volume of gas escape in the event of a leak.
• Oil-immersed transformer terminations.
• Submarine cable splices—described in Chapter 7—
can be flexible splices of the same diameter as the
Extruded-dielectric terminations at voltages of 230 kV
cable to aid in installation, or rigid splices, which are
and lower are generally completely self-contained; they
usually used for shallow water field repairs (~< 165 ft,
do not require a pressurizing source and, therefore, do
or 50 m, depending on manufacturer and laying
not require alarms. Extruded-dielectric terminations at
methods). Rigid joints cannot withstand bending
345 kV and above typically have a dielectric fluid (SF6
during laying and usually have lower tension-
gas or a liquid) that is pressurized and requires alarms
withstanding capabilities compared to flexible splices.
to indicate low pressure that would signal loss of fluid.
However, installation time for rigid joints is
considerably less than field flexible joints.
The fluid supply for pipe-type, SCFF, and GIL systems
• Anchor joints are used in special locations within is integral with the cable fluid supply, so it is monitored
each type of underground cable system. Their pur- as part of the cable system.
pose is to prevent movement of the conductor in situ-
2-15
Figure 2-11 shows a cable termination. Chapter 8 removed (with the cable de-energized) when the user
describes termination components that are specific to performs a jacket integrity test. The link box also may
individual cable types. contain sheath voltage limiters to limit transient over-
voltages that might damage the cable jacket or the shield
Outside North America, many self-contained cables are interruption within the splice. Chapter 8 provides more
terminated within a compact substation by either a gas- detail and shows photos of link boxes.
immersed switchgear (GIS) ter mination directly
interfacing with gas-insulated metal-clad bus bar or by 2.9.4 Fluid-handling Equipment
transformer terminations directly into a fluid-insulated Pipe-type and SCFF cables require equipment to main-
chamber attached to the transformer. GIS terminations tain fluid pressures in design ranges during all changes
employ cast resin insulators, which are usually of the in loading and ambient temperatures.
“ d r y - t y p e ” ( t h e y c o n t a i n n o i n s u l at i n g f l u i d ) .
Transfor mer ter minations employ cast resin or High-pressure fluid-filled cables require at least one
porcelain insulators. pressurizing plant, with a reservoir tank ranging from
2000 gal (7570 l) to 15,000 gal (56,780 l) or even larger.
2.9.3 Link Boxes Special relief valves maintain required pressure ranges
As described in several subsequent chapters, the bond- during fluid expansion and contraction. Pressures are
ing and grounding of extruded-dielectric and SCFF recorded and often transmitted remotely. An annuncia-
shields/sheaths is an important consideration to reduce tor panel monitors a series of alarms. Since maintaining
shield/sheath losses and increase cable ampacity. Con- pressure is so critical for integrity of the cable system,
nections from the cable shield/sheath are generally made most plants have a backup pressurizing source such as a
through a link box, which has bolted links that can be nitrogen-driven pump or are connected to an emergency
Figure 2-11 Extruded-dielectric composite termination.
2-16
power supply to be used if local distribution power The cathodic protection system places a small negative
is lost. dc voltage on the pipe to insure that current flows to the
pipe if the coating is damaged rather than away from the
Most pressurizing plants are housed in self-contained pipe carrying ionic metal. That dc voltage is isolated
aluminum-paneled buildings, which can be approxi- from ground by the pipe corrosion coating. However,
mately 12-ft (3.6-m) wide and 45-ft (13.7-m) long. the pipe should have a low-impedance path to ground
for fault currents to prevent dangerous voltages on the
High-pressure gas-filled and self-contained gas-filled pipe. Polarization cells, resembling substation batteries,
cables require a monitoring cabinet with alarms to indi- are used for this purpose, as are their solid-state equiva-
cate high or low pressures and a regulator with a few lent known as an isolator surge protector. Chapter 10
cylinders of nitrogen to clear an alarm (much larger describes cathodic protection of cable pipes.
quantities would be needed to feed a leak until repairs
are made). The nitrogen pressure fluctuates with 2.9.6 Monitoring
changes in cable loading and ambient temperatures; no Cable systems can have active monitoring in addition to
external reservoir is required, as thermal expansion is fluid monitoring for the terminations and fluid-handling
accommodated by the compressible nature of the gas. systems. Such monitoring includes the following:
GIL cables require a monitoring cabinet with alarms to • Temperature monitoring: Most cable systems include
indicate high or low gas density to detect gas leaks. The thermocouples at several locations on the pipe, duct,
gas pressure fluctuates with changes in cable loading or cable itself for directly buried cables. Distributed
and ambient temperatures; no external reservoir fiber optic temperature monitoring has become more
is required. usual in North America since the mid-1990s but is
not universally installed elsewhere. An optical fiber is
SCFF cables for land installations generally have small placed in the cable outer shielding, or in/on the duct
pressure reservoirs placed every few miles (several km) or cable pipe. A special optical time domain reflecto-
to accept fluid expansion and contraction. A distin- meter (optical radar) measures backscatter from laser
guishing feature of the SCFF pressurizing system is that pulses introduced into the fiber to provide a tempera-
it can be passive, requiring no pumps or power sources, ture profile along the length of the line. Chapter 12
with the exception of low-voltage telemetry of alarms. describes the distributed temperature sensing system.
The pressure reservoirs contain gas-filled flexible metal- • Corrosion monitoring: Pipe-to-earth voltage poten-
lic elements. The reservoir is typically capable of accom- tial may be monitored to determine the effectiveness
modating up to 50-gallon (200-l) fluid expulsion from of the cathodic protection system.
the cable system. These reservoirs require alarm circuits • Dynamic rating: Cable temperature, ambient earth
to indicate out-of-range pressures and low fluid levels. temperature, and circuit loading may be monitored
An exception is long submarine cables and some under- and results relayed to a computer that determines
ground applications, which have large pressurizing allowable future loading based on recent history.
plants at one or both ends. These resemble plants for Chapter 11 addresses dynamic rating.
high-pressure fluid-filled cables except the reservoir tank
has a vacuum over the liquid rather than a low-pressure • Partial discharge: Some utilities are beginning to
nitrogen blanket. monitor partial discharge level on splices and
terminations of extruded-dielectric cables in an
Chapter 9 addresses cable hydraulic considerations, effort to obtain early warning of incipient failures
including calculations and equipment descriptions. (CIGRE 2001).
• Other: Chapter 15, “Operation and Maintenance,”
2.9.5 Pipe-type Cathodic Protection and describes dissolved gas analysis and other testing that
Grounding may be done on a cable system, either periodically to
The pipe for pipe-type cables must be protected from assess potential cable system degradation or specifi-
corrosion by a high-quality anti-corrosion coating. cally in event of a problem with the cable.
Since it is important to avoid leaks, pipes also have a
cathodic protection system installed. The system may 2.10 SELECTING CABLE TYPES
consist of distributed sacrificial anodes or, more com- Underground transmission cables account for less than
monly, an impressed-current system with a rectifier pro- 1% of all transmission line miles installed in North
viding the cathodic protection voltage.
2-17
America and less than 5-10% in smaller more densely • Cable capacitance is much higher than that for over-
populated countries. Cables are almost always more head lines, with implications including the following:
expensive than overhead lines—sometimes significantly - Charging current to supply the capacitance
more expensive—and cables can impose constraints on heats the conductor, reduces the amount of
utility system operation. Cables are therefore generally allowable real-power transfer, and increases
installed in the following situations: cable system losses.
• Overhead lines are not feasible (for example, down- - The cable is a significant source of reactive
town areas in major cities or long water crossings). VARs which must be absorbed by the utility sys-
• Improve reliability (for example, to avoid overhead tem or by shunt reactors. In some networks, this
crossings of other overhead lines). is of benefit in compensating the inductive reac-
tance of the system.
• Preserve areas of special scenic beauty.
- Passage of charging current through the series
• Safety, such as airport runway zones. inductance of other transmission lines or
• Other parties pay for the undergrounding (for exam- through transformers can generate excessive
ple, to improve the value of a development). voltage rise, particularly on energization or in
the event of unexpected load rejection.
If undergrounding is considered, many factors should - Surge arrester energy duty is higher because of
be evaluated to determine which cable type would be the stored energy in the cable capacitance.
preferred. A few considerations and differences among
- Cable capacitance can cause problems with har-
major cable types, are summarized in Table 2-1. For this
monic resonance and transient overvoltages.
comparison, extruded-dielectric and SCFF cables are
assumed to be installed in ducts in typical North Ameri- - Overvoltages caused by capacitive current can
can practice. limit system restoration following a whole or
partial blackout; shunt reactors may be neces-
2.10.1 System Considerations sary to aid in system restoration.
Values for many cable parameters are significantly dif- - Capacitive cable links stiffen the network by
ferent from those for overhead lines, and values differ increasing system stability compared to induc-
among cable types. These differences can have a major tive links.
effect on operation of a utility’s transmission system. • Cables have a lower reactance than overhead lines;
Principal considerations include the following (Stewart therefore, the cable may “hog” load when installed in
et al. 1986): parallel with an overhead line, even though the cable
will have a lower rating.
Table 2-1 General Comparison Among Cable Types

Characteristic Cable Type
Extruded-Dielectric Pipe-Type SCFF GIL
History in North America
First operated at 138 kV Mid-1970s 1932 1926
First operated at 345 kV 2001 1964 1971 (525 kV) 1971
Highest voltage in service, 2006 345 kV 345 kV 525 kV 765 kV
Electrical parameters See Table 2-2
Typical insulation thickness,
1.0 in. (25 mm) 0.6 in. (15 mm) 0.6 in. (15 mm) 10 in. (250 mm)
345-kV cable
Typical electrical stress at con-
290 V/mil (11.5 kV/mm) 430 V/mil (17 kV/mm) 380 V/mil (15 kV/mm) 70 V/mil (2.7 kV/mm)
ductor, 345-kV cable
Typical outside diameter, 345-kV
5.5 in (140 mm) 3.5 in. (89 mm) 4 in. (102 mm) 18-25 in. (457-635 mm)
cable
Typical ampacity, large conduc-
1200-1500 A 1000-1200 A 1200-1500 A 2000-4000 A
tor
Restoration time, 345-kV cable 3 weeks or longer 4 weeks or longer 4 weeks or longer 2 weeks or longer
failure (a failure can take much
longer to repair if it is in an inac-
cessible location, replacement
cable must be already available in
stock, etc.)
2-18
Chapter 16 describes system parameters in detail. technical document (CIGRE 2004) provides an excellent
summary of transition requirements and design.
A summary of typical electrical characteristics for
345-kV transmission lines is given in Table 2-2. Extruded-dielectric cables have simple terminations that
do not require reservoirs or alarms, at least through
2.10.2 Installation Considerations 230 kV. Terminations can be composite construction,
Suitability of an underground cable versus an overhead which is much lighter than a porcelain unit, and can be
line—and evaluation of one cable type versus another— placed on riser poles without need for a fenced-in area.
depends strongly on installation conditions and must be This is a definite advantage to extruded-dielectric cables
evaluated on a case-by-case basis. Note that excavation for rural and suburban areas. Figure 2-12 shows a
for an underground cable differs from an overhead line; 138-kV riser pole that has a spare phase for rapid con-
the entire the route must be accessed and excavated (or a nection in event of failure of one of the operating
directional drill performed) for an underground cable. phases.
Only occasional tower footings are required for over-
head lines so they can span small wetlands, for example. Although there have been a few pipe-type installations
Access for underground cables is required along the on riser poles at voltages up to 115 kV, as shown in Figure
route during installation and also should be maintained 2-13, these are uncommon. A fluid-filled line requires a
for inspection and possible repairs; overhead lines may
require access only to tower locations. Chapters 12 and
13 provide detailed discussion of cable installation
design and construction.
Transition requirements are significantly different for

the different cable types and are an important consider-
ation when evaluating potential applications. A CIGRE
Table 2-2 Typical Electrical Characteristics—345-kV

Overhead Line and Underground Extruded-Dielectric
(2000 kcmil) and Pipe-Type Cables (2500 kcmil), 60 Hz
Underground
Extruded-
Overhead dielectric Underground
Parameter Line XLPE HPFF
Shunt capacitance,
0.0189 0.279 0.462
µf/mi
Figure 2-12 138-kV riser pole for XLPE cable (courtesy
Shunt reactance, AEP).
0.140 0.0095 0.0057
megaohm-mi
Series inductance,
1.56 0.234 0.167
mH/mi
Series reactance,
0.587 0.088 0.063
ohms/mi
Charging current,
1.42 21.0 36.5
A/mi
Dielectric loss per
Negligible 0.42 5.3
phase, kW/mi
Reactive charging
power, three-phase 0.849 12.6 22.8
MVA/mi
Capacitive energy,
2.3 33.2 55.0
kJ/mi
Surge impedance,
287 29 19
ohms
Surge impedance
415 4110 6875
loading, MVA
Typical rating,
2500 1400 1250
Amperes
Figure 2-13 115-kV riser pole for HPGF cable (courtesy
of South Carolina Electric & Gas).
2-19
large pressurizing plant at least at one end, with distribu- important differences among the cable systems them-
tion power supply, alarm connections, and so on. selves. Considerations include the following:
Terminals and pressurizing plants are almost always
placed in a fenced-in compound. HPGF cables require a Power transfer: Cables have a lower current rating than
control cabinet at one end of the line, with alarms. Both overhead lines because heat from the cables must pass
HPFF and HPGF lines also require cathodic protection through the earth before reaching the ultimate heat
equipment; impressed-current systems require a power sink—air or water. Therefore, a cable system may
supply for the rectifier. The equipment is placed in require more than one cable per phase. Daily load shape
stations or small transition compounds. has a stronger influence on underground cable rating
than on overhead line rating. Cables have a much longer
SCFF cables also have heavy porcelain insulators. How- thermal time constant because of their burial in the
ever, they are sometimes installed on platforms attached earth, giving the advantage that short-term emergency
to terminal towers. In many applications, they are termi- overloads can be much higher than those for overhead
nated by GIS terminations, either horizontally or verti- lines. Chapter 11 discusses these topics in greater detail.
cally into compact metal-clad bus bar.
Reclosing: Many faults on overhead lines are self-clear-
GIL lines have been placed in transition compounds, ing (for example, due to lightning or a tree falling on a
but most installations are within stations. They do not line). Utilities, therefore, have automatic reclosing on
require reservoirs but do require alarm systems for overhead lines. Faults on underground cables are per-
gas pressure. manent, so it is not prudent to reclose on a cable.
Reclosing on pipe-type cables is not recommended
A few additional basic considerations related to cable because of resulting additional cable damage, possible
installation are summarized in Table 2-3 by cable type. pipe breach and resulting fluid loss, and possible circu-
Many of the comments are subjective and may not lation of carbonized liquid. Similar problems can occur
apply to all installation conditions. for reclosing on SCFF cables.
2.10.3 Considerations for Cable Operation Reclosing might be permitted on solidly grounded
Underground cable systems differ from overhead lines extruded-dielectric lines if the cable shield/sheath is
in several areas related to cable operation, and there are designed to carry the system fault currents. Reclosing
on a specially bonded circuit risks the flash-over of
Table 2-3 General Considerations – Cable Installation for Different Locations
Condition Extruded-dielectric, SCFFa Pipe-type GIL
Location
Rural (note that an
Trench cross-section would resemble Would probably need to be
access road is needed Direct burial (with mechanical pro-
urban, without paving. Need extensive above ground; require
the full length for con- tection) may be possible to reduce
termination equipment including power mechanical protection. Need
struction and mainte- cost and increase ampacity
supply and alarms alarms.
nance/repair)
Would need to be in a trough
Suburban Direct burial or duct system Similar to urban
or in a protected area
Duct system; slower installation
Would need to be in a trough
Urban than pipe-type (utility tunnels are Rugged, more rapid installation
or a tunnel
common overseas)
HPFF would probably not be suitable
Ideal for substations, espe-
Ideal for short substation lengths unless a pressurizing plant were already
Substation cially for connection to gas-
because of simple accessories present. HPGF could be a suitable alter-
insulated substations
native through 138 kV.
Lengths up to a mile (1.6 km) feasi-
Lengths up to 1.5 miles (2.4 km) have Not suitable for long direc-
Long directional drills ble, depending on cable
been installed tional drills
size/weight and installation in duct
Factory or at-sea splices required Must have an “at-sea” splice every 1.5
Long underwater cross- every several miles for extruded- miles (2.4 km) or less; therefore, not Not suitable for long underwa-
ings dielectric; every 30-40 miles (48-64 really suitable for long underwater cross- ter crossings
km) for SCFF ings
a. The SCFF system would be very similar to the extruded-dielectric system in terms of installation, but it would have the addi-
tional complexity of requiring fluid reservoirs and alarms, and SCFF cables are not as amenable to installation on riser
poles.
2-20
damaged sheath voltage limiters (SVLs), consequent Dielectric fluid: Many users of pipe-type cables have
violent displacement of manhole covers, and a break- operated their systems for decades with no fluid leaks,
down of the joint shield interruption. Even if reclosing but a few other utilities have incurred a relatively large
is allowed, only a single reclosing would be permitted number of leaks. EPRI has developed leak detection
because of a potentially damaging temperature rise and location methods (Williams 1983), and utilities have
caused by multiple reclosings. Also, reclosing on lines refined such methods and have improved their mainte-
with vaults in populated areas is not recommended since nance practices to greatly reduce leak rates. Neverthe-
reclosing on extruded cables may cause fires or explo- less, the possibility of a dielectric fluid leak is an
sions in splice vaults when the initial fault is in the vault. environmental concern for pipe-type and SCFF cable
systems. As described in Chapter 15, there are methods
Reclosing on hybrid overhead and underground lines to reduce the amount of fluid that can be lost, but any
should be given special consideration, because loss of leak can still be considered a maintenance problem and
l i n e r e l i ab i l i t y m ay r e s u l t i f n o t p e r m i t t e d . can result in an outage until repairs are made (Ghafu-
Implementation of special protection schemes to rian et al. 1999).
discriminate between faults in overhead versus
underground sections may be required. Single-pole Heat: Cables generate heat, and the earth surface tem-
reclosing could be considered. perature might be elevated a few degrees at full load
because of cable heat. The only environmental effect
Outage rates, durations: Outages on overhead lines can reported has been premature germination of seeds
be relatively frequent, but they are usually short dura- (EPRI 1975).
tion because of automatic reclosing practices and the
speed with which lines can be repaired. Outages on Magnetic fields: The magnetic field right over a cable
underground cables are much less common, but the out- trench can be higher than that for an overhead transmis-
age duration is significantly longer—a month is possi- sion line even though the cable magnetic field dimin-
ble, or even longer if replacement cable must be ishes with distance much more rapidly that for an
manufactured. Insufficient data exist to compare outage overhead line. Pipe-type cables have the lowest magnetic
rates among cable types with sufficient confidence to fields because of the closeness of the three phases and
favor one cable type versus another. Note that all cable the attenuation due to the steel pipe. GIL can have low
types other than extruded-dielectric can have outages fields because the enclosure is solidly grounded and the
due to hydraulic problems as well as electrical ones. One circulating current is out of phase with, and almost
advantage of a pressurized cable system is that the con- equal to, the conductor current. Methods such as phase
trol system responds to loss of pressure following third- positioning can be used to reduce the magnetic fields
party damage, whereas in an extruded cable, water is from extruded-dielectric and SCFF cables. Chapter 16
likely to enter unnoticed and may cause extensive dam- addresses magnetic fields in greater detail.
age in the future depending on the effectiveness of
water-blocking materials. 2.10.5 Cost Considerations
Cable systems are almost always more costly than over-
2.10.4 Environmental Considerations head lines. Because the cable’s lower ratings or reliability
A few environmental concerns may be considered when considerations may require multiple conductors per
evaluating overhead versus underground installations phase, there is the need to trench or bore every part of
and when evaluating differences among underground the route, and the cost of the high-quality insulated con-
cable types: ductor and its installation can be high. Costs are very
specific to the application: for example, voltage, power
Visual: Cables may be buried and out of sight, but tran- transfer level, required redundancy, route, and obstruc-
sition stations are still needed at each end of the line. As tions. It is possible for the price of a given cable system
noted in Section 2.10.2, extruded-dielectric cables may to vary by a factor of two or greater from one applica-
be placed on riser poles, but other cable types will usu- tion to another, depending principally on civil works
ally require transition compounds. In addition, a per- and installation conditions. The problem of quantifying
manent 20- to 25-foot (6- to 7.5-m) easement is required direct and indirect costs leads to subjective compari-
for access for maintenance and repairs; this easement sons. Cost comparisons are usually based on the direct
cannot have woody vegetation that would impede vehi- cost of equipment and installation. Indirect costs of
cle access and that might remove moisture from the soil obtaining regulatory approval and the time delay to the
and cause cable overheating. project are usually excluded, although these may be con-
siderable. The proper approach to provide even a bud-
2-21
getary cost estimate is to define the power transfer AEIC Specifications

requirements, investigate the potential line locations, The Cable Engineering Committee of AEIC writes and
evaluate the alternative cable types and determine the updates power cable specifications and guides for utility
required conductor size, determine the civil require- use and serves as a forum for exchanging technical
ments, and obtain budgetary cost estimates from ven- information.
dors. Overheads and contingencies are then added to
those costs to develop the budgetary cost estimate. The AEIC Cable Engineering Committee consists of 25-
30 AEIC member utility engineers who meet twice a
Just as it is not possible to provide general comments on year to address the specific needs and problems experi-
cable system cost, it is not possible to provide a general enced in utility cable systems. Their specifications and
ratio of underground cable costs to overhead lines costs. guides have gained wide acceptance because of a focus
Not only do cable costs vary, but the overhead line costs on long-term performance and reliability.
vary as well depending on such factors as location, ter-
rain, and number of angles and deadend towers. Ratios When a specification is due for revision, an open forum
as low as 2:1 are possible (extremely expensive overhead is conducted, so that the experience of suppliers, manu-
line construction), and ratios higher than 20:1 have been facturers, researchers, and cable users can be reviewed
reported (CIGRE 2004). The calculation of the differ- and new knowledge incorporated in the next revision.
ence in cost for the particular application—for example, Since the group is rather small and has a common pur-
after evaluating required conductor size and installation pose, the revisions are produced in a timely manner.
conditions—is the only fair comparison.
Cable Specifications CS1 through CS4 are for cable
2.11 INDUSTRY SPECIFICATIONS AND designs using laminated paper insulation materials.
STANDARDS Specifications CS8 and CS9 cover extruded-dielectric
cables. An appendix in each specification lists all
2.11.1 Utility Specifications industry standards referenced in the document. Specifi-
cation CS31 is for low- and high-viscosity, pipe-filling
Many specifications written by power utilities are mate-
liquids used in high pressure pipe-type circuits. AEIC
rial and construction specifications with an increasing
specifications are usually referred to by their number
trend toward performance specifications. Industry stan-
and year of latest issue—e.g. CS1-90. In all cases, the
dards and specifications are used as a basis for the utility
user should ensure that the latest edition of the specifica-
specification. A main feature of the utility specification
tion is applied.
is to precisely describe which combination of materials
and construction methods is to be selected from the ICEA Standards
industry standards and specifications, as well as to The ICEA is a group of power cable suppliers who
clearly describe the application. Specific requirements write and update standards for power cable manufac-
for items such as conductor size, transmission capacity, turers. ICEA participation is mainly by North
voltage class, testing, packaging, shipping, installation American suppliers, although their international
layout, and commissioning are added as desired. The affiliates also contribute indirectly. They do not prepare
objective is for the utility to thoroughly describe the standards for laminated insulation cables or for power
application requirements, to provide proposers with a cable accessories.
clear, uniform description for bidding, while preserving
the opportunity for them to provide an optimum and There is an effective consultative arrangement between
innovative solution for the particular application. AEIC and ICEA, known as UPCSTAC (Utility Power
Cable Standards Advisory Committee), whereby bilat-
The industry specifications and standards usually refer- eral discussions and reviews can lead to optimum speci-
enced in North America are those produced by the fications and standards for both groups. Only extruded
Cable Engineering Committee of AEIC (Association of insulation cables are considered. Some AEIC specifica-
Edison Illuminating Companies), ICEA (Insulated tions (e.g., CS8) are supplements to ICEA standards, a
Cable Engineers Association), and IEEE (Institute of legacy from earlier times when ICEA was the only devel-
Electrical and Electronics Engineers). Some IEC (Inter- oper of extruded insulation cables in North America.
national Electrotechnical Commission) standards are
also referenced. Guides and Recommended Practices Some ICEA standards have also received accreditation
are sometimes referenced, but their applicability to con- by ANSI (American National Standards Institute).
tractual matters needs to be considered carefully.
2-22
IEEE Standards AEIC CS2-97 (6th Edition—March 1997). Specification

IEEE does not develop power cable standards but does for Impregnated Paper and Laminated Paper Polypro-
develop them for power cable accessories such as joints pylene Insulated Cable, High Pressure Pipe Type.
and terminations. Numerous related Guides and Rec-
ommended Practices are also prepared. AEIC CS3-90 (3rd Edition—October 1990). Specifica-
tions for Impregnated-Paper-Insulated Metallic
Some IEEE standards have also received accreditation Sheathed Cable, Low Pressure Gas-Filled Type.
by ANSI (American National Standards Institute).
AEIC CS4-93 (8th Edition—January 1993). Specifica-
Other Specifications tions for Impregnated-Paper-Insulated Low and
As more extruded-dielectric transmission cable (69 kV Medium Pressure Self-Contained Liquid Filled Cable.
and above) continues to be installed in North America,
at ever higher voltages, there has been greater interest in AEIC CS6—Replaced by CS9 for >46 kV
following the experience of the international manufac-
turers and installations. This trend continues to spur AEIC CS7—Withdrawn and replaced by CS9 for
interest in international specifications and in resolving >46 kV
the differences between AEIC specifications and ICEA
standards compared to those developed internationally. AEIC CS8-00 (1st Edition). Specification for Extruded
Dielectric Shielded Power Cables Rated 5 through
There have been many more HV and EHV extruded 46 kV. Note: CS8-00 replaces CS5 and CS6 for 5
insulation power cables manufactured and installed through 46 kV Cables.
internationally than in North America, mainly driven by
higher population densities and land costs in large over- AEIC CS9-06 (1st Edition). Specification for Extruded
seas cities. Cable and cable testing standards, therefore, Insulation Power Cables and their Accessories Rated
have been actively developed by IEC, JIC (Japan Indus- above 46 kV through 345 kVac. Note: CS9-06 replaces
trial Committee), and a variety of European countries CS6 for 69 kV Cables and CS7 for greater than 46 kV.
(now being harmonized as CENELEC [European Com-
mittee for Electrotechnical Standardization] standards). AEIC CS31-95 (2nd Edition—December 1995). Specifi-
cations for Electrically Insulating Pipe Filling Liquids
Industry specifications and standards have led to many for High-Pressure Pipe-Type Cable.
improvements in cable quality. The industry is continu-
ously searching for improvements such as better cable AEIC CG1-96 (3rd Edition—April 1996). Guide for
system testing and quality assurance, a more complete Ap plication of AEIC Maximum In su lation
understanding of the various cable operating condi- Temperatures at the Conductor for Impregnated-Paper-
tions, new materials with better performance character- Insulated Cables.
istics, more compact designs, lower lifetime costs, etc.
As new knowledge is gained, the specifications will AEIC CG3-2005 (2nd Edition—January 2005). Guide
be revised and improved. for Installation of Pipe-Type Cable Systems.
2.11.2 Specification Listing AEIC CG4-97 (2nd Edition—January 1997). Guide for
A list of transmission cable-related AEIC specifications Installation of Extruded Dielectric Insulated Power
and ICEA and IEEE standards and guides, current as of Cable Systems Rated 69 kV Through 138 kV.
2006, follows.
AEIC CG5-2005. Underground Extruded Power Cable
Readers should check with the issuing organization for Pulling Guide. AEIC CG6-05 (2nd Edition—Novem-
the latest versions. ber, 2005).
AEIC Specifications
AEIC CG6-2005. Guide for Establishing the Maximum
AEIC CS1-90 (11th Edition—October 1990). Specifica-
Operating Temperatures of Extruded Dielectric Insu-
tions for Impregnated Paper-Insulated Metallic-
lated Shielded Power Cables (1st Edition)
Sheathed Cable, Solid Type.
AEIC CG8-2003 (2nd Edition—August 2003). Guide
for Electric Utility Quality Assurance Program for
Extruded Dielectric Power Cables.
2-23
ICEA Standards and Guides IEEE C2 (2007). National Electrical Safety Code
ICEA T-24-380 (1994). Guide for Partial Discharge (NESC).
Test Procedure.
2.11.3 Contact Information
ICEA P-45-482 (1999). Short Circuit Performance of Other specifications and reports are generally available
Metallic Shielding and Sheaths of insulated Cable. and can be of great use to the cable design engineer.
Sources of the foregoing specifications and standards, as
ICEA T-26-465/NEMA WC 54 (2001). Guide for Fre- well as other cable related information, are given below:
quency of Sampling Extruded Dielectric Power, Con-
trol, Instrumentation, and Portable Cables for Tests. American National Standards Institute (ANSI)
11 West 42nd Street, 13th Floor
ICEA T-25-425 (2003). Guide for Establishing Stability New York, New York USA 10036
of Volume Resistivity for Conducting Polymeric Com- Website: www.ansi.org
ponents of Power Cables.
The Association of Edison Illuminating Companies
ICEA T-28-562 (2003). Test Method for Measurement (AEIC)
of Hot Creep of Polymeric Insulations. P.O. Box 2641
Birmingham, Alabama 35291-0992
ICEA S-108-720 (2004). Extruded Insulation Power Telephone: (205) 257-3839
Cables Rated Above 46 through 345 kV. Fax: (205) 257-2540
IEEE Standards and Guides Email: secretaryaeic@bellsouth.net
IEEE 535 (1993). Guide for Selecting and Testing Jack- Website: www.aeic.org
ets for Underground Cables.
American Society of Testing and Materials (ASTM)
IEEE 1406 (1998). Guide to the Use of Gas-In-Fluid ASTM International
Analysis for Electric Power Cable Systems. 100 Barr Harbor Drive
P.O. Box C700
IEEE 404 (2000). Standard for Extruded & Laminated West Conshohocken, PA USA 19428-2959
Dielectric Shielded Cable Joints Rated 2.5 kV - 500 kV. Telephone: (610) 832-9585
Fax: (610) 832-9555
IEEE 400 (2001), Guide for Field Testing & Evaluation Website: www.astm.org
of the Insulation of Shielded Power Cable Systems.
Conseil International des Grands Réseaux Électriques
IEEE 575 (2001). Guide for the Application of Sheath- (CIGRE)
Bonding Methods for Single-Conductor Cables and the 21 rue d’Artois
Calculation of Induced Voltages and Currents in F 75008 Paris, France
Cable Sheaths. Telephone: +33 (1) 53 89 12 90
Fax: 33 (1) 53 89 12 99
IEEE 1425 (2001). Guide for the Evaluation of the Email: publications@cigre.org
Remaining Life of Impregnated Paper- Insulated Trans- Website: www.cigre.org
mission Cable Systems.
Electric Power Research Institute (EPRI)
IEEE 82 (2002). Test Procedure for Impulse Voltage 3420 Hillview Avenue
Tests on Insulated Conductors. Palo Alto, CA USA 94304 -1338
Telephone: (650) 855-2000
IEEE 48 (R2003). Standard Test Procedures and Website: www.epri.com
Requirements for Alternating-Current Cable Termina-
tions 2.5kV - 765kV. Institute of Electrical and Electronics Engineers (IEEE)
IEEE Operations Center
IEEE 400.2 (2004). Guide for Field Testing of Shielded 445 Hoes Lane
Power Cable Systems Using Very Low Frequency. Piscataway, NJ USA 08854-4141
Telephone: (732) 981-0060
Fax: (732) 981 1721
Website: www.ieee.org
2-24
Insulated Cable Engineers Association (ICEA) Telephone: (703) 351-5252

P.O. 1568 Fax: (708) 739-6672
Carrollton, GA USA 30112 Website: www.nastt.org
Website: www.icea.net
ICEA and NEMA Standards may be ordered from: National Association of Corrosion Engineers (NACE)
Global Engineering Documents 1440 South Creek Drive
15 Inverness Way East Houston, Texas USA 77084-4906
Englewood, CO USA 80112 Telephone: (281) 228-6200
Website: www.global.ihs.com Fax: (281) 228-6300
Website: www.nace.org
International Electrotechnical Commission (lEC)
IEC Regional Centre for North America (IEC-ReCNA) National Association of Pipe Coating Applicators
446 Main Street, 16th Floor (NAPCA)
Worcester, MA USA 01608 8570 Business Park Drive
Telephone: (508) 755-5663 Shreveport, LA USA 71105
Fax: (508) 755-5669 Telephone: (318) 227-2769
Website: www.icc.ch Fax: (318) 222-0482
IEC standards may also be ordered from the Website: www.napca.com
American National Standards Institute (ANSI). See
earlier listing. National Electrical Manufacturers Association (NEMA)
1300 North 17th Street, Suite 1847
National American Society for Trenchless Technology Rosslyn, VA USA 22209
(NASTT) Telephone: (703) 841-6200
1655 N. Fort Meyer Dr., Suite 700 Fax: (703) 841-5900
Arlington, VA USA 22209 Website: www.nema.org
2-25
REFERENCES Cable Links.” Transactions of the Canadian Electrical

Allam, E. M. 1982. Development of Low-Loss 765 kV Association, Engineering and Operating Division. Vol.
Pipe-type Cable. EPRI Report No. EL-2196. January. 22. Part 5.
Allam, E. M., J. H. Cooper, and J. F. Shimshock. 1982. Endersby, T. M., B. Gregory, and S. G. Swingler. 1992.
“Development and Long Term Testing of a Low Loss “The Application of Polypropylene Paper Laminate
765 kV High Pressure Oil Filled Cable.” CIGRE paper Insulated Oil Filled cable to EHV and UHV Transmis-
21-06. Paris. sion.” CIGRE. Paper. 21-307. Paris.
Allam, E. M., A. L. McKean, and F. A. Teti. 1988. Endersby, T. M., S. J. Galloway, B. Gregory, and N. C.
“Optimized PPP-Insulated Pipe Type Cable System for Mohan. 1993. “Environmental Compatibility of
the Commercial Voltage Range.” IEEE Transactions on Supertension Cables.” IEE Third International Confer-
Power Delivery. Vol. 3. No. 2. ence, Publication 382. pp. 71. London.
Bartnikas, R., and K. D. Srivastava. 1987. Power Cable EPRI. 1975. Study of Environmental Impact of Under-
Engineering. Waterloo, Ontario. Sanford Educational ground Electric Transmission Systems. EDAW, Inc. San
Press. Francisco, CA. May.
Bjorlow-Larsen, K., M. Del Brenna, J. Kaumans, R. Foxall, R. G., K. B. Larsen, and G. Bazi. 1984. "Design,
Meier, M. Kirchner, and P. Argaut. 2003. “Large Manufacture, and Installation of a 525 kV Alternating
Projects of EHV Underground Cable Systems.” Paper Current Submarine Cable Link from Mainland Canada
A.2.5. Jicable. to Vancouver Island." CIGRE paper 21-04.
CIGRE. 2001. Construction, Laying and Installation Fukawa, M., T. Kawai, Y. Okano, S. Sakuma, S. Asai,
Techniques for Extruded and Self Contained Fluid Filled M. Kanaoka, and H. Yamanouchi. 1996. “Develop-
Cable Systems. Technical Brochure 194. October. ment of 500 kV XLPE Cables and Accessories for Long
Distance Underground Transmission Line – Part III:
CIGRE. 2001. Partial Discharge Detection in Installed Electrical Properties of 500 kV Cables.” IEEE Transac-
HV Extruded Cable Systems. Technical Brochure 182. tions on Power Delivery. Vol. 11. No. 2. April.
April.
Ghafurian, R., J. Dominguez, A. Santini, and C. Sobel.
CIGRE. 2003. Gas-Insulated Lines (GIL). Technical 1999. ”New Advances In Mitigating Environmental
Brochure 218. February. Impact Of Pipe-Type Cables.” IEEE Transactions on
Power Delivery Vol 14, No. 2 April.
CIGRE. 2004. General Guidelines for the Integration of
a New Underground Cable System in the Network. Work- Graneau, P. 1979. Underground Power Transmission.
ing Group B1.19. Technical Brochure No. 250. August. John Wiley & Sons. New York.
CIGRE. 2005. Special Bonding of High Voltage Power Grzan, J., R. V. Casalaina, E. I. Hahn, and J. O. C. Kan-
Cables. Technical Brochure 283. October. sog. “The 345-kV Underground/Underwater Long
Island Sound Cable Project.” IEEE Transactions on
Couderc, D. et al. 1996. “Development and Testing of a Power Delivery Vol 8, No. 3 July.
800-kV PPLP-insulated Oil-filled Cable and its Accesso-
ries.” CIGRE. Paper. 21-22/04. Kaminaga, K., T. Harada, M. Ono, T. Kojima, Y. Sekii,
and M. Marumo. 1986 “Research and Development of
Crowley, E., J. E. Hardy, and L. R. Horne. 1982. "Devel- 500 kV XLPE Cables.” Conference Record of 1986
opment Programme for the Design, Testing and Sea Tri- IEEE International Symposium on Electrical Insula-
als of the British Columbia Mainland to Vancouver tion. Washington, DC. June 9-11.
Island 525 kV Alternating Current Submarine Cable
Link." CIGRE paper 21-10. Minemura, S., et al. 1990. “Completion of Installation
of 500kV PPLP-Insulated Self-Contained Oil Filled
Crowley, D. J., and R. G. Foxall. 1983. “Planning, Cable Along Seto Ohashi Bridge for Honshu-Shikoku
Design and Construction of a 500-kV Transmission Transmission Line.” Sumitomo Electric Technical
System to Vancouver Island Including Long Submarine Review, 29.
2-26
Moore, G. F. 1997. Electric Cables Handbook. Third Sakuma, S., and N. Shinagawa 2000, “Experience on
Edition. Blackwell Science. ISBN 0-632-04075-0. 400-500 kV XLPE Insulated Cables and Their Accesso-
ries.” IEEE Transactions on Power Apparatus and Sys-
Occhini, E., et al. 1978. “Self-Contained Oil-Filled tems. Vol. 1. January.
Cable Systems for 750- and 1100-kV Design and Tests.”
CIGRE paper 21-08. Stewart., J. R., D. D. Wilson, and J. A. Williams 1986.
“System Implications of Underground Cable Systems,”
Ohki, Y, and Yasufuku, 2002. “The World's First Long- I E E E T & D C o n f e r e n c e. A n a h e i m , C A .
Distance 500 kV XLPE Cable Line. 3. Underground September 15-19.
Apparatus.” IEEE Electrical Insulation Magazine. Vol.
18. No. 4. July-Aug. Tanaka, T., and A. Greenwood. 1983. Advanced Power
Cable Technology. CRC Press. Boca Raton, Florida.
Peschke, E. and R. von Olhaussen. 1999. Cable Systems
for High and Extra High Voltage. Pirelli Kabel und Sys- Watanabe, Y. et al. 1998. “Development of New 500 kV
teme GmbH. Berlin. ISBN 3-89578-118-5. Laminated Paper Insulated Self-Contained Oil-Filled
Cable and its Accessories.” IEEE Transactions. Paper 86
Ray, J. J. et al. 1974. “525 kV Self-contained Oil-filled SM 390-9.
Cable Systems for Grand Coulee Third Powerplant.
IEEE Transactions on Power Apparatus and Systems. Williams, J. 1972. “First Underground 345 kV Gas-
Vol.PAS 93. Jan/Feb. Insulated Transmission System Energized,” Transmis-
sion and Distribution. May.
Roughley, T. H. et al. 1974. “Design and Installation of
a 138 kV, High-Pressure, Gas-Filled Pipe Cable Utiliz- Williams, J., S. Kozak, and T. Rodenbaugh. 1983. “Leak
ing Segmental Aluminum Conductors.” IEEE Transac- Location Methods for HV Underground Cables.” IEEE
tions on Power Apparatus and Systems. Vol. 93. Transactions on Power Apparatus and Systems. Vol.
March/April. PAS-102. Pp. 2029-2037. July.
2-27
CHAPTER 3 Basic Design

Considerations
Author: Brian Gregory, Cable Consulting International Limited
Reviewers: John H. Cooper, Power Delivery Consultants, Inc.
Dennis E. Johnson, POWER Engineers, Inc.
The purpose of the chapter is to explain how cable circuits transmit power and how the
components common to different cable constructions are designed. The chapter provides
a common understanding and theoretical basis for the specialist chapters on the design of
cable systems, accessories, and installations for either ac or dc transmission.
The basic theory is given in the order likely to be encountered in the design of a cable cir-
cuit, starting at the foundation subjects of cable system design and the role of the electro-
magnetic field components. Explanations follow for the interdependence of circuit power
rating with cable voltage and current. The chapter then describes how the required cur-
rent is used in the ampacity calculations to obtain heat generation and dissipation and
thus ultimately the cable conductor size.
The cable design procedure is described for each layer, working outwards from the con-
ductor, through the shields and insulation to the integral ground return conductor, radial
moisture barriers, metallic containment sheath, anticorrosion jacket, and shipping reel.
At the higher transmission voltages, the need to maximize the cable reel length is the key
driver in the development of high design stress insulation.
Detailed conductor designs and insulation designs are considered, together with calcula-
tions of dc resistivity, electrical stress, capacitance, and rates of aging. Cable impedance is
predominantly capacitive, albeit magnetic field linkages and circuit geometry strongly
influence conductor ac resistance and inductive reactance. To evaluate the performance
of the cable circuit in the transmission network, it is usual to insert values of cable imped-
ance, ac resistance, and reactances into a network model; as a result, the basic calcula-
tions of these parameters are described.
Brian Gregory is the Technical Director and a founder member of

Cable Consulting International Ltd (CCI), incorporated in 2001. He
received an honors degree in Electrical and Electronic Engineering from
the City University, London in 1968. He received professional training
as a student apprentice with AEI Supertension Cables from 1963 to
1968. In 1969, he joined BICC Supertension Cables as a Development
Engineer for cables and accessories, becoming the Chief Development
Engineer in 1981. In 1992, Mr. Gregory became the Chief Engineer of
BICC Supertension and Subsea Cable Systems, with responsibilities for supertension
cable technology worldwide. He has extensive expertise in the research, development,
design, manufacture, testing, and commercialization of land and submarine cable sys-
tems, ac and dc up to 525 kV, insulated with extruded XLPE, EPR, fluid impregnated
LPP and paper, mass-impregnated paper, and LPGF and HPGF paper. He has recently
researched the thermomechanical behavior of XLPE and HPFF cables in pipe and duct-
manhole systems for EPRI, producing design guides and new calculation methods. He is
an inventor and has authored over 45 technical papers and books. He is a Fellow of the
IEE, a Member of the IEEE, a Member of CIGRE, a Chartered Engineer, and a FEANI
registered European Engineer.
3-1
Chapter 3: Basic Design Considerations EPRI Underground Transmission Systems Reference Book
3.1 INTRODUCTION 3.2.2 Circuit Cross Sections

The chapter considers the basic principles involved in The majority of cable circuits are underground.
system design, cable electromagnetic fields, power rat- Figure 3-2 shows a selection of circuit types installed
ing, current rating, cable design, and circuit parameters. underground. Similar cable configurations are employed
Both alternating current (ac) and direct current (dc) cir- for installation above ground and in tunnels. The dia-
cuits are considered. gram shows duct-manhole and pipe systems, which his-
torically have been preferred in North America. Direct-
The topics are presented in the order likely to be encoun- buried and tunnel circuits have evolved in other parts of
tered when planning a new cable circuit, starting with the world. Duct-manhole and pipe systems have advan-
the circuit layout and passing through design to specifi- tages in cities of fast installation and less disruption to
cation of the cable system. Section 3.2 describes cable traffic. Direct-buried systems have higher ampacities.
system design, with information on circuits, cable types, Tunnel systems have more direct routes and good pro-
cable components, accessories, ancillary equipment, the tection from third-party disturbance.
hydraulic and pneumatic scheme, the bonding scheme,
the thermal design, the thermomechanical design, the DC transmission circuits require two parallel cables. AC
installation design, and operation, maintenance and transmission circuits require three cables. For purposes
testing. Section 3.3 describes the electromagnetic field. of comparison, the cables in Figure 3-2 are shown
Sections 3.4 and 3.5 outline the steps for calculating installed at a common minimum depth. In consequence,
power and current ratings. Section 3.6 considers basic the centerlines of the larger diameter cables are at a
aspects of cable design and specification. Section 3.7 dis- greater distance from the surface. It is usual to specify a
cusses the electrical parameters of a cable system. minimum depth of burial to the top of the cable for
urban and rural sections of the route. Depth provides
In this chapter, the equations are derived in the SI unit increased protection from “dig-in” and plow damage.
system (Système International d’Unitès) based on kilo- Increased depth also decreases the magnetic field at the
grams, meters, and seconds (MKS). Where possible surface, but has the disadvantage of decreased ampacity.
numerical examples are given first in U.S. Customary
Units followed by the appropriate SI primary or second- The following types of installation and cable are shown
ary unit. This follows the approach in the latest AEIC in Figure 3-2:
specification CS9 (AEIC 2006). • Sheathed cables in ducts:
Extruded crosslinked polyethylene (XLPE) or ethyl-
3.2 CABLE SYSTEM DESIGN
ene propylene rubber (EPR)
3.2.1 Circuit Layout Self-contained fluid-filled (SCFF) or gas-filled
(SCGF)
The circuit layout in Figure 3-1 is a single-core,
extruded-dielectric, solidly grounded system for ac High-temperature superconducting (HTS); refriger-
transmission. Although simple in concept, each compo- ated by liquid gas
nent is designed to perform special functions under nor- • Pipe-type, high-pressure fluid–filled (HPFF) or gas-
mal and transient operating conditions. filled (HPGF)
Figure 3-1 Cable system components for ac transmission (courtesy of CCI).
3-2
EPRI Underground Transmission Systems Reference Book Chapter 3: Basic Design Considerations
• GIL (gas-insulated line) insulated with an N2/SF6 XLPE. AC GIL cables and ac and dc HTS cables are at
gas mixture. present cables for special applications, the former having
demonstrated commercial service experience, and the
For comparison purposes, the GIL circuit is shown bur- latter presently gaining experience. New types of dc
ied; however, the majority of applications are in tunnels extruded XLPE cables are presently being tested and
or above ground in power plants. have the prospect of application at the highest transmis-
sion voltages.
In a pipe-type system, it is normal to install the three-
phase cables in a carbon steel pipe. This is permissible 3.2.3 Cable
because the resultant magnetic field surrounding the Figure 3-3 identifies the common cable components of
cables is substantially reduced, thereby limiting the an extruded XLPE cable and an HPFF cable with taped
heat generated by hysterisis and eddy currents to an insulation. The common components are the power
acceptable level. Single cables derive no such benefit and conductor, conductor shield, insulation, insulation
must be installed in nonmagnetic ducts, which are pref- shield, the shield and ground return conductors, the
erably also non-conducting. The ducts referred to in radial moisture barrier, and the jacket. The prime
Figure 3-2 are fiber-reinforced epoxy (FRE), which are function that separates a cable from air-insulated equip-
rigid and supplied in short lengths of 20 ft (6 m), or ment, such as an overhead line, is the complete contain-
polyethylene (PE), preferably a high-density grade, ment of the voltage and electrostatic field within the
which are flexible and supplied in typical reel lengths of insulation shield. Cables may, therefore, be installed in a
300–600 ft (100–200 m). PVC ducts are also selected for wide range of environments and weather conditions
many applications. without “shorting” the conductor to ground during nor-
mal operation.
Figure 3-2 shows circuits with one cable per dc pole or
ac phase. When higher power levels are required, the In addition to the power conductor, each cable has inte-
conductor cross-sectional area and cable spacings are grated auxiliary conductors. These are: (a) the shield
first increased. When larger conductor sizes become conductor, (b) the ground return conductor, and (c) the
impracticable to manufacture, two or more cables may ground continuity conductor (for special bonding situa-
be provided per dc pole or per ac phase. Another option tions). In the XLPE cable in Figure 3-3, the shield and
is to install backfill with improved thermal properties. ground return are combined into one conductor,
When natural heat dissipation does not provide ade- whereas in an HPFF cable, they are separate.
quate ampacity, forced-cooling may be selected.
The shield conductor is always part of the cable con-
The cable type with the highest prospective transmission struction. Its role is to collect the insulation charging
rating is ac GIL followed by dc SCFF, ac extruded current and carry it back to the positions where the cir-
XLPE, ac SCFF, ac HPFF, ac extruded EPR, and dc cuit is grounded, either at the joints or terminations, as
Figure 3-2 Circuit cross sections for dc and ac transmission (courtesy of CCI).
3-3
illustrated in Figure 3-4. The capacitive charging cur- the charging current from the complete cable length and
rents are distributed inch by inch along the length of the to limit the I2R heating to an acceptable level (1 or 2°C).
cable. The charging current in an ac cable is steady state This temperature rise is not included in the normal
and flows all the time that the cable is voltage energized. ampacity calculation for a short cable length, but must
In a dc cable, the charging currents are transient and be included for long-length submarine cables. The maxi-
comprise the inrush current that flows only when the mum current flow occurs at the end of the cable and is
cable is switched in and out of the circuit. The flow from halved when both ends are grounded, as shown in Fig-
each inch of insulation is radially outwards through the ure 3-4.
semiconducting insulation shield and bedding layer,
which have a comparatively high electrical resistivity. A pipe-type cable has the advantage that the metallic
The current is collected by the low resistivity shield con- skid wires contact each other at random intervals and
ductor, which carries the cumulative current to a ground transfer the capacitive currents between the three phase
connection at one or both ends of the cable or circuit. cables. The 120° phase difference in current results in
The area of the shield conductor must be sized to carry
Figure 3-3 Common cable components. Left: extruded XLPE cable. Right upper: core of HPFF: pipe-type
cable (kraft paper insulation). Right lower: HPFF pipe-type cable in pipe (LPP insulation) (courtesy of CCI).
Figure 3-4 Longitudinal cable section showing combined shield and ground return conductors (courtesy of CCI).
3-4
their cancellation at points of contact, thus longitudinal the electrical insulation and the calculation of their
flow is minimal. operating temperature.
The ground return conductor is always part of the con- 3.2.5 Ancillary Equipment
struction of a single-core cable, but not of a pipe-type The ancillary equipment of a cable system includes:
cable in which the role is performed by the steel pipe.
When the cable experiences a fault, the ground return • link boxes for bonding (extruded and SCFF systems),
conductor carries the current back to the ground con- • pilot cables for system telemetry,
nections at a termination, or at a grounded joint posi- • temperature-monitoring equipment (optional),
tion. In each type of bonding scheme, at least one end of
the cable shield conductor and ground return conductor • pumping and monitoring equipment for high-pres-
must always be connected to ground, or these conduc- sure, fluid-filled, pipe-type systems,
tors will rise to an appreciable voltage, damaging the • fluid reservoirs and pressure-monitoring equipment
cable and posing a serious safety risk to personnel. for low-pressure, fluid-filled, self-contained systems,
• partial discharge* monitoring equipment for accesso-
The role of the external ground continuity conductor, ries in extruded systems (optional), and
shown in Figure 3-1, is to safely transmit the ground
return current from a ground fault in other transmission • cryogenic equipment (HTS systems).
equipment back through the cable system to the ground
connection at the start of the cable circuit. The external *Note: A partial discharge (PD) is the occurrence of a
ground continuity conductor is installed as part of the spark within a small air- or gas-filled void (gap or split)
end-point and mid-point bonded system designs, in in the insulation. The spark is formed by the liberation
which the internal ground return conductor is intention- of the capacitive charge stored in the void when it
ally interrupted at an accessory to prevent flow of circu- breaks down. The name “partial discharge” distin-
lating currents. The design of solidly-bonded and cross- guishes the event from the complete electrical failure
bonded circuits requires the ground conductor within of the cable, in which all of the energy in the insulation
the cable to be grounded at both ends, and this then per- is discharged.
forms the role of ground continuity conductor, thereby
reducing cost. However, it is common in North America 3.2.6 Hydraulic and Pneumatic Scheme
for utilities to install the ground continuity conductor to The hydraulic scheme ensures that HPFF cables and
serve as a very strong path to the substation ground. SCFF cables are provided with dielectric fluid to achieve
their designed dielectric strength during heating and
3.2.4 Accessories cooling without experiencing extremes of pressure. Pres-
The basic accessories shown in Figure 3-1 consist of sure ensures that contraction voids can never form in an
splices to connect cable lengths together and termina- SCFF cable and that prospective voids in an HPFF
tions to interface the cable with other electrical equip- cable are prevented from partial discharging. The
ment at the ends of the circuit. Outside North America, designer calculates the size of external fluid reservoirs to
splices are named “joints.” Within North America, “ter- absorb the volume of fluid thermal expansion when the
minations” are commonly referred to as “potheads.” cable is heated and to re-supply the fluid volume when
Outdoor terminations may also be called “sealing the cable contracts. In HPFF systems, the hydraulic
ends.” A termination is the generic name for all methods scheme may also be designed to cool the cable by circu-
of connecting the cable to other items of equipment, lating the dielectric fluid. A gas-filled cable system is
whether outdoor, indoor, gas-insulated bus bar or liq- also maintained at sufficient nitrogen pressure to pre-
uid-filled transformers. The accessories are designed to vent partial discharging. The compressible nature of gas
perform each of the functions of the cable, with the permits it to absorb thermal expansion by a natural
additional requirement that they must be assembled increase in pressure according to the Combined Gas
under site conditions, these being less controlled than in law; thus external expansion reservoirs are not required.
the cable factory. The accessories form geometric dis-
continuities in the route, at which thermomechanical 3.2.7 Bonding Scheme
forces are concentrated, requiring them to be protected The bonding scheme grounds dc and ac cables at one
against movement by the provision of constraint sys- or more locations along the route. In an ac system,
tems in the joint vault and at the terminations. Unlike the alternating magnetic fields from the power currents
the cable, the accessories have a three-dimensional vari- induce a voltage in each of the adjacent cables and
ation in shape, thereby complicating both the design of auxiliary conductors. The magnitudes of the voltages
3-5
are designed to be below specified limits by reducing directly into the ground, to protect the cable from third-
the cable centerline spacings and the cable lengths party damage by laying the ducts at depth and encasing
between joints. them in concrete, and to cross obstructions such as
roads, rivers, and other services, using solutions such as
The act of solidly bonding a circuit, as shown in cable bridges, horizontal directional drilling (HDD),
Figure 3-1, forms a number of conducting loops thrust-bores, or tunnels.
between the shield/ground conductors in adjacent
cables. An additional loop is also formed by the conduc- 3.2.11 Operation, Maintenance, and Testing
tivity of the ground. The voltages drive circulating cur- When the cable system is designed, the electrical param-
rents through the cable ground return conductors, eters necessary to calculate the performance of the over-
which produce unwanted heat and significantly reduce all transmission system are provided to the electric
ampacity. In a specially bonded circuit, current flow is utility, as described in Chapter 16. In some applications,
prevented by interrupting the ground return conductors the constraints imposed by the electricity network
and shields at the accessories. The interrupted shield require that the cable dimensions and system design be
gaps experience unacceptably high transient voltages changed. This may occur, for example, in an all-cable
during circuit switching and lightning strikes, and must network, to reduce capacitance and charging current at
be protected by connecting across them suitably rated the sacrifice of cable diameter and optimum ampacity.
nonlinear resistors called SVLs (sheath voltage limiters).
The SVLs are voltage sensitive and “short” the insulated The satisfactory performance of a new design of cable
gap to limit the magnitude of the transient voltage. and accessories is validated before supply by performing
When the transient has passed, the SVL returns to high a formal qualification test (type tests)—for example, to
resistance and effectively becomes an “open-circuit.” standards: ICEA-S-108-720-2004 (ICEA 2004), AEIC
CS2 (AEIC 1997), AEIC CS9-06 (AEIC 2006), IEEE
3.2.8 Thermal Design Standard 404-2000 (IEEE 2000), IEEE Standard 48
The system thermal design ensures that the maximum (IEEE 2003), IEC 60840 (2004), IEC 62067 (2001). In
possible current is carried by the cable system without addition to qualification tests, long-term formal
exceeding the specified continuous operating tempera- prequalification tests are performed when service experi-
ture of the conductor and insulation. The designer ence is limited—e.g., for XLPE cable systems above
selects the conductor size, the spacing between cables to 150 kV to AEIC CS9 (AEIC 2006) and IEC 62067
limit mutual heating, and a stabilized backfill with low (2001). The quality of the cable and accessories is veri-
thermal resistivity. fied at the end of manufacture by performing “shipping
tests” (routine tests, factory tests, production tests) at
3.2.9 Thermomechanical Design 100% frequency of product and by more onerous “sam-
The system thermomechanical design protects the cable ple tests” performed at a reduced frequency. A formal
and accessories from experiencing any excessive forces acceptance test, which is also called a commissioning or
and movements that are generated when the cable is hand-over test, is performed on-site on the completed
heated, cooled, and load cycled. Design solutions are cable system to verify that the installation workmanship
selected to suit the method of cable installation. For is satisfactory, that no installation damage has occurred,
example, in direct-buried or close-cleated systems, the and that the circuit is safe to connect to the network.
cable is rigidly constrained. In tunnels or on trays, the Condition assessment tests are performed at periodic
cable is unconstrained in wave or snake patterns. In a intervals during an installed cable system’s life to iden-
pipe or duct-manhole system, the cable is semicon- tify deterioration and permit preventative action to be
strained and is allowed to expand laterally (Zenger et al. taken. Diagnostic tests are performed to investigate the
2006). In duct-manhole installations, methods of con- cause and extent of a malfunction.
straint are required in joint vaults to protect the joints
from movement. If the differential thermomechanical 3.3 ELECTROMAGNETIC FIELD
forces cannot be withstood by constraining the adjacent The electromagnetic field (EMF) surrounding the cen-
cable, either the cable system must be redesigned, or spe- tral conductor is divisible into two components, the
cial accessories, such as anchor joints, must be installed electrostatic field and the magnetic field. These fields are
to replace the conventional straight joints. produced by both a dc and an ac cable circuit. In the
former case, the fields are of constant magnitude and
3.2.10 Installation Design are non-time varying. In the latter case, the fields alter-
Installation design specifies the methods and equipment nate in magnitude with time at power frequency. Knowl-
necessary to install the cable into prelaid ducts or edge of the electromagnetic field is sometimes of interest
3-6
at the planning and approval stage of a circuit. At the Definition of Electrostatic Field
detailed design stage of the circuit, the fields are inextri- By definition, an electrostatic field strength, E, of
cably embedded within the equations for the design of 1V/m produces a force of 1 newton (0.225 lbs.f) on a
the cable insulation, ac heat losses, cable spacing, and particle carrying a charge of 1 coulomb, as shown in
induced voltage. Equation 3-3. The direction of the field is the path taken
by a unit positive charge, which is away from positive
This section provides a qualitative account of the distri- and toward negative potential.
bution of electrostatic and electromagnetic fields in the
F=QE (N) 3-3
vicinity of a cable circuit and an outline of their role in
cable design. The section offers a foundation for Where:
Section 3.6.3 on insulation design, Section 3.7.2 on F = Force (N).
inductance, Chapter 10 on grounding, and Chapter 16 Q = Electric charge (C).
on cable system considerations. E = Electrostatic field strength (V/m).
Distribution of Electrostatic Field
3.3.1 Electrostatic Field
As shown in Figure 3-5, the distribution of the electro-
The electrostatic field is produced by voltage and com- static field is completely contained within the cable con-
prises the electrostatic field strength, E, and the electro- struction by the insulation shield. Ri is the insulation
static flux density, D, which are related by the dielectric inner radius and Ro the outer radius. The external elec-
permittivity, ε, as given in Equations 3-1 and 3-2. The trostatic field for all types of cable is zero. The shape of
term “electrostatic field” usually refers to the electro- the field in Figure 3-5 is called “Laplacian,” and is deter-
static field strength, because this is the prime mover in mined by the externally applied voltage and the insula-
producing movement of charged particles. Electrostatic tion geometry and properties. In a dc cable at constant
field strength is also commonly named “gradient” voltage, the electrostatic field is of constant magnitude
because it is the gradient of voltage, V, with distance and continuous with time. In an ac cable, the field alter-
through the insulation, and “stress” because of the simi- nates in magnitude and direction at power frequency.
larity to mechanical stress.
Each type of cable insulation has a limiting electrostatic
D=εE (C/m2) 3-1
field strength (stress), which together with the magnitude
ε = εo εr (F/m) 3-2 of applied voltage, determines the insulation thickness.
Where:
E = Electrostatic field strength (V/m).
D = Electrostatic flux density (C/m2).
ε = Absolute permittivity of insulation (F/m).
εo = Permittivity of free space: 8.854 × 10-12
(F/m).
εr = Relative permittivity of insulation (ratio).
Symbols for electrostatic field strength are E or G, the

former being used in this chapter. The U.S. unit of elec-
trostatic field strength is V/mil (one mil is a thousandth
of an inch). The SI primary unit of electrostatic field
strength is MV/m; however, the secondary unit kV/mm
is generally preferred in the cable industry because cable
insulation is thin, being typically less than 1.4 in.
(35 mm).
Relative permittivity is also known as the “dielectric

constant” (and, in the 1992 Edition of the Green Book,
as “specific inductive capacity [SIC]”). In this edition of
the Green Book, both the terms “relative permittivity”
and “dielectric constant” will be used. The term “spe-
Figure 3-5 Electrostatic field distribution in the cable
cific inductive capacity (SIC)” will not be used. insulation (courtesy of CCI).
3-7
Effects of Electrostatic Field The basic equations that quantify electrostatic stress,
The electrostatic field in a dc cable produces the follow- insulation design, dielectric loss, and capacitance are
ing effects: considered later in this chapter, and their applications
• A transient flow of inrush current when the cable is are considered in detail in Chapters 11 and 16.
switched on and off voltage.
3.3.2 Magnetic Field
• An increase in the mobility of charge carriers present
The magnetic field is produced by flow of current and
in the insulation, resulting in an apparent reduction
comprises the magnetic field strength, H, and magnetic
in insulation resistivity in the high-stressed region
flux density, B, which are related by the magnetic per-
adjacent to the conductor. This reduces the magni-
meability of the surrounding medium, μ, as shown in
tude of the electrostatic field adjacent to the conduc-
Equations 3-4 and 3-5. The relative permeability for the
tor shield and increases it at the insulation shield.
following examples and diagrams, in which no ferrous
• The accumulation and entrapment in XLPE insula- material is present, is taken to be μr = 1.
tion of trapped charge in concentric “halos” adjacent
to the conductor and insulation shields, which is B = μΗ (T) 3-4
called “space charge.” The space charge is of oppo- μ = μo μr (H/m) 3-5

site polarity to the adjacent shield, thereby distorting Where:
the shape of the electrostatic field and increasing elec- H = Magnetic field strength (A/m).
trostatic stress. Some XLPE materials used in ac B = Magnetic flux density (T).
cables exhibit a significant reduction in the dc break- μ = Magnetic permeability of the surrounding
down voltage. Special types of XLPE insulation with medium (H/m).
a reduced space-charge effect have been developed μo = Magnetic permeability of free space:
for dc application. 4π × 10-7 = 1.257 × 10-6 (H/m).
μo = (Magnetic permeability of free space =
The alternating electrostatic field in an ac cable pro- 0.383 × 10-6 (H/ft)).
duces the following effects: μr = Relative permeability of surrounding
• Steady-state alternating flow of capacitive charging medium (ratio).
current through the insulation is supplied from the
conductor and thereby increases the I2R heat gener- The term “magnetic field” can refer to either the mag-
ated in the conductor. The path of current flow netic field strength, H, which is normally the first
through the insulation is expressed by a capacitive parameter to be calculated by engineers, or more usually
reactance, XC (ohms), which in circuit diagrams is to the magnetic flux density, B, which is the prime
connected in shunt across the supply and load. A key mover in producing both movement and induced volt-
component of XC is the insulation capacitance, C age in adjacent current carrying conductors. B is some-
(farads), which is a measure of the quantity of electric times named “magnetic induction.”
charge, Q (coulombs), stored in the insulation. In
Figure 3-5, the charge Q and voltage V are both pro- The unit of flux density, B, in the SI system is the tesla.
portional to the area under the curve, which is related The unit of flux density in the U.S. EMU (electromag-
by the identity Q = CV. netic unit) system is the gauss, this being still in use. The
conversion is 1 tesla = 104 gauss.
• Heat generation in the insulation by dipole rotation,
which increases the cable temperature and limits Definition of Magnetic Field
ampacity. The heat is called “dielectric loss.” The magnetic flux density, B, is defined to be 1 tesla
• A limit to the useful length of an ac circuit. At the when a force of repulsion of 1 N/m (0.069 lbs⋅f/ft) is
critical circuit length, the conductor is entirely heated experienced by two long parallel conductors spaced 1 m
by the flow of charging current and is unable to carry (39 in.) apart, each carrying 1 A in the opposite direc-
any load current. tion, as shown in Equation 3-6. The convention is for
3-8
the field to rotate clockwise, as shown in Figure 3-6, B = Magnetic flux density (T).
when the current flows away from the observer (Max- A = Area of the loop perpendicular to the
well’s cork-screw rule). The direction of current is flux (m2).
shown to be inwards by the X sign, which is visualized t = Time (s).
to be the flight feathers of an arrow of current, and out- Φ = Total flux (BA) linked with the loop (Wb).
wards by the • sign, which is visualized to be the point
of the arrow. The direction of force and movement is The SI unit of magnetic field strength, H, is the A/m.
given by Fleming’s Left Hand rule. The force in Equa- The U.S. unit is the oersted. The conversion is 1 A/m =
tion 3-6 is sometimes called the “Lorentz” force. 0.0126 oersted. The magnetic field strength, H, is some-
times called the “magnetizing force.”
F=BIl (N) 3-6
Where: Distribution of Magnetic Field

F = Force on current carrying wire (N). Figure 3-6 shows a simple “go” and “return” cable cir-
B = Magnetic flux density at wire (T). cuit, representing either a dc transmission system or an
I = Current in wire (A). ac single-phase distribution system. In a dc cable carry-
l = Length of wire (m). ing constant current, the magnetic field is continuous in
magnitude with time. In an ac cable, the field alternates
A uniform flux density, B, of 1 tesla, when linked to in magnitude and direction at power frequency. The
a perpendicular wire loop encompassing an area of direction of force between the two parallel conductors
1 square meter, will induce an electromotive force, e, of remains one of repulsion because the fields produced by
1 Volt when the magnetic flux density is reduced to zero, the two conductors alternate in synchronism.
at uniform rate, in a time, t, of 1 second, as shown in
Equation 3-7. The direction of voltage is given by Flem- The magnetic field strength, H, is present both inside
ing’s Right-Hand rule. and external to the cable construction and reduces to
zero with distance. In special situations, external shield-
e=−
d
(B A ) = − d Φ (V) 3-7
i n g c a n b e e m p l oy e d t o c o n t a i n t h e fi e l d ( s e e
Chapter 16). The distribution of magnetic field is in two
dt dt
Where: parts—that internal to the conductor, H i , and that
e = Electromotive force (V). external to it, He.
Figure 3-6 Magnetic field around a “go” and “return” dc cable circuit at a centerline
spacing S = 2D, where D is the cable outer diameter (courtesy of CCI).
3-9
Figure 3-7 is a plot of the magnetic field strength, H, Effects of Electromagnetic Field
along the axis joining the cable centers, showing the The alternating magnetic field in an ac system has the
individual fields from each conductor (dotted lines) and property expressed by the Faraday-Neumann law of
the resultant field (hard solid lines). Transmission-class inducing an electromotive force (voltage) in any parallel
cables are typically spaced at 12 in. (300 mm) to 20 in. conductors, such as metallic cable sheaths and ground
(500 mm) centerline spacing, which for a cable with conductors, both within the cable construction and in
diameter D = 6 in. (150 mm) is a spacing to diameter parallel cables. This induction is minimal in a dc cable
ratio of 2D–3.3D. The spacing shown in Figures 3-6 and system because the rate of change of current with time is
3-7 is 2D. Superimposition of the individual fields small and is only attributable to ripple current and to
from each cable increases the magnitude of the resultant load fluctuations.
field in the space between them (hard solid line) and
reduces it outside them, where the field reduces to zero Special ac cable constructions such as GIL cable and
with distance. HTS cable are surrounded by a high-conductivity cylin-
drical conductor. The induced voltages in these ground
The field inside the conductor, Hi, (shaded grey) rises to return conductors permit high current to be circulated
a maximum of 100% at the conductor surface. In the with minimal additional heat being generated. The cable
example shown, the field external to the conductor, He, shields and ground conductors are solidly bonded to
falls from 100% to 50% before it exits the cable. The each other and to ground at each end of the circuit, with
field at the center of the opposite cable has fallen to 13% minimal I 2 R heating. The circulating currents are of
in the 12 in. (300 mm) spacing. closely similar magnitude to the current in the power
conductor and, in accord with Lenz’s law, are in the
The magnitude of the external magnetic field at ground opposite direction. In consequence, the resultant exter-
level above either an ac or a dc system can be reduced by nal magnetic fields outside these special cable designs
reducing the centerline spacing of the cables compared are effectively zero since the two fields cancel.
to the depth of installation and providing shielding. The
pipe-type cable system has a minimal external magnetic In conventional ac cables of extruded and SCFF con-
field because the cables are in close proximity to each struction, the shield and ground conductors have an
other, the ferromagnetic pipe captures the magnetic appreciable resistance that, in a solidly bonded system,
flux, and eddy currents induced in the pipe reduce the limits the magnitude of circulating current. The degree
magnetic field produced by the cables. The gas-insulated of cancellation of the external field is still significant.
line (GIL) cable and most constructions of supercon- The disadvantage is that additional heat is generated in
ducting cable also exhibit a minimal external field, He. the auxiliary conductors, thereby derating the cable and
Figure 3-7 Magnetic field distribution in the lateral direction around two cables (courtesy of CCI).
3-10
requiring an increase in the area of the power conductor. • An increase in current density at the conductor sur-
The solution is to select a specially bonded system in face, thereby increasing conductor heating (skin
which circulating currents are prevented from flowing. effect). Under ac operation, the shape of the field
inside the conductor, Hi, shown in Figure 3-7,
In end-point and mid-point bonded systems, the shield changes from the straight-line distribution for dc cur-
and integrated ground conductors are interrupted by rent to a concave curve (not shown).
insulated gaps incorporated into the joints and termina-
• Non-uniformity in current density across the conduc-
tions called sheath interrupts. In a cross-bonded system,
tor diameter (proximity effect), thereby increasing
the route is divided into “major sections,” each consist-
conductor heating. This is indicated in Figure 3-7 by
ing of a group of three “minor sections.” At the joints
the asymmetry of the resultant internal conductor
between the minor sections, the shield/ground conduc-
field, Hi, (the position of zero field is shifted away
tors are transposed in position by being taken across
from the conductor center, and the shaded areas
one phase cable to the parallel cable in the next section.
under the curves are unequal).
The result is that each shield/ground conductor circuit
experiences an induced voltage of equal magnitude in
The basic equations that quantify the magnetic field,
each section, but of 120° time-shifted phase. The series
inductance, heat loss, and circuit parameters are consid-
connection of the three transposed minor section
ered later in thi s chapter and in applicati on i n
shield/ground conductors permits the resultant voltage
Chapters 11 and 16.
between the ends of each major section to be cancelled.
The terminations at the ends of the circuit and the joints
at the major sections, therefore, may be solidly bonded 3.4 POWER RATING
to each other and to ground without flow of circulating The first step in designing a cable system is to obtain the
current. On the other hand, an uninterrupted path, to power rating, which is usually specified by the utility. In
carry fault current from one end of the circuit to the the majority of cases, the transmission voltage is known
other, exists along the cable shield/ground conductors. because the cable is inserted into an existing network. In
A consequence of specially bonded systems is the the special cases of submarine dc circuits and long-
absence of a circulating current, thus the external mag- length ac circuits—e.g., 60 miles (95 km)—it is possible
netic field is not reduced in magnitude. for the cable designer to vary both the transmission
voltage and current levels and achieve an optimum
The external field may be reduced by increasing the power rating at minimum cost.
depth of cable burial, reducing the centerline spacing,
using a triangular formation, and/or installing metal With knowledge of the power and voltage, the second
cover plates or conducting loops above the cables (End- step is to calculate the current I. The cable is then
ersby et al. 1993). Methods of reducing the external designed in conjunction with the ampacity calculation.
magnetic field are described further in Chapter 16. In a dc cable, all the heat is generated by I2R ohmic heat
in the conductor. In an ac cable, a significant proportion
In an ac system, the alternating magnetic field also of the heat is generated by ac losses. For example, in
induces a “back-EMF” in the power conductor itself, 345-kV HPFF and 400-kV SCFF paper-insulated cable,
producing the following effects: this may be up to 30–50% of the total. In both ac and dc
• An impedance to the flow of current and to the cables, the conductor size is progressively increased until
the operating temperature is equal to or below the spec-
amount of useful power that the circuit may carry.
ified conductor design temperature.
In circuit design, this increase in the cable impedance
is in series with the conductor resistance and is
In the following section, it is assumed that the specified
expressed as the inductive reactance, XL, (ohms). A
power can be carried by one cable per dc pole or ac phase.
key component of XL is the inductance, L, (henrys),
It is quite usual for two cables per phase to be selected,
which quantifies the volume of magnetic flux linking
either to limit the size of the cable, or to provide continu-
the conductor both from its own conductor and from
ity of supply during planned and unplanned outages.
parallel cables. In Figure 3-7, the flux linkages are
proportional to the areas under the H curves, being
In calculating the current to match the power rating, the
linked by the identity B = μoμrH.
dc system is simplest and is considered first.
3-11
3.4.1 Direct Current (DC) Cable Systems PM

A dc system has two conductors—one “go” and one
IM = (kA) 3-8
V
“return”—as shown in each of the five circuit configura-
tions in Figure 3-8. Voltage is maintained constant by PB
IB = (kA) 3-9
the ac/dc converter station. In a current-source con- 2V
verter station, the polarity of the supply voltage is
reversed to reverse the power flow. In a voltage-source PM = VI M (MW) 3-10
converter station, voltage polarity is constant, and
power reversal is achieved by the power electronics. In PB = VI M + VI M = 2VI M = 2PM (MW) 3-11
both converters, current magnitude varies with power Where:
demand as in an ac network. PM = Power capacity of a monopole (MW).
PB = Power capacity of a bipole (MW).
DC cable is more efficient than an ac cable and can V = Pole voltage to ground (kV).
carry significantly higher current levels for the same IM = Pole current of a monopole (kA).
cable size. The disadvantages are that ac/dc converter IB = Pole current of a bipole (kA).
stations are more expensive than ac transformers and dc
transmission links cannot easily be interconnected to The type of cable system, Figure 3-8, is determined by
form a network. DC systems become economical for whether the ac/dc converter station has a monopole or
high power links, long circuit lengths, and connecting ac bipole output. A bipole system consists of two single
networks that are incompatible. cables of +V and −V to ground and has twice the power
rating of a monopole.
DC Current and Power
The current is calculated from Equation 3-8 for a mono- Monopole Cable
pole system and from Equation 3-9 for a bipole system. A monopole cable system, Figure 3-8a and b, consists of
When the ampacity of the cable is specified, the maxi- two power conductors, which carry the same current
mum transmission power is given by Equations 3-10 magnitude:
and 3-11.
Figure 3-8 DC cables in monopole and bipole configurations (courtesy of CCI).
3-12
• A high-voltage (HV) insulated “go” conductor tial discharge repetition rate under dc voltage is very
(pole). This may be at positive or negative voltage, low, being determined by the long natural time con-
depending on the converter station type and mode stant in which the void recharges. The rate of deterio-
of operation. ration of the dc insulation is similarly reduced.
• A low-voltage (LV) insulated “return” conductor at
3.4.2 Alternating Current (AC) Cable Systems
ground voltage. (Note that the ground conductor
requires to be insulated as a significant voltage drop The ac system consists of three “go” conductors, which
exists along its length). This may be incorporated as are designated phase 1, phase 2, and phase 3. Transmis-
an outer concentric conductor into the same cable sion systems carry equal magnitude currents in each
construction as shown in Figure 3-8a, or supplied as phase, which are 120° phase displaced. The three cur-
a separate LV cable, as shown in Figure 3-8b. In some rents sum to zero at the transformer wye point, and so
submarine installations, seawater has been used as the neutral return conductor used in unbalanced distri-
the “return” conductor. bution systems is unnecessary.
Bipole Cable The advantages of ac compared to dc transmission sys-

A bipole cable system, Figure 3-8c, d, and e, consists of tems are that an interconnected, meshed network can be
a minimum of two high-voltage insulated power con- extended to supply varying power demands over a wider
ductors (poles), both carrying the same magnitude of area, and that transformers permit the voltages to
current, with one operated at positive voltage and one at be easily transformed down and connected to distribu-
negative. The integrated cable design, Figure 3-8c, is tion networks.
limited by overall diameter to low transmission voltages
and smaller conductor areas, because the insulation AC System Voltage
between the inner conductor and concentric conductor The ac cable, in common with all items of equipment, is
has to be increased in thickness to withstand the com- specified by the system voltage, V, which is the line volt-
bined 2V voltage, and insulation is required over the age between overhead line conductors. In a cable sys-
outer concentric conductor to withstand V to ground. tem, each phase cable is isolated by the shields. The
At the higher transmission voltages, it is usual to select voltage experienced by the insulation is restricted to the
two separate monopole cables, as shown in Figure 3-8d. phase voltage to ground, Vg, which is 58% of the phase
To maintain continuity of supply during maintenance or voltage V (see Equation 3-12). Vg is used by the cable
failure, an LV insulated return conductor may either be designer to calculate the insulation thickness and to
built into each monopole cable, Figure 3-8a, or be laid determine test voltages.
as a separate cable, Figure 3-8e. This permits the bipole
V
system to operate in monopole at half power, with the Vg = = 0.58 V (kV) 3-12
other pole switched out. In normal operation, the LV 3
conductor carries negligible current. Where:
DC System Voltage Vg = rms phase voltage—i.e., phase-to-ground
A dc cable with the same insulation thickness as an ac (kV).
cable is capable of operating at approximately twice the V = rms line voltage—i.e., phase-to-phase
ac system voltage. In consequence, the dc design stress is (kV).
significantly higher than the ac stress. For example, the
design stress in an SCFF cable may be increased from The ac current is always stated as a root mean square
12–15 kV/mm for ac operation to 25–30 kV/mm for dc (rms) value. The rms current has the same I2R heating
operation, permitting the insulation thickness to be effect as a dc current of the same magnitude. In conse-
approximately halved (Arkell and Parsons 1981; Arkell quence, the system line voltage, V, and the phase voltage,
and Gregory 1984; Arkell et al. 1984). The reasons are: Vg, are also stated as rms quantities, because these val-
ues are linked by Ohm’s law, as shown in Equation 3-13.
• The ac voltage and stress are stated as rms values,
whereas the peak ac voltage is 40% higher and is Vg = I ⋅ Z (kV) 3-13
more indicative of dc insulation capability. Where:
• The partial discharge energy in a void under ac volt- I = rms phase current (kA).
age is forced to repetitively discharge energy at Z = Wye impedance (one phase to wye point)
greater than twice the ac power frequency. The par- of the load (Ω).
3-13
Table 3-1 gives the system voltages, V, in use in North t = Time (s).
America, the phase voltages, Vg, and the lightning test θ = Instantaneous angle of rotation at time
voltage, V P. The lightning voltage is called the basic t (deg).
impulse level (BIL). The ratio of BIL to phase voltage
significantly reduces as the system voltage is increased The three phase conductors 1, 2 and 3 have the same volt-
and so is important for cable design. The insulation in a age waveform, but each is displaced by one third of a
fluid-filled paper cable is designed primarily on the cycle, 120°. The instantaneous voltages in phases 2 and 3
stress at BIL. This produces a large performance margin are given by Equations 3-16 and 3-17. These voltages may
in normal operation at phase voltage and permits work- be expressed in phasor form, as shown in Equation 3-18,
ing stresses to be increased at the higher transmission taking the phase 1 voltage Vg1 as reference.
voltages. The insulation in an extruded cable is primarily
designed on the stress at phase voltage, with a perfor- v θ 2 = Vg sin (θ + 120 ) (kV) 3-16
v θ 3 = Vg sin (θ + 240 ) (kV)

mance margin being present at impulse voltage.
3-17
AC Voltage and Current Waveforms
In an ac system, the instantaneous system phase volt- Where:
age, vt, and current, it, sinusoidally alternate between θ = 360 f.t degrees.
positive and negative at the power frequency, f, which is
60 Hz in North America. The voltage at time, t, is given Vg.1 = Vg ∠(0)
by Equation 3-14, as a function of phase voltage Vg.
The cyclic waveform continuously repeats itself at inter- Vg.2 = Vg ∠(120)
vals of the periodic time f-1 seconds. Equation 3-14 is Vg.3 = Vg ∠(240) 3-18
given in radians, where one complete cycle occurs in 2π
(6.283) radians. For convenience, this is converted to Where:
degrees in Equation 3-15. θ is the time-varying angle of v θ 1 = Instantaneous rms voltage at angular rota-
rotation, where one cycle is θ = 360° and occurs in the tion θ in phase 1 (kV).
periodic time f-1. θ = Instantaneous angle of rotation at time
t (deg).
v t 1 = Vg sin (2π.f .t ) (kV) 3-14
The expressions for the waveforms of current in phases
v θ 1 = Vg sin (θ) (kV) 3-15 1, 2, and 3 are given in Equation 3-19 in which current
Where: lags voltage by the phase angle, ϕ. The phase angle, ϕ,
v t 1 = Instantaneous rms voltage at time t in between voltage and current results from the storage of
phase 1 (kV). inductive and/or capacitive energy by the load, which is
f = Frequency of generation (Hz). separate from the dissipation of power in the form of
true work done. Inductive energy causes the current to
lag system voltage (the mechanical equivalent of induc-
Table 3-1 System Voltages and Basic Impulse Voltage tance is inertia). Capacitive energy causes the current to
Levels in North America lead system voltage (the mechanical equivalent of
System Voltage Ratio
capacitance is elasticity). Current is a vector quantity
Lightning possessing magnitude and direction. The vector summa-
Line Voltage Phase Voltage Impulse Level BIL ⎛ BIL ⎞
Vg VP ⎜ ⎟ tion of the three phase currents (I1 + I2 + I3) at the wye
V ⎜ V ⎟
(kV rms) (kV rms) (kV) ⎝ g ⎠ point of the load is zero.
69 40 350 8.8
115 66 550 7.9 I1 = I ∠(0 − ϕ)
138 80 650 8.1
I 2 = I ∠(120 − ϕ)
161 93 750 8.2
230 133 1050 7.9 I 3 = I∠(240 − ϕ) (kA) 3-19
345 199 1300 6.5 Where:
525 303 1550 5.1 I1 = Current, rms, in phase conductor 1 (kA).
765 442 2050 4.6 ϕ = Phase angle by which I lags Vg (deg).
3-14
AC Current and Power When the cable is of long length, the cable capacitive
The apparent power, PA, (MVA) required to be carried charging current IC must be added to the load current I
by the cable system is specified by the utility. The rela- by vector addition to calculate the ampacity, noting that
tionship between real power, P, (MW), apparent power, the phase angle of I C leads the in-phase resistive load
PA, (MVA) and reactive power, PR, (MVAr) is shown in current by 90°. The method to calculate charging cur-
Figure 3-9. In the diagram, the apparent power, PA, rent is provided later in this chapter. Empirically, recal-
leads the real power, P, by the angle ϕ, where cos ϕ is the culation is unnecessary when IC is < 3% of I, because the
power factor of the load. resultant temperature rise is ≤ 5°C. This typically occurs
at circuit lengths of 20 miles (32 km) for a 345-kV XLPE
The current, I, is calculated from Equation 3-20 and is cable and 10 miles (16 km) for a 345-kV LPP cable.
used to determine the temperature rise and size of the
conductor. The relationship between PA , P, and the 3.5 CURRENT RATING
power factor (cos ϕ) is given in Equation 3-21, permit-
ting the alternative calculation of I in Equation 3-22. 3.5.1 Continuous Steady-State Rating
PA The sizing of the cable to carry the required current
I= (kA) 3-20 requires the calculation of the temperature rise of the
3.V conductor Δθ, the cable losses Wc, and the effective
P = PA cos ϕ (MW) 3-21 thermal resistance of the cable and surrounding envi-
ronment, Tt. Typical design temperature limits θ from
P IEC 60287 and IEC 61443 (IEC 1994, 1999) are given in
I= (kA) 3-22 Table 3-2.
3.V cos ϕ
Where:
PA = Apparent power to be carried (MVA).
P = Real power to be carried (MW).
ϕ = Phase angle between I and V (deg).
cos ϕ = Power factor of load (per unit). Table 3-2 Design Temperature Limits
Design Temperature
When the ampacity of a given cable is already known, Limits
the maximum apparent power PA that can be carried by (°C)
the circuit is given by Equation 3-23 and real power P Normal Short
Operation Circuit
by Equation 3-24. Application Type Item (θ) (θ2)
LDPE 70 130
V
PA = 3 Vg I = 3 I = 3.VI (MVA) 3-23
Extruded
HDPE 80 160
3 XLPE 90 250
Insulation EPR 90 250
V
P = 3Vg I cos ϕ = 3 I cos ϕ SCFF 90 250
3 Taped SCGF 85 160
= 3.VI cos ϕ (MW) 3-24 HPFF 85 200
PVC 70–80 200
Jacket Extruded LDPE 70 150
HDPE 80 180
Copper and
90 ≤ 350
Aluminum
Conductor
Lead alloy 90 210
Lead 90 170
Conductor
Compression 90 250
Exothermic
Connection 90 250
weld
Figure 3-9 Relationship between real and apparent power. Soldered 90 160
3-15
The equivalent electrical diagram for heat dissipation Δθ = Maximum allowable temperature rise (°C).
from a simple dc cable circuit is shown in Figure 3-10. Δθm = Temperature rise due to mutual
The heat W c is produced by the I 2 R conductor loss; heating (°C).
there are no ac-related losses. In the circuit diagram, the Wc = Heat losses generated in the conductor
heat loss is represented by the constant current genera- (W/m).
tor Wc. Tt = Total thermal resistance, conductor to
ground surface (K.m./W).
The mutual heating effect by the return cable is repre-
sented by the temperature rise of the soil Δθm. Δθm may The I 2 R heat generated in the conductor is given in
also result from radiation and convection when cables Equation 3-27 and is combined with Equation 3-26 to
are installed in air. The temperature rise due to mutual give the current I in Equation 3-28. Equation 3-28 is the
heating is added to θa, the ambient temperature of the basis for all rating calculations and permits the maxi-
ground surface. In the circuit diagram, the combined mum current to be calculated at the maximum permissi-
temperature rise above the reference datum of 0°C is ble conductor temperature θ.
represented by the constant voltage source (θa + Δθm).
WC = I 2 R θ (W/m) 3-27
The total thermal resistance to the flow of heat out-
wards from the conductor to the ground surface is rep- (θ − (θ a + Δθ m ) )
resented in the circuit diagram by the resistor Tt. In this
I= (A) 3-28
R θ Tt
diagram, the cable is assumed to be carrying constant
magnitude current 24 hours per day—i.e., a load factor Where:
of 100%. The thermal capacitances of the cable and soil I = Current (A).
are thus not modeled. Rθ = AC resistance of conductor at the maxi-
mum permissible conductor temperature,
The maximum permissible temperature rise of the con- θ (Ω/m).
ductor Δθ is solved by Ohm’s law, as shown in Equa-
tion 3-25—i.e., Δθ = WcTt, where Δθ = (θ−(θa + Δθm)), The complete current rating equations, which include ac
which is rearranged in Equation 3-26. losses and load factors of less than 100% are given in
Chapter 11 and in references IEC 60287 (IEC 1994),
Δθ = WC Tt = (θ − (θa + Δθ m )) (°C) 3-25 Neher and McGrath (Neher and McGrath 1957), and
WC =
(θ − (θa + Δθm ) ) (W/m) 3-26
Anders (Anders 1997, 2005).
Tt 3.5.2 Short-Circuit Rating

Where: The duration of the short circuit is set in relation to the
θ = Maximum operating temperature of con- clearance time of the circuit breaker (typically 3–
ductor (°C). 5 cycles) and more usually to that of the secondary pro-
θa = Ambient temperature (°C). tection. In some cases, times of up to between 1 and
Figure 3-10 Equivalent electrical circuit for cable heat dissipation (courtesy of CCI).
3-16
3 seconds are specified. The flow of short-circuit current In many cable designs, the area of the ground return
is along the power conductor to the fault, and the return conductor has to be increased to limit the temperature
flow is along the shield conductor/metallic sheath. In to the design limit. The cross-sectional area of an annu-
this short period, there is insufficient time for heat to lar layer of copper wires can usually be increased up to
flow radially from each conductor, and so adiabatic the limit set by the factory stranding machine. High-
heating of the conductor alone is considered (i.e., no resistance sheaths, such as lead-alloy or stainless steel,
heat loss). can be supplemented with a layer of copper wires.
The limiting temperature is determined by the adjacent 3.5.3 Emergency Rating

insulation and jacketing materials, the conductor con- Cable systems may carry an increased current following
nection method, and the mechanical forces that may be an unplanned outage elsewhere in the transmission sys-
present. International standards—such as IEC 60840 tem. The duration of the increased current is specified
(IEC 2004), IEC 62067 (IEC 2001), and IEC 61443 by the operator to limit, or prevent, thermal and ther-
(IEC 1999) in Table 3-2—give values for the first two momechanical deterioration to the cable system.
items, but not together with the simultaneous applica-
tion of short-circuit mechanical force, which must be The practice in North America is to calculate an “emer-
allowed for case by case. gency rating” in which the cable system is allowed to
operate above the normal continuous temperature θ
Equation 3-29 is used to determine the required cross- given in Table 3-2 and up to a specified emergency tem-
sectional area A at the temperature limit θ 2 , given in perature for a maximum specified time period. For
Table 3-2. The calculation is performed separately for example, specification AEIC CS9 (AEIC 2006) for
the power conductor, the shield conductor/metallic extruded insulation cables specifies an emergency opera-
sheath, the external ground continuity conductor (if tion time of no more than 72 hours (3 days) on average
present), the bonding leads connected at accessories, per year during the design life, without exceeding
and the links in link boxes. The temperature rise of the 216 hours (9 days) in any 12-month period. The maxi-
power conductor, in the majority of cases, is below the mum emergency temperature is specified to be 105°C for
design limit, and so the conductor area does not need to XLPE cables up to 345 kV. 105°C is conditional upon
be increased. Depending upon the magnitude of the cur- XLPE systems above 150 kV demonstrating satisfactory
rent and duration of the short circuit, this is also often performance in the prequalification tests to Standard
true for metallic sheaths when the thickness is deter- IEC 62067 (2001) with the addition of 90 loading cycles
mined by mechanical performance—for example, a cor- at 105°C, or at a higher temperature. The use of EPR
rugated extruded aluminum sheath with a thickness of cables is shown up to 138 kV, for which the emergency
120–160 mils (3–4 mm) on an SCFF cable, and a corru- operating temperature is specified to be 105°C. The lat-
gated welded copper sheath with a thickness of 30 mils ter is allowed to be increased to 130°C, if mutually
(0.7 mm) on an XLPE cable. agreed and if verified by qualification and prequalifica-
tion tests. Specification AEIC CS2 (AEIC 1990) for
⎛ ρθ .k.t ⎞ impregnated paper and LPP HPFF cable gives maxi-
A = I ⎜⎜ ⎟⎟ . 109 (mm2)
⎝ d.c. (θ2 − θ1 ) ⎠
3-29
mum emergency operation temperatures of 100°C and
105°C for durations of 300 and 100 hours per year,
Where: respectively. The average emergency operation duration
A = Cross sectional area of the conductor per year over the life of the cable system is limited to
(mm2). 20%—i.e., to 60 and 20 hours.
I = Short-circuit current, rms equivalent value
(kA). Outside North America the practice of some operators
θ2 = Short-circuit temperature limit (°C). is to recognize that, prior to emergency loading, the cir-
θ1 = Starting temperature prior to short circuit cuits are carrying a continuous “pre-load” current
(°C). (sometimes named a “pre-fault” current), this being less
d = Density of the conductor material (kg/m3) than the continuous rated current. In consequence, the
c = Specific heat of the conductor material pre-load conductor temperature is lower than the limit-
(J/Kg/°C). ing value of θ, Table 3-2. Equation 3-28 shows that the
ρθ = dc resistivity of the conductor material at ratio of pre-load temperature rise to rated continuous
θ2 (Ωm). temperature rise is proportional to the square of the
t = Duration of short circuit (s). ratio of pre-load current to rated current. For the case
k = ac to dc resistance ratio (ratio). of an XLPE cable with an inherently low dielectric loss,
3-17
a 60% pre-load current results in a conductor tempera- 3.6.1 Conductor

ture rise of only 36% of the design value. This provides The majority of conductors at transmission voltages of
a significant margin to transmit emergency load with- 69 kV and above are cylindrical in shape to ensure a
out exceeding the specified continuous temperature, smooth base for the conductor shield and a uniform
thereby safeguarding the cable system. The operator electrical stress distribution in the adjacent insulation. A
specifies a range of pre-load currents as a proportion of selection of conductor types is shown in Figure 3-11.
the rated continuous current (e.g., 0, 30, 60, 75, and
85%) and a range of short-time durations typically Stranded Round Wire Conductors
from 3 minutes to 24 hours, these being applicable to The majority of conductors are formed from layers of
both cables and to overhead lines. For a cable system, stranded copper or aluminum wires, although other
12 hours is considered to be a reasonable period for an types are also available. The helical winding of individ-
operator to redistribute and reduce current flow. The ual wires into successive layers gives the conductor
conductor temperature is not allowed to exceed the val- mechanical flexibility and stability in shape. The wind-
ues of θ in Table 3-2, thereby safeguarding the cable ing process is called “stranding” or “cabling,” and the
system from thermal and thermomechanical deteriora- wire conductor is called the “strand.”
tion, although the conductor connections in joints and
terminations must be capable of carrying the increased The individual wires are produced in a wire-drawing
current density. process by being pulled through lubricated, circular,
tungsten die-tools in successive operations from a rod
During the comparatively short period of emergency ~0.4 in. (10 mm) in diameter to typically 0.08–0.12 in.
loading, the high thermal capacity of the cable and its (2–3 mm). The final process for a copper wire is to be in-
surroundings results in the temperature rise being tran- line annealed, in which the wire is heated to high tem-
sient without reaching a steady-state temperature. The perature in a non-oxidizing atmosphere and then
method of current rating is given in Chapter 11. cooled. Annealing removes the effects of work harden-
ing during wire drawing—i.e., the electrical resistivity is
restored to its minimum value, and the copper is re-soft-
3.6 CABLE DESIGN
ened and bends more readily, since a copper wire bends
This section considers basic aspects of cable design and both by elastic and plastic deformation.
specification working outwards from the conductor.
The equations for circuit parameters that are related to Cable factories have a range of wire diameters to select
a single cable are also given—e.g., conductor resistance from, but for efficiency will generally manufacture a
and insulation capacitance. The calculation of cable given conductor area from one size only. The conductor
inductance is considered separately in Section 3.7.2, is started with one straight wire at its center. Successive
since this is also determined by the geometric configura- layers of round wires are then applied normally in oppo-
tion of the adjacent cables. site directions of lay, left hand, and right hand etc. Each
layer is usually “died-down” to a controlled diameter by
being pulled through a die. If one segment of a Milliken
conductor is being stranded, it will also be “compacted”
by being pulled through powered shaped rollers. The
stranding process partially re-hardens the conductor
and increases the resistance. Published tables of conduc-
tor resistance usually include a factor that increases the
maximum permissible resistivity to allow for this
unavoidable increase.
The dieing-down and compacting processes increase the

filling factor Ff, which is the ratio of the nominal area of
metal AN to the apparent overall area of the conductor
AA , as shown in Equation 3-30. Compacting the con-
ductor significantly reduces the conductor diameter and
more importantly minimizes the overall cable diameter
and cost. The overall conductor diameter DA is given in
Equation 3-31.
Figure 3-11 Stranded conductor types (courtesy of CCI).
3-18
A N = Ff A A (mm2) 3-30 The conversions between areas in kcmil and mm2 are:
• 1 kcmil = 0.5067 mm2
4 AN
DA = (mm) 3-31 • 1000 kcmil = 506.7 mm2
π Ff • 1 mm2 = 1.974 kcmil
Where: • 1000 mm2 = 1974 kcmil
AA = Apparent cross section of a conductor
(conductor and free space) (mm2). It is good practice to manufacture the outermost layer
AN = Nominal cross section of metal in a stan- of wire from one continuous length without in-process
dard conductor (mm2). joints such that the risk of breakage and subsequent
DA = Diameter over manufactured conductor penetration of the insulation is prevented. The outer
(mm). layer is usually heavily died down to provide a smooth
Ff = Filling factor, ratio of nominal to apparent bed for the conductor shield and accurately achieve the
conductor areas (per unit). specified diameter, which is especially important for an
extruded cable, because the conductor has to pass
The achievable filling factor Ff varies according to the through a close-fitting extrusion die without damage.
type of conductor, the degree of compaction, and the
efficiency of the stranding machine. Typical Ff values The types of stranded conductor are listed below
are as follows: using U.S. terminology, with alternative terminology
• Uncompacted concentric conductors: 0.75 in brackets.
• Compressed conductors: 0.8 Concentric Conductor (Uncompacted Conductor)
• Compacted round conductors: 0.9 The strand is laid up uncompressed and uncompacted
and so the individual wires retain their circular shape.
• Shaped wire (keystone and flat) segmental conduc- The filling factor is low at 0.75, which increases conduc-
tors: 0.95–0.97 tor flexibility. This conductor is unusual for transmis-
• Solid cylindrical conductors: 1.0 sion-class cables because it increases the cable overall
diameter and cost.
The unit of conductor cross-sectional area in the United
Compressed Conductor (Died-Down Conductor)
States is kcmil. The definition of a “circular mil” is:
Figure 3-11a shows the conductor in an extruded XLPE
cable. The outer layer has been deliberately flattened for
One circular mil is equivalent to the metal area of a
the XLPE application to provide a smooth outer layer
0.001 in. (0.0254 mm) diameter wire, which is an area of
and to reduce “fall-in” of the semiconducting shield.
0.785 × 10-6 in.2 (507 × 10-6 mm2).
The inner wires have been compressed lightly and have
substantially retained their circular shape.
1000 kcmil is equivalent to the metal area of 0.785 in.2
(506.7 mm2). Compact Round Conductor (Heavily Compacted Circular
Conductor)
The U.S. kcmil system permits the outer diameter D C A very high degree of compaction of 0.9 is achieved by
of the stranded conductor to be calculated from Equa- the use of powered rollers, such that the wires are com-
tion 3-32, and the metal area A kcmil to be calculated pacted into square, keystone shapes. These are compar-
from Equation 3-33, with knowledge of the number of atively stiff conductors.
wires N and the wire diameter Dw.
Hollow-Core Conductor (Circular Stranded Hollow-Core
Conductor)
N
DC = D W (in.) 3-32 A central duct is provided in an SCFF cable for the pas-
Ff sage of dielectric fluid and in a SCGF cable for the pas-
sage of gas. The stranded conductor is laid up onto an
Akcmil = 1000NDw2 (kcmil) 3-33 open spiral steel duct. The wire layers are lightly com-
Where: pressed (died down).
Akcmil = Area of metal (kcmil).
Compact Segmental Conductor (Milliken Conductor)
N = Number of wires
The size of the stranding machine limits the nominal
DW = Diameter of one wire (in.)
area of a concentric conductor to less than 3000 kcmil
DC = Diameter over conductor (in.)
(1600 mm2). Larger areas are formed from stranded seg-
Ff = Conductor filling factor (ratio).
ments with a number of pre-spiraled, sector-shaped seg-
3-19
ments. Figure 3-11b shows an HPFF four-segment Segmental Circular Conductor Hollow-Core (Conci
conductor, and Figure 3-11c shows a five-segment con- Conductor)
ductor in an XLPE cable. Conductor sizes of up to The keystone-shaped segments have the advantage that
7000 kcmil (3500 mm2) have been installed in Japan with they lock together and form a self-supporting duct. The
the prospect of 8000 kcmil (4000 mm2) being possible. A duct diameter can prospectively be increased in diame-
Milliken conductor is a stranded segmental conductor ter to, say, 1.4 in. (35 mm) such that the radial thickness
in which the segments are insulated from each other. of a large conductor size such as 4000 kcmil (2000 mm2)
This has very significant advantages for conductor sizes may prospectively be reduced to less than 13 mm. The
of greater than 1600–2000 kcmil (800–1000 mm2) ac resistance due to skin effect is reduced to a similar
because the ac resistance in a concentric conductor due value to a stranded Milliken conductor.
to skin and proximity effects significantly increases the Solid Circular Conductor
resistance when the radius exceeds the skin depth (the Solid cylindrical aluminum conductors up to 3000 kcmil
depth in which the major proportion of the current (1600 mm2) have been manufactured for use in XLPE
flows). For example, the ac resistance for a 4000-kcmil insulated cable. Aluminum has the advantage compared
(2000-mm2) concentric hollow-core conductor would be to copper of an increase in the ratio of skin depths of
30% higher than the dc resistance, whereas a Milliken 28% such that ac resistance is less affected by skin effect,
conductor reduces this to 13%. The disadvantages of a a reduced bending stiffness of 54% of copper, albeit a
Milliken conductor are increased stranding costs and significant proportion of conductor bending is by plas-
reduced output since more stranding passes are tic deformation, and a reduced weight, with a density of
required, and conductors for long-length submarine 30% of copper. The disadvantage is that the dc resis-
cables cannot be continuously laid up without making tance is 65% higher than copper. An aluminum conduc-
joints in each segment. tor having the same dc resistance as a copper conductor
Hollow-Core Compact Segmental Conductor (Milliken becomes economically attractive when the metal cost by
Hollow-Core Conductor) weight is less than 50% of copper.
Figure 3-11d shows a hollow-core Milliken conductor in
Conductor Resistance and Cross-Sectional Area
an SCFF cable. For conductor sizes up to 5000 kcmil
When manufacturing conductors and calculating
(2500 mm2), it is usual to support the segments on a pre-
ampacities, the value of electrical dc resistance for a par-
formed spiral duct with an internal diameter of 0.5 in.
ticular conductor size should always be taken from a
(12 mm). For larger conductor sizes, the stranded seg-
published conductor table—for example, ASTM B 8-04
ments can form a self-supported duct of 1 in. (25 mm)
(ASTM 2004) and IEC 60228 (IEC 2005). The pub-
internal diameter. Hollow-core conductors in SCFF
lished resistance is a guaranteed maximum value that
cables have the advantage in an ac system that, for a given
must not be exceeded and that is verified by measure-
area, the radial thickness of the conductor is reduced in
ment in the cable shipping test (routine test). The pub-
comparison to the non-hollow-core conductors used in
lished conductor resistances are based on a higher value
XLPE cables and so have reduced ac resistance.
of electrical resistivity than those given for copper and
Stranded Shaped Wire Conductors aluminum in the International Annealed Copper Stan-
A closely similar manufacturing process to round wires dard (IACS), as shown in Table 3-3. In the conductor
is followed for the dieing down of shaped wires—e.g., tables, the value of IACS resistivity is increased to allow
flat wires and keystone segments. for the lay factor, which is the increase in helical wire
length compared to a straight axial length. For example,
Segmental Circular Conductor (Conci Conductor)
in ASTM B 8-04, the values of conductor resistance
The application of preformed segments or flat wires has
allow for increments of 2% up to 2000 kcmil, 3% up to
the advantages of an increase in the conductor filling
3,000 kcmil, 4% up to 4000 kcmil, and 5% up to
factor to ~99%, thereby reducing the conductor diame-
5000 kcmil. The values in IEC 60228 include similar, but
ter and the volume and cost of the outer cable layers,
undisclosed, lay factors varying typically from 2 to 4.5%
and reducing the volume of the impregnating com-
and additionally includes factors for typical construc-
pound in an MI (mass impregnated paper) dc cable. The
application of preformed segments has the advantage
that the continuous stranding of sizes of circular con- Table 3-3 IACS Reference Resistivities
ductor larger than 3000 kmil (1600 mm2) in a single Resistivity at 20°C Temperature
pass is made possible for long-length dc land and sub- IACS (%) ρ Coefficient α
marine cables. Conductor Conductivity (μΩ m) (per °C)
Annealed Copper 100 0.017241 0.00393
Aluminum 60.6 0.02845 0.00403
3-20
tion changes. Importantly IEC 60228 refers to each con- Table 3-4), because these are based on the geometry of
ductor size, such as 2000 mm2, as being a “nominal uncompacted round wires. The ASTM table is included
area.” Thus, if conductor resistances are calculated from in this chapter for reference only, because it gives the
first principles and IACS resistivities alone, the resis- values of resistance applicable to all conductor types in
tances will be too low. North America up to 5000 kcmil, and it permits accu-
rate geometrical calculation of GMR. However, ASTM
Examples of the ASTM kcmil conductor sizes for con- Class B conductor constructions themselves are not
centric conductors are given Table 3-4. For this type of used for transmission-class cables, because the outer
uncompacted conductor, ASTM specify the area of diameter is excessively large, the outer surface does not
metal cross-section in kcmil, the number of wires, the provide a smooth bed for shields, and they are more dif-
diameter of each wire, the conductor diameter, the mass, ficult to water block for cables with extruded insulation.
and the resistance. Conductors that are compressed or
compacted have smaller diameters, which vary between The following tables of conductor dimensions are exam-
manufacturers. Examples are shown in Tables 3-5, 3-6, ples only. Apart from the achievement of the specified
and 3-7. The equivalent metric conversions are included minimum resistance, transmission-class cables do not
in the tables. have standardized conductor dimensions (e.g., segment
designs, wire sizes, and degree of compaction). Manu-
The geometric mean radius (GMR) is added in facturers have significantly different designs of conduc-
Tables 3-4, 3-6, and 3-7. GMR is used in the calculation tor, wire size ranges, stranding machines, and dies.
of conductor internal inductance and circuit inductance Manufacturers change from concentric to segmental
(see Section 3.7.2). GMR is a mathematical function constructions at different conductor sizes, examples
that is dependent upon the number of wires, the distance being given in Section 3.7.3, which considers the ac
between each wire, and the overall conductor shape. resistance of conductors. The approximate outer diame-
ter of the conductor DA may be calculated for any con-
The only standardized conductor dimensions are those struction from Equation 3-31 using the associated list of
of the ASTM Class B concentric conductors (see filling factors Ff.
Table 3-4 Parameters for ASTM Class B Concentric Conductors in North America
Diameter DC Resistance at 20°C
Individual Strands Overall Diameter R Geometric
No. of
Size Wires DW DC Copper Aluminum Mean Radius
kcmil mm2 N in. mm in. mm μΩ/ft μΩ/m μΩ/ft μΩ/m in. mm
250 127 37 0.0822 2.0877 0.575 14.608 42.3 139.0 69.4 227.7 0.221 5.613
300 152 37 0.0900 2.2860 0.630 16.002 35.3 116.0 57.8 189.6 0.242 6.147
350 177 37 0.0973 2.4714 0.681 17.297 30.2 99.1 49.5 162.4 0.262 6.655
400 203 37 0.1040 2.6416 0.728 18.491 26.4 86.6 43.4 142.4 0.280 7.112
450 228 37 0.1103 2.8016 0.772 19.609 23.5 77.1 38.5 126.3 0.296 7.518
500 253 37 0.1162 2.9515 0.813 20.650 21.2 69.5 34.7 113.9 0.312 7.925
550 279 61 0.0950 2.4130 0.855 21.717 19.2 63.0 31.5 103.4 0.330 8.382
600 304 61 0.0992 2.5187 0.893 22.682 17.6 58.1 28.9 94.8 0.345 8.763
650 329 61 0.1032 2.6213 0.929 23.597 16.3 53.5 26.7 87.6 0.359 9.119
700 355 61 0.1071 2.7203 0.964 24.486 15.1 49.5 24.8 81.4 0.372 9.449
750 380 61 0.1109 2.8169 0.998 25.349 14.1 46.2 23.1 75.8 0.385 9.779
800 405 61 0.1145 2.9083 1.031 26.187 13.2 43.3 21.7 71.2 0.398 10.109
900 456 91 0.1215 3.0862 1.094 27.788 11.8 38.7 19.3 63.3 0.424 10.770
1000 507 91 0.1280 3.2512 1.152 29.261 10.6 34.8 17.3 56.8 0.446 11.328
1250 633 91 0.1172 2.9769 1.289 32.741 8.46 27.8 13.9 45.7 0.499 12.675
1500 760 91 0.1284 3.2614 1.412 35.865 7.05 23.1 11.6 37.9 0.547 13.889
1750 887 127 0.1174 2.9820 1.526 38.760 6.05 19.8 9.91 32.4 0.592 15.037
2000 1013 127 0.1255 3.1877 1.632 41.453 5.29 17.4 8.67 23.5 0.633 16.078
2500 1267 127 0.1403 3.5636 1.824 46.330 4.28 14.0 7.01 22.9 0.708 18.045
3000 1520 169 0.1332 3.3833 1.998 50.749 3.56 11.7 5.83 19.1 0.775 19.766
5000 2534 217 0.1518 3.856 2.581 65.557 2.18 7.15 3.57 11.7 1.001 25.534
3-21
Compressed conductors (see Table 3-5) for transmis- Heavily compacted round wire conductors (see Table 3-6)
sion-class cables are typically used up to 1500–2000 are seldom used in transmission-class cables above
kcmil (800–1000 mm2), above which compact segmental 1000 kcmil. Compact round wire conductors comprised
conductors are normally selected for ac current applica- of shaped wires (flat or key-stone) can be offered up to
tions, because these have the advantage of a significantly 5000 kcmil (2500 mm2); however, for ac current applica-
reduced ac/dc resistance ratio. Some manufacturers that tions, they are unlikely to be competitive for conductors
do not have suitable conductor laying-up machines may above 1500–2000 kcmil (800–1000 mm2).
continue to offer compressed round conductors up to
3000 kcmil (1600 mm2); however, these will have either a Compact segmental (Milliken) conductors (see Table 3-7)
reduced ampacity or larger conductor size than the typically start at or above 1500–2000 kcmil (800–
competing segmental design (see Section 3.7.3). 1000 mm2), and are offered for ac current applications up
to the largest conductor sizes, which are normally
5000 kcmil (2500 mm 2 ), with special applications of
7000–8000 kcmil (3500–4000 mm2) being possible.
Table 3-5 Typical Compressed Conductor Parameters in
North Americaa
No. of Overall Diameter

Size Strands DC
kcmil mm2 N in. mm Table 3-6 Typical Compact Round Conductor Parameters in
250 127 37 0.558 14.173 North Americaa
300 152 37 0.611 15.519 Geometric
350 177 37 0.661 16.789 Size No. of Overall Diameter Mean Radius
400 203 37 0.706 17.932 kcmil mm2 Strands in. mm in. mm
450 228 37 0.749 19.025 250 127 37 0.520 13.2 0.203 5.156
500 253 37 0.789 20.041 300 152 37 0.570 14.5 0.222 5.639
550 279 61 0.829 21.057 350 177 37 0.616 15.6 0.240 6.096
600 304 61 0.866 21.996 400 203 37 0.659 16.7 0.257 6.528
650 329 61 0.901 22.885 450 228 37 0.700 17.8 0.273 6.934
700 355 61 0.935 23.749 500 253 37 0.736 18.7 0.287 7.290
750 380 61 0.968 24.587 550 279 61 0.775 19.7 0.302 7.671
800 405 61 1.000 25.400 600 304 61 0.813 20.7 0.317 8.052
900 456 61 1.061 26.949 650 329 61 0.845 21.5 0.329 8.357
1,000 507 61 1.117 28.372 700 355 61 0.877 22.3 0.342 8.687
1,250 633 91 1.250 31.775 750 380 61 0.908 23.1 0.353 8.966
1,500 760 91 1.370 34.798 800 405 61 0.938 23.8 0.365 9.271
1,750 887 127 1.480 37.592 900 456 61 0.999 25.4 0.389 9.881
2,000 1013 127 1.583 40.208 1,000 507 61 1.060 26.9 0.413 10.490
a. DC resistances are the same as those in Table 3-4. a. DC resistances are the same as those in Table 3-4.
Table 3-7 Typical Compact Segmental Conductor Parameters in North America

DC Resistance at 25°C Geometric Mean
Size No. of Overall Diameter Copper Aluminum Radius
kcmil mm2 Strands in. mm μΩ/ft μΩ/m μΩ/ft μΩ/m in. mm
1000 505 148 1.152 29.2 10.8 35.4 17.7 58.1 0.449 11.405
1250 693 148 1.289 32.7 8.67 28.4 14.2 46.6 0.502 12.751
1500 760 148 1.412 35.9 7.23 23.7 11.9 39.0 0.550 13.970
1750 887 244 1.526 38.8 6.19 20.3 10.2 33.5 0.594 15.088
2000 1013 244 1.632 41.5 5.42 17.8 8.89 29.2 0.635 16.129
2250 1140 244 1.730 43.9 4.82 15.8 7.91 26.0 0.675 17.145
2500 1267 244 1.824 46.3 4.34 14.2 7.12 23.4 0.710 18.034
3000 1520 364 1.998 50.7 3.61 11.8 5.92 19.4 0.778 19.761
3500 1773 364 2.159 54.8 3.10 10.2 5.08 16.7 0.841 21.361
4000 2027 364 2.309 58.6 2.71 8.9 4.44 14.6 0.899 22.835
3-22
Table 3-8 gives the specified standard conductor sizes Electrical resistance increases linearly with temperature
and their resistances from Standard IEC 60228 (IEC i n t h e wo r k i n g r a n g e s c o n s i d e r e d a s g i v e n i n
2005) for up to 2500 mm2, the closest North American Equations 3-36 and 3-37. The coefficients of increase of
size being 5000 kcmil. resistance for copper and aluminum with temperature
are given in Table 3-3. These and similar equations are
The dc resistance R of a solid conductor of length L and used in the calculation of ampacity. ASTM and IEC
area A is given in Equation 3-34. Reference resistivities give tables of temperature correction factors to permit a
ρ for copper and aluminum base metal without factors resistance measured at shipping test to be corrected to
for stranding or work hardening are given in Table 3-3. the reference temperature, which is usually 20°C for IEC
Table 3-3 shows that aluminum has 60.6% of the con- standards and 25°C for AEIC and ICEA standards.
ductivity of copper.
R t = R o (1 + α o Δ θ) (Ω/m) 3-36
ρL
R= (Ω) 3-34 θ = θo + Δθ (°C) 3-37
A
Where: Where:
R = Resistance of conductor length L and area Rt = Resistance of conductor per unit length at
A (Ω). temperature t (Ω/m).
ρ = Electrical resistivity of conductor wire at Ro = Resistance of conductor per unit length at
reference temperature θo (Ωm). reference temperature (Ω/m).
L = Effective length of conductor (m). αo = Temperature coefficient of resistance at
A = Cross-section of metal in a solid conductor reference temperature (per unit).
(m2). Δθ = Temperature rise (°C).
θo = Reference temperature (°C).
The nominal area of the conductor A N may be deter-
mined approximately from Equation 3-35. However, the
apparent electrical resistivity ρ A calculated from 3.6.2 Conductor Shield
Table 3-8 has to be increased by factors allowing for The conductor shield (screen) is applied over the conduc-
stranding and work hardening. It is recommended that tor and is in close contact with the inner surface of the
the apparent resistivity be calculated from the resistance insulation, which is the highest stressed part. Its purpose
of a similar cable construction in the appropriate table, is to smooth the outer rippled surface of the conductor
whether ASTM or IEC. wires into a cylinder and thereby eliminate local concen-
trations of electric stress, called “stress raisers.” In a cable
ρA L with extruded insulation, the semiconducting shield is
AN = (mm2) 3-35
R bonded to the insulation such that a gaseous void cannot
be formed during expansion and contraction, as would
Table 3-8 International Standard Conductor Sizes and
occur if the insulation is extruded directly onto the con-
Maximum Resistances, IEC 60228 ductor. In a cable with taped insulation, the interface
between the shield and insulation tape is always filled
Maximum DC Resistance
(μΩ m-1) or (μΩ km-1) at 20°C with the insulating medium, thereby preventing forma-
Nominal Area Stranded Solid tion of voids and any possibility of partial discharging.
mm2 kcmil Copper Aluminum Copper Aluminum
150 296 124.0 206.0 124.0 206.0 3.6.3 Insulation
185 365 99.1 164.0 101.0 164.0 The cable electrical insulation is also called the “dielec-
240 474 75.4 125.0 77.5 125.0 tric,” which is a material that, when voltage is applied,
300 592 60.1 100.0 62.0 100.0 stores electric charge, without the free flow of electricity
400 790 47.0 77.8 46.5 77.8 and the generation of heat that occurs in a resistor.
500 987 36.6 60.5 - 60.5
630 1244 28.3 46.9 - 46.9 The requirements of cable insulation for transmission
800 1579 22.1 36.7 - 36.7 class cable are:
1000 1974 17.6 29.1 - 29.1 • To withstand high electrical stress for a nominal
1200 2369 15.1 24.7 - 24.7
design life of 40 years.
1600 3158 11.3 18.6 - -
2000 3948 9.0 14.9 - -
• To perform under a wide range of temperatures. In
2500 4935 7.2 12.7 - - normal operation, this is typically 15°C to 90°C, but
3-23
low winter ambient temperatures and short-term mum design stress and others allowing the manufacturer
emergency operation extend this in some applications to select the stress and to validate it by performing long-
to −20°C to +130 °C. term prequalification tests and qualification tests.
• To be sufficiently flexible to permit bending to small
radii during manufacture, installation, and service Different types of insulation have different limiting volt-
life—for example, in the thermomechanical system age conditions. Examples are given below:
designs that are unconstrained, or semi-constrained • Extruded XLPE cable in an ac system. The limiting
such as HPFF and duct-manhole systems. condition occurs under elevated ac voltage, both
• To be manufactured to the very highest quality when hot and when load cycled.
standards such that defects and contaminants • Self-contained fluid-filled cable in an ac system. The
are excluded. limiting condition occurs under lightning impulse,
when hot.
The maximum stress is the key parameter in determin- • HPFF fluid-filled cable in an ac system. The limiting
ing the insulation radial thickness. The performance of
condition occurs under lightning impulse, when hot.
the cable insulation is validated by formal long-term
prequalification (PQ) and qualification tests (type tests) • Self-contained gas-filled cable in an ac system. The
performed on installations of full-sized cable-containing limiting condition occurs under elevated ac voltage.
accessories. The cable system must withstand without This requirement is excluded from most test specifi-
failure or deterioration each elevated voltage, tempera- cations to preserve the cable from unrepresentative
ture, and loading condition. conditions. The second limiting condition occurs
under lightning impulse, when hot.
The phase voltage Vg and the lightning impulse voltage • GIL cable in an ac system. The limiting condition
VP are given in Table 3-1 for the common transmission occurs under lightning impulse.
voltages in North America. The ratio V P /V g is deter- • Extruded XLPE cable in a dc system. The limiting
mined by the insulation co-ordination of the electricity
condition occurs under elevated dc voltage,
network under lightning strike and results in a high
particularly during transient tests, such as impulse
ratio at 69 kV and low ratio at 765 kV. The ratio has
superimposed on dc and dc polarity reversal, follow-
important consequences on insulation design, because
ing heating.
in general lower system voltage cables are designed to
withstand the impulse lightning voltage, thus giving a • Fluid-filled self-contained cable in a dc system.
natural performance margin at working voltage, and at The limiting condition occurs under lightning
higher system voltages, when the cable is designed on impulse and upon impulse superimposed on dc volt-
impulse voltage, the performance margin at normal age, when hot.
working voltage is reduced. • Mass-impregnated paper-insulated cable in a dc
cable. The limiting condition occurs at elevated dc
The application of a 69-kV insulation design stress to a voltage, when cooling after a load cycle.
345-kV cable would result in an uneconomic system,
having an excessively thick insulation, a large cable A cable has axial symmetry and so has no longitudinal
diameter, a short reel length, and a large number of component of stress. This is not so in an accessory in
joints. The development of higher operating stresses was which the diameter of the shields and insulation is
essential to make 345-kV cable systems practical. greater than those in the cable, thereby introducing lon-
gitudinal components of stress. At the interface between
The practice in North America for pipe-type cable is to the accessory and the cable, the magnitude of resultant
specify conductor dimensions and insulation thickness, stress is resolved into a longitudinal component (tan-
as given in AEIC CS2 (AEIC 1997). The practice for gential stress), which is usually the weaker direction,
extruded XLPE cable is to specify for each transmission and a normal component (i.e., at 90°), which is usually
voltage a maximum stress at the conductor shield, AEIC the stronger direction. The magnitude of longitudinal
CS9 (AEIC 2006) and ICEA S-108-720 (ICEA 2004) stress dictates the shape of an accessory in an HPFF or
and at the insulation shield AEIC CS9. For example, at SCFF system. Accessories in extruded systems can
345 kV, the maximum stress limit at the conductor shield operate at higher interfacial stresses, providing no air
is 350 V/mil (14 kV/mm) and at the insulation shield voids are present and the interface remains undisturbed
150 V/mil (6.0 kV/mm). The practice outside North in service.
America varies, with some authorities specifying maxi-
3-24
Electrical Stress Distribution lation layers. When stress is to be calculated in the next
The concentric geometry of a cable produces a nonlin- layer B, the new value of dielectric permittivity εrN = εrB
ear voltage distribution, Vr, at a variable radius r as is inserted. In consequence, the magnitude of stress
shown in Equation 3-38. increments in magnitude at the interface, as given in
Equation 3-42. The stress at the interface is inversely
⎛R ⎞ proportional to the ratio of permittivities, such that the
ln⎜ o ⎟
Vr = V ⎝
r ⎠ stress is always highest in the low permittivity insulation.
(kV) 3-38
⎛R ⎞
ln⎜⎜ o ⎟⎟ Typical values of relative permittivity of insulating mate-
⎝ Ri ⎠ rials used in cables and accessories are given in Table 3-9.
Relative permittivity is also named specific inductive
In the context of insulation design, electrical stress is capacity (SIC) and dielectric constant. The ratios of the
defined as the gradient of the voltage distribution, as permittivities to air (or any gas) show that the stress
shown in Equation 3-39. The variable stress Er at radius increases experienced by an air void are very high.
r in a single type of insulation (called a “single permit-
V
tivity” insulation) is given in Equation 3-40. The stress is Er N =
inversely proportional to radius. Equation 3-40 is the ⎡ 1 ⎛ R ⎞ 1 ⎛ R oB ⎞⎤
main equation used in cable design since most cables r N ε rN ⎢ ln⎜⎜ oA ⎟⎟ + ln⎜⎜ ⎟⎟⎥
have single permittivity insulation. ⎣ ε rA ⎝ R iA ⎠ ε rB ⎝ R iB ⎠⎦ N
(kV/mm) 3-41
dV
Er = (kV/mm) 3-39 ⎛ε ⎞
dr E B = E A ⎜⎜ A ⎟⎟ (kV/mm) 3-42
V ⎝ εB ⎠
Er = (kV/mm) 3-40
⎛R ⎞ Where:
r ln⎜⎜ o ⎟⎟ ErN = Electrical stress at radius r in material layer
⎝ Ri ⎠ N (kV/mm).
Where: rN = Radius in material N at which E is to be
Er = Stress at radius r (kV/mm). calculated (mm).
Vr = Voltage at radius r (kV). RiA = Inner radius of insulation A (mm).
V = Voltage applied across the insulation (kV). RoA = Outer radius of insulation A (mm).
r = Radius at which E and V are to be calcu- N = Number of material layers of different per-
lated (mm). mittivity.
Ri = Inner radius of insulation (mm). εrA = Relative permittivity of insulation A
Ro = Outer radius of insulation (mm). (per unit).
εrB = Relative permittivity of insulation B
The units of stress E follow those of the applied voltage (per unit).
V. In the United States, the voltage is entered into the εrN = Relative permittivity of insulation N
equation as kV and the radii as inches. The unit of stress (per unit).
is kV/in., which is numerically equal to V/mil (volts per
thousandth of an inch). The permittivity ε is part of the equation for capaci-
tance C, as shown in Equation 3-43. Permittivity quanti-
In the SI system, voltage is entered as kV and radii as fies the ability of the insulation to store electric charge Q
mm, and so the unit of stress is kV/mm. The conversion when an ac or dc voltage V is applied, as shown in
is 1 kV/mm = 25.4 V/mil. Equation 3-44. Permittivity ε is the magnitude of the
capacitance C of an elemental cube when the dimen-
Equation 3-41 gives the stress in a multiple-layer insula- sions A and d are of unit dimension, as shown in Equa-
tion—for example, in an EPR molded one-piece joint on tion 3-45. Equation 3-46 is the fundamental relationship
an XLPE cable. It is necessary to insert the dielectric per- that relates flux density D to stress, where D = Q/A and
mittivity εrN = εrA when stress is calculated in layer A. E = V/d. For convenience, the absolute permittivity,
Within layer A, stress is proportional to radius. The ε = εo εr, is separated into the permittivity of free space
expression in brackets remains constant for all the insu-
3-25
εo and the relative permittivity of the particular insula- Examples of Calculation of Stress in an AC Cable
tion εr. In the following section, the XLPE insulation thickness,
insulation shield stress, void stress, and impulse stresses
εA are calculated for a selected manufacturer’s design stress
C= (F) 3-43
d of 282 V/mil (11.1 kV/mm), which is below the limit
of 350 V/mil (14 kV/mm) specified in AEIC CS9
Q = CV (C) 3-44 (AEIC 2006):
⎛A⎞ • System voltage, V: 345 kV
ΔQ = ε ⎜ ⎟ ΔV = ε ΔV (C) 3-45
⎝d⎠ • Phase voltage, Vg: 199 kV
• BIL: 1300 kVp
D=εE (C/m2) 3-46
• Conductor size: 2500 kcmil (1250 mm2)
Where:
C = Capacitance (F). • Radius over conductor shield, Ri: 0.92 in. (23.4 mm)
Q = Electrical charge (C). • Insulation type: Extruded XLPE insulation
ΔQ = Elemental charge (Q). • Insulation design stress, ERi: 282 V/mil (11.1 kV/mm)
V = Voltage (V).
ΔV = Total difference across cube (V). Note that the metric system of units is shown in the
D = Electrostatic flux density (C/m2). equations below, where insulation dimensions are in
E = Electrostatic stress, also called electro- mm, the voltage in kV, and the resultant stress in
static field strength (V/m). kV/mm. The equations can be used without modifica-
A = Area of the face (m2). tion for the U.S. system of units by inserting the insula-
d = Distance between faces (m). tion dimensions in in., the voltage in kV, and the
ε = Absolute permittivity = εo εr (F/m). resultant stress in V/mil. Alternatively, the conversion is
εo = Permittivity of free space: 8.854 × 10-12 25.4 V/mil = 1 kV/mm (or 100 V/mil = 3.93 kV/mm).
(F/m).
εr = Relative permittivity (from Table 3-9). Insulation Thickness, t, and Insulation Radius, Ro
The maximum stress Emax occurs at the conductor shield
radius Ri as given in Equation 3-47. This equation is
rearranged in Equation 3-48 to give the insulation thick-
Table 3-9 Relative Dielectric Permittivities of Some Cable
and Accessory Insulations
ness t = Ro−Ri.
Relative Stress Raising Vg
Insulating Permittivity Factor in Adjacent
εr
E max = E Ri =
Type Material Air Void
⎛R ⎞
Gas (all) Air 1.0 1.0 R i ln⎜⎜ o ⎟⎟
Insulating PIB-DDB-
2.2–2.4 2.2–2.4 ⎝ Ri ⎠
fluids siliconea
(kV/mm) or (V/in.) 3-47
Extruded XLPE 2.3–2.5 2.5
Extruded EPR 2.5–3.5 3.5 t = (R o − R i )
Stress control Molded CSPb 8–10 10
⎡ ⎛ Vg ⎞ ⎤
= R i ⎢exp⎜⎜ ⎟⎟ − 1⎥ (mm) or (in.)
Blends of
Stress control metal oxide 15–25 25 3-48
and polyolefin ⎣⎢ ⎝ R i E max ⎠ ⎦⎥
Taped FF LPP laminate 2.6–2.8 2.8
Taped GF Kraft paper 3.4 3.4 Result:
Taped FF Kraft paper 3.5 3.5
• t = 1.06 in. (27 mm).
a. PIB: polyisobutene, an insulating liquid used in HPFF
paper and LPP cables. • Ro = 1.98 in. (50 mm).
DDB: dodecyl benzene, an insulating liquid used in LPFF
paper and LPP cables. Stress at Insulation Screen, ERo
b. CSP: chlorosulphonated polyethylene, an elastomer (rub- The stress at the outer shield is required to check whether
ber) with high relative permittivity used in some extruded it is too high for an accessory to operate satisfactorily.
cable accessories.
3-26
Fro m ab ov e, t h e v a l u e s o f E R i = 2 8 2 V / m i l tion 3-50 to obtain the impulse stress at the conductor

(11.1 kV/mm), Ri = 0.92 in. (23.4 mm), and Ro = 1.98 in. shield ERip and into Equation 3-49 (where ERi = ERip) to
(50 mm) are entered into Equation 3-49. obtain the insulation shield stress, ERop.
⎛R ⎞ E Rip = E Ri
VP
E Ro = E Ri ⎜⎜ i ⎟⎟ (kV/mm) or (V/mil) 3-49
Vg
(kV/mm) or (V/mil) 3-50
⎝ Ro ⎠
Result:
Result: ERo = 131 V/mil (5.2 kV/mm).
• ERip = 1842 V/mil (73 kV/mm).
This stress is satisfactory, being below the 150 V/mil • ERop = 856 V/mil (34 kV/mm).
(6 kV/mm) AEIC CS9 (AEIC 2006) for service-proven
accessories and below 200 V/mil (8 kV/mm) for accesso- These are within the performance capability of an XLPE
ries on high-stress systems. cable. A manufacturer will normally add a 10% design
Stress in a Void at the Insulation Screen, ERI margin when calculating impulse stress to allow for vari-
The calculation of stress in an air void is performed to ations in manufacture and test conditions. Similar calcu-
assess the vulnerability of the accessory to distress occur- lations are performed for stress in the accessories.
ring at the interface with the cable insulation. The stress E-Ratio Cable Design
at the peak of the voltage waveform is to be calculated. In special applications, the minimum economical con-
ductor diameter is determined from Equation 3-51.
From the above, ERo = 131 V/mil (5.2 kV/mm). From Example applications are for conventional cables carry-
Table 3-9, εrXLPE = 2.3 and εrVOID = 1. ing low current at high transmission voltage (for exam-
⎛ε ⎞ ple, in small hydroelectric or pumped storage plants),
E RI = E Ro ⎜⎜ r XLPE ⎟⎟ (kV/mm) or (V/mil) and GIL cables carrying high current through large-
⎝ ε rVOID ⎠ diameter tubular bus bar conductors. The equation
gives the thinnest insulation and smallest cable diameter
Result: for a given insulation design stress. If a smaller conduc-
tor is selected, the insulation thickness and cable diame-
• Stress in void: ERI = 300 V/mil (11.8 kV/mm) ter have to be increased to keep the design stress within
(rms stresses). the limit. This optimum conductor radius condition is
• Peak stress is 2 greater: ERI = 424 V/mil called the “e” ratio design, or the minimum stress
(16.7 kV/mm). design, Emin. It occurs when the ratio of outer radius Ro
to inner radius Ri is equal to 2.718. Equation 3-51 per-
The void stress is twice the value at which a thin air gap mits the minimum conductor radius and nearest stan-
would exhibit partial discharge. An inadequate cable- dard nominal conductor area to be calculated.
to-accessory interface, operating at the shield stress of Example 1
131 V/mil (5.2 kV/mm), would be vulnerable to distress. The e-ratio conductor radius and size are calculated for
Example recommendations would be to select a design the following application:
of accessory with a high interfacial pressure, to specify
jointer training, and to obtain evidence of qualification • System voltage, V: 345 kV
testing. These conclusions are common to all cables and • Phase voltage, Vg: 199 kV
accessories at the 345-kV voltage level. The breakdown • Insulation type: Extruded XLPE insulation
strength of air at different pressures and gap thicknesses
is obtained from the Paschen curves (Dakin et al. 1974). • Insulation design stress, ERi: 282 V/mil (11.1 kV/mm)
Impulse Stress at Conductor Shield and Insulation Shield, ⎛R ⎞ ⎛ Vg ⎞

ERip and ERop At E min ; ⎜⎜ o ⎟⎟ = e ; R i = ⎜⎜ ⎟;
⎟
The stress at the lightning impulse BIL level is required ⎝ Ri ⎠ ⎝ E min ⎠
to check that the insulation has sufficient thickness to R o = (R i e ); t = (e − 1)R i = 1.718R i ;
satisfactorily pass a qualification test. From the above
example, ERi = 282 V/mil (11.1 kV/mm). From Table 3-1, ⎛ Vg ⎞
the 345-kV phase voltage Vg = 199 kV and the BIL Rti = 1.718 ⎜⎜ ⎟
⎟ 3-51
peak voltage Vp = 1300 kV. These are entered into Equa- ⎝ E min ⎠
3-27
Where: d u c t o r s t r e s s t o b e i n c r e a s e d f ro m 2 8 2 V / m i l
Emin = minimum conductor stress design = ERi. (11.1 kV/mm) to 350 V/mil (14 kV/mm), the minimum
e = ratio of the optimum conductor radius economic conductor size has been reduced from 1500
condition = 2.718. kcmil (800 mm2) to 900 kcmil (500 mm2) (see Tables 3-5
and 3-8). The diameter over the insulation has been
Result: reduced from 3.84 in. (97 mm) to 3.1 in. (79 mm).
• Ri = 0.71 in. (17.9 mm). Stress in a DC Cable
• Ro = 1.92 in. (48.7 mm). Depending upon the cable type, the dc design stress will
be typically 25–40 kV/mm for normal dc operation,
• t = 1.21 in. (30.8 mm).
which is significantly higher than the 12–18 kV/mm
• ERo = 104 V/mil (4.1 kV/mm). (from Equation 3-49) range for a similar ac cable. The lightning impulse
design stress is in the same range of 75–100 kV/mm for
The smallest possible radius over the insulation of this both dc and ac cables.
345kV cable is thus 1.92 in. (48.7 mm). The optimum
minimum shield radius is 0.71 in. (17.9 mm). It is Different insulant types have widely varying dc resistivi-
now necessary to calculate the size of the conductor. ties ρ, as shown in Table 3-10 (Hampton et al. 2000).
Allowing a shield thickness of 0.06 in. (1.5 mm) and a The table shows that greater than a 100 times increase in
conductor filling factor of 0.8 (from Section 3.6, we resistivity is produced in the same material when a tem-
see that a 0.8 filling factor relates to a compressed con- perature gradient of 40°C exists, and between two mate-
ductor), this gives an overall conductor diameter of rials such as EPR joint insulation and XLPE cable
(2 × (0.71 in.−0.06 in.)) = 1.30 in. (33 mm). From insulation. These differences are of great significance in
Tables 3-5 and 3-8, the nearest standard size is a the design of accessories for dc systems. The stress
1500 kcmil (800 mm2) conductor. Generally, it would be increase at an interface between two materials is calcu-
uneconomical to supply a cable with a smaller conductor lated from Equations 3-41 and 3-42 by substituting
size because the insulation radius would have to be permittivity for insulation conductivity, which is
increased to limit the stress to the design value. If this the reciprocal of resistivity, ρ. If the second layer has a
size is larger than needed to carry the current—i.e., the 102 increase in resistivity, it will experience a 102 increase
conductor does not reach 90°C—then a 1500-kcmil alu- in stress.
minum conductor with the same shield diameter would
be selected in place of a copper conductor. The 1500- In a cable insulation, dc resistivity ρ is not a constant
kcmil aluminum conductor would have an equivalent value and reduces typically 2–5 times with stress and
area to a 1000-kcmil (500-mm2) copper conductor. 200 times with temperature. The effect of stress on an
Example 2
unheated cable is to reduce the stress at the conductor
The stress at the insulation shield is comparatively low at shield and to increase it at the insulation shield, com-
104 V/mil (4.1 kV/mm), and so the possibility exists of pared to the ac capacitive distribution in Figure 3-5. At
allowing this to be increased by increasing the conductor full load, a 5 to 15°C temperature drop is present across
shield design stress from 282 V/mil (11.1 kV/mm) to the the insulation, which further reduces the stress at the
maximum conductor design stress of 350 V/mil inner shield such that it falls below that at the outer
(14 kV/mm) allowed in AEIC CS9 (AEIC 2006). The shield; this phenomenon is called “stress inversion.” In a
following dimensions are calculated from Equations 3-51 dc cable, the insulation at the outer shield is designed to
and 3-49: be of equal electrical strength and to be equally mechan-
Result: Table 3-10 Typical Values of Insulation Resistivity

• Ri = 0.57 in. (14.4 mm). Stress Resistivity (1012 Ωm)
E ρI
• Ro = 1.55 in. (39.3 mm).
Insulation (V/mil) (kV/mm) 20°C 80°C
• t = 0.98 in. (24.8 mm). 250 10 300 3
Kraft paper
• ERo = 128 V/mil (5.1 kV/mm). 510 20 300 4
510 20 5000 300
LPP
The cable design has remained acceptable, because the 760 30 5000 300
calculated stress at the insulation shield of 128 V/mil EPR 130–180 5–7 40–500 3–5
(5.1 kV/mm) is less than the 150 V/mil (6 kV/mm) limit XLPE AC 130–180 5–7 200 1
given in AEIC CS9 (AEIC 2006). By allowing the con- XLPE DC 130–180 5–7 8000 50
3-28
ically protected as the inner shield. DC cables are con-

1
sidered in greater detail in Chapter 7. I C = Vg = j. Vg 2π f C (A/ft)
XC
Insulation Capacitance
MVAr = V g I C = V g 2π f C
Capacitance per unit length C is an important circuit 2
parameter in both ac and dc systems. In an ac system, (MVAr/ft per phase) 3-54

the capacitance determines the magnitude of the total
charging current demanded by the cable length, which is Where:
added in quadrature to the load current. In long cable C = Capacitance (F/ft).
circuits, the capacitance may have to be reduced, or εo = Permittivity of free space: 2.70 × 10-12
compensated, to carry the load current without over- (F/ft) (8.854 × 10-12 (F/m)).
heating the cable. The cable manufacturer may be εr = Relative permittivity of insulation, see
required to guarantee that a specified capacitance or Table 3-9 (ratio).
charging current will not be exceeded—for example, in a XC = Capacitive reactance at frequency f (Ω-ft).
predominantly capacitive network in which inductive f = Power frequency (cycles/s).
compensation is essential, and in a long-length subsea IC = Charging current (A/ft).
link in which excess cable capacitance will limit the j = The operator, − 1, giving a phase advance-
power that can be traded. In a dc system, capacitance ment to Ic of 90°(dimensionless).
determines the insulation time constant RC, the voltage MVAr= Reactive power per phase (MVAr/ft per
rise during a polarity reversal and the energy stored in phase).
the cable ( E = CV 2 2).
Relative permittivity is also named specific inductive
Capacitance is the measure of the cable insulation’s abil- capacity (SIC) or dielectric constant.
ity to store charge (and hence flow of ac capacitive cur- Example of Calculation of Capacitance, Capacitive
rent) per unit voltage, as shown in Equation 3-44. The Reactance, and MVAr
capacitance for an elemental parallel plate capacitor is The capacitance, reactance, and current are first calcu-
given in Equation 3-45. In the discrete elemental annu- lated per foot (meter) and are then calculated for a
lar rings forming the cable insulation, the capacitance is 10-mile (16-km) circuit length. The following cable
proportional to the circumference, which is small at the dimensions are taken:
conductor shield and large at the insulation shield. The
charge Q remains constant for each elemental capacitor. • System voltage, V: 345 kV at 60Hz
Thus the voltage drop at the conductor is large and, at • Phase voltage, Vg: 199 kV
the insulation shield, small, as shown in Equation 3-38. • Conductor size: 2500 kcmil (1250 mm2)
The total capacitance is the sum of the elemental capac-
itances shown in Equation 3-43. The capacitance for • Radius over conductor shield, Ri: 0.92 in. (23.4 mm)
single permittivity, concentric cable insulation is given • Radius over insulation shield, Ro: 1.99 in. (50.5 mm)
by Equation 3-52. The equations can be used for both • Insulation type: Extruded XLPE insulation
U.S. and metric units without change, providing the
appropriate unit is taken for the absolute permittivity of • Relative permittivity, εr: 2.3
free space εo, the values of which are given below. Thus • Circuit length: 10 miles (16 km)
the unit of C is (F/ft) if εo is (F/ft), and the unit for C
becomes (F/m) if εo is (F/m) and so on for the parame- The dimensions are entered into Equations 3-52, 3-53,
ters XC and IC. and 3-54. The single-phase MVAr = Vg I.
2π ε o ε r
C= (F/ft) 3-52 Result:
⎛R ⎞ C = 50.5 × 10-12 F/ft (165 × 10-12 F/m).
ln⎜⎜ o ⎟⎟ •
⎝ Ri ⎠ • XC = 52.4 × 106 Ω-ft (16.0 × 106 Ω-m).
• IC = 3.79 × 10-6 kA/ ft (12.4 ×10-6 kA/m).
1
XC = (Ω-ft) 3-53 • MVAr = 0.754 × 10-3 MVAr/ft per phase
2π f C (2.48 × 10-3 MVAr/m per phase).
• C for 10 miles (16 km) = 2.67 μF.
• X for 10 miles (16 km) = 993 Ω.
3-29
• I for 10 miles (16 km) = 0.2 kA. shield radius Ri , then the heat loss is directly propor-
• MVAr for 10 miles (16 km) = 39.8 MVAr per phase. tional to phase voltage Vg. A 345-kV cable would have
five times the loss of a 69-kV cable. This is not the case,
as the design stress is increased at higher transmission
The charging current and reactive power are of signifi-
voltages. Equation 3-56 shows the pessimistic case when
cant magnitude on the 10-mile circuit. If required, they
voltage is increased on the same size of cable (i.e., con-
could be reduced, by increasing the insulation thickness.
stant insulation thickness and capacitance). The heat
Insulation Losses loss is proportional to the square of phase voltage,
AC voltage generates heat in the insulation as a conse- V g 2 —i.e., the heat loss in a 345-kV cable would be
quence of dipole rotation. The heat loss is represented 25 times higher than a 69-kV cable. In practice, the
mathematically by the inclusion of a resistance in parallel dielectric loss is proportional to VgN, where N is between
with the capacitance. The phasor diagram, Figure 3-12, 1 and 2, depending upon the cable dimensions. The
shows that the cable charging current IC is not in exact equations can be used for both U.S. and metric units
quadrature with the voltage V. The charging current IC is without change, providing the appropriate units are
at angle ϕ to voltage V and at angle δ to the current of a taken as given below:
perfect capacitor, iC. The ‘loss current’ is IC cos ϕ, which
WD = Vg E R1 (2π) f εo ε r π
2
is in-phase with the applied voltage V. The angle δ (radi-
ans) is a measure of the quality of the insulation—the (MW/ft) or (MW/m) 3-55
smaller the angle, the higher the quality—and is called
WD = Vg 2π f C δ
2
the dielectric loss angle (DLA). DLA is routinely mea- (MW/ft) or (MW/m) 3-56
sured in the factory acceptance test on each reel of cable
Where (in U.S. units):
by such instruments as a transformer ratio arm bridge
WD = Dielectric loss (MW/ft).
and a Schering bridge. The loss angle in transmission
Vg = Phase voltage (kV).
class cables is very small, and so,
E = Insulation stress at the conductor shield
δ ≅ tan δ ≅ sin δ ≅ cos ϕ (V/mil).
R = Radius of conductor shield (in.).
since δ = (90°−ϕ) and δ < 0.5° (0.009 radians) for most f = Power frequency (cycles/s).
cables DLA is also named ‘tan delta’, ‘cos ϕ’, and εo = Permittivity of free space: 2.70 × 10-12
‘power factor’. DLA may also be stated as a percent- (F/ft).
age—e.g., 0.23% is a DLA of 0.0023. The loss current is εr = Relative permittivity of insulation, see
commonly written as ‘IC δ’ or ‘IC cos ϕ’. The heat loss, Table 3-9 (ratio).
WD, is then ‘V IC δ’ or ‘V IC cos ϕ’. C = Capacitance (F/ft).
δ = Insulation dielectric loss angle, see
The equation for heat loss, WD, can be arranged in two Table 3-11 (radians).
identical ways. Equation 3-55 is the optimistic case,
showing that if the insulation is designed always to have Where (in metric units):
a constant design stress E and common conductor WD = Dielectric loss (MW/m).
Vg = Phase voltage (kV).
E = Insulation stress at the conductor shield
(kV/mm).
R = Radius of conductor shield (mm).
f = Power frequency (cycles/s).
εo = Permittivity of free space: 8.854 x10-12
(F/m).
εr = Relative permittivity of insulation, see
Table 3-9 (ratio).
C = Capacitance (F/m).
δ = Insulation dielectric loss angle, see
Table 3-11 (radians).
Equation 3-55 shows that, for a given transmission

voltage, frequency, and cable dimensions, the loss is pro-
Figure 3-12 Dielectric loss angle δ portional to ‘εr δ’, called DLF (dielectric loss factor).
of insulation (courtesy of CCI).
3-30
Table 3-11 gives the ranges of ε r, δ, and DLF values ρI ⎛ DO ⎞

for each type of insulation. The lower values were mea- RI = ln⎜ ⎟
2π ⎜⎝ D i ⎟⎠
(Ωm) 3-57
sured at operating temperature on short cable lengths of
30–60 ft (10–20 m) of one particular cable size and
design stress and on insulation from one grade and ρ I = ρθ=0 exp(− α θ) (Ωm) 3-58
ρ I = ρ E=0 exp(− β E ) (Ωm)

batch of material. The DLF increases with temperature
for each of the insulations shown and is also stress 3-59
dependent. When performing ampacity calculations, it Where:
is good practice to select the higher DLF value, based RI = Resistance of the insulation per unit
on values from IEC 60287 (IEC 1994). This will reduce length (Ωm).
the risk of dielectric thermal runaway and failure occur- ρI = Resistivity of the insulation (Ωm).
ring at a “hot spot” in service. The percentage ρθ=0 = Base resistivity extrapolated to zero tem-
heat losses given in the table are based on the higher perature (Ωm).
DLF values. ρE=0 = Base resistivity extrapolated to zero
stress (Ωm).
Dielectric loss in fluid-impregnated kraft paper cables is DO = Outer diameter of the insulation (mm).
comparatively high. The availability of the low-loss, Di = Inner diameter of the insulation (mm).
high-operating-temperature dielectrics LPP and XLPE E = Electric stress (kV/mm).
has made a major impact on increasing cable thermal θ = Temperature (°C).
efficiency and ampacity. α = Temperature coefficient of resistance
Insulation Resistance (per °C).
The insulation resistance is omitted from the circuit dia- β = Stress coefficient of resistance (mm/kV).
gram of the ac cable, as being much greater (≥ 106) than
the capacitive shunt reactance. The electrical resistance of cable insulation is difficult to
measure because the current is low and the resistance
Insulation resistance is an important parameter in a dc high. The insulation current can be calculated from
cable because it is needed in the calculation of design I = Vg/RI, where RI is obtained from Equation 3-57. For
stress and insulation time constant. Equation 3-57 gives example, taking the dimensions of the example ac
the electrical resistance of an annular ring of dc insula- 345-kV 2500-kcmil XLPE cable, a resistance of 200 ×
tion of outer diameter DO and inner diameter Di for a 106 MΩ. and applying a 100-kV dc system voltage, the
constant value of resistivity ρI. Expressions for the tem- resistance is greater than 80 × 106 MΩ.ft (24 × 106 MΩ.m),
perature dependence of resistivity and stress dependence and the current is less than 0.0012 μA/ft (0.004 μA/m).
are given in Equations 3-58 and 3-59 Typical values of After energizing the dc cable, a long time must be allowed
resistivity ρI are given in Table 3-10. for the transient charging current to reduce to zero value;
this occurs after greater than ten time constants (RC).
The 345-kV XLPE cable has a time constant of greater
Table 3-11 Typical Range of Values of εr, DLA, and DLF for
than 1–10 hours—thus a period of greater than 0.4–4
Various Cable Insulations days is required before measurement of steady-state resis-
tive current can be attempted. The resistance reduces
Relative
Permittivity DLA DLF % Heat
with temperature and stress, as given in Equations 3-58
Insulation Type εr δ x 1000 Losses and 3-59, such that measurement of apparent resistance
Ethylene becomes possible at higher insulation temperatures and
100
propylene rubber 2.5–3.5 0.0035 8.8–12.3 voltages.
(reference)
(EPR)
Fluid-impregnated 0.0020–
kraft paper
3.4–3.6
0.0024
6.8–8.6 70 Extruded insulation and particularly XLPE is suscepti-
Impregnated
ble to the space-charge phenomenon in which halos of
laminated charged molecules accumulate with time and stress in
0.0006–
polypropylene 2.6–2.8 1.6–3.9 32 the proximity of the conductor and insulation shields,
0.0014
paper laminate
(LPP)
thereby distorting the stress distribution and the calcu-
High-density
lation of apparent resistivity.
0.0001–
polyethylene 2.3–2.5 0.23–2.5 20
(HDPE)
0.001 Insulation Design Life
Crosslinked The design life of the cable system is the nominal useful
0.0001–
polyethylene 2.3–2.5
0.001
0.23–2.5 20 life under normal voltage and temperature operation.
(XLPE) The life is typically given as 40 years, which originated
3-31
from the accounting period of a person’s working life. • A long XLPE cable system containing joints and
Shorter periods of 25 and 30 years may also be speci- subjected to load cycling may give a value of n > 12.4
fied. The factors that affect insulation aging are: to 17.
• Cumulative time at operating voltage and numbers • An XLPE cable or accessory containing air voids
of switching surges, lightning strikes, and through- may give an n value of 9.
fault transients. • A small sample of XLPE cable containing deliber-
• Magnitude and duration of “standing” over-voltages ately punctured shields may give an n value of 4.2.
and harmonics in the system.
• Loss of pressure in the dielectric fluid or gas in a fluid AEIC CS9-06 specifies a shipping test on a 345-kV cable
or gas-filled cable. of 1 hour at 2.0Vg, based on limiting the maximum test
stress to 700 V/mil (28 kV/mm). Equation 3-60 calculates
• The presence of defects or damage that weaken the that by taking the ratio of 40 years to 1 hour and the
insulation or increase the stress—e.g., in bending ratio of the test stress to service stress as 2:1, an n value
damage to tapes and shields in HPFF and SCFF of 18.4 is derived. Lower values of n, representative of
cables; shield protrusions, particulate contamination insulation defects can be inserted into Equation 3-60; an
and voids in extruded XLPE and EPR cables; and n of 17 reduces the comparative life to 15 years life, and
particulate contamination in GIL cables. an n value of 13 reduces the comparative life to 1 year.
• Duration and magnitude of conductor temperatures
greater than the published normal operating temper- En t = k ((V/mil)n. hour) 3-60
ature (e.g., > 90°C for extruded XLPE cables). ⎛ 1⎞ ⎛1⎞

• Contamination of the insulation by foreign materials, log10 E = ⎜ − ⎟ log10 t + ⎜ ⎟ log10 k
⎝ n⎠ ⎝n⎠
usually following puncture or deterioration of the
(dimensionless) 3-61
cable jacket—e.g., moisture entering a taped
paper/LPP cable or extruded XLPE cable, air enter- Where:
ing the dielectric fluid in a fluid-filled cable, precipita- E = Reference stress at conductor shield
tion of particulate contamination in a fluid-filled (V/mil).
cable and a low resistivity contaminant entering a t = Time (hour).
fluid-filled cable, such as a lubricating oil or bitumi- n = Life exponent (dimensionless).
nous compound. k = A constant ((V/mil)n hour).
Example Calculation of Insulation Life
The effect of electrical aging on full-sized cables with
extruded or taped insulation, over a limited range of ac Example 1
stresses, may be assessed by the empirical “inverse A test loop of cable and accessories withstands t1 = 10
power law,” as shown in Equation 3-60 (Attwood et al. days of load cycling at 1.73 Vg. What will be the equiva-
1998). The relationship is based on elevated voltage lent service life t2 at normal voltage Vg? The life expo-
breakdown tests with endurance times to failure in the nent chosen for a cable system undergoing load cycles is
range of 1 minute to 5 years. Equation 3-61 is plotted as n = 13. The values are entered into Equation 3-62.
a straight line of log 10 time to failure against log 10 • Service stress: E2 (V/mil).
breakdown stress to determine the slope −n, which is
called the life exponent. The simple model pessimisti- • Test voltage: 1.73 Vg (kV).
cally assumes that the value of n does not change with • Test stress, E1: 1.73 E2 (V/mil).
time—i.e., that a threshold stress does not exist under • Life exponent, n: 13 (dimensionless).
which further aging is absent. Care is needed in the
selection of n as this depends upon the mechanism of • Time to failure, t1: 240 (hours).
aging and the magnitude of the design stress (Dissado • Equivalent service life, t2: solve for this
and Fothergill 1992). n
⎛E ⎞
Examples of n values are taken from Smith et al. (2003) t 2 = t1 ⎜⎜ 1 ⎟⎟ 3-62
and from AEIC CS9-06 (AEIC 2006): ⎝ E2 ⎠
• XLPE cable subjected to breakdown testing at con- Result: Equivalent time, t2 = 0.3 × 106 hours (34 years).
stant temperature in short 33-ft (10-m) cable lengths
with no joints may give a high value of n = 25.
3-32
Example 2 • Defect stress raiser: 2 (ratio).

A cable passes a 1-hour factory test. Routine material
• Design stress on conductor shield: 282 V/mil.
sampling shows defects to be present on the conductor
shield, with a geometric stress raiser of × 2. A cable sam- • Life exponent of defect, n: 9 (dimensionless).
ple is to be submitted to a special HV breakdown test.
What test stress E1 on the defect is required to produce Result:
failure in a time t2 = 1 hour? What is the equivalent • Stress on defect at service voltage:
stress and voltage needed on the cable to achieve this? E2 = 2 × 282 V/mil: 564 V/mil (22.2kV/mm).
T h e d e s i g n s t r e s s o f t h e c ab l e i s 2 8 2 V / m i l • Time t2 to produce failure at E2:

(11.1 kV/mm). From Smith et al. (2003), a typical 76 × 103 hours (8.7 years).
breakdown stress of E1 = 2540 V/mil (100 kV/mm) on a
defect produces cable failure in t1 = 0.1 hour. The life This example shows that the factory routine visual
exponent chosen for voids and defects is n = 9. The val- examination of the cable was more important in detect-
ues are entered into Equation 3-62. ing the defect than the factory acceptance test, which
would have permitted the defect to enter service and fail
• Stress on defect for failure in 1 hour, E2: solve for this. in approximately 9 years.
• Test voltage to produce failure in 1 hour: solve for this.
Insulation Thermal Aging
• Defect breakdown stress, E1: 2540 (V/mil). Cable insulations are designed to operate for a substan-
• Time to failure at 2540 V/mil, t1: 0.1 (hour). tial part of their service life close to, or at, the design
temperatures in Table 3-2. The life of kraft paper, insu-
• Life exponent, n: 9 (dimensionless).
lating fluid, and extruded insulation is shortened at
increased temperature. At elevated temperatures, paper
Result:
tapes discolor, become brittle, and prospectively frac-
• Stress on defect to produce failure in 1 hour: ture. For this reason, the time of the factory drying pro-
E2 = 1967 V/mil (77 kV/mm). cess is limited at maximum temperature. Antioxidants
• Stress on conductor shield = 1/2 × 1967 = 983 V/mil are provided in polymeric insulation to protect the insu-
(39 kV/mm). lation from deterioration in the presence of air and heat
at the elevated temperature in the crosslinking process
• Test voltage, per unit = ratio of conductor shield and at the operating temperatures through service life.
stresses = (983/282) = 3.5 Vg. At elevated temperature in the presence of air, XLPE
• Test voltage = 3.5 × 199 kV: 697 kV. insulation eventually discolors and becomes brittle.
The calculation shows that a voltage significantly above The effect of short-term operation at greater than the
the capability of most factory HV test sets would be design temperature is estimated from the Arrhenius
required. The test voltage could be reduced and the time relationship in Equation 3-63. The aging rate constant k
on test increased, or the test could be performed on a varies directly with temperature and follows the Arrhe-
1 to 3 MV ac test set in an HV laboratory. When test nius aging law, shown in Equation 3-64. For example,
failure is achieved, the new time and breakdown stress the mechanical and electrical life of paper cable is con-
can be added to the calculation to improve accuracy for ventionally taken to be 40 years at 85–90°C. Using the
future predictions. “8°C Rule” originally from paper cable and transformer
insulation studies, at elevated temperatures up to 105°C,
Example 3
the insulation will experience an accelerated aging of
If the cable from Example 2 entered service, what is its
50% reduction in mechanical strength for each 8°C rise
life prognosis, t2? The life exponent chosen for voids and
in temperature. A closely similar aging relationship is
defects is n = 9. The values are entered into Equa-
taken for extruded insulation.
tion 3-62.
• Stress on defect in service, E2: solve for this. S(t1 ) = S(o) exp(- kt1 ) 3-63
• Service life of defect, t2: solve for this. ⎛ B⎞

• Test voltage to produce failure: 697 kV. k = k T = A exp⎜ − ⎟ 3-64
⎝ T⎠
• Time to failure at 697 kV, t1: 1.0 (hour).
Where:
• Defect breakdown stress, E1: 1967 (V/mil). S(o) = Reference value of aging parameter at time
t = 0 (appropriate dimension).
3-33
S(t1) = Value of aging parameter at time t = t1 ends of the span by overheating and at the center of the
(appropriate dimension). span from erosion by sparking between the outer layers
k = Aging rate constant (per hour). due to longitudinal voltage rise.
kT = Aging rate constant at temperature T
Impulse Surge Propagation in AC and DC Cables
(per hour).
An instantaneous radial voltage appears across insu-
T = Absolute temperature (K).
lated layers during the passage of switching surges and
A = Constant determined experimentally
lightning impulses. The voltage is proportional to the
(per hour).
ratio of the surge impedances of each layer and is not
B = Constant determined experimentally (K).
reduced by the longitudinal conductivity of the layers.
To prevent spark erosion of the insulation shield, a high
3.6.4 Insulation Shield
conductivity shield conductor is applied in intimate con-
The insulation shield performs the same function as the tact either directly with the insulation shield or with con-
conductor shield. Additionally the shield must be of ducting bedding/cushioning tapes applied in between.
robust construction and be protected from indention
and thermal damage by the outer cable layers occurring 3.6.5 Ground Return Conductor
during manufacture and in service. Binding and bedding
The function of the ground return conductor is to
tapes are applied in the case of taped insulation and
carry fault current during through-faults, return fault
cushioning tapes in the case of extruded insulation. It is
current when the cable is punctured, circulating currents
essential that the protective tapes are radially conduct-
in ac operation, and capacitive current collected by the
ing such that the insulation capacitive currents can pass
insulation shield conductor.
radially outwards to the shield conductor.
Conductor Area for Short-circuit Current
Shield Conductor
The cross-sectional area A of the ground return conduc-
The function of the shield conductor is to back up the
tor is determined by the ampacity calculation, which
insulation shields with a high conductivity layer during
ensures that the limiting temperatures given in Table 3-2
the normal events of ac voltage operation and impulse
are not exceeded during the short time in which short-
surge propagation.
circuit current flows. The principles of the short-circuit
rating method are described in Section 3.5.2, and the
Separate shield conductors are provided in those cables
area is calculated by Equation 3-29. The resistivity of
that either do not have an integrated ground return con-
the conductor material ρθ, is temperature dependent
ductor to collect capacitive currents, as in an HPFF
and is calculated in Equation 3-36 using reference 20°C
cable, or have an inner polymeric moisture barrier
values from Table 3-3.
sheath, as in some XLPE cables and in some HPGF
cables. In most SCFF and XLPE cable types, the shield Cable System Fault
conductor function is performed by the metallic sheath Some utilities require that the cable and accessories be
or the integrated ground return conductor. contained within a conducting envelope that in the
event of the cable being punctured (spiked) will return
Normal Operation in an AC Cable
the current for the duration of the fault. This is to assist
The shield conductor in SCFF and most extruded in the efficient operation of the fault protection system.
cables backs up the insulation shield by collecting the The radial dimension and area of some ground return
capacitive current from the insulation, as shown in Fig- conductors have to be increased to prevent local melting
ure 3-4, and returning it to the bonding position at the by the radial power arc following an accidental puncture
adjacent joints. The current is then returned by the by a third party, or following a puncture to deliberately
ground return conductor to the ground positions at the ground the cables and make them safe prior to mainte-
terminations and at certain straight joints. The HPFF nance work.
cable system is an exception to this because the skid
wires of the three adjacent cables contact each other, Circulating Currents
and the return currents cancel, each being 120° phase In a solidly grounded, ac cable system, the magnitude of
displaced. The capacitive current is of significant magni- the circulating currents can be 40% of the rated power
tude—for example, 7.5 A for a 2000 ft (600 m) span current, producing a significant temperature rise in the
length of the example XLPE 345-kV 2500-kcmil integrated ground return conductor. In ac GIL cables
(1250-mm2) cable. In an XLPE cable, the extruded semi- and in HTS cables, the circulating current is the same
conducting insulation shields are not designed to carry magnitude as the rated power current. The magnitude
high longitudinal current. The consequence of doing so of the circulating current is proportional to the load
is prospective failure of the primary insulation at the current. Unlike the fault current, the circulating current
3-34
is present all the time that the cable carries load. It is

important that the bonding connections between the
accessories and the cable are designed to take the circu-
lating current without overheating and burning out. The
calculation of the heat produced by circulating current
is included in the ampacity calculation for the cable in
Chapter 11.
Radial Moisture Barrier
In some designs of XLPE cable, a metallic sheath is not
employed. In such cables, shown in Figure 3-3, the func-
tion of radial moisture barrier is performed by a thin, Figure 3-13 Principal stresses in a metallic sheath or pipe
overlapped, metallic foil. The foil is preferably bonded to (courtesy of CCI).
the jacket to provide strength during bending and to
improve corrosion resistance. The foil is normally applied
over a cushioning or water-blocking tape; it is important distance from the neutral axis and so is increased in
that this tape has sufficient radial conductivity to prevent large diameter cables and changes from a compressive
sparking during transient voltage propagation. stress on the inner surface of the bent sheath to a tensile
stress on the outer surface. Thermomechanical stress,
Metallic Sheath shown in Equation 3-70, is highest in rigid thermome-
The metallic sheath is a watertight barrier, which also chanical systems because the sheaths are constrained
combines the roles of a conductor for insulation capaci- from thermal expansion, for example, by direct burial,
tive currents, circulating currents in a solidly grounded or by close-cleating. The magnitudes of bending stress
sheath, and the short-circuit current. The sheath with- and axial stress are proportional to the effective elastic
stands the internal pressure in fluid- and gas-filled bending modulus E B and axial modulus E A of the
cables and the external pressure in a subsea cable. sheath material and construction, being lower in lead
Extruded sheaths are typically 80–160 mils (2–4 mm) alloy sheaths and in corrugated sheaths, and higher in
thick. Welded aluminum or copper sheaths are typically plain sheaths and in pipes.
less than 80 mils (2 mm) thick, and a welded stainless
steel sheath is typically less than 40 mils (1 mm) thick. The following equations are applicable to both metric
An aluminum sheath has the capability of withstanding and U.S. units. Metric units are shown. The insertion of
high internal and external pressures. A lead alloy sheath dimensions in (in.), elastic modulii E in (psi), moment
has a low elastic limit and acts as a semiflexible dia- M in (lb-in.) pressure p in (psi), and second moment of
phragm that progressively creeps and increases in diam- area I in (in.4) produces mechanical stress σ in (psi).
eter. Metal reinforcing tapes are applied over lead alloy
sheaths to contain the internal pressure of fluid-filled
and gas-filled cables. In a subsea cable in deep water, σ=
1
2
[
(σ1 − σ2 )2 + (σ2 − σ3 )2 + (σ3 − σ1 )2 ]
the inner core is formed into a fully filled cylindrical
shape to support the lead sheath in withstanding the (N/m2) 3-65
external pressure.
σ1 = P (N/m2) 3-66
The magnitude of the combined mechanical stress σ, P.D

shown in Equation 3-65, in the sheath or reinforcing σ2 = (N/m2) 3-67
tapes must be below the elastic limit. More usually a
2t
proof stress is specified as sheathing metals such as lead σ 3 = σ 3 '+ σ 3 " (N/m2) 3-68
alloy, aluminum, and copper and reinforcing tapes such
as stainless steel and tin-bronze have imprecise elastic ⎛ y.M ⎞ ⎛ y.E B ⎞
limits. Figure 3-13 shows the principal stresses that
σ3 ' = ± ⎜ ⎟ = ±⎜ ⎟ (N/m2) 3-69
⎝ I ⎠ ⎝ R ⎠
con tribute to σ: the radia l stre ss, σ 1 , show n in
Equation 3-66, and circumferential hoop stress σ 2 , σ 3 " = ± ( E A .α .Δθ ) (N/m2) 3-70
shown in Equation 3-67, both formed by the pressure
difference, and the longitudinal stress σ3 Equation 3-68, Where:
formed by the bending stress σ 3 ′, and the thermome- σ = Combined stress in the metallic sheath or
chanical stress σ3″. The magnitude of the bending stress, reinforcing tapes (N/m2).
shown in Equation 3-69, is directly proportion to the σ1 = Radial principal stress across sheath
(N/m2).
3-35
σ2 = Circumferential principal stress in sheath peratures, as shown in Table 3-2. The grade also has a
(N/m2). high shore hardness, which gives a degree of protection
σ3 = Longitudinal principal stress in sheath against termite attack. The resistance to water of PE is
(N/m2). excellent; however, some moisture diffuses through the
σ3′ = Component of longitudinal stress due to jacket, risking electrolytic corrosion of the metallic
bending (N/m2). sheath or foil moisture barrier. Improved performance is
σ3″ = Component of longitudinal stress due to achieved by coating a metallic sheath with a suitable
thermomechanical force (N/m2). bituminous compound, and ensuring that the jacket is
P = Pressure difference across sheath (N/m2). fully bonded to a metallic foil, with no air gaps being
D = Diameter of sheath (m). present. PE contains a small percentage of carbon black
t = radial thickness of sheath (m). to give resistance to UV light; however, the black finish
y = Radial distance from neutral axis to posi- also increases temperature rise by absorption of solar
tion in sheath (m). heat. Disadvantages of PE are:
M = Bending moment applied to sheath (Nm). • The stiffness of the cable is increased, especially at
I = Second moment of area of sheath (m4).
low temperature.
EA = Effective modulus in axial loading of
sheath (N/m2). • The locked-in thermal contraction strain from manu-
EB = Effective modulus in bending of sheath facture applies significant contraction forces onto
(N/m2). bonding connections and sheath seals at accessories.
R = Bend radius of sheath to neutral axis (m). • Combustion is sustained in the presence of air, such
α = Coefficient of thermal expansion of sheath that it is unwise to repeatedly apply test voltages to
(per °C). locate a fault on a PE jacket in air.
Δθ = Temperature rise of sheath (K).
Extruded polyvinyl chloride (PVC) jackets are extruded
3.6.6 Jacket from special grades of filled and plasticized PVC, which
The extruded jacket is internationally recognized as generally render them softer than PE and largely elimi-
being the most important layer in determining the lon- nates shrink-back, albeit they are more susceptible to
gevity of a cable, because it provides the main barrier to penetration during installation in hot climates. PVC
corrosion and to water ingress. However, the jacket is installed in waterlogged conditions also exhibits a lower
comparatively thin, for example, 0.13–0.3 in. (3.3– dc resistance, which may increase difficulties in fault
8 mm), is subjected to significant abrasion during instal- location. Standard grades of PVC produce dense black
lation, is softened by solar heating, aged by UV light, acidic smoke in a fire, which is unacceptable in a build-
and is vulnerable to attack by termites and vermin. In ing or in the proximity of electronic equipment. Low
some indoor applications, it is also required to have fire- halogen grades of PVC, which have reduced smoke and
resistant properties, either in not spreading a fire or in fire-spread properties, are available for application in
not liberating corrosive gasses or dense smoke. Care is buildings. PVC has a higher permeability to moisture
required in selecting a fire performance grade for heavy than PE, making it more important to apply a protec-
transmission class cables, because the mechanical and tive coating on internal metallic sheaths. Low halogen
water-resistant properties may be compromised. grades are generally less robust and have reduced resis-
tance to moisture when applied to large transmission
Extruded polyethylene (PE) has good all-round proper- class cables and would not be first choice for direct
ties for buried and water-immersed situations. The pre- burial on a specially bonded system.
ferred grades of PE have a density value midway
between medium- and high-density polyethylene Resistance of a jacket to termite attack is gained by
(MDPE and HDPE), with a typical melting point of selecting a high-density grade of PE with a smooth fin-
115°C. Linear low-density grades (LLDPE) are now ish, the inclusion of an environmentally acceptable
available with similar high-temperature properties. insecticide, or the application of an overall layer of hard
Higher-density grades of HDPE are available with melt- nylon in a co-extrusion process.
ing points up to 135°C; however, these have increased
vulnerability to stress cracking under conditions of Specially bonded cables and accessories are at greater
bending, abrasion, notching, and exposure to sunlight. risk of accelerated electrolytic corrosion, because of the
The MDPE/HDPE grade is suitable for installation at presence of a 60-Hz induced voltage to ground. This
higher ambient temperatures than PVC and LDPE and voltage difference is not present in a solidly bonded
operation at higher short-term emergency loading tem- cable system. It is essential that the jacket has excellent
3-36
insulating properties and that the system is subjected to DO = Diameter over outer turn of cable (m).
routine maintenance tests, typically comprising a dc DH = Diameter of hub (m).
withstand voltage of 5 kV dc for 1 minute. To facilitate k = Factor to give minimum hub diameter for
voltage testing in dry environments, it is good practice a given cable construction (value).
to apply over the jacket a semiconducting graphite coat-
ing, or a thin bonded layer of co-extruded semiconduct- 3.7 CABLE ELECTRICAL PARAMETERS
ing compound.
For short-length circuits, the cable is represented by the
lumped parameters shown in Figure 3-14. RAC is the
3.6.7 Cable Length
total conductor resistance. L comprises the internal and
A key objective in developing higher insulation design self-impedance of the conductor, and the mutual imped-
stresses is the need to transport economic lengths of ances linked with the parallel phase cables. Gi is the
cable and so reduce the number of joints. Equation 3-71 insulation conductance, this being the reciprocal of
shows that, for a given reel, the length of cable, L, is resistance. Ci is the insulation capacitance. For trans-
approximately inversely proportional to the cable diame- mission-class ac cables, the insulation conductance Gi is
ter DC. The size of the reel and hence the cable length are neglected as being very low.
limited by the specified minimum cable-bending radius
on the hub, the road bridge clearance above the outer The lumped parameters are applicable to both ac and to
diameter of the reel, the road highway restriction on reel dc cable systems, depending upon the time-varying con-
width or diameter, and the department of transporta- tent of voltage and current. In an ac system, the sinusoi-
tion’s limit on weight. Typical minimum hub diameters dal variation is at power frequency and produces the
for self-contained cables are 20–24 × DC. The cable is inductive reactance XL and the capacitive susceptance
wound onto the reel in the factory and is then subjected Bi, Equations 3-80 and 3-84, which are proportional to
to the factory acceptance HV test, which requires that frequency, and the conductor ac resistance RAC, which is
both ends of the cable be unwound 16–33 ft (5–10 m) to proportional to a power of frequency.
permit HV test connections to be assembled. This length
is made accessible from the inner cable end by the inclu- In a dc system in steady-state operation, each cable
sion of a helical ramp built adjacent to an internal face is represented by its series dc resistance. However, four
of the reel flange. Equation 3-71 does not include length circuit parameters are required to characterize time-
reductions for ramp width, packing density, or lay angle. varying behavior:
The typical maximum dimensions of a reel carrying
transmission-class cable are W = 8 ft (2.5 m), Do = 15 ft • Voltages and currents produced by the harmonic
(4.7 m), and k = 20. The equivalent maximum length content of ripple from the ac/dc converter station.
for a 6-in. (150-mm) diameter cable with the reel axis • Circuit charging and discharging times.
crosswise to the road, ignoring weight restrictions, is
3142 ft (958 m). In special applications, the length is
increased by turning the reel axis parallel with the road
to increase W.
The following equations are applicable to both metric

and U.S. units. Metric units are shown. The insertion of
dimensions in (ft) produces cable length L in (ft).
2
⎛ ΔR ⎞ ⎛ k D C ⎞
L = π W⎜⎜ ⎟⎟ ⎜ + 1⎟
⎝ D C ⎠ ⎝ ΔR ⎠
(m) 3-71
ΔR = 0.5(D O − D H ) (m) 3-72
D H = k DC (m) 3-73
Where:
L = Length of cable on reel (m).
W = Internal width of reel (m).
ΔR = Radial depth of cable turns on hub (m). Figure 3-14 Cable lumped parameters and equivalent two-
DC = Diameter of cable (m). port network (courtesy of CCI).
3-37
• Fault transients. in Equation 3-80. This is a vector addition with the sca-
• Inductive and capacitive energies stored in the system. lar magnitude given in Equation 3-81 and phase differ-
ence ϕ as given in Equation 3-82.
For medium-length circuits, the cable is modeled as a ‘T’ Z = R AC + j X L (Ω) 3-79
circuit in which G i and C i are connected at the mid-
point with ‘R AC and L’ being split into ‘0.5R AC and X L = 2πfL (Ω) 3-80
0.5L’ on either side, or as a ‘PI’ circuit in which ‘Gi and
C i ’ are split into ‘0.5G i and 0.5C i ’ to provide the
two legs, with ‘RAC and L’ bridging them.
Z= (R 2
AC + XL
2
) (Ω) 3-81
⎛ X ⎞
For long circuits, the cable is either divided into many
ϕ = tan −1 ⎜⎜ L ⎟⎟ (degrees) 3-82
distributed lumped sections or expressed as functions of
⎝ R AC ⎠
the propagation constant, γ, and the characteristic
impedance, ZC, as given in Equation 3-74. The equations Where:
for voltages and currents are in hyperbolic functions Z = Series impedance, circuit length (Ω).
(Nasar 1990). Series impedance Z and shunt admittance RAC = AC conductor resistance (Ω).
Y are obtained from Equations 3-79 and 3-83. XL = Inductive reactance (Ω).
ϕ = Phase angle between XL and RAC
(degrees).
γ= (ratiom))
Z Y (per ZC =
Z
(Ω) 3-74 f = Power frequency (Hz).
Y L = Cable inductance (H).
Where:
γ = Cable propagation constant (per meter). Parallel Cable Admittance Y
ZC = Cable characteristic impedance (Ω). Admittance Y is the reciprocal of impedance Z and is a
Z = Cable unit series impedance (Ω/m). mathematically more convenient way of expressing
Y = Cable unit parallel admittance (mho/m). shunt connected components. The insulation admit-
tance Y, shown in Equation 3-83, comprises the vector
3.7.1 Cable as a Two-Port Network addition of conductance Gi (reciprocal of the insulation
resistance, RAC) and the insulation susceptance Bi,
The circuit is represented as a “black-box” by its four
shown in Equation 3-84, (reciprocal of the insulation
generalized circuit constants (a, b, c, d), which is
capacitive reactance Xc). This is a vector addition with
referred to as a “two port network,” as shown in
the scalar magnitude given in Equation 3-85 and phase
Figure 3-14. The black box representation conveniently
difference ϕ given in Equation 3-86.
relates the sending end voltage and current to those at
the receiving end, as shown in Equation 3-75. Y = G i + j Bi (mho) 3-83
VS = a VR + b I R (V) Bi = 2πfC (mho) 3-84
I S =c VR + d I R
(G )
(A) 3-75
Y= + Bi
2 2
i (mho) 3-85
The transmission line constants a, b, c, and d are given
in Equations 3-76, 3-77, and 3-78, the series impedance ⎛B ⎞
Z and the shunt admittance Y of the circuit length. ϕ = tan −1 ⎜⎜ i ⎟⎟ (degrees) 3-86
⎝ Gi ⎠
⎛ YZ ⎞
a = d = 1+ ⎜ ⎟ (ratio) 3-76 Where:
⎝ 2 ⎠ Y = Insulation admittance, circuit length
(mho).
b=Z (Ω) 3-77
Gi = Insulation conductance (mho).
⎛ ⎛ YZ ⎞ ⎞ Bi = Insulation capacitive susceptance (mho).
c = Y⎜⎜1 + ⎜ ⎟ ⎟⎟ (mho) 3-78 ϕ = Phase angle between Bi and Gi (degrees).
⎝ ⎝ 4 ⎠⎠ C = Cable capacitance (F/m).
Series Cable Impedance Z Capacitive Susceptance of Insulation Bi
The series impedance Z of the cable length given in The capacitance C from Equation 3-52 times length is
Equation 3-79, which comprises the phasor addition of entered into Equation 3-84 to obtain Bi.
RAC in series with the inductive reactance XL, as shown
3-38
Conductance of the Insulation GC current in conductor 2 is counterclockwise. In between

The conductance G = Rc-1. Rc is calculated from Equa- the two conductors, the fields are in the same direction
tion 3-57 divided by circuit length. and so add to produce a higher magnitude resultant
external field, He. Outside the loop, the two fields rotate
3.7.2 Inductance L in opposing directions and subtract to produce a lower
Inductance is a measure of the degree to which a time- magnitude H.
varying current is impeded by the magnetic field. Induc-
tance L is defined as the ratio of the magnetic flux link- Figure 3-7 shows the distribution of the field external to
ages Φ to current I, as shown in Equation 3-87. one conductor H e. The shape of the field H around a
filament of current is directly proportional to the
Φ current and inversely proportional to the radial distance
L= (H/m) 3-87
r, as shown in Equation 3-90. At each radius, the field
I
Hr produces a local flux of density Br shown in Equa-
Faraday showed that an electromotive force (emf) is tion 3-91, which is proportional to the magnetic perme-
induced in a circuit by a changing magnetic field. Lenz ability μ of the local material. In the absence of a ferrous
showed that the emf is in a direction that will drive a material, μ is equal to the permeability of free space.
current, preventing the flux from changing. These The flux density Br at radius r from current I is given in
observations are expressed mathematically in the Fara- Equation 3-92 and is a thin annular ring of flux. If an
day-Neumman law, as shown in Equation 3-88. It can adjacent loop of wire of area A and unit length is placed
be seen that the magnitude of the back-emf e is the nearby, a total quantity of flux rings Φ will be “cap-
product of circuit inductance L and the rate of change tured” and are said to be linked to the loop. The total
of current. In an ac circuit, current varies continuously quantity of flux Φ is directly proportional to A and to
at power frequency. In a dc circuit, ripple currents are the average flux density BAV, as shown in Equation 3-93.
present, but are of low magnitude.
I
dΦ dI Hr = (A/m) 3-90
e=− = −L (V/m) 3-88 2π r
dt dt
B r = μH r (T) 3-91
Where:
L = Circuit inductance per unit length (H/m). μI
Φ = Flux linkages per unit length (Wb/m). Br = (T) 3-92
I = Current producing the magnetic field (A). 2π r
e = Back-emf per unit length (V/m). Φ = BAV A (Wb/m) 3-93
t = Time (s).
Where:
In an ac system, the rate of change dI/dt of the sinusoi- Hr = Magnetic field strength at r (A/m).
dal current waveform (It sin2πft) is 2πf, thus the back- r = radial distance from line current (m).
emf is, e = 2πf L I. Division by I yields the inductive Br = Flux density at r (T).
reactance XL (Ω/m), as shown in Equation 3-89. BAV = Average flux density in the loop (T).
A = Area per unit length (m2/m).
e μo
XL = = 2πf L (Ω/m) 3-89
= Magnetic permeability of free space =
I 4π.10-7 (H/m).
Where:
XL = Inductive reactance per unit length (Ω/m). A value of inductance L is ascribed to each individual
f = Frequency of generation (Hz). phase conductor. The inductance L of one phase or pole
includes both self-inductance due to linkages with its
The following sections first derive the geometric distribu- own current and mutual inductance due to linkages with
tion of the magnetic field and then the equations for the the currents in the two other phases or pole.
inductance of a dc circuit and of ac three-phase circuits.
Figure 3-7 shows the distribution of field strength H in a
Magnetic Field Distribution simple two-conductor circuit. H is also representative of
A diagram of the resultant magnetic field of strength H the distribution of flux density, as H = B/μ. The areas
surrounding a simple “go” and a “return” cable is illus- underneath the curves of H are proportional to the flux
trated in Figure 3-6. By convention, the field around the linkages Φ. Each conductor current provides two contri-
“inwards” flowing +ve current in conductor 1 is taken to butions of flux linkages with itself, the internal flux link-
be clockwise and that around the “outward” flowing –ve
3-39
ages inside the conductor Φi (shaded area) and the flux μ

linkages external to the conductor, Φe (white area under
ΣΦ = I =0.5 I. 10 −7 (Wb/m) 3-100
8π
the broken line). In Figure 3-7, the only part of Φe
linked with the loop is between the conductors. These ΣΦ
flux linkages are represented by the conductor internal Li = = 0.5. 10 −7 (H/m) 3-101
I
inductance Li and the external inductance Le, as shown
Where:
in Equation 3-94.
J = Current density in conductor (A/m2).
L = Li + Le (H/m) 3-94 I = Total current in conductor (A).
ΣIr = Current within radius r (A).
Where:
R = Outer radius of the solid conductor (m).
Li = Inductance internal to the conductor
r = Variable radius (m).
(H/m).
μ = Magnetic permeability of free space, =
Le = Inductance internal to the conductor
4π.10-7 = 1.257 × 10-6 (H/m).
(H/m).
μo = (Magnetic permeability of free space =
Conductor Internal Inductance 0.383 × 10-6 (H/ft)).
dr = Radial thickness of annular shell (m).
Inductance of a Circular Solid Conductor
Br = Flux density at r (T).
A circular solid conductor of unit length, outer radius
Φdr = Flux in shell dr at r, per unit length
R, and carrying a uniform, constant dc current I
(Wb/m).
will have a constant current density J, as shown in
ΣΦ = Total flux in conductor, per unit length
Equation 3-95. The magnitude of current Ir in a smaller
(Wb/m).
cylinder of radius r within the conductor is given by
Equation 3-96. The cumulative flux density Br at r due Inductance of a Circular Hollow-Core Conductor
to the contributions from each cylinder of current Hollow-core conductors are used in SCFF, GIS, and
within it, is given by Equation 3-97. The quantity of flux HTS cables. A metallic sheath is also a hollow-core con-
Φdr at r in an elemental shell, of thickness dr and unit ductor. Internal inductance Li is obtained by the same
length, is given in Equation 3-98. Φdr is linked with the method and is given in Equation 3-102. The use of a
internal volume (πr2 × 1) and, for the purpose of calcu- hollow-core conductor significantly reduces the magni-
lating inductance, is redistributed in proportion to the tude of internal inductance and, in an ac system, the
volumes to link it with the total conductor, as shown in back-emf. When the inner radius equals zero, Li
Equation 3-99. The latter equation is integrated to the becomes the same value as given in Equation 3-101 for a
outer radius R to obtain the cumulative flux linkages solid conductor. Equation 3-102 also gives the internal
ΣΦr inside the conductor, as shown in Equation 3-100, inductance of a metallic sheath.
from which the internal inductance Li, Equation 3-101,
Li =
is derived. This supports the experimental finding that
μ ⎡ 4⎛ ⎛ R ⎞ ⎞⎤
Li is a constant value for a solid conductor irrespective ⎢0.25R o − (R o R i ) + R i ⎜⎜ 0.75 − ln ⎜⎜ i ⎟⎟ ⎟⎟⎥
4 2
of conductor radius, which is 0.5 × 10-7 (H/m). (

2π R o − R i
2
)
2 2
⎢⎣ ⎝ ⎝ R o ⎠ ⎠⎥⎦
(H/m) 3-102
I
J= (A/m2) 3-95 Where:
πR 2 Ro = Outer radius of solid hollow core conduc-
tor (m).
r2
ΣI r = IJ = I 2 (A) 3-96 Ri = Inner radius of solid hollow core conduc-
R tor (m).
⎛ μI ⎞ External Inductance of a Go and Return Circuit

Br = ⎜ 2 ⎟
r (T) 3-97
The dc system comprises one go and one return conduc-
⎝ 2πR ⎠ tor. The external flux linked with one conductor is Φe.
⎛ μI ⎞ This is obtained by integrating flux density B, shown in
Φ dr = ⎜ 2 ⎟
r dr (Wb/m) 3-98 Equation 3-103, from the surface of one conductor of
⎝ 2πR ⎠ radius R to the center of the opposite cable at spacing S.
The external inductance Equation 3-104, is then
r2 ⎛ μ I ⎞ ⎛ μI ⎞ 3 obtained from the identity in Equation 3-87.
Φ dr = 2 ⎜ 2 ⎟
r dr =⎜ 4 ⎟
r dr (Wb/m)
R ⎝ 2πR ⎠ ⎝ 2πR ⎠
3-99
3-40
μI 1 μI ⎛ S ⎞ For circular conductors comprising a low number of

S
Φe = ∫
2π R R
dr = ln⎜ ⎟ (Wb/m)
2π ⎝ R ⎠
3-103 wires (less than 61) or shaped (noncircular) conductors,
the GMR is calculated by Equation 16-6 in Chapter 16
with the dimensions shown in Figure 16-3. Values of
μ S S GMR for different conductor sizes and wire numbers
Le = ln = 2 ln . 10 −7 (H/m) 3-104
2π R R are given in Tables 3-4, 3-6, and 3-7. The internal induc-
tance is calculated from Equation 3-107.
Maximum external inductance occurs when the cables
are widely spaced. However, successive increases in ⎛ R ⎞ −7
spacing have a progressively reducing effect on induc- L i = 2 ln⎜⎜ ⎟⎟ 10 (H/m) 3-107
tance because the log term shows the majority of the ⎝ m⎠
R
flux is captured at close spacings. The minimum induc-
tance occurs when the cables are installed touching. Equation 3-108 is a simple and sufficiently accurate
equation for internal inductance of stranded conduc-
Total Inductance of a DC Cable tors. A factor k is introduced in place of the constant of
The total inductance per conductor, internal and exter- 0.5 for a solid conductor. The internal inductance of the
nal, is obtained by Li from Equation 3-101 and Le from stranded conductor is k × 10 -7 H/m, from which the
Equation 3-104, giving Equation 3-105. GMR can be calculated.
⎛ ⎛ S ⎞⎞ ⎛ ⎛ S ⎞⎞
L = ⎜⎜ 0.5 + 2 ln⎜ ⎟ ⎟⎟10 −7 (Wb/m) 3-105 L = ⎜⎜ k + 2 ln⎜ ⎟ ⎟⎟10 −7 (H/m) 3-108
⎝ ⎝ R ⎠⎠ ⎝ ⎝ R ⎠⎠
Where: Where:
L = Inductance per unit length of one conduc- k = Internal inductance of a stranded conduc-
tor (H/m). tor (10-7 H/m).
S = Centerline spacing between dc “go” and
“return” (m). Typical values of GMR and k for different conductors
R = Radius of conductor (m). are given below.
Equation 3-105 shows that internal inductance Li has • Circular stranded, ≥ 37 wires: GMR = 0.769R,
the biggest impact when the cables of diameter 5 in. k = 0.525
(127 mm) are installed touching—e.g., Li is 13% of the • Circular stranded, ≥ 61 wires: GMR = 0.773R,
total inductance—having a conductor radius R of 1 in. k = 0.515
(25 mm) and centerline spacing of 5 in. (127 mm). L i • Circular solid conductor: GMR = 0.779R, k = 0.500
reduces to 7% of total when the cables are spaced at
24 in. (610 mm). • Hollow-core stranded, 12 mm duct: GMR = 0.826R,
k = 0.382
The internal inductance of the solid conductor of Li =
0.5 × 10-7 H/m in Equation 3-105 can be incorporated Conductors with a small number of wires have a lower
into the (ln S/R) term by modifying R to become GMR and a higher internal inductance. Table 3-4 con-
R exp(−1/4) = 0.779R, which is called the geometric firms that circular conductors with 61 wires or more,
mean radius, Rm. The geometric mean radius (GMR) is with areas of 550 kcmil (300 mm2) and above are closely
the radius of a fictitious, thin-walled tubular conductor similar to, and may be treated as, a solid conductor, hav-
that has no internal flux linkages. Equations 3-105 and ing a GMR of 0.779R and a k of 0.5. If the hollow-core
3-106 are identical. conductor becomes a thin shell, as in some supercon-
ducting cables, the GMR approaches unity, and k
⎛ ⎛ S ⎞ ⎞ −7 approaches zero—i.e., the internal flux linkages and
L = ⎜⎜ 2 ln⎜⎜ ⎟⎟ ⎟10 (H/m)
⎟ 3-106 inductance are negligible.
⎝ ⎝ m ⎠⎠
R
Total Inductance of an AC Cable
Where:
General Case: Asymmetric Formation
Rm = Geometric mean radius (GMR) of a solid
conductor (m). The external inductance Le of each cable depends
upon the relative configuration of the three-phase cables
1, 2, and 3. Typical configurations are shown in
3-41
Figure 3-15. The general mathematical case is the asym-

⎛ ⎛ SS ⎞ ⎞
metric spacing in Figure 3-15c.
L 3 = ⎜ k + 2 ln⎜ 2 3 ⎟ − j 1.73 ln⎛⎜ S2 ⎞⎟ ⎟10 −7
⎜ ⎜ R ⎟ ⎜ S ⎟⎟
In the calculation, it is necessary to include the 120° ⎝ ⎝ ⎠ ⎝ 3 ⎠⎠
phase displacement between each of the three currents. (H/m) 3-111
A physical return conductor is not provided in a bal- Where:
anced three-phase system, and this introduces a pro- S1 = centerline spacing between cables 1 and
spective conundrum in calculating the flux linked with a 2 (m).
“go and return loop.” One solution, Figure 3-16, is to S2 = centerline spacing between cables 1 and
form three “go and return loops” by positioning a ficti- 3 (m).
tious neutral return conductor N at a long distance from S3 = centerline spacing between cables 2 and
the circuit, at distances x1, x2, and x3 from each cable. In 3 (m).
the integration of the three contributions of flux cap-
tured in loop “1−N” etc., the terms containing “x” dis- Direct-buried cable systems generally use symmetrical
appear from the equation, leaving simply the ratios of flat cable formations to suit the trench floor, as shown in
cable spacing to conductor radius (S/R). Figure 3-15b and less frequently equilateral (trefoil)
shown in Figure 3-15e, the latter usually being with the
Another equally valid approach is based on the mathe- cables touching. Duct and pipe systems generally use
matical identity that the addition of the currents in any right-angled triangular (Figure 3-15d) and equilateral
two phases produces a current that is exactly in triangular formations to suit the duct bank. The sym-
antiphase (180°) with that in the third. Thus a three- metry permits the general equations to be simplified
phase system can be visualized as a simple “go and using the spacing relationships given below:
return” circuit, as shown in Figures 3-6 and 3-7 in which
the current flows down one phase cable and returns in • Equilateral triangle: S = S1 = S2 = S3
the other two, thereby providing loops with which to • Right-angled triangle at cable 2: S = S1 = S3,
capture the flux and calculate inductance. S2 = S 2
The resulting inductance equations for each cable are • Equal flat spacing, 1, 2, 3: S = S1= S3, S2 = 2S
given in Equations 3-109, 3-110, and 3-111. These are
the equations for the inductance of each of three phase In the general equations for asymmetric formation, the
cables with an unequal spacing (the asymmetrical trian- third term in the inductance equations is preceded by
gular spacing shown in Figure 3-15c). the operator j (representing a 90° phase shift), showing
that, in addition to impeding the flow of current by pro-
⎛ ⎛ SS ⎞ ⎞ ducing an inductive reactance, the cable configuration is
L1 = ⎜ k + 2 ln⎜ 1 2 ⎟ − j 1.73 ln⎛⎜ S1 ⎞⎟ ⎟10 −7 also acting as a transformer in which an in-phase volt-
⎜ ⎜ R ⎟ ⎜ S ⎟⎟
⎝ ⎝ ⎠ ⎝ 2 ⎠⎠ age is induced in each conductor. The in-phase voltage
produces a transfer of power between the cable phases.
(H/m) 3-109
For this reason, the cables in some long circuits are
transposed at each joint position to balance the power
⎛ ⎛ SS ⎞ ⎞
L 2 = ⎜ k + 2 ln⎜ 1 3 ⎟ − j 1.73 ln⎛⎜ S3 ⎞⎟ ⎟10 −7 transfer. This is readily achieved in the joint vaults in a
⎜ ⎜ R ⎟ ⎜ S ⎟⎟
⎝ ⎝ ⎠ ⎝ 1 ⎠⎠ cross-bonded circuit by transposing the physical posi-
tion of the phase cables in the next section.
(H/m) 3-110
Figure 3-15 Cable configurations in a dc circuit and three-phase ac circuits (courtesy of CCI).
3-42
The average inductance per phase Lphase is used for the Equispaced Flat Formation (Figure 3-15b)
circuit equations in Chapter 16, which is the average of The inductances of the three cables in the Flat forma-
Equations 3-109, 3-110, and 3-111, and is given in tion are given in Equations 3-113, 3-114, and 3-115. The
Equation 3-112. The third term has disappeared, show- inductances of the two outer phases, Cables 1 and 3, are
ing that the averaged power transfer between phases is equal. The inductance of the center phase Cable 2,
zero. The right-hand term of Equation 3-112 is essen- shown in Equation 3-114, is the same as the two-cable
tially the same as that incorporated into the calculation case, Equation 3-105. Cable 1 experiences the trans-
of inductive reactance X L (j 2πf L phase) for the single- former effect of power transfer, with an in-phase voltage
point bonded system described in Chapter 16, in Equa- contribution of magnitude (I. 2πf 0.12 × 10-6) V/m,
tion 16-5. GMDϕ is the geometric mean distance of the whereas Cable 3 experiences an opposing voltage of the
three conductors, (3 S1 S 2 S 3 ), and GMRϕ is the geomet- same magnitude. The voltage contribution is 41.3 V per
ric radius of one conductor as previously defined. 1000 A per 3000-ft (914-m) span.
⎛3 S S S ⎞ −7 The expression for average inductance Lphase is the same

L phase = k + 2 ln⎜ 1 2 3 ⎟ 10
⎜ R ⎟ as given in Equation 3-105, where S 1 = S 2 = S and
⎝ ⎠ S3 = 2S.
⎛3 S S S ⎞ −7 ⎛ ⎛S 2 ⎞ ⎞
= 2 ln⎜ 1 2 3 ⎟ 10 L1 = ⎜ k + 2 ln⎜⎜ ⎟ + j 1.2 ⎟ 10 −7 (H/m)
⎟ 3-113
⎜ Rm ⎟ ⎜ ⎟
⎝ ⎠ ⎝ ⎝ R ⎠ ⎠
⎛ GMD φ ⎞ −7 ⎛ ⎛ S ⎞⎞
= 2 ln⎜ ⎟ 10 (H/m) 3-112 L 2 = ⎜⎜ k + 2 ln⎜ ⎟ ⎟⎟10 −7 (H/m)
⎜ GMR ⎟ ⎝ R ⎠⎠
3-114
⎝ φ ⎠ ⎝
Where:
⎛ ⎛S 2 ⎞ ⎞
Lphase = Inductance per unit length of one conduc-
L 3 = ⎜ k + 2 ln⎜⎜ ⎟ − j 1.2 ⎟10 −7 (H/m)
⎟
3-115
tor (H/m). ⎜ ⎟
k = Internal inductance of a stranded conduc- ⎝ ⎝ R ⎠ ⎠
tor (10-7 H/m). Equilateral Triangular Spacing (Trefoil) (Figure 3-15e):
S1 = Centerline spacing between cables 1 and Equation 3-116 is obtained by entering S = S1 = S2 = S3
2 (m). into the general equations. The inductance is found to
GMDϕ=Geometric mean distance (3 S1 S 2 S 3 ) of be equal for each phase cable with no power transfer
cable formation (m). between phases. Equation 3-116 is also obtained when
R = Radius of a conductor (m). the values for S are entered into Equation 3-112 for
GMRϕ= Rm, Geometric mean radius of conductor average circuit inductance Lphase, confirming that the
≅ R exp(−1/4) = 0.779R (m). expression for average inductance represents an equilat-
eral cable formation.
⎛ ⎛ S ⎞⎞
L = ⎜⎜ k + 2 ln⎜ ⎟ ⎟⎟10 −7 (H/m) 3-116
⎝ ⎝ R ⎠⎠
3.7.3 Conductor Resistance, R
A conductor in a circuit carrying an ac current experi-
ences a magnetic field that produces a nonlinear distri-
bution of current density and an increase in I2R heating.
The theoretical representation of this phenomenon is
complex (CIGRE 2005a). A simple explanation is given
in this section.
By Lenz’s law, a time-varying magnetic field will induce

a voltage (a back-emf) in a solid conductor in a direc-
tion that will drive a current to produce an opposing
Figure 3-16 Circuit loops 1-N, 2-N, and 3-N to fictitious field. The conductor experiences both its own internal
return N (courtesy of CCI). field and the field from the adjacent conductors. Eddy
3-43
currents will be set up to oppose both these fields. Eddy When the conductor carries ac current, the rate of
currents, therefore, flow within the solid conductor and, change of flux linkages with time produces a higher
depending on interwire resistance, will attempt to flow back-emf at the center filament. This opposes current
within a stranded conductor. In some parts of the con- flow at the center and reduces the local current density.
ductor, the eddy currents will oppose the flow of main In consequence, the current density is increased near the
power current, reducing the current density, and in oth- conductor surface. As a general principle, the current
ers will reinforce it and increase the density. The abbre- filaments tend to carry more current in a position where
viated explanation is that the back-emf opposes the flow they experience minimum internal flux linkages and
of power current and so reduces the local current den- hence minimum back-emf, this being the region of high-
sity such that the current density is increased in the est field. An equilibrium of minimum impedance to cur-
parts of the conductor experiencing low back-emf. rent flow is reached, when the increase in resistive
potential drop IR from the increased flow of current,
Resistance of a Conductor Carrying DC Current, R starts to exceed the reduction in back-emf.
The values for dc resistance are obtained from pub-
lished standards such as ASTM B 8-04 (ASTM 2004) The concentration of current at the outer surface is
and IEC 60228 (IEC 2005). Examples are given in called “skin effect.” The increase in total I2R heat loss in
Tables 3-4, 3-7, and 3-8.
the conductor is represented by the skin effect factor ys.
Resistance of a Conductor Carrying AC Current, The concentration can be expressed in terms of the ‘skin
RAC depth’ δ, as given in Equation 3-118. The skin depth is
The ac resistance, shown in Equation 3-117, includes taken to be the radial depth at which the current density
factors ys and yp to increment the dc resistance. The fac- falls to 37% (1/e) of the 100% value at the surface. This
tor ys increments the dc resistance for skin-effect heat- is equivalent to all of the current being concentrated
ing. The factor yp increments the dc resistance for within the outer radial depth δ, at a constant current
proximity-effect heating. The calculations of ys and yp density of 50%.
are based on the solutions of Maxwell’s electromagnetic
wave equations. The solutions incorporate Bessel func- When the radius of a conductor is increased in excess of
tions and are provided in published works as a list of the skin depth, the full benefit of increased area is not
mathematical functions, each related to a specific range gained in ampacity. The skin effect factor ys increases in
of conductor diameters. The equations are listed in IEC magnitude with conductor diameter, power frequency,
60287 (IEC 1994) and, for conventional conductor size and metal conductivity. Skin effect is reduced when the
cables, in Chapter 11. The rigid tubular conductors in conductor is heated to operating temperature, thereby
GIL are of larger diameter than conductors in conven- reducing conductivity; and when copper is replaced by
tional cables, requiring the selection of a different math- aluminum, which possesses lower conductivity; and
ematical function within the list (CIGRE 2003). when the radial wall thickness is reduced by use of a
hollow-core conductor.
R AC = R (1 + y s + y p ) (Ω/m) 3-117
ρ
Skin Effect ys δ = 503 (mm) 3-118
The dc current density inside a conductor is uniform,
f
and so the field strength Hr and flux density Br increase Where:
linearly from the center, as shown in Figure 3-7 and δ = skin depth (mm).
expressed in Equation 3-97. The flux linkages with a fil- ρ = electrical resistivity (μΩ-m).
ament of current near to the center are higher than f = power frequency (Hz).
those linked with a similar filament on the outer surface.
Table 3-12 Conductor Sizes at Which Skin Effect Becomes Appreciable

60 Hz 50 Hz
Conductor Temperature Skin Depth Conductor Area Skin Depth Conductor Area
Material (°C) in. mm kcmil mm2 in. mm kcmil mm2
20 0.34 8.5 1250 630 0.37 9.3 1500 800
Copper
90 0.38 9.6 1500 800 0.42 10.6 2000 1000
20 0.43 11.0 2000 1000 0.47 12.0 2500 1200
Aluminum
90 0.49 12.0 2500 1200 0.54 13.6 3000 1600
3-44
The conductor sizes at which the radius exceeds the skin lated by IEC 60287 ampacity equations. Enameled wire
depth for circular copper and aluminum conductors at Milliken conductors almost eliminate the ac skin effect.
60 and 50 Hz are calculated from Equation 3-118 and
Proximity Effect yp
shown in Table 3-12. The table illustrates the conductor
sizes at which a change to the higher-efficiency Milliken The external magnetic field produced by the adjacent
stranded segmental conductor would improve ampacity. cables is inversely proportional to distance, as shown in
In practice, the upper limit for a concentric conductor is Figure 3-7 and expressed in Equations 3-90 and 3-92.
based on a precisely determined ac resistance value On the near side of the reference conductor, this rein-
within the ampacity calculation and is balanced by the forces the internal field and on the far side reduces it.
increased manufacturing cost of the Milliken conductor. The back-emf modifies the current density. The current
Experience shows that, for non-pipe-type applications, density will be increased on the near side and reduced
manufacturers generally start to offer Milliken copper on the far side. The resultant increase in total I2R heat
conductors for 60-Hz applications at conductor sizes of loss in the conductor is represented as an apparent
1500 kcmil (800 mm2) and above and for 50-Hz applica- increase in total conductor resistance, by the proximity
tions at 2000 kcmil (1000 mm2) and above. Selection of effect factor yp, shown in Equation 3-117.
an aluminum conductor permits circular stranded con-
ductors to maintain efficiency for another two standard The proximity effect becomes significant when cables
conductor sizes—i.e., at 60 Hz to 2500 kcmil (1200 mm2) are installed touching each other; however, it rapidly
and at 50 Hz to 3000 kcmil (1600 mm2). reduces as spacing is increased. Stranded segmental
conductors significantly reduce proximity effect as the
Skin effect is reduced significantly by the use of the thin layer of separating tapes present between each laid-
stranded segments in Milliken conductors. In a Milliken up segment forces the current to take a helically rotating
conductor, shown in Figure 3-11, the current in each path around the conductor, irrespective of fluctuations
wire is constrained to follow the path of the wire as it in the flux from adjacent cables. The proximity effect
rotates helically about the central axis of the segment. factor is less than the skin effect factor for uninsulated
The length of wire occupying the outer surface is equal wire Milliken conductors.
to the length on the inner surface. Overall, each wire Skin and Proximity Effect in Pipe-Type Cable
experiences an average back-emf along its length and so The geometrical formation of the three cables in the pipe
maintains a more uniform current density and reduced is asymmetric (unlike the photograph in Figure 3-3),
I2 R heating. The presence of interstrand resistance in being off the pipe centerline and being in irregular trefoil
the form of poor contact and naturally occurring oxide or cradle shape. In service operation, the cable configu-
layers between wires, constrains the current to follow rations change as they form lateral sinusoidal and heli-
the helical path of the wire. However, a proportion of cal patterns. The common feature is that the cables will
the current will transfer radially outwards at points of always be closer to one side of the pipe than the other.
good contact. The unbalanced magnetic fields from the cables induce
eddy currents on the steel pipe, which increase the fields
The efficiency of the Milliken conductor can be signifi- experienced by the conductors, thereby increasing the ac
cantly improved by the use of insulated wires and to a loss. An empirical increase of 50% in the skin and prox-
lesser extent by uni-lay conductor constructions. The use imity effect factors is added into Equation 3-119 (IEC
of an enameling process is generally preferred to an oxi- 60287, 1994).
dizing process because it is more consistent and less haz-
ardous in manufacture. CIGRE Brochure 218 (CIGRE R AC = R (1 + 1.5( y s + y p ) ) (Ω/m) 3-119
2003) compiles RAC/RDC ratio measurements at 40–90°C
of different Milliken segmental conductors in extruded
XLPE cables from Okada (Okada et al. 1999) and Table 3-13 RAC/RDC Ratios for Milliken Conductors in XLPE
Bourgeat (Bourgeat et al. 1999). These are summarized Cables (courtesy of CIGRE)
in Table 3-13. In the latter publications, it is stated that Conductor Conductor RAC/RDC Ratios
the ac/dc ratios for uninsulated wires in XLPE cable Uninsulated Insulated Wires
conductors are: (1) higher than in other cable types, Wires Oxidized Enameled
being attributed to reduction in interwire resistance by kcmil mm2 (%) (%) (%)
compaction in XLPE extrusion; (2) the ratios varied 2400 1200 115 - -
widely and inconsistently between different conductor 2800 1400 119 - -
constructions; and (3) the ratios are higher than calcu- 3200 1600 124 107 103
4000 2000 124–134 - 104
5000 2500 135 107.5 102
3-45
3.7.4 Positive, Negative, and Zero Phase Zero Phase Sequence Impedance
Impedances This is the impedance to the flow of current in a fault to
The sequence impedances are used in the calculation of ground. The current has the opportunity to flow back
currents and voltages during an asymmetric short cir- through both the internal and external ground conduc-
cuit—for example, when one phase conductor faults to tors and through the ground itself. The inductance of
ground. The currents and voltages that result are used the cable is of different magnitude because the 120°
to calculate: phase displaced flux linkages are no longer present. The
ground return impedance represents the distributed cur-
• The ac voltage induced in the cable’s integrated rent flow paths, which are determined by the ground
ground return conductor. In a specially bonded sys- resistivity and flux linkages. Published equations are
tem, the sheath voltage limiters (SVLs) are selected to available to calculate the inductance of a cable with a
withstand this voltage (IEEE 1988, CIGRE 2005b). ground return—for example, Carson (1926) and Pollac-
This produces significant heat in the SVL and is the zek (1926).
most onerous condition that they experience.
• The magnitude of the voltage induced in adjacent Cable system short-circuit performance is described in
alarm, relaying, and communication cables. The detail in Chapter 16, “Cable System Considerations.”
induced voltage is reduced to an acceptable level by,
for example, increasing the conductivity of the power 3.7.5 Characteristic Impedance
cable shield/ground return conductors and adding a The characteristic impedance Zc is the relationship
conducting or magnetic shield to the pilot and com- between the peak current and voltage present in a high-
munication cables (Mendez et al. 1998 and Gregory speed transient pulse, such as a normal switching surge
et al. 1999). or lightning impulse as it propagates along the cable. In
an ideal loss-free cable with nonfrequency-dependent
In the calculation method, the asymmetric short-circuit parameters, the characteristic impedance is simplified to
current can be resolved into three components: Equation 3-120. L is the inductance in coaxial mode,
• The positive sequence currents are of equal magni- shown in Equation 3-121, and capacitance C is given in
tude, 120° phase displaced, and rotate in the normal Equation 3-52. The speed of propagation is given in
time sequence (shown as counterclockwise on a pha- Equation 3-122, and permittivity values are given in
sor diagram). Table 3-9. In a cable system, the characteristic imped-
ance is used to calculate the transient currents and volt-
• The negative sequence currents are of equal magni- ages for the selection of the peak residual voltage (PRV)
tude, 120° phase displaced, but rotate in the opposite of the SVL, and the BIL levels for the cable jacket, the
time sequence (clockwise). bonding lead insulation, the shield interruption gaps,
• The zero sequence currents are of equal magnitude and the link box insulation. Design methods and limits
and rotate in the normal time sequence (counter- are given in industry standards such as IEEE 575 (IEEE
clockwise), but are exactly in phase with each other. 1988) and CIGRE Brochure 283 (CIGRE 2005b). The
subject is discussed further in Chapter 16.
Cable manufacturers are required to calculate the posi-
tive, negative, and zero phase impedances for insertion L
into the system network model. The induced voltages ZC = (Ω) 3-120
C
present in a specially bonded cable circuit are frequently
calculated using the positive sequence impedance ⎛R ⎞
because these voltages are solely dependent on known L = 2 ln⎜⎜ o ⎟⎟10 −7 (H/m) 3-121
cable parameters. This assumes that a three-phase fault ⎝ Ri ⎠
has occurred, not the usual single-phase fault to ground;
c
however, the method produces similar voltages and v= (m/s) 3-122
results in a safe design. εr
Positive and Negative Phase Sequence Impedance Where:
A cable system is isotropic, and so the positive and neg- ZC = Characteristic impedance of layer (Ω/m).
ative impedances are equal. The impedances are the L = Inductance in coaxial mode (H/m).
standard cable circuit impedances given in the C = Capacitance of layer (F/m).
Section 3.7 equations. εr = Relative permittivity of insulation (ratio).
v = Speed of surge (m/s).
c = Speed of light: 9.8 × 108 ft/s, (3 × 108 m/s).
3-46
REFERENCES CIGRE. 2005b. CIGRE Technical Brochure 283.

AEIC. 1997. AEIC CS2. “Pipe-Type Cable.” Special Bonding of High Voltage Power Cables.
Working Group B1-18.
AEIC. 2006. CS9-06 “Specifications for Extruded Insu-
lation Power Cables and Their Accessories Rated 69 Dakin, T. W., G. Luxa, G. Oppermann, J. Vigreux,
through 138kV.” First Edition. Association of Edison G. Wind, and H. Winkelnkemper. 1974. “Breakdown of
Illuminating Companies. Gasses in Uniform Fields Paschen Curves for Nitrogen
Air and Sulfur Hexafluoride.” ELECTRA. Vol. 32.
Anders, G. J. 1997. Rating of Electric Power Cables. pp. 61–82. January.
IEEE Press. McGraw Hill. ISBN 0-7803-1177-9.
Dissado, L. A. and J. C. Fothergill. 1992. “Electrical
Anders, G. J. 2005. Rating of Electric Power Cables in Degradation and Breakdown in Polymers.” Chapter 14.
Unfavorable Thermal Environment. IEEE Press. Wiley IEE Materials and Devices Series 9. Peter Peregrinus
Inter-Science. ISBN 0-471-67909-7. Ltd. ISBN 0 86341.1967.
Arkell, C.A. and A. F. Parsons. 1981. “Insulation Endersby, T.M., S. J. Galloway, B. Gregory, and N. C.
Design of Self Contained Oil-Filled Cables for DC Mohan. 1993. “Environmental Compatibility of
Operation.” TD Conference. Paper 81 TD 689-9. Supertension Cables.” p. 71. IEE Third International
Conference. Publication 382. London.
Arkell, C.A. and B. Gregory. 1984. “Design of Self Con-
tained Oil-Filled cables for UHV DC Transmission.” Gregory, B., J. Monteys, S. Baris, and J. M. Mendez.
CIGRE. Paper 21-07. Paris. 1999.“Installation of 220 kV XLPE Cables in a Tunnel
Installation Whilst Minimising Electromagnetic Induc-
Arkell, C.A., B. Gregory, and J. E. Hawkes. 1984. tion in Communication Cables.” Paper A4.7. Jicable.
“Design and Testing of 270 kV DC Self Contained, Oil
Filled Cable for the Land Section of the UK-France Hampton, R. N., F. Chang, and S. B. Hobdell. 2000.
Connection.” IEEE/PES. Transmission and Distribu- “What Happens to Materials under HVDC?” CIGRE
tion Conference. Paper 84 T&D 302-6. Kansas City. paper P2-001. Paris.
ASTM. 2004. ASTM B 8-04. “Standard Specification ICEA. 2004. ICEA S-108-720-2004. “Standard for
for Concentric-Lay-Stranded Copper Conductors, Extruded Power Cables Rated Above 46 Through
Hard, Medium-Hard, or Soft.” ASTM International. 345kV.” Insulated Cable Engineers Association
April. Inc. July.
Attwood, J.R., B. Gregory, M. Dickinson, R. N. Hamp- IEC. 1994. IEC 60287 1994-12 and Amendments.
ton, and R. Svoma. 1998. “Development of High “Calculation of the Current Rating Part 1, Current
Stress HV and EHV XLPE Cable Systems.” CIGRE. Rating Equations (100% load factor) and Calculations
Paper 21-108. of Losses.”
Bourgeat, X., J. Santana, A. Fustier, and P. M. Dejean. IEC. 1999. IEC 61443 1999-07. “Short-circuit Tempera-
1999. “A New Method for Power Cables AC Resis- ture Limits for Electric Cables with Rated Voltages
tance.” Paper B8.1. Jicable. Above 30 kV (UM = 36 kV).”
Carson, J. R. 1926. “Wave Propagation in Overhead IEC. 2001. IEC 62067 2001-10. “Power Cables with
Wires with Ground Return.” Bell System Technical Extruded Insulation and Their Accessories for Rated
Journal. Vol. 5. Voltages above 150 kV (Um = 170 kV) up to 500 kV
(Um = 550kV).”
CIGRE. 2003. CIGRE Technical Brochure 218.
Gas Insulated Transmission Lines (GIL). CIGRE Joint IEC. 2004. IEC 60840 2004-04. “Power Cables with
Working Group 23/21/33.15. February. Extruded Insulation and Their Accessories for Rated
Voltages above 30 kV (Um = 36 kV) up to 150 kV
CIGRE. 2005a. CIGRE Technical Brochure 272. Large (Um = 170 kV).”
Cross-Sections and Composite Screens Design. Working
Group B1.03. June.
3-47
IEC. 2005. IEC 60228: 2005. “Conductors of Insulated Neher, J. H. and M. H. McGrath. 1957. “The Calcula-
Cables.” tion of the Temperature Rise and Load Capability of
Cable Systems.” AIEE Trans. Vol 76. Part 3.
IEEE. 1988. IEEE ANSI/IEEE Std 575-1988. pp. 752–772.
“IEEE Guide for the Application of Sheath-Bonding
Methods for Single-Conductor Cables and the Okada, M., T. Karasaki, H. Furuhata, N. Imajo, M.
Calculation of Induced Sheath Voltages and Currents Mitant, and S. Aakuma. 1999. “Transmission
in Cable Sheaths.” Capacity Design of Underground Power Cables.” Paper
A 7.1. Jicable.
IEEE. 2000. IEEE Std 404-2000. “IEEE Standard for
Extruded and Laminated Dielectric Shielded Cable and Pollaczek, F. 1926. “Uber das feld einer unendlich lagen
Joints Rated 2500 V to 500,000 V.” wecshselstromdurchflossen Einfachleitung.’ ENT3.
IEEE. 2003. IEEE Std 48-1996(R2003). “IEEE Smith, C., S. J. Galloway, B. Gregory, S. D. Lloyd, and
Standard Test Procedures and Requirements for D. Notman. 2003. “The Development of an Ultrasound
Alternating-Current Cable Terminations 2.5 kV Quality Monitoring Process for the Manufacture of
Through 765 kV.” Enhanced Reliability HV and EHV XLPE Cables.”
Paper A.2.6. Jicable. Versailles.
Mendez, J. M., S. Barris, J. Monteys, and B. Gregory.
1998 “220kV XLPE Cables in a Tunnel Installation: Zenger, W., S. J. Galloway, and B. Gregory. 2006.
Minimising Electromagnetic Induction in Communica- “Thermo-Mechanical Design of XLPE Insulated HV
tion Cables.” IEE Conference. Cyprus. and EHV Cables Installed in Duct-Manhole and Pipe
Systems.” Paper B1-111. CIGRE. Paris.
Nasar, S.A. 1990. Electric Power Systems. McGraw Hill.
ISBN 0-07-045917-7.
3-48
CHAPTER 4 Cable Construction:

Extruded Dielectrics
Authors: Brian Gregory, Cable Consulting International Ltd.
Axel Schlumberger, Southwire HV Solutions
Ernesto Zaccone, Prysmian Cables and Systems
Reviewer: Pierre Argaut, Silec Cable
This chapter gives a comprehensive account of all aspects of extruded cable technology to
introduce the prospective user to cable design, construction, and manufacture. The con-
structions and tests are described with reference to North American specifications AEIC
CS9-06 and standard ICEA S-108-720-2004 and to international standards IEC 62067
and IEC 60840.
North America experienced a cable milestone in 2006 with the installation of its first
long-length 345-kV extruded crosslinked polyethylene (XLPE) circuits containing joints.
Worldwide, EHV XLPE cable systems have been progressively installed in Japan and
Europe starting in the 1990s, and are now the preferred choice for transmission at 345 to
500 kV. In doing so, XLPE cables have displaced the dominant position of cables with
fluid-filled taped insulation. XLPE has the benefits of low loss, low capacitance, and
absence of dielectric fluid. Ethylene propylene rubber (EPR) has benefits in HV applica-
tions of improved bending flexibility, temperature performance, and moisture resistance.
The advances in, and barriers to, extruded cable technology are described—from the first
synthesis of thermoplastic polyethylene in 1933 to the present dominance of the thermo-
sets XLPE and EPR. Accessories posed the final barrier to EHV cable applications; their
design is described in Chapter 8. The properties of XLPE and EPR are described in
detail, because these significantly influence manufacturing techniques and cable con-
structions. This chapter includes descriptions of cable manufacture and in particular of
the available methods of continuous vulcanization (CV). Also included is a description of
the wide choice of cable outer constructional finishes and their merits. The chapter
includes detailed information on selection of insulation design stress and its dependence
on manufacturing quality control in minimizing insulation inclusions, shield protrusions,
and extrusion defects. Methods to relate design stress to service life are covered.
Brian Gregory is the Technical Director and a founder member of

Cable Consulting International Ltd (CCI), incorporated in 2001. He
received an honors degree in Electrical and Electronic Engineering from
the City University, London in 1968. He received professional training
as a student apprentice with AEI Supertension Cables from 1963 to
1968. In 1969, he joined BICC Supertension Cables as a Development
Engineer for cables and accessories, becoming the Chief Development
Engineer in 1981. In 1992, Mr. Gregory became the Chief Engineer of
BICC Supertension and Subsea Cable Systems, with responsibilities for supertension
cable technology worldwide. He has extensive expertise in the research, development,
design, manufacture, testing, and commercialization of land and submarine cable sys-
tems, ac and dc up to 525 kV, insulated with extruded XLPE, EPR, fluid impregnated
LPP and paper, mass-impregnated paper, and LPGF and HPGF paper. He has recently
4-1
Chapter 4: Cable Construction: Extruded Dielectrics EPRI Underground Transmission Systems Reference Book
researched the thermomechanical behavior of XLPE Ernesto Zaccone is currently with the
and HPFF cables in pipe and duct-manhole systems for R & I ( Re s e a rc h a n d I n n ovat i o n )
EPRI, producing design guides and new calculation department of Prysmian Cables and
methods. He is an inventor and has authored over 45 System (formerly Pirelli Cables) based
technical papers and books. He is a Fellow of the IEE, a in Milan, Italy. He obtained a degree
Member of the IEEE, a Member of CIGRE, a Char- in Electrotechnical Engineering in
tered Engineer, and a FEANI registered European 1969. During the same year, he joined
Engineer. the power cable industry, where he was
engaged in a number of different activities such as
Axel Schlumberger is the General projects in high-voltage testing laboratories and high-
Manager of Southwire High Voltage voltage cable system design. In particular, he was
S ol u t io ns a n d i s res po ns i ble for involved in the study, design, and realization of some
Southwire Company’s underground major EHV (Extra High Voltage) extruded cable
transmission cable business. He holds projects. He was also the author of several studies on the
a Master’s Degree in electric energy application and impact of EHV cables when integrated
systems and power apparatus from the in transmission systems. Currently he is involved in a
University of Karlsruhe in Germany. number of innovative underground cable projects and in
His career in transmission cable systems began as a sales international standardization activities. Mr. Zaccone is
engineer for Cortaillod Cossonay Cable in Switzerland, the Chairman of the CEI CT 20 Italian standardization
covering transmission cable markets and projects in the body for power cables. He is the Co-convenor of the IEC
Middle East and Southeast Asia, with voltage levels up TC20 WG 16, which is responsible for the international
to 400 kV. Later, he joined Southwire Company near standardization of high-voltage cables and their accesso-
Atlanta, Georgia. There, he has played a key role in the ries, and is an active member of CIGRE and IEEE.
design, development, and implementation of South-
wire’s XLPE transmission cable systems business. Mr.
Schlumberger is a Member of the IEEE Power Engi-
neering Society and CIGRE.
4-2
EPRI Underground Transmission Systems Reference Book Chapter 4: Cable Construction: Extruded Dielectrics
4.1 INTRODUCTION 4.1.1 Objectives

Three types of extruded insulation cables are in service This chapter is written to inform utility engineers about
worldwide in the HV to EHV voltage range. In order of the selection of extruded cable types and the witnessing
discussion in this chapter, they are PE (thermoplastic and acceptance of tests.
polyethylene), XLPE (crosslinked polyethylene) and
EPR (crosslinked ethylene propylene rubber). In North The chapter describes the differences between the
America, the designations HV and EHV describe trans- generic types of insulating materials, cable designs, and
mission circuits in the voltage range of greater than manufacturing equipment with respect to their merits
46 kV and up to 345 kV (Figure 4-1), while worldwide and individual characteristics. It is left to the utility
this extends to 500 kV (Figure 4-2). HV and EHV cable engineers to decide which types are the most suitable for
types are distinguished from lower-voltage types by their particular applications. In describing the cable
their high power ratings, larger conductor sizes, higher manufacturing processes, it is helpful to show examples
insulation design stresses, increased insulation thick- of past manufacturing defects. At the higher transmis-
nesses and transition from a mass-produced commodity sion voltages, manufacturers are well aware of the qual-
to a custom-designed product. In particular, the electri- ity defects that may occur and exercise great diligence
cal design stress is significantly increased at 230 kV and and responsibility to both the purchaser and to their
above, requiring that the highest standards of manufac- own shareholders in ensuring that defects either do not
turing quality and materials supply be applied to occur or that their QA system detects and rejects them
achieve the required 40-year service life. at the earliest possible stage of manufacture.
The chapter also provides an account of the technologi-

cal barriers encountered and overcome, so that the util-
ity engineer has an appreciation of the design and
quality control measures that are built into present-day
HV and EHV cable design and manufacture. The engi-
neer may reasonably ask a prospective manufacturer to
explain the quality control measures that are in position
to achieve and maintain HV and EHV cable quality.
4.1.2 Technological Position of Extruded Cables

The first long-length 345-kV XLPE (crosslinked poly-
ethylene) cable circuit with joints to be installed in the
Figure 4-1 United States’ first 345-kV XLPE, 1750 kcmil United States was commissioned in Connecticut in 2006
(800 mm2), lead sheath cable with joints, 2006 (courtesy (see Figure 4-1). Five other 345-kV projects are in plan-
of Silec Cable). ning and installation, with a total combined single-cir-
cuit length of 64 miles (103 km). This is a key landmark
in the evolution of extruded cable technology in the
United States, which commenced with the first synthesis
and applications of polyethylene for telegraphy and
power cable 73 years ago.
Although deceptively simple in construction, modern

EHV XLPE cables have only been made possible by
major developments in dielectrics science, polymer chem-
istry, clean materials, manufacturing equipment and
accessories. Significant technological barriers have been
overcome, including the delay caused by the high failure
rate of URD (underground residential distribution) cable
from water treeing and the time needed to develop and
prove compatible accessories (see Figure 4-3).
Figure 4-2 World’s first 500-kV 1600 kcmil (800

mm2) XLPE cable in 1988 (courtesy of J-Power
Systems).
4-3
seded the previous pre-eminent designs of HPFF and

SCFF cables. There is now no easy way of turning back,
since the manufacturing capacity for extruded cables is
increasing and that for SCFF cables reducing, such that
the demise of SCFF cable and factories can be extrapo-
lated. There now follows a learning period in which ser-
vice experience and knowledge are accumulated. The
technical author Peter Graneau wrote in 1980, “the gen-
eral acceptance of new and more economical systems usu-
ally has to wait until quite a few installations have been in
successful operation for 10-20 years” (Graneau 1979).
Figure 4-3 Joint development lagged cable development For the highest-voltage category of XLPE cable at 230 to
by 10 years in Japan.
500 kV, this period began in mid-1995, when the first cir-
cuits with large-conductor XLPE cables and joints were
installed. All being well, the year 2015 will pass unnoticed
The main drivers for the application of extruded HV as a quiet testament to sound cable engineering.
and EHV cables have been:
• Environmental compatibility, arising from absence of 4.1.3 Contents
the fluid impregnant that is an essential component This chapter describes cable constructions in relation to
in the established designs of paper and LPP cables the new North American specifications for XLPE and
and hence complete elimination of the leaks associ- EPR extruded cable. The first recommended point of
ated with them. reference for the utility engineer is AEIC CS9-06 (AEIC
• Reduced fire risk in tunnels and buildings arising 2006) for cables and their accessories, which refers to,
from absence of any hydrocarbon fluid that could and complements, ICEA-S-108-720-2004 (ICEA 2004)
spread a fire to other circuits and other equipment. for the cable alone. The second points of reference are
to the international test specifications IEC 62067 cover-
• Reduced maintenance arising from the absence of
ing cables and accessories at >150 to 500 kV and to IEC
pressurizing equipment and the need to monitor and
60840 covering 30 to 150 kV, also for cables and acces-
refurbish it.
sories. Both specifications are now entering the mainte-
• Reduced capacitive reactance and dielectric losses, as nance stage. On the request of IEC, CIGRE TB 303
compared to paper and LPP cables. (CIGRE 2006) has been published, giving recommen-
• Economic advantages for long-length submarine dations for the evolution of these standards. Accessories
applications: for ac applications, reducing charging are regarded as being equally, if not more, important in
current magnitude and eliminating the need to feed achieving circuit reliability than the extruded cable, and
dielectric fluid, and for dc applications, eliminating so the three key specifications advocate the “systems
mass-impregnated paper cables and increasing oper- approach” in which one manufacturer is responsible for
ating temperature and ampacity. the design of the system and for the manufacture and
assembly of the cable and accessories. This chapter con-
The main thrust of the development has been to siders those parts of the cable design that are required
increase the design stress of extruded cables such that to be compatible with the accessories. The designs of
the thickness of the insulation, overall cable diameter, the accessories themselves are considered in detail in
cable span length, and most importantly, service life reli- Chapter 8.
ability are competitive with those of the highly evolved
paper and LPP cables. Although the development of The chapter compares the advantages and technological
“high-stress” XLPE has succeeded in significantly challenges for each cable type with respect to the EHV
reducing insulation thickness, HV and EHV cables are applications in the voltage range of 230 kV and above.
still slightly thicker than their paper and LPP counter- The comparison is still relevant to lower-voltage HV
parts. The difference is now sufficiently small for XLPE cables; however, historically these cables have had lower
cables to be selected in preference for major EHV design stresses, which render them more tolerant of less
projects and to leave little doubt that diameter reduc- stringent cleanliness levels and geometric tolerances. The
tions will become possible when further confidence is inevitable future cascade of the higher EHV design
gained in accessory performance. stresses to the lower-voltage cables will require the appli-
cation of the associated EHV materials and manufactur-
It is now clear that extruded cables have become the pre- ing techniques.
ferred type for all transmission voltages and have super-
4-4
Sections 4.2 through 4.13 summarize the evolution of • Mechanical robustness, which provides protection
PE, XLPE, and EPR extruded cable types that have during installation and in service.
made possible service applications up to 500 kV, and • Low dielectric permittivity, which gives low capaci-
describe the technological barriers that have been over-
tive reactance and low charging current.
come by improvements in manufacturing equipment
and materials supply. • Exceptionally low dissipation factor, which mini-
mizes dielectric heating.
Sections 4.14 through 20 describe the component parts • Low dielectric loss factor, which provides energy sav-
of the cable construction with reference to the North ings.
American and international specifications and, in par- • Simple homo-polymer chemistry, high purity, and
ticular, give the reasons for the key specification require-
absence of fillers, which result in a low-cost insulation
ments. The components are considered with respect to
compared to other polymers.
their dimensions and the manufacturing equipment.
This is followed by a description of the manufacturing
4.2.2 Challenges
quality control measures necessary for a 230- to 500-kV
circuit to consistently achieve long service life. • Low melting point, which limits the normal operat-
ing temperature to 70°C and hence limits the ampac-
Section 4.21 describes quality control and industry spec- ity, requiring a larger conductor size.
ifications. Section 4.22 discusses shipping of cable. • Comparatively stiff (inelastic) cable at normal and
low ambient temperatures, increasing difficulty in
Section 4.23 describes the methods of calculating the bending, especially for jointing.
electrical design stresses necessary to achieve a 40-year
life.
• Minimal resistance to partial discharging; as a result,
high standards are essential in quality control for
The evolution of extruded cable technology is consid- manufacture of the material and the cable and for the
ered with respect to the order in which the materials assembly of the accessories.
were first developed and extruded: the thermoplastic, • High thermal expansion, requiring allowances in the
homo-polymer polyethylene (PE), the thermo-set, cable construction.
crosslinked homo-polymer polyethylene (XLPE), and • Susceptibility to premature electrical failure by water
the thermo-set, crosslinked co-polymer, ethylene-propy- treeing.
lene rubber (EPR). Each insulant is described, together
with its advantages and challenges and this is followed
• Susceptibility to premature ageing by oxidation,
by an account of the technological advances that were requiring the inclusion of antioxidants.
necessary to make present-day 345-kV cables possible.
The technological barriers and the solutions give the 4.3 GENERAL DESCRIPTION OF
reader an appreciation of the design and manufacturing POLYETHYLENE INSULATION
complexities and quality control measures that are built The monomer, ethylene, is the simple hydrocarbon mol-
into present-day extruded cable designs and manufac- ecule C2H4, which is a gas at NTP (normal temperature
turing equipment to produce high-class insulation. and pressure). A double bond joins the two carbon
atoms. The act of polymerization opens one of the
4.2 PE CABLE: BENEFITS AND CHALLENGES bonds, such that a very long chain molecular backbone
The benefits and challenges of PE for cable applications is formed, a short length of which is shown in Figure
are as follows. 4-4, left. Polyethylene is a thermoplastic material, which
upon the application of heat becomes a liquid that can
4.2.1 Benefits be molded and, if required, remolded again.
• Low melting temperature, which gives ease of extru- Low-density polyethylene (LDPE), produced by the
sion. high-pressure polymerization process, is a high-
• Low viscosity of the liquid melt, which permits filtra- molecular-weight molecule with many side chains, as
tion to high purity. depicted diagrammatically in Figure 4-4, right. The side
• Exceptionally high electrical strength. chains prevent close packing of the molecules, which
reduces the density, reduces the degree of crystallinity,
• Electrical strength attained without need of a liquid increases the amorphous volume, and reduces the
impregnant. melting temperature to 105-110°C. For the purpose of
4-5
Figure 4-4 Left: chemical backbone of polyethylene. Right: Long side chains attached to the LDPE backbone
(courtesy of CCI).
brevity, a temperature of 107°C is taken in this chapter In uncrosslinked LDPE, the lamellae also form “crystal-
to be indicative of the melting point. lites,” which have the appearance of the spherical head
of a dandelion flower in seed. Each thin frond is formed
LDPE has the following important properties for the by a long, thin lamellae, radiating from a common cen-
manufacture of extruded cable insulation. It can be: tral nucleating point. The structure of the lamellae
frond is similar to a long strip of carpet. Long chain PE
• Extruded at low temperature (>120°C).
molecules fold concertinalike, back and forth parallel to
• Filtered at low temperature in the liquid phase to each other and perpendicular to the frond in a forma-
remove particles. tion similar to woven parallel tufts of wool in a carpet.
• Crosslinked to form XLPE, because the tips of the The chains are held parallel by the weak attraction
side chains are terminated with a hydrogen atom that forces. The thin spaces between the fronds contain
is available to participate in chemical crosslinking. amorphous PE. Some of the molecular chains pass from
one frond to the other through the amorphous region,
In North America, LDPE cable insulation is also binding it together, such that at room temperature the
referred to as high-molecular-weight polyethylene amorphous regions cannot easily unravel and flow. A
(HMWPE). range of crystallite diameters exist, with small crystal-
lites filling the gaps between larger ones. The crystallites
At room temperature, LDPE is an engineering solid, interfere with the clear transmission of light such that
possessing stiffness—i.e., low elasticity, resulting from its below the melting point temperature PE becomes trans-
semicrystalline structure. The crystalline structure com- lucent white.
prises lamellae, which are regions in which some parallel
sections are available to be attracted together out of the The volume of PE between adjacent crystallites is amor-
amorphous state by Van der Waal’s force. The lamellae phous—i.e., the molecules do not form a pattern. The
form because they are a lower energy structure. Figure small volume of amorphous PE has the opposite prop-
4-5 shows the chaotic distribution of lamellae in XLPE. erties to the rigid crystallite structure and behaves as a
The image was taken by an electron microscope; note viscous gel. The mixed morphology gives PE the proper-
that the scale bar of 500 nm is the wavelength of light. ties of plasticity—i.e., creep with time, in which, upon
the application of load, the material initially exhibits a
high modulus (inelasticity), but will slowly deflect under
time and load. PE exhibits viscoelastic behavior in
which deflection under constant load increases with
time until an equilibrium deflection is attained. The
crystallite structure produces a high elastic modulus
(low elasticity), and so at temperatures under 70°C, high
stress is required to reveal the viscoelastic effect. Above
70°C, the deflection increases and the time constant of
the insulation reduces. The insulation also shows
increasing “compression set”—i.e., it does not return to
its original shape.
When the temperature is increased, the small crystallites

lose their shape and melt into the amorphous gel,
increasing the proportion of amorphous material. The
Figure 4-5 Lamellae in the etched surface of XLPE from
scanning electron microscope (courtesy of CCI).
rate of softening and increase in plasticity of the bulk
PE become perceptible from an engineering viewpoint
4-6
above 60°C, accelerating until the majority of crystal- molecular spines can pack close together under Van der
lites have dissolved at the nominal LDPE melting point Waal’s force of attraction to form a highly semicrystal-
of 107°C. LDPE then becomes a clear transparent liq- line structure. This gives HDPE high mechanical
uid with the viscosity of a medium syrup. Figure 4-6 strength and stiffness, a reduced resistance to cracking,
shows liquid PE that is cooling. The PE in the meniscus and a high melting temperature (135 °C); however, it
has cooled to below 107°C. The crystalline structure is cannot be crosslinked.
forming and has lost its transparency, becoming trans-
lucent white, in a similar way to the core of a PE cable The expansion curve for LDPE shown in Figure 4-7 is
cooling after extrusion. identical to XLPE up to its melting point of 107°C. The
volumetric expansion for HDPE is the same as LDPE,
Melting from the semicrystalline structure to a liquid but is distributed up to its higher 135°C melting point.
produces a particularly high rate of volumetric expan- The volumetric expansion of PE and XLPE through the
sion, very much greater than normal solids, as shown in 80 °C temperature rise to 107 °C is 12.5%, which is 20
Figure 4-7. times greater than the 0.6% value for aluminum. The
normal operating temperature specified for an LDPE
High-density polyethylene (HDPE) does not have the cable is comparatively low at 70°C, which limits the vol-
side chains, shown in Figure 4-4 right, and so the long umetric expansion significantly to 4%. In comparison,
the volumetric expansion of an XLPE cable at its 90°C
operating temperature is 7.5%. These values illustrate
the significant dimensional changes that have to be
allowed in cable manufacture during cooling and in ser-
vice operation when the cable construction is heated and
particularly when heated to an emergency operating
temperature of 105°C. (It should be noted that the lin-
ear radial expansion of an XLPE cable is up to 33%
greater than indicated by the above volumetric compari-
sons because the insulation in a cable is constrained
from expanding lengthwise.)
The gradient of the line in Figure 4-7 is the coefficient of

volumetric expansion. The coefficient of expansion is
constant for “EPR 2,” which is a grade with low crystal-
linity. The four other materials, and in particular the
three PEs, show significant increase of coefficient of
expansion with temperature due to crystalline “melt-
Figure 4-6 PE granules melted into liquid (courtesy ing”. The long chain molecules in PE are initially held
of CCI). tightly packed together in the lamellae structure. As the
crystallites melt, the lamellae release the molecules into
Figure 4-7 Volumetric thermal expansion of PE with temperature (courtesy CCI).
4-7
the lower density amorphous phase to cause an immedi- emitting heat, light, and sound to varying degrees. The
ate volumetric expansion. Conversely, when liquid PE is breakdown voltage is given by the Paschen characteris-
cooled, the growth of the crystallites requires a contrac- tic for the particular gas, gap thickness and pressure (or
tion of the bulk volume. When extruding or molding PE gas density). Ideally the breakdown stress for a perfect
and XLPE, the high rate of thermal contraction may insulation would be constant with thickness, but this is
produce internal contraction voids, unless external pres- not so for a gas, due to the mobility and acceleration of
sure is applied and the rate of cooling is reduced. charge carriers within it. The breakdown stress for air
with gap depth at constant atmospheric pressure is given
Electrically, polyethylene, like other insulators, has a in Figure 4-8 (CIGRE 1974) and illustrates the rapid
wide energy gap between the valence bands and the reduction in breakdown stress with increasing thickness.
empty conduction band. The energy gap for PE is wide, For example, a laminar gap of 0.12 mil (0.003 mm or
at 7.6-9.0 eV, and thus forms an excellent, low-dielectric- 3 μm) at atmospheric pressure has a breakdown stress of
loss insulation (Dissado and Fothergill 1992). The 3800 V/mil (150 kVp/mm), and a gap of 0.8 mil
60-Hz relative permittivity is low at room temperature, (0.02 mm or 20 μm) has a tenfold reduction to a stress
being 2.3, and reducing to 2.1 at 90°C (Bartnikas and of ~510 V/mil (20 kVp/mm). Allowing for the conver-
Srivastava 2000). The 60-Hz dissipation factor (dielec- sion to an rms voltage by a multiplication factor of 0.71,
tric loss angle) is very low, being 0.0002 in the range 20- the latter becomes 360 V/mil (14 kV/mm). The stress in
90°C (Bartnikas and Srivastava 2000). The low dissipa- the adjacent PE insulation (multiplication by a factor of
tion factor, which is beneficial in ac applications, is 2.3, the reciprocal of the permittivity ratio of PE to air)
attributed to the deep electronic charge traps of is 155 V/mil (6 kV/mm). This illustrates that, within the
~1 to1.5eV, which immobilize charge carriers and so recommended AEIC stress limit for a 345-kV cable of
minimize conduction and loss (Bartnikas and Srivastava 356 V/mil (14 kV/mm), a 0.8 mil (20 μm) air-filled, lami-
2000). Charge traps are associated with the interface nar void at atmospheric pressure will emit discharge.
between the crystalline and amorphous phases (Dissado
and Fothergill 1992). Under dc voltage, the charge traps The partial discharge releases the capacitively stored
are detrimental to the insulating properties, because charge Q, which is proportional to the voltage difference
they accumulate sufficient charge to increase the local V across the void (see Equation 4-1). Note that Q is
electrical field and so significantly reduce the cable related to the magnitude of the partial discharge, mea-
breakdown strength under dc and impulse voltage. This sured in pC, recorded by a partial discharge detector,
phenomenon is called “space charge,” which at ambient but is not equal to it. The detector actually measures the
temperature remains permanently trapped. The charge milli-volt change in voltage appearing across the cable
carriers are only released when the temperature is raised shields, and compares it to the voltage change produced
close to the melting point and the semicrystalline struc- by an external PD calibration charge injected across the
ture merges into the amorphous background. cable insulation (Kreuger 1989). Equation 4-2 shows
that the energy W is proportional to the square of the
Polyethylene has little resistance to partial discharge electrical stress G in the void. Thus as the design stress
(PD) under ac voltage. Partial discharge is the electrical of a typical 345-kV cable is 350 V/mil (14 kV/mm) and is
breakdown of gas within a void in the form of a spark, approximately twice that of a 69-kV cable design stress
Figure 4-8 Electrical breakdown stress of air with gap at atmospheric pressure. Left: linear/log. Right: log/log
(courtesy of CCI).
4-8
of 150 V/mil (6 kV/mm), the energy liberated in the Alternative production methods, developed in the
same void will be four times greater. Under ac voltage, 1960s, made possible the manufacture of higher-density
the partial discharge is repetitive, with typically 3 to 4 grades (HDPE) using a catalyst in a low-pressure reac-
pulses each half cycle; thus at 60-Hz frequency, approxi- tor. However, the base LDPE for extruded PE and
mately 500 pulses per second, each of energy W, will XLPE HV and EHV cables has continued to be pro-
impinge on the PE surface. PE has a poor resistance to duced using the original high-pressure reactor process,
high temperature and is readily pitted, dissociated and because this material is inherently clean. Also, the mate-
burnt, thereby eventually forming an incipient electrical rial is capable of being chemically crosslinked, because it
“carbon tree,” which will grow through the insulation. has long, branched side chains, providing available sites
for reaction, which are not present with HDPE.
Q = CV (C) 4-1
Where: 1937: Commercial manufacture of PE commenced, and
Q: Electrical charge (C). quantities were supplied to the British Insulated Cable
C: Capacitance of void (F). Company and the Telegraph Construction and Mainte-
V: Voltage difference across the void (V). nance Company in the United Kingdom and the
DuPont Company in the United States. Its use began in
1 1 the United States in 1942.
W = QV = QG d (J)
2 2
1939: A 1-nautical-mile length of telegraph wire was PE
1 1
W= CV2 = G2 εo εr Ad (J) extruded for sea trials in the United Kingdom.
2 2
4-2
1939: In the war years, LDPE was put into full produc-
Where:
tion, being demonstrated to be a low-loss, low-capaci-
W: Energy liberated by PD in void (J).
tance dielectric, especially suitable for telephony, radio
G: Electrical gradient (stress) across the void (V/m).
transmission, and radar, up to the then maximum fre-
d: Thickness of void (m).
εo: Permittivity of free space (F/m). quency of 3 GHz. LDPE was also used to insulate 2-kV
εr: Relative permittivity of the gas in the void (unity) power supply wires in aircraft electronics, having a suit-
able operating temperature range of -40 ° C to 85 ° C,
(ratio).
although widespread application in aircraft was hin-
A: Area of void (m2).
dered by its high calorific value and susceptibility to
burn. Extrusion and jointing equipment was developed
Polyethylene contains a “free volume” between the long
for LDPE cable.
chain molecules. The presence of microscopic filamen-
tary channels passing through the structure has been
1945: In the years after the war, long-length submarine
postulated by Bahder (Dissado and Fothergill 1992).
telephone cables were insulated with PE, including
Polyethylene is capable of holding some liquids and gas-
trans-Atlantic links.
ses in solution, particularly when hot—for example, 200
ppm by weight of water at STP (standard temperature
1947: Extruded PE insulation was applied to the URD
and pressure).
(underground residential distribution) power cables in
the United States, at 15 kV (and by 1955 at voltages up
4.4 LANDMARKS IN PE CABLE TECHNOLOGY to 35 kV). In the first applications, PE insulations were
extruded directly onto the conductors of power cables in
4.4.1 Evolution the same way as telephone and radio frequency cables
1933: Polyethylene was first synthesized by M. W. Perrin by using “mono extrusion” (a single extruder). To a
and J. C. Swallow of Imperial Chemical Industries UK, greater extent than communication cables, power cables
producing a white powder. Ethylene gas was polymer- produce I2R conductor heating, which caused the poly-
ized in a high-pressure process at 1400 bar and 170°C. A meric insulation to expand off the conductor, forming
trace chemical from an earlier experiment was found to an electrically stressed air gap, initiating partial dis-
be necessary to initiate the reaction. The resulting pow- charge and failure. It should be recalled that, at that
der was reheated and cooled, producing a “white waxy time, conductor shields were not universally used in
solid,” which was later measured to be at the low density paper cables. However, paper has a low coefficient of
end of the PE spectrum, in the range of 57.1-57.7 lb/ft3 expansion, and any gaps present were generally filled
(915-925 kg/m3). with insulating compound or oil, and impregnated
paper has good (self-extinguishing) PD resistance. The
4-9
failure mechanism was investigated using the early par- 1960: Appreciable quantities of PE (HMWPE) insulated
tial discharge detection apparatus described by F. H. URD cables of 5-46 kV had already been installed as part
Kreuger (Kreuger 1964). This apparatus was made pos- of the U.S. postwar house-building program. In the next
sible by the developments of high-gain thermionic tube ten years, these quantities were further increased, mainly
amplifiers and cathode ray oscilloscope tubes before at 15 and 25 kV, until in 1972, PE cables were displaced
and during World War II. by volume installation of XLPE cable. The PE and XLPE
cables were supplied without lead sheaths (now referred
Semiconducting (semi-con) tapes were developed and to as “wet application” designs) in the incorrect belief
applied to shield the conductor, such that they became that PE insulation, in addition to being more robust than
amalgamated to the insulation by the heat of the PE paper insulation, was unaffected by water.
extrusion. In service, the semi-con and PE insulation
expanded together and eliminated gap formation. The 1962: Concentric dual die extrusion heads were devel-
presence of a heat-bonded, continuous and smooth oped, which permitted the simultaneous dual extrusion
shield interface has remained a prime requirement for of the conductor shield and PE insulation in one pass,
extruded cables. minimizing the opportunity for surface damage or con-
tamination. The designs of the new dies were signifi-
Later, it was found that the sharpness of the tape edges cantly more complex, and the different output rates of
were electrical stress raisers, which provided initiation the two extruders required to be carefully synchronized
sites for vented, water tree growth, causing large numbers both with each other and the conductor line speed. The
of service failures. The stress-raisers also directly limited quality of the interface was greatly improved; however,
the design stress for higher voltage cable applications. it was found that holes (skips or holidays) in the
extruded shield could pass hidden by the covering layer
Notionally smooth extruded semiconducting shields of insulation. A present-day example of a shield skip is
were developed by mono-extruding the shield onto the shown in Figure 4-9. Visual checks were possible
conductor in one process pass (i.e., the conductor was through the transparent insulation, if kept hot, but, by
passed reel to reel), followed by extrusion of the insula- their nature, the checks were not continuous. A continu-
tion layer in a second process. Experience showed that ous, on-line “skip detector” was developed, which mon-
that interface, although improved, was still electrically itored the electrical resistance between the tip of the
vulnerable, because it suffered from contamination, extruder die and the conductor (Tanaka and Green-
mechanical scuff damage, and voids at the interface. wood 1983).
1960: The locations of discharging defects were found in An alternative approach, still used today with one type of
URD cables (22 kV) in the United Kingdom and in the EPR cable, is to extrude a high-permittivity stress control
Netherlands (Kreuger 1964) by the development of in- shield in place of the semi-con conductor shield. The
line partial detection equipment, both for purposes of ratio of shield to insulation permittivity is sufficiently
research and routine process testing. Voids were formed high to reduce the stress at the conductor surface to a
by shield defects or by bubbles that escaped the early level that does not produce PD. The shield has sufficient
designs of insulation extruder. After the insulation layer insulating properties after extrusion to permit it to with-
had been mono-extruded onto a tape shield, or dual stand an in-line voltage test similar to the jacket spark
extruded, the unshielded core was passed through the
PD detector, in which voltage was applied to the con-
ductor and reels. The core was passed through pairs of
ground electrodes comprising either metal wheels or
water baths. The electrodes were connected to form a
balanced bridge detection circuit, which enhanced the
detection sensitivity by semi-canceling background
noise. A logic circuit discriminated the presence of void
PD from spurious noise by the detection of a repeat PD
when the void passed over the second electrode. This
technique was also used on fully tape-screened cables
and relied upon the high longitudinal impedance of the
outer shield at the high PD frequencies, electrically seg-
regating the detection electrodes. It was thus possible to Figure 4-9 Hole in conductor shield viewed through
locate voids and either repair the cable or study the transparentized insulation (courtesy of CCI).
voids for research.
4-10
test, thereby ensuring that any prospective holidays are stress at the conductor shield interface. In practice,
detected, prior to the extrusion of the insulation. emission shields have been superseded by improvements
in supersmooth semi-con screens (Champion et al.
In the dual-extrusion technique, the insulation outer 1993). The disadvantage of the dual-tandem process was
shield comprised a coating of graphite lacquer, backed that visual inspection revealed a morphological
by a graphite-loaded semiconducting tape. A final pro- interface between the two insulation layers, and the
cess test was applied to the finished cable, comprising a insulation interface was still vulnerable to particulate
voltage withstand test and a partial discharge test. Many contamination, cooling, and void formation. The
such cable constructions successfully entered service. insulation interface was also vulnerable to macro
Figure 4-10 shows a graphite and tape shield on a PE damage if the soft insulation surface contacted the inner
insulated cable that was extruded in 1960 for long-term edge of the close-fitting die tip in the second dual-head.
evaluation by connection in a substation of the 100-kV
dc England-France submarine cable link. One disadvan- By 1972, the annual installation of LDPE cables in the
tage was that the act of bending the cable during installa- United States had significantly fallen in volume and was
tion risked breaking the adhesion of the graphite layer, largely superseded by XLPE.
thus producing an electrically stressed void.
The French cable industry successfully persevered in the
1963: The dual-tandem extrusion process was developed development of higher-voltage LDPE, installing the
in which one dual-die head applied the conductor shield first 63-kV cables in 1962. High volumes of cable were
together with a thin layer of insulation approximately installed, and the cable voltages were progressively
80-120 mils (2-3 mm) thick. The core then immediately increased, finally installing 400-kV LDPE cable 23 years
passed through a second dual-die head in which the later. Good service performance was experienced, which
bulk of the PE insulation layer and the outer semicon- may be attributed, in part, to the development of clean
ducting insulation shield was extruded. The two insulation and to the protection from water by lead-
dualhead dies were in close tandem proximity, typically alloy sheaths and later by foil laminate sheaths. A failure
18 in. (450 mm) apart, and were connected by a tube rate of less than 0.2/100 km/year was reported by
that protected the intermediate insulation surface from Clavreul and Reignier in 1993, with diagnostics showing
contamination. The connecting tube was longitudinally that 90% of the damages were caused by water penetra-
split to facilitate access for setting-up at the start of tion (Clavreul and Reignier 1993).
extrusion. This technique was successfully used into the
1990s to manufacture 275-kV and experimental 400-kV French cable landmarks include the following:
cables. Alternatively, the first thin layer of PE insulation • 1962: 63-kV LDPE cable was installed.
could be replaced by a high-permittivity insulating layer,
called an “emission shield,” to significantly reduce the • 1967: 90-kV LDPE cable was installed.
• 1969: 225-kV LDPE cable was installed of 295 ft
(90 m) length, comprising a 1500 kcmil (800 mm2)
conductor, 0.9 in. (23 mm) LDPE insulation, semi-
conducting shields of carbon-filled PE, and lead-
alloy sheath and PVC (polyvinyl chloride) jacket.
Produced in a dual-tandem VCV (vertical continuous
vulcanizing) line of 164 ft (50 m) height (Tanaka and
Greenwood 1983).
• 1975-1980: HDPE cables were introduced up to
225 kV and permitted the operating temperature to
be increased from 70°C to 80°C.
• 1985: First 400-kV PE cable was put into service, fol-
lowed by 26 circuits at 12 sites, totaling 20 cable miles
(32 cable km).
• 1990: First 500-kV LDPE was supplied and installed
in China.
Figure 4-10 100-kV dc PE cable with graphite and The application of PE ceased in France in 1994 after 32
taped insulation shield (courtesy of CCI). years, when extruded XLPE cables were introduced at
all voltages.
4-11
4.4.2 Water Trees the manufacturing process. Hot steam diffuses into the
In 1971, within ten years of installation in the United soft insulation and produces microscopic voids of 0.1-2
States of high numbers of 15- and 25-kV extruded cable mil (2-50 μm) diameter with concentrations of 1.6 per
circuits, a rapidly growing failure rate developed. This mil 3 (10 5 per mm 3 ) in a halo shape at mid-insulation.
was mainly in early constructions of PE cable, but also Steam lines are no longer acceptable for HV cables, but
occurred in the newer XLPE cables (Bartnikas and until recent years, some were still in use in different
Srivastava 2000). The failures were attributed to the pre- countries for making URD cable.
viously unknown mechanism of water tree growth (also
called electrochemical tree growth). Within one year, Modern wet designs of MV cables (46 kV and lower) are
water tree failures were also reported in Japan and sub- far less susceptible to failure by water trees, because the
sequently in other countries worldwide. Water tree number and size of stress raisers have been significantly
growth was also found to occur in EPR cables, although reduced by the development of shield compounds with
the fillers make EPR opaque, rendering the task of low ion content and very smooth surfaces, and low-con-
counting trees more difficult. Eventual experience dem- taminant-level insulating compounds and extrusion
onstrated a superior resistance of EPR to water treeing processes. Tree retardant grades, TRXLPE, have been
compared to other dielectrics. introduced for MV cables in wet applications. The great-
est risk of accelerated tree growth is at the conductor
As the failure rate grew, massive R&D resources were shield, and so now most stranded conductors are water
committed to understand and solve the problem. The blocked by a filling compound. The rate of ingress of
contributory causes of the problem are now sufficiently moisture from the ground is either reduced by the use of
well understood for engineering solutions to have been an MDPE/HDPE jacket in place of a PVC jacket or
taken and proven by long-term qualification tests and eliminated by use of a metallic foil water barrier.
service experience. However, tree growth is complex,
and scientific consensus has not converged to agree on a The water tree problem had the detrimental effect of
definitive mechanism. The initial part of the mechanism slowing the development of extruded cables for trans-
is likely to be the concentration of water molecules by mission voltages. However, the beneficial effects were:
dielectrophoretic force at electrical stress raisers. Dielec- (a) accelerating the development of improved extrusion
trophoretic force is the force experienced in an alternat- equipment, cleaner insulation, and smoother shields; (b)
ing electric field by an uncharged particle of different demonstrating the benefits of stringent qualification
permittivity to the insulation, such as a water molecule. and long-term tests before widespread application; (c)
Tree-shaped, water-filled micro-channels grow in the the further development and application of radial mois-
direction of the electric field, and eventually an electrical ture barriers and (d) greatly increasing the fund of
carbon tree is initiated, with electrical failure following. research knowledge on polymeric cables.
Many stress-raising sites were freely available in early
extruded cables in the form of: (a) asperities (protru-
sions) on the surfaces of taped conductor shields, which
are focal points for “vented tree” shapes (Figure 4-11,
left), and (b) particulate contaminants present in the
insulation, which are focal points for “bow-tie” trees
(Figure 4-11, right). Contributory causes were found to
be ionic salts that could migrate into the PE insulation
in aqueous solution from the extruded semi-con shield
materials and from the surrounding ground. Accelerat-
ing causes of tree growth in both PE and XLPE cables
were: (a) an insulation moisture level of approximately
80% of the saturation level, (b) an electrical design stress
of greater than 51 V/mil (2 kV/mm), and (c) a medium
to high operating temperature.
The “normal method” of manufacturing XLPE cables

was in a steam-filled CV line. These cables were vulnera- Figure 4-11 Vented and bow tie trees in MV XLPE cable
(courtesy of Blackwell Publishers).
ble, because water-filled microvoids were produced by
4-12
4.5 XLPE CABLE: BENEFITS AND cable size, die dimensions, shield thickness, and
CHALLENGES extruder output rate.
The following benefits and challenges of XLPE cable • The extrusion run time is dictated by the time at
are in addition to those described for PE in Section 4.2. which “ambers” (heat-oxidized PE) are released by
the extruders. Ambers initially pass through the fil-
4.5.1 Benefits ters into the insulation, and if the run time is too
• Increase in the cable operating temperature from long, they risk blocking and bursting the filters.
70°C for PE to 90°C, thus increasing allowable tem- • The pressure applied to the cable in the extruder and
perature rise from a 20°C ambient by 40% and CV tube increases the risk of the melt flowing into
ampacity by 18%. the conductor between the wires (“fall-in”).
• Increase in the emergency operation temperature to • In-process scrap is increased, because a significant
105°C. length of unusable cable core has to pass through the
CV tube at the start of an extruder run before full
• Increase in the short circuit temperature to 250°C. pressure can be applied and the required cable dimen-
sions are achieved. Similarly the trailing length of
4.5.2 Challenges cable core is unusable, because CV pressure and core
The challenges are a consequence of crosslinking and geometry are lost when the trailing end passes
illustrate the sophistication required of modern cable through the extruder.
manufacturing processes: • The extruded melt has more time to flow off the con-
• The crosslinking process requires a synchronized CV ductor (“droop”) and so lose the circular, concentric
(continuous vulcanizing) line; this is a large machine shape.
requiring significant capital investment. • The recirculated heat transfer mediums in the CV
• The crosslinking process is comparatively slow and tube require cleaning processes to remove
reduces factory output. (a) crosslinking by-products that have diffused from
the core and (b) gels formed by partial crosslinking of
• The crosslinking process is a chemical reaction that
silicone fluids by diffused peroxide.
increases dependence on high standards in material
and manufacturing quality control. • The CV tubes in VCV (vertical continuous vulcaniz-
ing) and in CCV (catenary continuous vulcanizing)
• By-products of crosslinking are produced in the insu-
lines require periodic cleaning to remove acetophe-
lation. An additional time-consuming degassing pro-
none, which condenses on the surface as a fat.
cess is required to: (a) remove moisture, (b) prevent
gas pressure assisting voids to pass through factory • The generic name “extruded XLPE cable” does not
HV test into service, (c) reduce the volume of meth- adequately describe the variations that exist in mate-
ane/ethane gas, thereby reducing fire risk in jointing rial formulations, purity, cable geometry, extrusion
vaults, and (d) reduce pressure differentials that line types, and process conditions. For example:
might displace accessory insulation. —Insulation grades are available with increased melt
• The high crosslinking temperature increases the insu- viscosities to reduce droop in CCV and MDCV
lation contraction strain that acts at accessories. (Mitsubishi Dainichi continuous vulcanizing) lines.
• The residual core temperature at the end of the CV —The heat transfer media in the CV tube tend to
process locks the take-up reel curvature into the insu- impregnate the XLPE insulation. These media
lation. A heat straightening process is required imme- include dissolved nitrogen gas in VCV and CCV
diately prior to the assembly of the accessories. lines, silicone oil in MDCV lines (lubricant) and in
some CCV lines (buoyancy aid), and water in the
• The increased risk of impurities and voids in the insu- earlier generation of “wet” VCV and CCV steam
lation requires increased process control and sample lines.
testing.
• The heat-sensitive crosslinking agent in the PE 4.6 GENERAL DESCRIPTION OF XLPE
“melt” risks premature crosslinking (scorch) down- INSULATION
stream of the extruder, both in the strainer pack and
The act of crosslinking long-chain molecules together
in the semi-con extrusion dies. One consequence is
changes the morphology of LDPE from a thermoplastic
that the melt cannot be filtered to such a high purity.
to a thermoset. A thermoset material is one that cannot
Another is that the process must be kept within a
be remolded into a different shape by the application of
“parameter window” calculated for the particular
4-13
heat. Up to the melting temperature of 107°C, the prop- EVA (ethylene vinyl acetate) crosslinkable resin to give
erties of XLPE retain similarities with those of LDPE. maximum flexibility (high elasticity) to match bending
Figure 4-7 shows that the volumetric thermal expansion properties when it is carbon filled with those of the
of XLPE is the same as LDPE, but above the melting XLPE insulation. The shield resin is filled with a high
point, the coefficient of thermal expansion reduces in carbon black loading of typically 30-60% to achieve a
magnitude and becomes constant. The morphology is stable, high plateau level of electrical conductivity. Mea-
still semicrystalline at room temperature. The presence surements show that the filled and crosslinked shield
of the crosslinks results in parallel concertinalike pat- compound behaves as an elastomer across the working
terns, called “lamellae,” as shown in Figure 4-5, instead temperature range in a manner similar to filled EPR
of round “crystallite” patterns. In XLPE, the amor- insulation.
phous material between the lamellae is elastomeric (a
rubber) and is no longer a viscous gel. The mixed morphology gives the XLPE insulation vis-
coelastic properties. Below the crystalline melting point
The significant effects of crosslinking LDPE are that of 107 ° C, XLPE has a high modulus (inelastic) and
above 107°C, although the material still becomes trans- requires a high load to produce viscoelastic deflection,
parent and clear, it changes to a soft elastomer (rubber) with a long time constant. For example, under 1.35 MPa
and not a liquid. XLPE remains in this elastomeric state tensile load at 70°C, XLPE reaches a plateau extension
up to the ultimate thermal degradation temperature of in a period of 500 hours. Under 0.2 MPa tensile stress at
250°C. Also, unlike thermoplastic LDPE, XLPE cannot 130°C, XLPE exhibits most extension in a period of 0.5
be dissolved by an aggressive hydrocarbon solvent such hours and reaches a plateau within a period of 20 hours.
as Xylene or Decaline. The solvent test can be used to
assess the degree of crosslinking of the cable insulation; Figure 4-13 shows the deflection at different times of the
however, the preferred quantitative tests are the tensile same XLPE tensile test sample with an initial gauge
elongation and hot set tests described later. length of 20 mm at 130°C, at which temperature XLPE
is transparent. The photograph on the left is the deflec-
Figure 4-12 shows the modulus of elasticity in tension— tion upon first application of load. The center view is the
i.e., the ratio of stress/strain. This is the initial modulus increased final plateau value after viscoelastic deflection.
immediately after applying load and before viscoelastic The right-hand view is the elastic retraction that has
deformation strain has occurred with time under stress. occurred after removal of load, which occurred in 0.25-
The modulus is high at ambient temperature (inelastic) 0.5 hours. It can be seen that almost perfect retraction
and falls at an increasing rate with temperature up to has occurred—i.e., that tensile hot-set is minimal.
107°C, above which it has a low and comparatively con-
stant modulus (high elasticity). In comparison, EPR is The tensile deflection of XLPE above the melting point
an elastomer (rubber) at all temperatures in the working is used as the basis for the routine factory hot-set test to
range and has less variation of modulus with tempera- check that sufficient crosslinking has been achieved in
ture. The semiconducting shield compound in an XLPE samples cut from the cable insulation. ANSI/ICEA T-
cable usually has an EBA (ethylene butyl acrylate) or 28-562-2003 (ANSI/ICEA 2003) requires that the die-
Figure 4-12 Modulus of elasticity for XLPE and EPR with temperature
(courtesy CCI).
4-14
cut dumbbell specimen be heated at 150 ° C under a within a premolded EPR accessory is heated to above
stress of 29 psi (0.2 MPa) for 15 minutes and the per- 107°C, the softened elastomeric XLPE will experience
centage hot creep elongation be measured. The load is increased compressive deformation that will also be
removed, and the sample is eventually allowed to cool to locked-in when the cable cools and contracts, resulting
ambient, when the percentage hot creep set residual in a reduction in the interfacial pressure in normal oper-
elongation is measured, both parameters being in rela- ation and increasing risk of partial discharge.
tion to the original gauge length at ambient temperature
of 1 in. (25 mm). ICEA S-108-720-2004 gives pass levels Taking into account the highly nonlinear morphology
of maximum elongation of less than 175% and hot set of XLPE in terms of radial expansion, modulus of elas-
less than 10%. The IEC specifications differ slightly in ticity, and crystalline memory, the consequence of oper-
that they require that the samples be heated to 200°C ating an EHV XLPE cable at emergency overload
for 15 minutes under the same stress of 0.2 MPa. The temperatures in excess of the normal operating tempera-
pass levels are that the elongation under load should be ture of 90 ° C is to significantly increase the risk of
less than the same value of 175%, but that the hot set mechanical deformation and electrical failure. It is
should be less than 15%. This test is used as a routine noted that ICEA S-108-720-2004 (ICEA 2004) limits
production test of the degree of crosslinking. Viscoelas- the emergency overload temperature to 105 ° C for
tic behavior may be approximately characterized as a greater than 138 through 345 kV and to 105 to 130°C
mechanical model with a perfect, elastic spring in paral- for greater than 46 through 138 kV. ICEA gives a cau-
lel with a dash-pot, or as an electrical model with a tionary note to utilities on not operating cables at ele-
resistor (R) in series with a capacitor (C), the time con- vat e d t e m p e r at u r e wi t h o u t ap p rova l f ro m t h e
stant being RC. manufacturer. The more recent AEIC CS9-06 (AEIC
2006) gives a more sensible engineering approach at this
If XLPE is held elastically deformed under mechanical time by limiting all XLPE and EPR cables to an emer-
load in the elastomeric state above 107 °C and is then gency temperature of 105 ° C. AEIC notes that only
cooled, the formation of the semicrystalline structure 130 ° C may be used for EPR cables in the range 46-
freezes the deformation and locks it in the deformed 150 kV with mutual agreement between purchaser and
state. The elastic memory of the locked-in strain can be manufacturer.
released by reheating XLPE again to above 107°C, this
being the principle of a heat shrink sleeve. Thus if an 4.6.1 Chemical Crosslinking Process
XLPE cable is heated to above 107°C during emergency The chemical crosslinking process for HV and EHV
loading and conductor thermomechanical sidewall XLPE cables is to use an organic peroxide. Dicumyl per-
forces produce insulation deformation in a bend (thin- oxide is preferred, because it has a low decomposition
ning), the deformation will be held locked into the insu- temperature of 140°C, thus giving a margin above the
lation permanently below 107°C. Similarly, if the cable
Figure 4-13 0.2 Mpa tensile stress on XLPE samples at 130oC. Left: Initial deformation.
Center: Plateau deformation. Right: Unloaded “hot set” (courtesy of CCI).
4-15
120-130°C extrusion temperature for the LDPE com- 3. Two CH3 radicals from item 2 combine to produce
pound. ethane.
4. The peroxide radical takes a hydrogen atom from the
The dicumyl peroxide molecule has two symmetrical antioxidant and produces the following by-products
parts joined by an oxygen-to-oxygen covalent bond. in the insulation:
This bond is inherently unstable with temperature.
a. Cumyl-alcohol
Upon heating to above 140°C, the O-O bond is broken.
The peroxide is said to decompose. The two peroxide b. Anti-oxidant inert by-products
entities are negatively charged ions called “free electron 5. Two free radical LDPE chains produced by items 1
radicals” in which the oxygen is highly reactive. The free and/or 2 combine to produce XLPE.
radicals have the ability to initiate crosslinking between
the hydrogen atoms at the ends of the LDPE side chains. The total weight of the by-products is equal to the initial
Hydrocarbon by-products are produced, some of which weight of the dicumyl peroxide in the compound, which
react with the antioxidant additive included to protect is typically 2-2.5%. Typical weights of the nongaseous
the XLPE from thermal aging by oxidation during by-products in XLPE are 1.4% cumyl-alcohol and 0.5%
extrusion and in-service life. High reactivity rates and acetophenone from a 2% addition of dicumyl peroxide
short process times are achieved at 160°C and higher. (Wartusch and Wagner 1977). The gaseous by-products
are methane, ethane, and water vapor, and these diffuse
A number of reactions are necessary, as detailed below, into the surrounding space with time. If the total mass
to remove the hydrogen atoms from the ends of the of the gas mixture is taken to be 0.1%, it would occupy a
LDPE side chains, leaving them chemically reactive to volume approximately 10 times greater than the insula-
join together and form XLPE (BICC 1997, pages 59- tion volume at STP. The water content of XLPE is
60). Each decomposition of a peroxide molecule pro- approximately 200 ppm (0.02%) at 160°C and 500 ppm
duces a by-product that is left in the insulation. Suffi- (0.05%) at 180°C (Ishikawa et al. 1994).
cient peroxide has to be added to the LDPE to achieve
the required crosslinking density and to allow for loss in The physics and chemistry of the extrusion and
efficiency due to parallel and secondary reactions. The crosslinking processes are simulated in Figures 4-14 and
chemical reactions are listed below. The reactions that 4-15. Figure 4-14 (left) shows granules of compound
are necessary to produce crosslinking are items 1, 2, and that have been melted into liquid PE and heated at
5. Eight by-products are produced, which are listed in 160oC to crosslink them. A steel weight applied gravita-
items 1 to 4: tional pressure to the granules to expel entrained air and
1. The peroxide radical extracts an H atom from the either expel or dissolve gas produced by crosslinking.
end of a side chain of LDPE. This makes the PE Figure 4-15 (left) shows that few bubbles remained. In
chain reactive (a free radical). The following by-prod- comparison, Figure 4-14 (right) shows granules at
ucts are left in the insulation: 160oC without the assistance of pressure from the steel
weight. Figure 4-15 (right) shows the sample after cool-
a. Cumyl-alcohol
ing and dissection. It can be seen that the PE granules
Cumyl-alcohol may also decompose with heat to
without external assistance do not consolidate, and the
produce:
gaseous by-products have generated many large voids
b. Alpha-methyl styrene and water.
2. The peroxide radical decomposes in the first stage by
losing an oxygen atom and a free-radical CH3 mole-
cule. In a second stage, the CH3 radical takes an H
atom from the end of the LDPE chain, causing the
PE chain to become a free radical. The following by-
products are left in the insulation:
a. First stage reaction:
i. Acetophenone
ii. Oxygen (to probably combine with the anti-
oxidant).
b. Second stage reaction: Figure 4-14 Granules of XLPE after melting at 160oC.
i. Methane Left: Melt after consolidation by a metal weight
(removed). Right: No weight applied. (Courtesy of CCI)
4-16
similar to those seen in leavened bread. Additionally, 4.7 LANDMARK DEVELOPMENTS IN XLPE
the XLPE sample containing voids in Figure 4-15 CABLE TECHNOLOGY
(right) has become discolored by oxidation, whereas the
void-free sample on the left has remained white. The 4.7.1 Worldwide Evolution
high thermal contraction that occurs when XLPE cools 1952: The first samples of XLPE were produced by A.
and crystallizes is shown by the formation of the deep Charlesby, who irradiated PE in a nuclear reactor at
concave meniscus in Figure 4-15 (left). Harwell, United Kingdom. It was noted that the new
material would no longer dissolve in hot solvents, and
In the manufacturing process, the pressure from the upon heating to its melting point, it became an elas-
extruder screw amalgamates and consolidates the gran- tomer instead of a liquid. The useful temperature range
ules to form a void-free melt. During the crosslinking could thus be increased. Cable makers installed experi-
process, it is necessary to apply an external pressure to mental crosslinking plants using high-energy electron
keep the gaseous by-products dissolved in the polymer. beams. The radiation technique was developed into
A pressure of typically 145psi (10 bar or 1 MPa) is commercial application for thin insulation wires and
applied, either by nitrogen gas, silicone fluid, or contact heat shrink sleeves, but practical difficulties precluded
pressure generated within a long-land die. After its continued development for thick-wall power cables in
crosslinking, it is necessary to maintain the external factory environments.
pressure until the XLPE insulation is sufficiently cool
for the crystalline morphology to have gained the 1954: Material suppliers and cable makers were
strength to keep the internal gas in solution. If the exter- researching an alternative chemical crosslinking route to
nal pressure is too low, or the core leaves the CV tube produce XLPE.
too hot, the gas creates bubbles in the soft insulation
similar to those produced when a bottle of carbonated 1954: Cable makers researched the type of manufactur-
water is opened. Bubbles form preferentially at micro- ing equipment necessary to chemically crosslink XLPE.
voids and particulate impurities. Existing thermoplastic PE cables were extruded hori-
zontally at approximately 120-160°C and then quickly
The dissolved gasses are encouraged to diffuse safely cooled to a solid, before the liquid could drip off the
from the cable by a subsequent de-gassing process (also conductor. Existing rubber cables were already being
called a drying process) in which the cable is gently continuously vulcanized (CV) with sulfur additives by
heated for a prolonged period to approximately 60- being extruded vertically downwards while being heated
80 ° C. If a sample of freshly extruded un-degassed and pressurized in direct contact with steam at tempera-
XLPE core of 0.9 in. (23 mm) thickness is quickly tures of approximately 200°C, within a vertical vulca-
heated to above 107°C, gas bubbles form in the insula- nizing tube (VCV line). Steam heating was adopted as
tion in the way that melting ice cubes of carbonated the standard method to crosslink XLPE cables. Fifteen
water liberate carbon dioxide bubbles. years later, steam curing became known as “wet extru-
sion” when it was discovered that steam produced
micro-voids in the XLPE, many of which were filled
with water.
Catenary continuous vulcanization (CCV) lines were

developed to overcome the VCV line limitations of the
high cost of building a tall tower and the limit that the
height sets to the maximum production output speed.
During passage through the CV tube, the cable has to be
heated to the crosslinking temperature, held at that tem-
perature for sufficient time to achieve the chemical reac-
tion, and then cooled for sufficient time to regain
mechanical strength; these times dictate the maximum
output speed.
In a CCV line, the extruder is mounted in a low height

tower and is angled downwards such that the insulated
Figure 4-15 Polyethylene melt cooled to room
temperature Left: Compressed sample. Right: conductor passes into a similarly downward angled CV
Uncompressed (Courtesy of CCI) tube. The curved tube is straightened into a horizontal
shape before slightly rising again to form an asymmetric
4-17
catenary; in this way, a total CV tube length of 100- making XLPE cable competitive with fluid-filled paper
250 m can be achieved within the height of the factory and cables.
roof. The PE melt insulation has a reduced tendency to
drip off the conductor compared to a horizontal CV 1965: 138-kV XLPE system was installed in Puerto
line, but still produces an eccentric, oval-shaped insula- Rico, the first such system in the North American
tion. The degree of ovality is acceptably low for thin region.
insulation on a large diameter conductor, but becomes
unacceptable when either the thickness is increased or 1965: XLPE cable was installed in the United States, fol-
the conductor diameter is reduced. lowed in the later 1960s by an installation in Canada.
1962: The “long land die” extrusion process was pat- Semiconducting shields were initially extruded from a
ented in the United States and became known as the graphite-loaded XLPE compound; however, this com-
Anaconda process. It was developed by R. F. Hinderer pound was stiffer and had inferior bending properties to
and J. A. Mullen. The PE melt is extruded into a stan- the XLPE insulation. Later graphite-loaded EVA com-
dard conical die, which previously had a cylindrical out- pounds were introduced to give better matched bending
let (called the “land”) of approximately 0.5-2 in. (10-50 properties to the XLPE insulation. It was found to be
mm) length. In the long-land die process, the land is difficult to exclude shield asperities (semiconducting
extended by a bolt-on die tube of approximately 30 ft protrusions into the insulation) of less than 6 mil
length (10 m). The process is a continuous molding pro- (150 μm) (Tanaka and Greenwood 1983, page 29).
cess in which uncrosslinked PE liquid insulation and
semiconducting shields enter the tube, and crosslinked 1972: First dry-cure line (no steam). This dramatically
insulation and shields leave the die tube to pass through reduced the size and number of voids, making possible
a loose-fitting water-cooled section. The die-tube is performance at higher electrical stresses and transmis-
heated to the specified crosslinking temperature, and sion voltages. Dry methods gave 100 ppm by weight of
heat is efficiently transfered by direct conduction to water content for an undegassed dry cured cable com-
raise the temperature of the insulation and conductor. pared to 200 to 300 ppm for a steam-cured cable (BICC
To prevent the soft insulation from sticking, a lubricant 1997).
is injected inwards through a porous section of tube
onto its surface. 1976: 138-kV XLPE cable was installed by Detroit Edi-
son.
Mitsubishi Dainichi refined the long land die continu-
ous vulcanizing process to produce the MDCV line, 1978: Swedish 230-kV XLPE cable was installed in
which is a modification of the HCV (horizontal continu- Libya.
ous vulcanization) and CCV (catenary continuous vul-
canization) processes to improve concentricity (Tanaka 1992: First 230-kV XLPE system was installed in the
and Greenwood 1983, page 47.) This process includes United States, in Florida.
two improvements: (a) an additional extruded poly-
meric/elastomeric semiconducting shield layer, not con- 2000: First 345-kV XLPE short cable installation with-
taining peroxide, is extruded over the PE cable core to out joints in the United States, in Boston.
limit the unwanted crosslinking of the lubricant from a
liquid to a viscous gel, and (b) the lubricant silicone oil, 2006: First 345-kV XLPE cable and joint installation in
having a high temperature and peroxide resistant perfor- the United States, in Connecticut.
mance, is injected through the die wall onto the outer
core layer. Tubes are now 40-60 ft (12-18 m) long. Con- 4.7.2 Japanese Landmarks in XLPE System
centricity is achieved by the use of a low-melt index Technology
XLPE compound (Bartnikas and Srivastava 1987). 1960-70: Japanese power utilities embarked on a long-
term national industry program to develop XLPE
1963: Installation of XLPE cables in the United States power cables at 66, 154, 275, and 500 kV. The key items
by GE and other companies. Material was licensed to in the program were development by the chemical
three material producers for general supply to the cable industry of super clean polyethylene and development
industry. Crosslinking changed PE from a thermoplastic by the cable industry of manufacturing equipment capa-
to a thermoset, which instead of melting at 107 ° C, ble of precision extrusion of superclean insulation and
becomes an elastomer up to 250 °C. This permits the supersmooth shield interfaces. Large quantities of
normal operating temperature of 70°C for LDPE cable 66-kV cables were initially extruded from steam-heated
and 80°C for HDPE to be increased to 90°C for XLPE, CV lines. The routes of 275-kV and 400-kV major trans-
4-18
mission circuits were planned many years in advance in ever, the conductors were small, and the circuit lengths
co-ordination with the Departments of Transport, such were short, with no joints. Longer lengths of larger con-
that, where possible, cable tunnels and major freeways ductor sizes with joints were being installed 10-15 years
were constructed together, thus providing direct access later in 1990-1995. Care is required in interpreting the
to cities. The following technological landmarks date of application, because long cable circuits take sev-
occurred in Japan (Sato et al. 1993): eral years to manufacture, install, and commission. For
• 1965: Tandem extrusion. example, the long-length 500-kV circuits in Japan that
commenced installation in 1996 completed commission-
• 1966: Dual tandem extruder group. ing in 2001.
• 1971: Superclean compound for 154-kV cables.
• 1972: Gas curing process and triple head developed. The successful installation of the first 380- and 400-kV
XLPE cable circuits in Europe was based on the studies
This reduced the void size restriction present with
of CIGRE working groups and in particular Working
steam cure from less than 2 mil (50 μm) to far less
Group 21.03, set up in 1990. It was recognized that
than 0.4 mil (10 μm). MDCV is approx 0.04 mil (1
progress from less than 150-kV system voltages to above
μm). (Tanaka and Greenwood 1983, page 48).
150 kV, and in particular at 230 kV and above, was a
• 1975: Continuous stable extrusion line. large technological extrapolation in terms of increased
• 1977: 154-kV cables installed, based on long term electrical stress, cable size, accessory performance, and
tests on lower voltage cables using the dry curing pro- thermomechanical force, with little service experience to
cess. base it on. Experience at the lower transmission voltages
• 1978: Extra-clean compound for the 275-kV cable on the whole had been good. However, isolated cases
existed of cable failures due to extrusion defects and of
system.
accessory failures associated in part with design and in
• 1982: Ultra-clean compound for the 500-kV cable particular with workmanship, installation methods, and
system. contracts with divided responsibilities. The objective of
• 1988: World’s first 500-kV XLPE cable-only circuit CIGRE WG 21.03 was to extend the range of IEC
installed (Figure 4-2). 60840 for routine and qualification tests on extruded
• 1993: 275-kV long-length cable circuit with joints cables from 150 to 400 kV. WG 21.03 studied the two
countries with leading technology and experience at
commissioned. Figure 4-16 shows a low-loss cable.
higher voltages: France with extruded LDPE/HDPE
• 2001: 500-kV long length cable circuit with joints cable, and Japan with extruded XLPE cable. Both coun-
commissioned. tries had developed a power law rationale to relate the
duration of long-term tests to service life for complete
4.7.3 Commercial Applications of XLPE Cable cable systems—i.e., cables with accessories in a configu-
Systems ration that simulated installation conditions. WG 21.03
Table 4-1 shows that commercial applications of XLPE published their recommendations in ELECTRA 151 in
cable at 220, 230, and 275 kV began around 1980. How- December 1993 and extended them to 500 kV in ELEC-
TRA 193 in December 2000. Electra 151 recommended
to utilities and manufacturers that development, rou-
tine, prequalification, and qualification tests be per-
fo r m e d . I n p a r t i c u l a r t h e d u r a t i o n o f t h e
prequalification test is specified to be 365 days. Utilities
specified ELECTRA 151 for a period of eight years until
superseded in 2001 by IEC 62067 (IEC 2001). AEIC CS
9 (AEIC 2006) specifies that the prequalification tests
shall meet the IEC 62067 test requirement.
In the 1980s and 1990s, it was found that dc high-

voltage, after-laying tests that had evolved with paper
cables were insensitive in the detection of gross mechan-
ical damage to the extruded cable and assembly errors in
Figure 4-16 Japan—first 275-kV XLPE, 5000 kcmil the accessories, such that failures occurred after weeks
(2500 mm2), insulated strand segmental conductor,
to months in service. Additionally the application of the
corrugated stainless steel sheath, long-length cable,
with joints, 1993 (courtesy of J-Power Systems). dc voltage was considered to be detrimental to both the
4-19
Table 4-1 Commercial Applications of XLPE Cable by Voltage, Conductor Size, and Circuit Length
Voltage Conductor Circuit Length Country of
kV Year kcmil mm2 ft m Joints Installation
138 1965 - - - - - Puerto Rico
1976 - - - - - USA
220 1981 1500 800 140 44 No Japan

1984 1250 630 6287 1917 - Sweden
1990 1250 630 577 177 No Canada
1992 4000 2000 727 222 No Japan
1992 5000 2500 170 52 No Japan
1995 3000 1600 154,860 47,202 Yes Qatar
230 1978 750 400 327 100 No Libya

1985 2500 1200 547 167 No USA
1992 2500 1200 - - No USA, Florida
1994 2500 1200 25,150 7667 Yes India
1997 1000 500 4757 1450 - USA
1998 4000 2000 108,267 33,000 Yes Singapore
1999 2000 1000 9247 2819 Yes USA
275 1979 1750 800 423 130 No Japan

1981 1750 800 4037 1231 - Japan
1986 4000 2000 5783 1763 - Japan
1990 5000 2500 39,623 12,078 Yes Japan
345 2000 1200 630 1757 536 No USA

2000 2500 1200 13,123 4000 Yes Iran
2000 5000 2500 70,863 21,600 Yes Taiwan
2006 1750 800 10,500 3200 Yes USA
380/400 1995 750 400 7073 2157 - Saudi Arabia

1997 3000 1600 69,770 21,267 Yes Denmark
1998 3500 1800 293 90 - Germany
1999 3500 1600 22,417 6833 Yes Germany
2000 5000 2500 36,743 11,200 Yes Saudi Arabia
2002 2500 1200 91,863 28,000 Yes Denmark
2004 5000 2500 83,990 25,600 Yes Spain
2005 5000 2500 68,897 21,000 Yes UK
2005 2500 1200 34,120 10,400 Yes Austria
500 1988 1500 800 623 190 No Japan

1993 1500 800 7873 2400 Yes Japan
1996 5000 2500 65,103 19,844 Yes Japan
2000 5000 2500 261,154 79,600 Yes Japan
cable, because of space charge accumulation, and to the fidence that XLPE cable systems could commence com-
accessories, because of application of high electrical mercial installation with a reasonable prospect of
stress formed by the large differences in dc resistivity. predictable good performance.
AC after-laying tests were shown to be effective, and
CIGRE WG 21.07 published test levels in ELECTRA Commercial 380- and 400-kV XLPE systems were
173 in 1997, which were also incorporated into IEC installed from 1995 onwards, with key events being the
62067. Utilities and manufacturers could then have con- first buried installation in Copenhagen, Denmark, in
4-20
1997 (Andersen et al. 1996) and two tunnel installations accessories, installation, manufacturing plant, and
in Berlin, Germany in 1998 and 1999 (Henningsen et al. materials warrants performing a new one-year prequali-
1998). The latter applications are of particular signifi- fication test. The tests are expensive and delay projects.
cance, because the utility BEWAG sponsored two series CIGRE, therefore, has published a revision of the quali-
of prequalification tests in Italy at the CESI Milan labo- fication procedures in TB 303 (CIGRE 2006). The revi-
ratories. These were based on the tests recommended in sion describes an “extension of qualification” test
ELECTRA 151 (CIGRE 1993). Six manufacturers took procedure for changes in designs in EHV cable systems
part in parallel tests in each series. Each manufacturer including their cables and accessories. The revision also
assembled a 100-m test loop with samples of each type recommends a prequalification test procedure for HV
of accessory and samples of each type of installation cable system design with high stresses when no earlier
method (buried, tunnel, and duct). The first test series experience is available.
showed that the accessories, and in particular the joint,
were a major weakness with only one manufacturer 4.8 ELECTRICAL DESIGN STRESSES FOR
completing the tests, which was due to the robust design XLPE SYSTEMS
of a pre-fabricated composite straight joint. The first
An EPRI technology review of XLPE EHV cables, 220
test series showed that factory-prefabricated joint insu-
to 500 kV, was performed in 2002. Figure 4-17 was com-
lation had a superior performance to hand-taped insula-
piled from the surveys of utilities and cable manufactur-
tion in resisting disturbance from both heat and stress
ers and from published data. Conductor shield and
deformation and from conductor thermomechanical
insulation shield stresses are increased at higher system
forces. Designs of prefabricated accessories were
voltage. This is driven by the need to ship a worthwhile
installed in the second test series, and these exhibited a
length on a reel and to reduce the number of joints. Reel
greatly improved performance. Even then, only one
lengths of 1600-5000 ft (500-1500 m) have been achieved
manufacturer completed the second test series without
up to 500 kV.
distress or failure. Another four manufacturers were
permitted to remake existing accessories, or insert new
Figure 4-17 shows data for all circuits, regardless of cir-
accessories (joints and terminations) to eventually and
cuit length and whether joints are installed. Figure 4-18
fully complete the tests. The sixth manufacturer with-
is for circuits containing joints.
drew after suffering repeat cable failures, which resulted
from defects in the extruded shield, but later progressed
to remake the cable at a different factory and then to
successfully complete a third test series.
The majority of utilities have since required evidence of

CIGRE 151, CIGRE 193, or IEC 62067 long-term
prequalification tests to demonstrate compatibility of
cable, accessories, and installation design. IEC then
amended standard IEC 60840 (IEC 2004) to include
combined tests for accessories with the extruded cables
for system voltages from 30 to 150 kV.
North American standards ICEA S-108-720-2004 and Figure 4-18 Conductor and insulation shield stresses
increase with transmission voltage.
AEIC CS9-06 were issued in 2004 and 2006, respec-
tively, and share the CIGRE and IEC approach of per-
forming qualification and prequalification tests. A
significant difference is that in Europe the manufacturer
is free to optimize and to take responsibility for the
design of the cable system, providing evidence of IEC
tests are supplied, whereas in North America, the stan-
dards also specify the cable design stress limits and the
dimensions of the cable, leaving the manufacturer with
the responsibility for quality control.
Manufacturers and utilities internationally have found Figure 4-19 Conductor and insulation shield stresses
difficulty in deciding when a change in cable design, are lower in circuits that contain joints.
4-21
The figures show: Figures 4-19 to 4-21 show the variation of conductor
• From 220/230 kV to 500 kV, the design stresses have and insulation shield stress with conductor size for
transmission voltages of 220, 345, 400, and 500 kV. As
nearly doubled.
would be expected, the conductor sizes are smaller at
• Conductor shield stresses are generally lower in cir- 220 kV, extending from 750 to 4000 kcmil (400 to
cuits that have joints installed, but there is little effect 2000 mm2), whereas at the higher voltage, they extend
on insulation shield stresses. from typically 1000 to 5000 kcmil (500 to 2500 mm2).
• The 400 and 500-kV stress levels represented the then
“state of the art”; the first major 400-kV system was Figures 4-20 and 4-21 show that, at small conductor
commissioned in 1997, and the first major 500-kV sizes, there is a wide difference between the conductor
system in 1996. and insulation shield stresses. At larger conductor sizes,
• The design stresses specified in the new standard the difference reduces. The conductor shield stress
reduces, and the insulation shield stress increases. The
AEIC CS9-06 are in accord with the international
reasons are:
experience shown in Figure 4-18.
• For a small conductor radius, the logarithmic equa-
Conductor and insulation shield stresses for a selection tion for stress distribution described in Chapter 3
of projects installed throughout the world are shown in produces a high conductor stress and a low shield
Table 4-2. stress. Chapter 3 also notes that it is uneconomic to
design a cable with an insulation-to-conductor-
Table 4-2 Design Stress and Insulation Thickness in Order of Transmission Voltage
Nominal Insulation Conductor Insulation shield

Voltage Conductor Year Thickness Stress Stress
[kV] mm2 In Service Country mm kV/mm kV/mm
500 2500 2000 Japan 27 14.6 8.1
400 1600 1997 Denmark 32 10.9 5.0
400 1600 1998 Germany 29.5 11.5 5.6
400 1600 1998 Germany 27 12.2 6.2
400 1600 1999 Germany 29.5 11.5 5.6
400 1600 1999 Germany 27 12.2 5.6
400 800 2000 Abu Dhabi 29 12.7 5.3
275 1200 1998 Australia 27 8.9 4.1
230 2000 1998 Singapore 24 7.2 4.0
230 1000 1999 USA 24 8.1 3.9
230 1200 2002 USA 27 7.5 3.4
220 1000 1999 Spain 24 7.8 3.7
220 1600 1999 Ireland 25 7.0 3.4
Figure 4-20 Variation in conductor and insulation shield Figure 4-20 Variation in conductor and insulation shield
stresses with conductor size at 220 kV. stresses with conductor size at 400 kV.
4-22
in the paper by Bahder et al. 1976 that voltage break-

down originates at imperfections in the insulation struc-
ture of the cable and results in local high-voltage
stresses sufficient to cause ionization. When this ioniza-
tion is of such magnitude as to produce an electron ava-
lanche, an electrical tree will develop and propagate
rapidly, constituting an electrical breakdown of the
insulation. The most critical imperfections, in decreas-
ing order of importance, are:
• protrusions at the conductor shield
Figure 4-21 Variation in conductor and insulation shield
stresses with conductor size at 500 kV. • contaminants in the insulation including conductive
particles and insulating particles having a signifi-
cantly higher conductivity than the insulation
shield-diameter ratio greater than the “e-ratio” of
2.72. To avoid having to select an unnecessarily large • voids that discharge during the application of the
conductor, manufacturers increase the conductor voltage.
shield stress to the highest level, which is 16 kV/mm
at 500 kV and 14 kV/mm at 400 kV. These high Further, dry curing reduces the size and number of
stresses are acceptable, because it is easier to produce voids per unit volume, but voids can still occur due to
a circular, smooth conductor of small diameter; on either abnormal conditions in materials and processing
the other hand the stress at the insulation screen is or high local electrical stress.
lower.
The voltage breakdown initiation is determined by the
• At larger conductor diameters, the logarithmic equa-
charges that flow back and forth, constituting an ioniza-
tion increases the insulation screen stress until the
tion current in the insulation in the regions of local
accessory limiting stress is reached at the cable/acces-
high-voltage stress. This current is similar to a gaseous
sory interface. At 400 and 500 kV, the accessory limit
discharge from a point electrode. The reader should
is in the range of 6-8 kV/mm. The logarithmic equa-
refer to the Bahder paper (Bahder et al.1976) for a
tion produces an equivalent stress at the conductor
detailed analysis of the failure mechanism of XLPE-
shield in the range of 11-15 kV/mm.
insulated transmission cables.
4.9 PROPERTIES OF XLPE CABLE A second breakdown mechanism, water treeing, applies
INSULATION to cables installed in wet locations, where moisture can
gain ingress into the insulation structure from the con-
4.9.1 Electrical Breakdown Mechanism ductor or by diffusion through the outer covering from
Short-time breakdown of a laboratory sample of XLPE the environment. The necessary conditions for the
is given as 20,000 V/mil (800 kV/mm) in Tanaka and development and propagation of fine, water-filled, den-
Greenwood 1983, page 21. The breakdown strength of dritic patterns in the insulation (water trees) are mois-
XLPE insulation in a cable adjacent to an electrical geo- ture in the insulation, ac voltage stress, imperfections in
metric stress raiser formed by a particle or asperity is the insulation that can serve as cavities for accumulating
given as 300 kV/mm in AEIC CS9 Appendix A, and this moisture, and areas of localized, high electrical stress.
value is given in Bartnikas and Srivastava 2000. Unlike electrical trees, water trees propagate very slowly
and bridge the insulation only after a number of years.
The need to monitor the conductor shield was consid- Water trees reduce the ac and impulse breakdown
ered to be essential to eliminate significant perturba- strengths of the cable. Failure generally occurs during,
tions (Bahder et al. 1976). or as a consequence of, switching or lightning impulse
surges when an electrical tree propagates from the
Substantial work has been performed to establish the extremity of the largest water tree.
mechanism of electrical breakdown of polymeric-insu-
lated cables—for example, Whitehead 1951, Kitchin Electrochemical treeing is defined as a form of water
1958, Mason 1959, McMahon 1963, Toriyama 1967, treeing in which the water contains solute ions that
Shiweck 1969, Bahder 1974, Bahder 1976, and St.Onge move under the action of an electrical field and are
1978. An excellent review is given by Dissado and detected within the wall of insulation or at an electrode
Fothergill (1992). Based on these studies and work per- surface after having passed through the insulation. Cop-
formed on models and cable samples, it was postulated per sulfide and copper oxide have been found in the
4-23
branches of such trees where cables have been installed nificant impurities), free of harmful voids, and, in the
in the vicinity of a chemical plant and immersed in case of EPR, the insulation must be processed to ensure
water rich in hydrogen sulfide. a homogeneous mixture of compounding ingredients
free of significant size agglomerates in order to retain
Water trees have been broadly classified as vented trees their intrinsic resistance to water treeing.
and nonvented (bow-tie) trees. Vented trees start their
growth at the interface of the insulation with the con- The rate of propagation of vented water trees appears to
ductor shield and insulation shield with an unlimited be proportional to the electric stress. Hence, higher
supply of air and water and can grow completely stress cables, with reduced insulation-wall thickness, will
through the insulation to bridge the two electrodes. exhibit larger water trees over a given period of time and
Nonvented trees start at an internal void or contami- shorter cable life, compared to lower stress cables,
nant and, in most cases, are limited in size. assuming that the cable design and operating conditions
permit the ingress of water into the insulation.
Although a number of theories have been proposed to
explain why water trees develop and propagate in In the design of medium- and high-voltage extruded-
extruded-dielectric cables, (e.g., immeasurably small dielectric cables, it is helpful to fill the strand interstices
partial discharge in channels that are incompletely filled to prevent ingress of water under any circumstances.
with water, cracking or distortion of the insulation due Crosslinked polyethylene-insulated, nonmetallic
to vaporization and expansion of liquid water, localized sheathed transmission cables with overall polyethylene
oxidative thermal degradation or decomposition, and and to a lesser extent PVC jackets, 110-140 mils (2.79-
fatigue failure due to repeated applications of compres- 3.56 mm) in thickness, have not exhibited failures due to
sive force in the very small, high-stress regions), no gen- water treeing. These cables, rated at 69 kV, have been
eral agreement has been reached to date, and there are operating at up to a maximum stress of 100 V/mil (4
continuing studies in this regard. kV/mm) since 1969, and at 115/138 kV at a maximum
stress of 150 V/mil (6 kV/mm) since 1975. One exception
Water in the conductor of a cable greatly accelerates the is a 115-kV, crosslinked polyethylene-insulated, non-
development and propagation of water trees. This may lead-covered, polyethylene-jacketed cable that failed
be attributed to the higher temperature of the conduc- after 15 years due to water treeing. This cable exhibited
tor, which acts to move the water into the insulation clear evidence of ingress of water into the conductor and
structure of the cable. This may be separate from or in under the outer covering.
conjunction with high amounts of impurities in the con-
ductor shield, which can be transported into the insula- It is noteworthy that in Europe, where the operating
tion by the water. stresses are generally higher, the practice is to employ a
metallic sheath or moisture barrier to prevent ingress of
A low-moisture diffusion jacket, such as polyethylene, moisture into the insulation structure of the cable. The
crosslinked polyethylene, or chlorinated polyethylene, is stress level at which a moisture barrier is required to
effective in retarding ingress of water from the environ- prevent failure in service life due to water treeing is
ment, aided by the higher temperature of the conductor dependent on many factors, including the thickness of
relative to the environment, which acts to retard the dif- the insulation. However the rate of tree growth is accel-
fusion of water through the jacket. The jacket and the erated at a stress of greater than 50 V/mil (2 kV/mm),
extruded insulation shield are more effective in this and so it is usual to apply a moisture barrier when this
regard than the insulation, since the insulation is in a stress is exceeded. AEIC CS9 (AEIC 2006) limits the
divergent electric field, and dielectrophoretic forces act insulation internal and external ac stresses to 100 V/mil
on uncharged but polarized molecules of water to (4 kV/mm) and 50 V/mil (2 kV/mm), respectively, if
attract the molecules of water toward the regions of there is not a metallic moisture barrier. Please see
greatest field intensity—i.e., through the insulation in “Additional Bibliography—Water Trees” in the list of
the direction of the conductor shield. references at the end of this chapter for an overview of
water treeing in extruded cable insulations.
Tree-retardant insulations, such as TRXLPE and prop-
erly formulated and mixed EPR insulations, are more 4.9.2 Electrical Properties
resistant to the development and propagation of water Typical electrical properties determined on crosslinked
trees than conventional crosslinked polyethylene and polyethylene-insulated high-voltage cables are listed in
noncrosslinked, high-molecular-weight polyethylene Table 4-3.
insulations. Insulations must be super clean (free of sig-
4-24
Table 4-3 Typical Electrical Properties for Crosslinked increases its molecular motion, melting crystallites of
Polyethylene Insulation Used in Transmission Cable polyethylene and adding to the amorphous fraction
Applications (Union Carbide Company n.d.). Crystallites of larger
Relative permittivity at 90°C 2.3 ratio and larger size are melted with increasing temperature,
DLA (dielectric loss angle) at 90°C 0.0001 radians which causes an increase in the rate of expansion until
AC breakdown stress at conductor at the crystallite melting is complete. Beyond this thermo-
V/mil dynamic melting point, the material is completely amor-
room temperature. (30-minute step test, >1270 (50)
(kV/mm)
40 V/mil (1.58 kV/mm) steps) phous. With further increase in temperature, thermal
Impulse breakdown stress at conductor. expansion increases at a constant rate characteristic of
V/mil
(1 x 50 μs wave at 90°C conductor tem- 3425 (135)
perature)
(kV/mm) the molecular structure of the material.
Volume resistivity at 20°C, ASTM D 991 >1 x 1013 ohm-m
The result of this effect is shown in Figure 4-22.
The dielectric constant does not vary significantly with Since organic materials expand much more than metals
temperature. The dissipation factor does not vary signif- in the temperature range of interest in this application,
icantly with temperature between 20 and 90°C. At a and since the thermal expansion of the insulation along
130°C conductor temperature, an increase in the dissi- the axis of the cable is negligible, this data, based on vol-
pation factor of the order of 40% may be observed. umetric expansion, can be used conservatively to calcu-
late the radial expansion of the insulation at the normal
The reduction in ac dielectric strength between room and emergency temperatures of the cable. Since the
temperature and 90°C is of the order of 25%. The reduc- extruded conductor and insulation shields are generally
tion in impulse strength between 90 and 130°C is of the based on the same or similar polymers as the insulation,
order of 40%. Volume resistivity decreases significantly with the addition of 30-40% carbon black, and because
with increasing temperature. the extruded shields comprise a small volume compared
to that of the insulation of high-voltage cables, it is pos-
4.9.3 Physical Properties sible to use the total volume of the insulation and
Typical physical properties for crosslinked polyethylene shielding structure of the cable conservatively for calcu-
insulation, determined on specimens from unaged high- lating the increase in diameter over the insulation shield
voltage cables, are listed in Table 4-4. under normal and emergency operating temperatures.
4.9.4 Thermal Properties Thermal expansion of the insulation structure of the

Thermal expansion, thermal conductivity, and specific cable under normal and, particularly, emergency operat-
heat are important thermal properties of the insulation ing conditions is an important consideration in the
that influence the design and operating characteristics design of metallic shields and sheaths and particularly
of extruded-dielectric cables. in the design of sealed metallic moisture barriers for
extruded-dielectric cables. Where the expansion of the
Thermal Expansion insulation structure at elevated temperatures cannot be
Crosslinked polyethylene is a homopolymer, and, as ac commodated by th e design of the me tallic
such, it is partially crystalline with a melting tempera- shield/sheath or moisture barrier, a resilient, compress-
ture of around 107°C. As discussed in Union Carbide’s
Kabelitems periodical, the addition of thermal energy in
the form of heat to the material in the solid phase
Table 4-4 Typical Physical Properties for Crosslinked

Polyethylene Insulation
Tensile strength 3200 (22) psi (MPa)
Elongation at break 540 %
1400
Tensile stress at 100% elongation, psi (MPa)
(9.7)
After air-oven test at 120°C for 168 hours
Tensile strength, variation 10 %
Elongation at break, variation 10 %
Low-temperature brittleness, ASTM D 746 -90 °C Figure 4-22 Thermal volume expansion of crosslinked
Hardness, ASTM D 2240 45 Shore D polyethylene insulation (courtesy of Union Carbide
Company n.d.).
4-25
ible semiconducting layer compatible with the insulation expressed as joules per gram per degree centigrade
structure may have to be incorporated between the based on the concept of mechanical equivalent of heat.
extruded insulation shield and metallic shield/sheath or Specific heat varies widely for different materials, and in
moisture barrier of the cable, or the design of the general, the specific heat of a material increases directly
extruded insulation shield must perform this function. with increases in temperature.
Thermal Conductivity Figure 4-24 shows the specific heat of crosslinked poly-
The thermal conductivity of the insulation structure of ethylene insulation as a function of temperature.
the cable, in conjunction with that of the other elements
in the thermal path from the extruded insulation shield The peak in the curve corresponds approximately to the
to ambient earth or air, affects the temperature rise of melting temperature of crosslinked polyethylene insula-
the current-carrying conductor, and hence the ampacity tion at which the structure changes from a crystalline-
of the cable under normal and emergency operating amorphous to a completely amorphous state. This
conditions. All other things being equal, if the thermal change in the state of crosslinked polyethylene insulation
conductivity of the insulation structure is higher, then necessitates an input of heat before its temperature can
the ampacity of the cable will be greater. The thermal exceed about 107°C, which favorably affects the transient
conductivity and specific heat of the materials that sur- ratings of crosslinked polyethylene-insulated cables.
round the conductor and the metallic shield/sheath of a
cable influence their transient temperature rise during
short circuits and hence their short-circuit current-car- 4.10 EPR CABLE: BENEFITS AND
rying capacities and short-time ampacities up to several CHALLENGES
tens of hours. The benefits and challenges of EPR cable are in relation
to those of XLPE, which has become a more universally
Figure 4-23 shows thermal conductivity of crosslinked applied insulation for general use.
polyethylene insulation over a wide range of tempera-
tures. Note that thermal conductivity is the reciprocal of 4.10.1 Benefits
the more commonly used term, thermal resistivity. • Increased flexibility at normal and low ambient tem-
peratures, improving ease of installation and jointing.
For the reasons cited previously, the thermal conductiv-
ity values shown in Figure 4-23 can be used conserva-
• Reduced flexibility and excellent behavior at higher
emergency operating temperatures.
tively for the extruded conductor and insulation shields
if such data are not available for these materials. • The mechanical properties of elasticity, thermal
expansion and thermal resistivity show less variation
Specific Heat with temperature.
Specific heat is defined as the heat capacity per unit
mass of a material. It can be expressed as the calories • Improved water tree performance.
required to change the temperature of one gram of • Special grades are suitable for some wet applications,
material one degree centigrade. Specific heat can also be both land and submarine, without a metallic radial
water barrier.
Figure 4-23 Thermal conductivity of crosslinked

polyethylene insulation (courtesy of Union Carbide Figure 4-24 Specific heat of crosslinked
Company n.d.). polyethylene insulation.
4-26
• The insulation shield can more readily be removed compared to XLPE are increased mechanical flexibility,
from the EPR insulation by the technique of heat smaller bending radius, greater short-term temperature,
stripping. and greater resistance to degradation by water trees.
• Special grades have improved resistance to partial
The designation EPR is the generic common name that
discharge.
refers to these two sub types:
4.10.2 Challenges • Ethylene-propylene rubber monomer copolymer
• The dielectric losses are significantly higher than (EPM, Figure 4-26)
those of XLPE and limit the practical application to • Ethylene-propylene rubber diene monomer terpoly-
cables of rated voltage 150 kV and below. mer (EPDM, Figure 4-27)
• The insulating material needs to be compounded in a
separate process, and this represents an additional Ethylene-propylene rubber monomer (EPM) is a copol-
cost. ymer of low density, completely saturated without C-C
double bonds, and crosslinked radically (i.e., by perox-
• In order to guarantee the reproducibility of the com- ides as polyethylene). EPM was the first polymer to be
pounding of several ingredients, a suitable character- adopted and found a rapid importance thanks to the
ization criterion is necessary. outstanding processing properties and resistance to
• A range of EPR compounds exists with significantly thermal aging.
different properties and costs, requiring care in selec-
tion for a particular application. The further development of this saturated rubber, incor-
• EPR insulation remains opaque and cannot be visu- porating dienes (EPDM), and having two lateral C-C
ally examined before and after aging or to verify the double bonds, allows traditional sulfur crosslinking as
integrity of the cable structure. well as peroxide crosslinking. Sulfur curing was soon
abandoned in the cable industry because of process lim-
itations and reaction with the copper conductor. Perox-
4.11 GENERAL DESCRIPTION OF EPR CABLE ide crosslinking was adopted, being more reactive and
Ethylene-propylene rubber (EPR) is a filled, crosslinked, efficient.
thermoset insulation suitable for use up to 138 kV
(150 kV in Europe), above which the dielectric loss The EPDM terpolymer was developed by DuPont to
becomes significant with voltage and progressively introduce sites of unsaturation in the polymer chain to
reduces the ampacity compared to an XLPE cable. EPR facilitate crosslinking of nonelectrical rubber products
is an elastomer (Figure 4-25), which has an amorphous by employing sulfur at relatively low temperatures with
structure throughout the working temperature range exposure to the atmosphere. With the advent of perox-
and, compared to XLPE, exhibits less variation in the
elastic modulus with temperature (Figure 4-12). The
short-term temperature for emergency operation is usu-
ally limited to less than 130°C, with a temperature limit
of 250°C for short-circuit ratings. The advantages of EPR
Figure 4-26 Structure of EPM (courtesy of
Prysmian Cables & Systems).
Figure 4-25 Slices cut from EPR cable Figure 4-27 Structure of EPDM(courtesy of Prysmian
insulation (courtesy of CCI). Cables & Systems).
4-27
ides for crosslinking EPM and EPDM, the diene mono- known relationship between crystallinity and perfor-
mer is no longer required for electrical insulation. mance in a state-of-the-art design high-voltage cable.
However, EPDM is largely, but not exclusively, used in
conjunction with peroxides for crosslinking electrical Ethylene propylene is capable of exhibiting a balance of
insulations, since it is the predominant polymer manu- physical and electrical properties suitable as an insula-
factured in large quantities in the United States for non- tion for high-voltage power cables when formulated
electrical rubber products. with a quality mineral filler and other carefully selected
ingredients controlled by stringent material specifica-
EPM copolymer and EPDM terpolymer are amorphous tions. It also needs to be mixed to ensure a very good
or partially crystalline elastomers. In the crystalline dispersion of the ingredients and to be free of harmful
state, the macromolecules (molecules orientated in a contaminants.
long chain) are arranged geometrically so that there are
elements of symmetry and high cohesive forces. The The term “EPR” will be employed in this book when
crystallinity of the polymer is a function of the ratio of referring to all ethylene-propylene polymers whether or
ethylene to propylene content and can be categorized as not they contain a diene monomer.
follows:
• Low crystallinity, 45-55% ethylene content 4.11.1 Formulation of EPR
The term “EPR-based” insulating compound, rather
• Medium crystallinity, 55-65% ethylene content
than EPR insulating compound, is more technically cor-
• High crystallinity 65-70% ethylene content rect, since a typical EPR high-voltage insulating com-
pound contains seven to nine chemical ingredients of
EPM copolymers or EPDM terpolymers with greater which EPM or EPDM constitute approximately 45-50%
than 72% ethylene content progressively lose their rub- by weight. The remainder of the compound consists of
bery properties and become a kind of modified poly- mineral filler, antioxidant, crosslinking agent, co-agents,
ethylene. Hence polymers with greater than 72% and processing agents.
ethylene content are not normally employed for electri-
cal insulations. The EPM or EPDM is the backbone of the insulation
compound and, to a large extent, determines the basic
In addition to the ratio of ethylene to propylene, crystal- physical and electrical properties of the insulation. The
linity of the polymer is affected by the selection of the mineral filler normally comprises from 25-35% of the
catalyst and reactor conditions used in forming ethylene compound by weight and serves to enhance its physical
propylene, since they affect the size and dispersion of properties. Treating the mineral filler with silane can
clusters of ethylene in the polymer structure. In addi- contribute to the physical and electrical stability charac-
tion, the crystallinity can be affected by type and teristics of the compound in water. The antioxidant pro-
amount of compounding ingredients, as discussed later. tects the EPM or EPDM against thermal decomposition
Crosslinking reduces crystallinity, and crystallites in due to high processing temperatures and also prolongs
EPM and EPDM melt at temperatures on the order of the life of the insulation in service when exposed to nor-
40°C. mal and emergency operating temperatures. The
crosslinking agent, activated by heat, affects crosslinking
In the amorphous state, the macromolecules have low by the formation of cross-bonds in the molecular chains.
cohesive forces and are in complete disorder. Amor- Crosslinking provides the physical and mechanical prop-
phous polymers or polymers with low ethylene content, erties essential for use as cable insulation. The processing
and hence low crystallinity, generally result in com- agents facilitate extrusion and improve the surface con-
pounds that are softer—i.e., they exhibit lower tensile dition of the insulation. The antioxidant, crosslinking
moduli, and have increased flexibility. Since lower mod- agent and co-agents, and processing agents are each less
ulus compounds tend to be more resistant to surface than 5% by weight of the compound.
discharge, these polymers may be desirable in EPR com-
pounds where the shield structure of the cable allows the In general, compound formulation (i.e., the selection of
insulation to be exposed to continual electrical dis- the polymer, filler, and other ingredients and the quanti-
charge. Such polymers may also be desirable for use in ties of each) is important in determining the physical
portable power cables or in applications where the and electrical characteristics of the insulation and the
cables will be continually flexed in service. Other than service performance of the cable. The mineral filler, gen-
this special cable design and application, there is no erally a silicone-treated clay of high purity, facilitates
processing, provides dimensional stability, and enhances
4-28
the mechanical properties of the compound. It is also tamination. However, in recent years, with increasing
largely responsible for the higher dissipation factor and use of the more crystalline polymers, which lend them-
dielectric constant, and hence dielectric loss, of EPR selves to pelletization, the pellet form has increased in
compounds compared to LDPE and HDPE insulations. use because of reduced handling and exposure to the
However, the processing or mixing of the compound is environment, and consequently, reduced risk of contam-
at least of equal importance to the compound formula- ination. Pellets can be used in conventional air-convey-
tion with regard to the service performance of the cable. ing equipment.
4.11.2 Mixing of High-Voltage EPR Insulation 4.11.3 Cable Applications

Compounds EPR insulation has been used in the United States for
The achievement of a high-quality, high-voltage EPR transmission cables at voltage ratings of 69 and
insulation compound depends on the following raw 115/138 kV since the late 1960s and early 1970s, respec-
material and mixing requirements: tively. This parallels the experience in Europe, where the
• The base polymer must be properly constituted and first 150-kV EPR-insulated cable was manufactured and
installed in 1975 (Figure 4-29).
free of conducting materials and other contaminants.
• All compounding ingredients must be of high quality EPR cables have seen widespread use up to 150 kV in
and free of conducting material and other contami- Italy, with some applications up to 220 kV, particularly
nants. in mountainous terrain, for example in Switzerland.
• Mixing of the polymer, mineral filler, and other com-
pounding ingredients must be such as to yield a An impressive amount of EPR medium-voltage subma-
homogenous compound that is free of conducting rine cables with a wet design (see Figure 4-30) have been
materials and other contaminants. installed since the late 1960s with an excellent service
record. A wet cable design is not provided with an
Stringent specifications, including intensive inspection
and acceptance testing for all raw materials, are neces-
sary to meet these requirements. Carefully controlled
storage and handling procedures must be observed to
further minimize the risk of contamination.
The ingredients should be mixed in an internal mixer

(generally known as a Banbury mixer) or a continuous
mixer. For optimal electrical and electrical-stability-in-
water characteristics, the Banbury mixer should be used
exclusively for high-voltage EPR insulation compound,
and the mixing cycle should be adhered to rigidly. A
continuous mixer lends itself to thorough cleaning; if
other compounds are mixed in the continuous mixer,
the mixer should be disassembled and completely
cleaned prior to mixing high-voltage EPR insulation
compounds. Final straining through screens with small
mesh openings is effective in dispersing agglomerates
and removing nondeformable contaminants. The screen
size for high-voltage EPR insulation should be no less
than No. 200 mesh and preferably higher mesh (smaller
size openings). Screenings should be done for each
group of ingredients and again for the final mixed com-
pound to achieve optimal dispersion. Figure 4-28 shows
a Banbury mixer.
The compound, when in its final mixed state for extru-

sion, may be in the form of either strips or pellets. Both
forms are suitable for use as high-voltage insulation,
Figure 4-28 Banbury mixer (courtesy of Farrel Co.).
provided that proper care is exercised to prevent con-
4-29
decreasing conductor size. The insulation thickness for

a 2000-kcmil (1013-mm 2 ) conductor, 150 kV cable
would be 800 mils (20 mm), with the insulation thick-
ness again varying inversely with the conductor size to
achieve essentially the same maximum stress.
The maximum conductor temperatures under normal,

emergency, and short-circuit operation, as set forth in
Table 1.9-1 of AEIC CS9-06, are 90, 105, and 250°C,
respectively. Essentially the same temperatures are used
in Europe for EPR-insulated high-voltage cables. Due to
the excellent thermal behavior, appropriate EPR com-
pounds are suitable for temperature under normal ser-
vices of 105°C and emergency temperatures of 140°C
Figure 4-29 150-kV 2000 kcmil (1000 mm2) EPR cable (Cinquemani 1996). AEIC CS9-06 (AEIC 2006) notes
(courtesy of Prysmian Cables & Systems). that only emergency temperatures of 130°C may be used
for EPR cables in the range 46-150 kV with mutual
agreement between purchaser and manufacturer.
4.12 LANDMARK DEVELOPMENTS IN EPR

CABLE
1960: Studies carried out by the German scientist Karl
Ziegler and by the Italian Giulio Natta during the 1950s
showed that some organometallic catalyst greatly
increased the speed of some reactions. This made possi-
ble the copolymerization of ethylene with polypropy-
lene, creating the saturated synthetic rubber known by
the generic designation of EPR. The invention of the
oriented catalyst polymerization process led the two
inventors to the Nobel Prize in 1963 (Natta 1963).
In 1957, polypropylene was first commercially produced

Figure 4-30 46 kV 500 kcmil (250 mm2)
EPR submarine
and marketed by Montecatini Italy as Moplen using
3/C cable (courtesy of Prysmian Cables & Systems). Zeigler-Natta catalysts to control the structure of the
polymer and to polymerize products that had previously
impervious metallic radial barrier, and so water is
been impossible.
allowed to enter the cable and contact the EPR insula-
tion. Note that in Figure 4-30 the metallic layer around
In 1960, the first ethylene-propylene rubber polymer
each EPR cable core is formed from a tinned copper
was commercially available from the Italian company
tape shield and is not a water barrier.
Montecatini.
AEIC CS9-06 and ICEA S-108-720-2004 set forth the
EPR was later produced by the Montecateni Company,
requirements and operating conditions for EPR-
Italy. The EPM co-polymer was further modified by
insulated transmission cables in the United States.
DuPont USA to become EPDM so that it could be
crosslinked using compounds of sulfur, the standard
IEC 60840 covers EPR cables for rated voltages up to
method for rubber cables. This permitted EPR to pro-
150 kV; IEC 62067 covers EPR insulated cables up to
gressively replace butyl rubber as an insulating and jack-
220 kV and higher, even though EPR is rarely used at
eting compound.
these voltages.
1963: The subsequent development in 1963 of the chem-
Operating stresses employed in Europe for EPR-insu-
ical crosslinking route became the standard method to
lated transmission cables are higher than in the United
produce EPR cables, which, as used below, refers to
States. Insulation thicknesses for 72.5-kV rated cables
EPM and EPDM co-polymers. EPR cables are a com-
range from 435 to 515 mils (11-13 mm), increasing with
petitor to XLPE cables up to 138 kV. Their increased
4-30
flexibility at normal ambient temperature makes them Maximum electrical properties of EPR at room temper-
easier to install, especially where space is restricted. ature are given in ICEA S-108-720-04 as:
Their flexibility increases slightly with temperature, but
that of XLPE significantly falls at 105°C to a lower Relative permittivity: 4.0
value, thus giving EPR a relative improvement in opera- Dielectric loss angle: 0.015
tion at emergency overload temperatures up to 130°C.
Although they also form water trees, the trees are gener- 4.13.3 EPR Physical Properties
ally smaller than those in XLPE, which has led to their Typical physical properties for Types I and II EPR insu-
installation at MV (15-35 kV) in three-core submarine lation, for cables rated through 138 kV, as described in
applications and in wet land applications at 150 kV. The ICEA S-108-720-04 are listed in Tables 4-6 and 4-7.
disadvantage of EPR compared to XLPE is that the
insulation is generally more expensive, it requires to be EPR exhibits minimal deformation over its operating
moderately filled to improve mechanical robustness, the temperature range. The hot-creep test is performed in
increased number of ingredients complicates quality accordance with ICEA Publication T-28-562. The pur-
control, and it has a significantly higher dissipation fac- pose of the test is to determine the degree of crosslink-
tor (dielectric loss angle), dielectric permittivity, and ing of the insulation.
loss factor, which reduces its attractiveness at applica-
tions of greater than 138-161 kV.
1960-present: EPR insulated cables continue to be Table 4-5 Typical Electrical Properties for EPR-insulated HV
installed in the United States. Cables (AEIC CS 6-87)
Relative permittivity at 90°C 2.75 ratio
4.13 PROPERTIES OF EPR INSULATION Dielectric loss angle at 90°C 0.0035 radians
AC breakdown stress at conductor,
V/mil
4.13.1 EPR Electrical Breakdown Mechanism (Ten-minute step test, 40 V/mil steps at 1050 (41)
(kV/mm)
room temperature)
The breakdown mechanism of EPR insulation is similar Impulse breakdown stress at conductor,
V/mil
to that of XLPE. Section 4.9.1 provides information on (1 x 50 us wave at 90°C conductor tem- 1850 (73)
(kV/mm)
this subject. perature)
Volume resistivity at 20°C, ASTM D 991 >1 x 1013 ohm-m
EPR-insulated 69- and 115/138/150-kV rated cables
manufactured in the United States and Europe have Table 4-6 ICEA-NEMA Minimum Requirements for Type I
employed concentric wire shields and overall polymeric EPR-Insulated HV Cables
jackets, which could make them susceptible to water Tensile strength 700 (4.8) psi (MPa)
treeing. For the cables operating at elevated gradient at
Elongation at rupture 250 %
conductor screen, it is also the practice in Europe to
Tensile strength after 168 hours at 121°C,
adopt a water-blocked conductor design in order to 75 %
ratio of unaged value
avoid water penetration and to improve the resistance to
Elongation at break after 168 hours at
water treeing. A properly formulated and processed 755 %
121°C, ratio of unaged value
high-voltage EPR-insulation compound is more resis- Hot-creep elongation after hot-creep test at
tant to the development and propagation of water trees 50 %
150°C +/-2°C, maximum
than crosslinked polyethylene insulation. Water treeing Hot-creep set after 150oC +/- 2°C maximum 5 %
is discussed in detail in the Section 4.9.1.
Table 4-7 ICEA-NEMA Minimum Requirements for Type II
4.13.2 EPR Electrical Properties EPR-Insulated HV Cables
Typical electrical properties for EPR-insulated high- 1200 psi
Tensile strength
voltage cables are shown in Table 4-5. (8.2) (MPa)
Elongation at rupture 250 %
The relative permittivity does not vary significantly with Tensile strength after 168 hours at 121°C, ratio
80 %
temperature. The dielectric loss angle at room tempera- of unaged value
ture is approximately one-half of the value at 90°C. Elongation at break after 168 hours at 121°C,
80 %
ratio of unaged value
The dielectric loss angle is strongly dependent on the Hot-creep elongation after hot-creep test at
50 %
150°C ±2°C, maximum
compound formulation; the value in the table applies to
Hot-creep set after hot-creep test at 150°C
a high-quality, low-filler-content EPR compound suit- 5 %
±2°C maximum
able for 69- to 150-kV cables.
4-31
4.13.4 EPR Thermal Properties tions. If all other things are held constant, the ampacity
Thermal expansion, thermal conductivity, and specific of the cable will be greater when the thermal conduc-
heat are important thermal properties of the insulation tivity of the insulation structure is higher. The thermal
that influence the design and operating characteristics conductivity and specific heat of the material that
of extruded-dielectric cables. surrounds the conductor and metallic shield/sheath of a
cable influence the transient temperature rise during
Thermal Expansion short circuits, and hence the short-circuit current-
Ethylene-propylene rubber insulations expand with carrying capacities and the short-time ampacities of
input of thermal energy in the form of heat. Due to the extruded-dielectric cables up to several tens of hours.
high mineral filler content in EPR insulation, which has
relatively negligible expansion in the temperature range Figure 4-32 shows the thermal conductivity of EPR
of interest, and the lower level and size of crystallites of insulation over a wide range of temperatures (EPRI
EPR insulation, the thermal expansion of EPR insula- 1978). The broad ranges of values at each temperature
tion is less than for LDPE, HDPE, and XLPE insula- reflect the variations in three EPR formulations that
tions, particularly at the higher temperatures. Figure 4- were tested. Line A in the plot is applicable to an
31, calculated from data in EPRI EL-938, shows the EPDM formulation suitable for higher-voltage applica-
range of volume expansion of EPR-insulation com- tions. The data for this plot were obtained by the same
pounds with temperature (EPRI 1978). method (ASTM 1980) as in Reference (EPRI 1978), but
in a different laboratory (Dynatech 1975).
The volume expansion of the higher-voltage, partially
crystalline EPR-insulation compounds lies at the top of The thermal conductivities in Figure 4-32 can also be
the EPR range that is commonly used for transmission applied to the extruded conductor and insulation
cables. shields if such data are not available.
Since the thermal expansion of EPR is similar to XLPE Specific Heat

in relation to that of the metals used in the conductor, The reader is directed to Section 4.9.4 for a discussion
shield, and sheath, please refer to Section 4.9.4 for a of specific heat and its significance, in conjunction with
more detailed explanation of the significance of thermal thermal conductivity in the short-circuit rating of con-
expansion with regard to cable design. ductors and metallic shields/sheaths and the short-time
ampacities of extruded-dielectric cables.
Thermal Conductivity
Thermal conductivity of the insulation structure of the The specific heat of a high-voltage EPR-insulation com-
cable, in conjunction with that of the other elements in pound was determined in the performance of EPRI
the thermal path from the extruded insulation shield to Project 1286-2 (EPRI 1983). The unpublished value
ambient earth or air, affects the temperature rise of the obtained in the investigation was 0.44 cal/gm °C over a
conductor carrying current and hence the ampacity of temperature range of 70-140°C.
the cable under normal and emergency operating condi-
Figure 4-31 Thermal volume expansion range of EPR Figure 4-32 Thermal conductivity of EPR insulation.
insulations.
4-32
4.14 CONSTRUCTION: CONDUCTOR General-purpose stranding machines cannot always

Conductors with sizes up to 7000 kcmil (3500 mm2) handle the wire sizes, shapes, and compacting tensions
have been employed in 66-kV XLPE cables in Japan, necessary to achieve a sufficiently smooth shape. A
with a maximum manufacturing capability of approxi- compounded semiconducting fabric binder tape may be
mately 8000 kcmil (4000 mm2). The transition size to wound tightly over the conductor to provide a strong
Milliken depends upon the maximum number of wires circular surface. The presence of the binder tape also
that any particular stranding machine can apply in one affects the handling characteristics of the cable. The
pass and upon the reduction in conductor area that can binder tape reduces the grip of the semiconducting
be achieved by the improvements in the ac resistance of shield onto the conductor such that the cable can be
the Milliken conductor, as described in Chapter 3. more easily bent into alignment for jointing (bending is
always difficult for thick-walled transmission class
XLPE cables suffer the disadvantage that the contrac- cables), but the insulation shrink-back at the insulation
tion that occurs during the cooling and crystallization cut will be greater during jointing.
of XLPE after the crosslinking process, compresses the
individual wires together and breaks the natural surface It has become common practice for direct-buried cables
oxide insulation. This substantially destroys the benefits to have blocked conductors to prevent the longitudinal
of segmental Milliken conductor in reducing the ac transmission of water in the event of either damage dur-
resistance in XLPE cables. Chapter 3 notes that the ing installation or in service. The conductor may be
ac/dc resistance of large Milliken conductors at 3500 blocked with either a free or tape-bound water-swellable
kcmil (1600 mm2) and above can be as high as 40%. For powder, a viscous compound, or an elastomeric com-
this reason, CIGRE recommends that the ac resistance pound. It is important that the materials are applied
of large conductor XLPE cables be measured as part of uniformly and do not disturb the smooth surface of the
the development and qualification tests. Insulated conductor.
strand conductors utilizing an enameled coating or an
oxidized finish have been developed for some XLPE Solid aluminum conductors are both inherently smooth
cable applications (Figure 4-33). The main benefit of and perfectly water blocked. However, they are stiffer
insulated strand conductors is that they overcome the and more difficult to handle, both in the cable factory
current-carrying limit imposed by the largest conductor and during installation, and are usually limited in size to
that a particular factory can strand. less than 2400-3200 kcmil (1200-1600 mm2).
It is important that the outer surface of the conductor is The conductor geometry has to be designed to be fed
formed from a layer of tight fitting and smooth wires into a close-fitting circular steel die in the CV line extru-
(Figure 4-34), which may be either died down or shaped sion head, which applies the conductor shield in a com-
to produce a nearly true cylindrical shape. This provides paratively thin layer of semiconducting compound. The
an accurate bed for the semiconducting conductor shield conductor must have a
and also prevents fall-in of the molten semiconducting • circular geometry to avoid abrasion with the die.
and XLPE insulating compounds into the conductor.
Figure 4-33 5000 kcmil (2500 mm2) insulated strand

conductor in 400-kV cable (courtesy of Prysmian Cables Figure 4-34 Smooth circular stranded conductor
& Systems) in a 132-kV XLPE cable (courtesy of CCI).
4-33
• constant diameter to avoid variations in the thickness • completely cover the irregular conductor surface with
of the semi-conducting shield. no holes present.
• uniform smooth finish to prevent variations in either • bond to the insulation such that they expand and
surface friction or surface geometry from causing contract radially as one body, thus preventing the for-
compound surging and thickness variations. mation of air-filled voids.
• clean surface, free of debris to avoid either swarf pen- • share the same bending flexibility as the XLPE insu-
etration through the shield into the insulation, or lation without cracking.
accumulations of debris from initiating shield protru- • be sufficiently robust to resist penetration by wires
sions into the insulation. due to conductor sidewall loading in route bends.
• dry construction free of wire-drawing lubricant or • be extrudable at output rates and die temperatures
liquid cleaning agents that could degrade the perfor- that are compatible with the XLPE or EPR insula-
mance of the semiconducting shield. tion.
Accurate control of circularity is more difficult with a • be chemically crosslinkable such that the shield
segmental Milliken conductor, because the shape is shares the same normal operating temperature of
formed from a number of separately stranded, prespi- 90°C, emergency overload temperature of 105°C or
raled segments that are wound together in a laying-up 130°C, and short-circuit temperature of 250°C, with-
machine. A 1600 kcmil (800 mm2) conductor consists of out losing electrical conductivity.
four segments that can be held firmly in a cylindrical
shape by a tight fabric binder tape. For larger sizes, six The requirement of the insulation shield is closely simi-
segments may be used, although some manufacturers lar, with the exception that it is required to resist the
prefer to form the conductor from an unequal number possibility of penetration and distortion by irregularly
of segments (i.e., five or seven), such that a shear plane shaped overlying layers such as shield wire conductors
across the diameter of the conductor is avoided. The and corrugated metal sheaths, subject to the forces of
segments lock together and keep the shape circular dur- thermal expansion and sidewall force in bends.
ing bending prior to insulation extrusion. Control of the
shape of each segment shoulder is difficult. If sharp, The interface between the conductor shield and the
these shoulders could protrude into the conductor XLPE insulation is the most critical part of the extruded
shield, and, if round, they could leave an interstitial gap core, because it experiences the highest electrical stress
requiring a filler of synthetic string, paper, or water- and the highest operating temperatures. In manufacture,
blocking tape. It is, therefore, important to significantly the shield extruder has the greatest difference in volu-
increase the thickness of the semiconducting screen in metric output compared to the insulation extruder and
comparison with that on a circular conductor such that is sensitive to the risks of: (a) “scorch” due to either too
protrusions and depressions are smoothed out and are long a dwell time or to too high a temperature, resulting
not present at the insulation interface. in premature curing and blocking in the die and large
protrusions on the shield surface, and (b) to “surging,”
The conductor is bent many times during manufacture, in which the line speed is too low and the extruder out-
factory transportation, and extrusion, and is thereby put is unstable, resulting in holes in the shield layer. The
exposed to the risk of loss of circularity, abrasion dam- conductor shield is also the most remote location for
age, and contamination from general factory debris. heat to reach during the crosslinking process and in
Protective measures are to apply at the stranding pro- cable manufacture is the most inaccessible for quality
cess a disposable barrier tape (this is removed upon inspection.
entry to the extrusion line) and to transport the conduc-
tor on suitably shaped rollers. The other provisions of AEIC CS9-06 and ICEA S-108-
720-04, as well as the manufacturer’s proven capabilities,
should also be taken into consideration when applying
4.15 CONSTRUCTION: CONDUCTOR AND Table 4-8 to any specific application.
INSULATION SHIELDS
The purpose of the conductor shield in an extruded AEIC CS9-06 (AEIC 2006) specifies that the shield shall
XLPE or EPR cable is to: be filled with “acetylene black” for XLPE insulation
• provide a smooth cylindrical electrically conducting design stresses above 200V/mil (8 kV/mm). AEIC CS9-
surface of near constant radius, such that stress rais- 06 specifies the maximum design stress for each trans-
ers formed by irregularities in the outer surface of the mission voltage and for conductor size ranges (Table
conductor are completely eliminated. 4-8). This shows that 115-through 345-kV cables have
higher design stresses and so must use acetylene black
4-34
Table 4-8 Internal and External Shield Stress Limits (courtesy of AEIC)a
Internal Shield Stress External Shield Stress Range of Insulation

Conductor Area Limit Limit Thickness
Rated Voltage
kV kcmil mm2 V/mil kV/mm V/mil kV/mm Mils mm
500- 240- 638 16.2
69 wet 100 4.0 50 2.0
4000 2000 638 16.2
500- 240- 390 9.9

69 dry 150 6.0 75 3.0
4000 2000 449 11.4
750- 400- 484 12.3

115 dry 200 8.0 100 4.0
4000 2000 547 13.9
750- 400- 579 14.7

138 dry 200 8.0 100 4.0
4000 2000 638 16.2
750- 400- 638 16.2

161 dry 225 9.0 100 4.0
4000 2000 728 18.5
1000- 500- 728 18.5

230 dry 275 11.0 125 5.0
5000 2500 827 21.0
1000- 500- 913 23.2

345 dry 350 14.0 150 6.0
5000 2500 996 25.3
a. The insulation thicknesses given in Table 4-8 have been derived such that neither internal nor external stress limts
are exceeded (see also CS9 Appendix E “Generic Nominal Insulation Thicknesses,” suggesting standard values for
each rated voltage, spanning the complete conductor size range without exceeding stress levels).
Table 4-9 Stresses Derived from AEIC CS7-93 Specified Insulation Thicknesses (courtesy of AEIC)
Rated Voltage Conductor Size Internal Stress External Stress Insulation Thickness
kV kcmil mm2 V/mil kV/mm V/mil kV/mm Mils mm
500- 103.7 4.1 40.5 1.6
69 253-1013 650 16.5
2000 84.5 3.3 47.5 1.9
750- 138.9 5.5 53.6 2.1
115 380-1520 800 20.3
3000 113.2 4.5 62.7 2.5
750- 160.2 6.3 59.4 2.3
138 380-1520 850 21.6
3000 129.7 5.1 69.9 2.8
shields. Table 4-8 shows that the 69-kV wet and dry of the asperities form electrical stress raisers, which ulti-
cables in AEIC CS9-06 have internal shield stresses of mately determine the design operating stress and life of
less than 200 V/mil (8 kV/mm). Table 4-9 shows that the an otherwise well-made cable.
full-wall 69- to 138-kV cables in the superseded AEIC
CS7-93 each have low stress; thus these cables could The cost of the semiconducting compound is high and,
now be manufactured with lower-cost “furnace black” in cables in which greater shield thicknesses are
shields. required, may equal or exceed the cost per unit length of
the extruded XLPE insulation. This is an incentive for
The carbon-black loading is required to achieve the the cable manufacturer to apply the minimum possible
specified ICEA-S-108-720 (ICEA 2004) resistivity for thickness. The minimum point (spot) thicknesses speci-
the insulation shield to be less than 500 Ωm and less fied in ICEA S-108-720-2004 (ICEA 2004) are given in
than 1000 Ωm for the conductor shield under both nor- Table 4-10.
mal and emergency operating temperatures. The IEC
limits are less than 500 Ωm for both shields. The thickness of the insulation shield is given in Table
4-11.
When extruded, the acetylene-black semiconducting
compound produces a super smooth finish with minute The shape of the outer layer of conductor wires is an
protrusions called “asperities.” The size and aspect ratio important factor in determining the insulation thick-
4-35
nesses are usually significantly greater than those in

Table 4-10 Conductor Shield Minimum Point Thickness
(ICEA S-108-720-2004, Table 3-1) ICEA S-108-720-2004 (ICEA 2004, Table 3-1). Typical
thicknesses are 60-100 mils (1.5-2.5 mm).
Conductor Area Minimum Point Thickness
kcmil mm2 mils mm
The effect of fall-in is reduced on both concentric and
250-550 127-279 16 0.41
Milliken conductors by wrapping the conductor with a
551-1000 279-507 20 0.51
strong, semi-conductive binder tape. The tape acts as a
1001-1500 507-760 24 0.61
pressure-retaining shell, which blocks the inward flow of
>1500 >760 30 0.76
the melt. The addition of the binder alters the mechani-
Table 4-11 Insulation Shield Minimum Point Thickness cal properties of the cable, because the grip of the
(ICEA S-108-720-2004, Table 5-1) extruded core onto the conductor is reduced. XLPE
Insulation Shield Thickness
cables with binder tapes are perceptibly more flexible
Minimum Insulation Diameter Minimum Maximum
during jointing, and when cut, exhibit a greater retrac-
in. mm mils mm mils mm
tion of locked-in stress from typically two to three times
0-2.0 0-50.8 40 1.02 80 2.03
the length of adjacent core.
>2.0 >50.8 40 1.02 100 2.54
The normal continuous operating temperature of the
conductor shield and insulation is 90°C. ICEA limits
ness, because a significant proportion of the semi-con
the emergency operating temperature to 105°C for
shield is forced by the high pressures in the extrusion
XLPE cables for greater than 138 kV, but allows 105 to
dies and in the CV tube in between the wires, as illus-
130°C for above 46 through 138 kV, but cautions use of
trated in Figure 4-35. The thickness in Figure 4-35 (bot-
130°C. AEIC restricts this to 105°C for all applications,
tom) is inadequate and would result in insulation stress
XLPE and EPR, but allows 130°C for 46- to 150-kV
raisers and probably shield holes, as shown in Figure
EPR only if mutually agreed between purchaser and
4-9. An additional risk is the presence of gaps between
manufacturer.
the wires, because this permits both the semi-con and
XLPE melts to flow into the center of the conductor,
which is called “fall-in.” 4.16 CONSTRUCTION: INSULATION
The construction of conductor sizes greater than 1500- 4.16.1 Insulation Thickness
2000 kcmil (800-1000 mm2 ) are usually changed from The selection of the thickness of the insulation for an
concentric to segmental (Milliken) to reduce the ac/dc extruded dielectric cable is an important part of the
resistance ratio, which is explained in Chapter 3. It is specification and design process. The conductor geome-
difficult to keep the surface of a Milliken conductor try, grade of shields and insulation used, extrusion and
smooth and circular, and so the average shield thick- curing processes, types of accessories, voltage and basic
insulation levels, use of a radial moisture barrier, prior
qualification, and operating experience, as well as the
target electrical properties—namely, the capacitance of
the cable—have to be taken into account when selecting
the thickness of the insulation for a particular project or
application.
Further, thicker insulation is not necessarily a better or

more conservative choice, because cables with a compa-
rably thicker insulation may result in larger diameter,
more expensive cables that may not be thermomechani-
cally compatible with existing duct systems or duct sys-
tems with a given diameter that are prevalent in North
America. Likewise, cables with thicker insulation experi-
ence increased radial thermal expansion in service, have
a higher thermal resistance, and are more difficult to
manufacture.
4.16.2 Calculation Methods

Figure 4-35 Top: correct conductor shield thickness.
For all practical purposes, a cable is a concentric cylin-
Bottom: inadequate thickness (courtesy of CCI).
drical capacitance, with the outer conductor shield sur-
4-36
face as its inner boundary and the inner surface of the As an example, a typical 138-kV XLPE cable with a
insulation shield as its outer boundary. More detailed 2000 kcmil (1000 mm2) copper conductor to ASTM B8,
formulae and considerations are given in Chapter 3. The a 6 mil (0.15 mm) conductor tape, a 50 mil (1.3 mm)
selection process for the insulation thickness in an insulation shield and 650 mil (16.5 mm) insulation will
extruded dielectric cable, however, will be based prima- yield the following electrical stress levels:
rily on the design stress at the conductor shield and the ECS = 165 V/mil (6.5 kV/mm)
impulse stress at the insulation shield, which represents
the critical interface with the stress control elements of EIS = 93 V/mil (3.7 kV/mm)
the accessory. EIMP = 762 V/mil (30 kV/mm)
The design stress at the conductor shield ECS is given by 4.16.3 Selection Process
the following formula. Based on a given conductor size to meet the ampacity
requirements of a given application and the premise that
Note that if the dimensions of conductor shield d and a cable with a thin insulation is preferable for its smaller
insulation diameter D are entered as inches and the volt- diameter, lower weight, longer delivery lengths, suitabil-
age V as kV, then the unit of stress ECS is V/mil (kV/in.) ity for smaller and/or existing ducts and pipes and com-
Similarly, if the dimensions are entered in mm and the parably lower cost, the insulation thickness is
voltage as kV, ECS will be kV/mm. determined based on the design stresses at the conduc-
tor shield and at the insulation shield. Further, past and
Stresses may be calculated in kV/mm and converted to present practice, regional and customer preference,
V/mil by multiplying with the factor 25.4. Or stresses in industry specifications and standards, and target electri-
V/mil may be coverted to kV/mm by multiplying by cal properties will be considered:
0.039.
Electrical Tests on Insulation
ECS =
2V0
D
[kV / mm] The cable will be subject to routine/production ac with-
stand testing according to AEIC CS9-06, ICEA S-108-
d ln 720-2004, IEC 60840, and/or IEC 62067 or a combina-
d 4-3
tion thereof. For small conductor cables at higher volt-
The design stress at the insulation shield is ages of 230 kV and above, dependent on the specified ac
withstand voltage, the test voltage should not result in a
2V0 stress at the conductor shield exceeding 630-690 V/mil
E IS = (25-27 kV/mm) to avoid the degradation of the insula-
D
D ln tion material during testing. The test stress at the con-
d 4-4 ductor shield is ECS divided by Vo and multiplied by the
test voltage.
The impulse stress at the insulation shield is
Range of Approval
V The manufacturer typically provides existing type or
E IMP = E IS IMP qualification test reports or certificates for cables, or
V0 4-5
cable systems, which employ the same or similar materi-
als and processes for their manufacture and assembly.
Where The electrical stresses should not exceed the stresses
ECS Electrical stress at the conductor shield used in the type or qualification tests.
(kV/mm).
Vo Phase-to-ground rms voltage over the insula- Stress Limits at the Insulation Shield for Accessory
tion (kV). Compatibility
d Diameter over the conductor shield (mm). As discussed in detail in Chapter 8, a variety of technol-
D Diameter over the insulation (mm). ogies are available for splicing and terminating extruded
EIS Electrical stress at the insulation shield dielectric cables, dependent on the voltage level, the
(kV/mm). application, and the manufacturer of the accessory. The
EIMP Electrical impulse stress at the insulation maximum permissible operating and impulse stresses at
shield (kV/mm). the insulation shield largely depend on the manufac-
VIMP BIL impulse voltage for the voltage class of turer’s design, prior experience, and tests. Therefore,
the equipment (kV). there are no absolute values in this regard that would
determine the minimum insulation thickness. The stress
levels at the insulation shield will, however, be relevant
4-37
for cables with large conductors for the selection of the nal and external stresses at the extremes of the conduc-
insulation thickness. tor range. The cable dimensions were based on ASTM
Class B conductors and ICEA S-108-720-2004 shield
State-of-the-art premolded accessory designs for XLPE thicknesses.
HV cables today are designed for insulation shield
stresses of 115 V/mil (4.5 kV/mm) up to 200 V/mil The AEIC CS7-93 thicknesses are commonly referred to
(8 kV/mm) for 400- and 500-kV systems. The impulse in North America as “full-wall” insulations, as opposed
stress levels at the insulation shield are in a range of to “reduced-wall” insulations that represent stress-
760 V/mil (30 kV/mm) up to 1200 V/mil (47 kV/mm) for based designs. Further, AEIC CS7-93 mentions stress-
the highest voltage designs. based designs and recommends the use of a metallic
sheath or other effective moisture barrier for cable
Stress limits in IEC standards do not provide recom- designs with an electrical stress at the conductor shield
mended insulation thicknesses or stress levels. They are of 150 V/mil (6 kV/mm) for 69-kV cables and 200 V/mil
performance standards and allow insulation thicknesses (8 kV/mm) for 115- and 138-kV cables. By comparison
based on the range of approval obtained through system with Table 4-8 for AEIC CS9-06, it can be seen that the
qualification tests (type tests) or long-term prequalifica- 138-kV cable with an 850-mil full-wall insulation thick-
tion tests. The present worldwide trend is for design ness, if provided with a metallic water barrier, is now
stresses of XLPE cable to increase towards those of permitted to operate as a 230- to 345- kV cable and rep-
fluid-impregnated paper cables, especially in recabling resents a very significant change in practice and in cable
applications. In North America the maximum design manufacturing technology.
stresses are now specified in AEIC CS9 (AEIC 2006),
Table 4-12. Care should be taken in the selection of a The new standard AEIC CS9-06 stipulates maximum
high design stress at phase voltage, because this will also stress levels by voltage class at both the conductor shield
increase the impulse stress. (internal stress) and the insulation shield (external
stress). The stress limits are given in Table 4-8, to which
AEIC Insulation Thickness and Stress Limits have been added the calculated insulation thicknesses
The superseded AEIC CS7-93 standard was a cable- for both the smallest conductor and the largest conduc-
only specification that has now been replaced by AEIC tor sizes.
CS9-06. AEIC CS7-93 stipulated insulation thicknesses
by voltage class, independent of the conductor size and Wet Cable Designs
cable geometry. Although these insulation thicknesses Wet designs designate extruded dielectric cables without
are now considered to be obsolete, they were widely a metallic sheath that acts as a radial moisture barrier.
adopted in North America for many years and so are Since transmission class cables operate under higher
recorded in Table 4-9 together with the calculated inter- stresses than distribution cables, the use of a metallic
Table 4-12 Comparison of Electrical Stresses at the Conductor Shield (courtesy of Southwire)
AEIC CS7 ICEA U.S. European Japanese
Voltage Class Full Wall AEIC CS9 S-108-720 Practice Practice Practice
kV kV/mm (V/mil) kV/mm (V/mil) kV/mm (V/mil) kV/mm (V/mil) kV/mm (V/mil) kV/mm (V/mil)
3.2-4.0 6 6 3.2-5.1 4.4-6.1 5.3-6.0
69
(81-100) (150) (152) (81-130) (112-155) (135-152)
4.4-5.4 8 8 4.4-7.0 5.4-7.6 5.8-6.8
115
(112-137) (200) (203) (112-178) (137-193) (147-172)
5.1-6.2 8 8 5.1-7.4 6.3-7.5* 7.0-8.5
138
(128-158) (200) (203) (130-188) (160-191) (178-216)
9 9 6.0-8.0 6.0-8.4 7.2-8.4
161 --
(225) (229) (152-203) (152-213) (183-213)
11 11 7.9-9.9 8.5-10.2 8.2-10.0
230 --
(275) (279) (201-251) (216-259) (208-254)
14 16 11-13 11.0-12.6 11.2-12.8
345 --
(350) (406) (279-330) (279-320) (284-325)
11.1-13.0 12.0-14.8
400 -- -- -- --
(282-330) (305-376)
13.8-15.0 14.8-16.7
500 -- -- -- --
(351-381) (376-424)
* Compact XLPE cables for replacement of 110-132kV paper cables in pipes operate at stresses of 10.3 kV/mm (262 V/mil).
4-38
moisture barrier is recommended for all transmission Capacitance and Dielectric losses
class cables including 69 kV. However, in the United The capacitance of a cable is dependent on the cable
States, in particular, there is long-term positive operating geometry and therefore the insulation thickness, as dis-
experience with wet XLPE and EPR designs, predomi- cussed in Section 3.7.3. When the insulation thickness is
nantly at 69 kV, but also up to 138 kV. It has to be noted decreased for a given conductor size, the capacitance
that these cables have conservative insulation thick- increases. Therefore, in particular for long cable installa-
nesses, as shown in Table 4-9, and are AEIC CS7 full- tions at high voltages, where the charging current and
wall- designs, resulting in insulation shield stress levels capacitance of the cable are a concern, the engineer may
mostly below 50 V/mil (2 kV/mm). Further, these cables increase the insulation thickness above the levels dic-
are typically installed in duct systems. Opinions differ tated by specifications, standards, or stress-level criteria.
whether these stress levels are too high for cables without
a radial moisture barrier. These cables have been used As given in Chapter 3, Equation 3-55, the dielectric
since the 1970s in the United States, as well as other losses WD in a cable of constant conductor size are pro-
countries, including Denmark, Switzerland, and Italy. portional to (E .εr. δ), the product of the design stress at
the conductor shield E, the relative permittivity εr and
AEIC CS9-06 and ICEA Standard S-108-720-2004 the dielectric loss angle δ . The loss factor of EPR is
specify only one design of “wet” cable—i.e., a cable appreciable compared to XLPE, but not to fluid-
without a metallic water barrier, which is at 69 kV. The impregnated paper (Table 3-11), which is of similar
external stress limit of 50 kV/mil (2 kV/mm) is suffi- magnitude. The comparatively low design stresses E
ciently low that the size and rate of water tree growth adopted at 138 kV and below for extruded cables, signif-
are acceptable. The resulting insulation thickness is 702- icantly reduce the dielectric loss and resultant tempera-
483 mil (17.8-12.3 mm). The stresses permitted in the ture rise to a small proportion of the conductor I2R loss
“dry” cable designs with metallic sheaths are signifi- such that EPR is competitive with XLPE.
cantly higher, such that if the wet design of 69-kV cable
is protected within a metallic barrier, it could be used as 4.17 CABLE CONSTRUCTION: OUTER LAYERS
a 138- to 161-kV cable.
Worldwide Practice 4.17.1 Bedding Tapes
Often referred to as European Practice and Japanese Bedding tapes are required by AEIC CS9-06 applied
Practice, stress levels specified and used outside North over the extruded core to protect the insulation shield
America are typically higher (see Table 4-12). With and insulation from thermal deformation or mechanical
extruded dielectric cables from most major manufactur- damage from overlying layers. The manufacturer is
ers sharing similar materials, technology, and processes required to state the radial expansion allowance
today, stress-based designs will become standard in designed into a particular cable construction. Bedding
North America in the future, especially as AEIC CS9 tapes are also called “cushioning tapes.”
(AEIC 2006) now specifies the maximum stress at each
transmission voltage. Manufacturers may use a compressible lightly foamed
rubber tape, creped paper tape, nonwoven swelling tape,
or a combination thereof. Figure 4-36 gives the percent-
Figure 4-36 Radial expansion of cylinders of cable insulation: LDPE, HDPE, XLPE, and EPR.
4-39
age radial expansion exhibited by different types of a large EHV cable, it is usual to apply a sandwich of
insulation and was derived from Figure 4-7 for the volu- both cushioning tapes and water-blocking tapes. Creped
metric expansion. For core-radius-to-conductor-radius paper or water-swellable tapes alone may suffice for an
ratios of greater than 2:1, the radial expansion allow- HV cable.
ance is obtained by multiplying the percentage radial
expansion directly by the core radius. For smaller ratios, The cushioning and water-blocking tapes are required
it is beneficial to calculate the reduction in radial expan- to be electrically conducting to transmit the capacitive
sion resulting from the presence of the conductor. Thus charging current from the XLPE insulation to the shield
a 50-mm radius XLPE core containing a 25-mm radius conductor. The tapes themselves must be protected from
conductor operating at 90°C will expand by 3.5%, which losing conductivity by being crushed or broken by core
requires a 1.75-mm radial expansion allowance. The expansion, remembering that once the core has cooled
same core operating at an emergency temperature of and contracted, there must remain an unbroken radial
110°C will expand by 7%, which requires a 3.5-mm path to the shield conductor.
radial expansion allowance.
Sufficient expansion allowance and thermal protection
If discrete shield conductors (also called neutral also have to be allowed for the temperature rise in other
conductors), such as stranded copper wires or tapes, are manufacturing processes. For example, the extrusion
to be applied, or a corrugated metallic sheath is to be temperature of an aluminum sheath is above 400°C and
applied, then an expansion allowance is required to of an HDPE sheath approximately 160°C.
protect the extruded core from deformation. Experience
has shown that lead-alloy sheathed cables are less The presence of poor radial contact may be revealed by
vulnerable partly because the inner surface of the sheath the presence of partial discharge “contact noise” during
is a smooth cylinder, and because the lead sheath and the routine HV process test. Contact noise (current-
polymeric jacket are capable of yielding under internal induced PD) from the cushioning and shield layers is as
pressure. However, it is recommended that cushioning equally unacceptable as voltage-induced PD from the
tapes be applied both to protect the core during the heat insulation because the PD sparks may eventually erode the
of lead sheathing and to reduce expansion forces to an insulation shield and lead to primary insulation failure.
acceptable level. Figure 4-37 shows a heat-distorted
insulation shield on a 132-kV XLPE cable that has been 4.17.2 Water-blocking Tapes
subjected to a high temperature rise during a short- The water-blocking function is a mandatory require-
circuit test for 1 s. The extruded core has expanded and ment in AEIC CS9-06 for XLPE insulated cables. Water
is heavily indented with the shape of the bedding tape may enter a cable laid in a flooded duct-manhole system
and the shield wires. AEIC CS9-06 requires that the or in a buried installation either during installation
bedding limits the deformation to less than 5% of the through the cable end caps or in service through third-
insulation thickness. party dig-in damage. Water may travel longitudinally
under the metallic water barrier and also along the con-
Sufficient radial thickness of compressible bedding ductor if the cable fails electrically. The electrical
tapes is required to be calculated. For example, if the stresses in AEIC CS9-06 (AEIC 2006) are now suffi-
compressibility of a tape is 70%; then for the previous ciently high at both the conductor and the insulation
90°C example, the tape thickness would need to be shield that eventual water tree growth and failure are
1.9 mm and for the 110°C example, 3.9 mm. Water- significant risks and may require replacement of the
blocking tapes are also required to be compressible. For complete cable span. The risk of water entry also exists
in the factory from either a water-cooling bath or from a
water-filled HV test termination.
Water-blocking tapes may be applied over the core and

between the layers of the conductor, such that water
ingress is restricted to a few meters length on each side.
This length is shorter than would be cut back in the
event of electrical failure to provide room for two
straight joints to be inserted. In contact with water,
water-blocking tape rapidly produces a gel that swells in
height by a factor of greater than ten to occupy all avail-
Figure 4-37 Radial expansion and indention of an able empty spaces and hence block longitudinal water
insulation shield during a short-circuit test flow. Figure 4-38 shows the local swelling of samples of
(courtesy of CCI). tape and yarn that have been activated by droplets of
4-40
water. In addition to swelling, the requirements are value advised for the insulation shield. The moisture
compatibility with the cable insulation and shields, uni- content of a factory-degassed, dry-cure XLPE insula-
form dispersion of the water-swelling agent, and ther- tion is typically less than 15 ppm, measured by the Karl
mal stability during the life of the cable. Water-blocking Fischer titration method. The semiconducting insula-
tapes ensure a uniform dispersion of the agent and are tion shield has a high moisture absorbency and so is a
more practical to handle than a powder in the cable fac- good indicator of the presence of moisture, with possi-
tory and during jointing. ble saturation values as high as 1000-2000 ppm. The
normal moisture content varies according to the mate-
Water-blocking tapes contain powder, cut fibers, or a rial and cable construction, and so statistical process
coating of polyacrylate. Polyacrylate is a super-absor- control (SPC) values should be recorded for informa-
bent polymer. Super-absorbent materials have been tion to establish an in-house pass level. A normal shield
applied in cable applications since the 1980s. Polyacry- from a factory sample typically has a moisture content
late is a hydrophilic crosslinked polymer. When dry, the of less than 150 ppm.
polymer chain is coiled and lined with carboxyl groups.
When wet, the carboxyl groups change to negatively 4.17.3 Metallic Shields and Sheaths
charged carboxylates, which force the chains to uncoil, Most extruded dielectric transmission cable designs
resulting in a gel with a rapidly expanding volume. In include a metallic sheath as opposed to a simple metallic
tape form, the material is usually combined with one or shield in the form of concentric neutral wires or helically
more nonwoven polyester layers. In the presence of applied copper tapes. The fundamental difference
water, a single tape produces a dramatic swell height of between a metallic sheath cable design and a metallic
up to 12 mm. British Standard tests require that the shield cable design, according to this definition, is that
achieved swell height be greater than 7 mm within a 10- the metallic sheath provides a radial moisture barrier
minute period. When dry, the material contracts and is and the metallic shield does not. This is the reason why
capable of being reactivated. metal shield cables are also commonly referred to as
“wet designs.”
Concern has been expressed that water-blocking tapes
may absorb moisture from the atmosphere in the cable Wet design cables have large insulation thicknesses and,
factory and transfer this via the insulation shield to the therefore, low insulation shield stresses, which are more
XLPE insulation. It is, therefore, important that the tolerant of a tendency to form water trees. The insula-
cable factory is dry and that the cable is sheathed soon tion thicknesses are listed in Appendix F of ICEA S-
after the water-blocking tapes have been lapped. Sam- 108-720 for 69-, 115-, and 138-kV cables. AEIC CS9-06
ples of the cable should be dissected during the routine requires that cables with a conductor shield stress of
process test and the moisture content of the shields and greater than 4 kV/mm and insulation shield stress of
the insulation measured. Specification BS 7970 (British 2 kV/mm shall be provided with a metallic moisture bar-
Standards Institution 2005) requires that the moisture rier to maintain dry insulation. AEIC only lists a wet
content of the tape be less than 70% of the equilibrium design cable at 69 kV.
value at 75% relative humidity and 20 o C. The XLPE
moisture content required in 09-16 2004 (Energy Net- Functions
works Association 2004) is less than 150 ppm, with no ICEA S-108-720-2004 requires that the metallic shield
or sheath provides a concentric conducting path for
insulation charging currents and leakage current, as well
as neutral current, phase unbalance current, fault cur-
rent, and surge current. If a metal foil water barrier is
fitted, it is in addition to the metallic screen conductor,
which should be designed to carry current alone.
The basic functions of metallic screens and sheaths

relate to the manufacturing, installation, and operation
of a cable and are as follows:
• Carry the capacitive and dielectric loss currents.
• Carry and withstand, thermally and mechanically,
the single-phase fault current.
Figure 4-38 Water-swellable tape and yarn • Carry the circulating sheath current in a solidly
(courtesy of GECA). bonded system and the residual circulating sheath
current in an unbalanced cross-bonded system.
4-41
• Provide a ground path in the event of cable damage have been found to kink and split and then to punc-
for the protection of personnel and to limit the step ture the insulation shield and insulation.
potential. • Helically applied metallic tape shields. Thin copper
• Mechanically protect the cable core. tapes have been found to be stretched during core
• Withstand the mechanical stresses during pulling and expansion, becoming wrinkled and split during con-
traction until finally a sharp edge punctures the insu-
installation.
lation. Sufficient bedding tapes are required to
• Provide a radial moisture barrier in the case of a absorb expansion and protect the core.
metallic sheath.
• Longitudinally applied and overlapped corrugated
• Provide a longitudinal moisture barrier in combina- tape. These tapes are capable of expanding and con-
tion with swellable tapes. tracting, and no adverse effects have been reported.
• Accommodate the thermomechanical expansion and However, caution should be exercised, because they
contraction of the cable core. are primarily used on lower voltage cables, and they
contain sharp inner edges. Protective bedding tapes
The roles of the shield conductor and ground conductor are required.
are detailed in Chapter 3, Sections 3.2.3 and 3.6.5. The • Flat strap shield. Double layers of thick copper tapes
primary requirement is sufficient cross-sectional area to have been found to prevent radial expansion of the
carry the specified short-circuit current and time, with- core, thereby indenting it and also stiffening the
out exceeding the specified short-circuit temperature. cable, reducing its flexibility in duct installations.
Short-circuit temperatures are given in Table 3-2, and These are not recommended for XLPE cables that
the basic method of calculation in Equation 3-28. The may operate near 130 °C.
detailed method is given in Chapter 11.
4.17.5 Wet Designs of Cable with Metallic Shields
The secondary requirement is for solidly bonded shield
Simple metallic screens, consisting of either concentric
circuits, the consequence of which is the flow of circulat-
neutral wires, as shown in Figure 4-39, or helically
ing current during normal operation and the generation
applied copper tapes, are used on many EPR cable
of losses. The relationship between the temperature rise
designs as well some XLPE designs, as discussed in Sec-
and shield resistance is checked to ensure, if freedom of
tion 4.15.
design permits, that an optimal minimum heat genera-
tion exists. The heat generation is partly a function of
Although these metallic screens are very cost-effective,
(I2R) and partly of (V2/R), where R is the shield resis-
they do not provide a radial moisture barrier, elevated
tance, V the voltage induced by the current in the pri-
short-circuit capability (depending on the total cross-
mary conductor and by the shield currents, and I is the
sectional area of the screen wires), or increased mechan-
shield current. Note that within limits the shield resis-
ical protection. Semiconductive swellable tapes are typi-
tance can be increased by reducing the cross-sectional
cally used between the cable core and the neutrals or
area and changing to a higher resistivity metal (for
copper tapes to provide a cushion layer and to prevent
example, from an aluminum to a lead-alloy sheath).
the longitudinal propagation of moisture in the screen
area. Also, a separator tape or a swellable tape is com-
4.17.4 Shield Conductors
AEIC CS9-06 refers to ICEA S-108-720-2004, which
lists shield types in North America.
Types of Metallic Shields
Appendix B of ICEA S-108-720-2004 provides the fol-
lowing useful commentary on performance experience
with different shield types, particularly under emergency
operating temperatures, to which the author’s experi-
ence has been added:
• Wire shields. Wire shields can imbed into the insula-
tion shield; as a result, suitable bedding tapes are
required to minimize the problem. There can also be
a problem if a voltage-equalizing tape is applied in
opposite lay, open spiral, over the wires. Some types Figure 4-39 “Wet” design of XLPE cable
(courtesy of Southwire).
4-42
monly used above the concentric neutrals and under the Aluminum forms an oxide layer on the surface, which
cable jacket, because neutrals are not encapsulated in may affect the electrical contact between the insulation
the jacket, as would be common for distribution cables shield and the sheath itself. The manufacturer designs
in North America. the bedding layer between the insulation shield and the
sheath accordingly and may use semiconductive or
4.17.6 Metallic Sheaths metal-woven tapes for this application. Further, alumi-
A variety of metallic sheaths are used for extruded- num sheaths are typically provided with an external cor-
dielectric transmission cables today: rosion protection layer consisting of a bituminous
compound. The thickness of an extruded aluminum
• corrugated extruded aluminum sheath sheath depends mostly on the extrusion equipment and
• aluminum or copper foil laminate the sheath diameter, and is in the range of 75 to 125 mils
• corrugated welded aluminum sheath (1.9 to 3.2 mm).
• smooth welded aluminum sheath Thus, corrugated extruded aluminum sheaths have a
• smooth welded copper sheath (special cable applica- very high short-circuit capability and provide excellent
tions) mechanical protection. The high conductivity has been
• corrugated welded copper sheath used to advantage in applications in which power cables
are installed adjacent to communication cables, because
• corrugated welded stainless steel sheath the sheath carries a high proportion of the return cur-
• lead-alloy sheath rent during a short circuit and so minimizes the external
magnetic field and the induced voltage (Gregory et al.
The least common types—bronze sheaths, smooth 1999). On the other hand, the high thickness and con-
extruded aluminum sheaths, and, to the extent that they ductivity of this metallic sheath increase the sheath eddy
qualify as metallic sheaths, longitudinally corrugated current losses. The losses are minimized by increasing
tape shields—are not further discussed in this chapter. the phase spacing and thence restore the current rating.
As with all corrugated sheath cables, the cable diameter
Corrugated Extruded Aluminum Sheath is larger than for other designs, and the soldering of,
Extruded aluminum sheaths (Figure 4-40) for power and the connections to, the aluminum sheath require the
cables were first introduced by Siemens in the 1940s and appropriate skills and procedures.
have been popular ever since. Corrugated extruded alu-
minum sheaths are preferred in Japan, China, and Sin- Aluminum or Copper Foil Laminated Sheath
gapore with XLPE cables. Aluminum laminated foil sheaths were first introduced
in Germany in the 1970s and adopted the design and
Machined billets of commercially pure aluminum are manufacturing process from telecommunication cables.
extruded over the cable core using a hydraulic press. In- Aluminum laminated sheaths have gained popularity
line with the extrusion press, the aluminum sheath is ever since and are dominant in Germany and widely
corrugated before the sheathed cable core is taken up on used in Europe at all voltage levels. Copper laminated
a manufacturing reel. sheaths are a variation using copper foil in lieu of alumi-
num foil and are preferred by some Swiss and a number
of U.S. utilities. Unlike telecommunication cables,
power cables are heavy and are required to withstand
high temperature excursions with radial and longitudi-
nal expansion; as a result, CIGRE published test recom-
mendations in ELECTRA 141 (CIGRE 1992). A new
CIGRE Working Group was set up in 2006 to prepare
updated recommendations to be based on manufactur-
ing and service experience.
Typically, copper concentric wire neutrals are applied to

the cable core over a layer of bedding tapes. In some
cases, aluminum neutral wires have been applied for alu-
minum laminated sheaths, or no neutral wires have been
used, as determined by short-circuit requirements. Note
Figure 4-40 Corrugated extruded aluminum sheath that a separate shield conductor is now required by
on 220-kV XLPE cable (courtesy of Jicable). AEIC CS9-06 to carry current when a longitudinally
applied metal foil moisture barrier is used. Another
4-43
layer of semiconductive tape is applied over the neutral The tape is formed to a round tube encapsulating the
wires. The laminated sheath consists of a 6 to 20 mil cable core; it is TIG (tungsten inert gas) welded, and
(0.15 to 0.51 mm) thick copper or aluminum tape that is then corrugated. Compared to its extruded aluminum
longitudinally applied to the cable. The copper or alu- sheath counterpart, the sheath is thinner and has a
minum tape is guided into a spiraled cone through welded seam.
which the cable passes, which forms it into cylinder with
Smooth Welded Laminated Aluminum Sheath
an overlapped tape seam. The copper or aluminum tape
Recently introduced in France as a replacement for lead
is supplied coated with a copolymer, and, as the cable
sheaths, the welded laminated aluminum sheath is gain-
jacket is extruded in-line with the application of the
ing popularity in Europe, because it incorporates some
metal tape, the temperature of extrusion melts the co-
of the advantages of laminated sheaths as well as corru-
polymer and bonds the metal tape to the jacket, or as
gated welded sheaths.
the name indicates, laminates it to the outer jacket. The
bending stiffness of the jacket, together with the layer of
As for the other sheath types, semiconductive or con-
bonded polymer, prevents the metal foil from wrinkling
ductive swellable bedding tapes are applied to the cable
and buckling when the cable is bent. The closure of the
core prior to the sheathing process. The aluminum tape
overlapped longitudinal seam is difficult to achieve with
is applied longitudinally and formed to a round tube
precision and so creates a possible entry point for mois-
encapsulating the core. The aluminum tape thickness is
ture. To achieve a moisture seal, most manufacturers
in the range of 30 to 100 mils (0.8 to 2.5 mm). The alu-
apply a hot melt adhesive prior to the closure of the
minum tube is then laser or TIG welded and coated with
overlap, in addition to the layer of co-polymer. Similarly
a laminating adhesive. The cable jacket is extruded in-
when new spools of tape are fitted to the end of the old
line with the sheath application and laminated to the
spool, the ends of the tape are welded metal-to-metal to
sheath, and thus allows bending of the smooth sheathed
ensure continuity of the moisture seal and to prevent
cable onto the take-up reel without buckling of the alu-
sections of tape from electrically floating and sparking
minum sheath.
in service.
The welded laminated aluminum sheath combines a her-
In most foil laminated sheath applications, the current is
metic design with a small diameter, and substantial short-
carried by a layer of round wire neutrals, which form the
circuit capabilities without the need for additional wires.
shield conductor. For applications with high short-
circuit currents, where the emergency operating temper- Corrugated Welded Copper Sheath
ature is limited to 105°C, one or two layers of flat strip First introduced in Germany by Kabelmetall, corru-
neutrals may be applied to increase the cross section of gated welded copper sheaths are used widely in Switzer-
the sheath, taking care not to restrict the radial thermal land and the United States at all voltage levels.
expansion of the core.
As with the other sheath types, semiconductive swellable
The cables are lightweight and have a comparably small bedding tapes are applied to the cable core prior to the
outer diameter. From an installation and operating per- sheathing process. The copper tape is applied longitudi-
spective, the laminate tape has to be exposed and electri- nally and formed to a round tube encapsulating the
cally bonded to the neutral wires at every splice or core, as shown in Figure 4-42. The copper tape thick-
termination. The thinness of the tape compared to the
jacket requires care to be taken. Mechanical protection
to the cable is provided by the jacket and not by the
metal laminate. For this reason, foil laminate cables are
preferred for semiprotected applications such as cable
tunnels and sealed ducts where the risk of third-party
damage and consequent water entry is reduced.
Corrugated Welded Aluminum Sheath
The corrugated welded aluminum sheath (Figure 4-41)
is similar to both the corrugated extruded aluminum
sheath already discussed and the corrugated welded
copper sheath discussed below. The material used is a
longitudinally-applied aluminum alloy tape with a
thickness between 40 and 80 mils (1 and 2 mm).
Figure 4-41 Corrugated welded aluminum sheath
4-44
ness is in the range of 25 to 50 mils (0.6 to 1.27 mm). Corrugated Welded Stainless Steel Sheath
The copper tube is then TIG welded and corrugated The applications for stainless steel sheathed cables are in
before the sheathed core is taken up on a reel. high-power, low-loss circuits, and in corrosive environ-
ments. Low-loss stainless steel sheaths are used in Japan
The corrugated welded copper sheath has a high short- in some 275-kV and 500-kV applications (Figure 4-16).
circuit capability and provides good mechanical protec- Their purpose is to achieve high ampacity by reducing
tion as well as good connectivity for bonding and sheath eddy current losses in cables installed in close
grounding. As with all corrugated sheath cables, the cable proximity—for example, when cleated together in tun-
diameter is larger than for other designs (Figure 4-43). nels. Thin stainless steel sheaths have high electrical
resistivity, which minimizes eddy currents, while provid-
ing excellent mechanical robustness and protection to
the cable. A concentric stranded copper wire conductor
is required inside the stainless steel sheath to act as a
shield and ground conductor for capacitive and short-
circuit currents. The shield wires are spaced to prevent
eddy current losses. The sheaths are cross bonded and
so do not carry longitudinal circulating currents. Stain-
less steel has also been selected in Australia and in the
Pacific Rim for installation in chemical plants where a
higher risk of corrosion may exist.
Lead-Alloy Sheath
Lead-alloy sheaths for power cables were first introduced
around 1880 in Switzerland. They have been widely used
for most cable applications and are still preferred in the
Middle East. They are still used in Europe and the
United States, although in declining quantities, while
Japanese manufacturers produce lead-sheathed trans-
mission cables for export only. The first 345-kV XLPE
cable in a long-length duct-manhole circuit with joints in
North America has a lead-alloy sheath (Figure 4-1).
Lead sheaths are made using one of several types of lead

alloys, such as those listed in ICEA S-108-720 Appendix
I, or Cu-Te lead alloy as commonly used in Europe.
Figure 4-42 Copper tube entering TIG welding head These lead alloys are more resistant to intergranular
(courtesy of Southwire). cracking due to vibration and heat cycling compared to
other alloys. The lead sheath is extruded onto the cable
core using a continuous screw type press. Some manu-
facturers add an anticorrosion layer. Lead sheath thick-
nesses range from 1.5 to 4 mm (60 to 160 mils).
Lead sheaths may not be sufficient to carry the short-

circuit current in some applications due to their higher
resistance and temperature limitations. For such appli-
cations, copper or aluminum concentric wire neutrals
may be added under the lead sheath to improve the
short-circuit capacity and reduce the overall weight.
Lead sheaths have been used for a long time in land and
submarine applications and, as a result, have a long
operating experience. They have excellent corrosion
resistance and are preferred in wet applications. Lead
Figure 4-43 Welded copper sheath (courtesy of has a low elastic limit and so lead sheaths readily creep
Southwire). under the action of ground subsidence and jacket retrac-
4-45
tion, thereby protecting the accessory sheath termina- methods have utilities interested in the actual distrib-
tions from experiencing high tensile loads. Lead sheath uted temperature profile of a cable in operation. In addi-
cables are heavy compared to other designs and in some tion, Distributed Temperature Sensing (DTS) enables
cases limit the lengths of cable that can be transported. utilities to implement dynamic rating systems when
Lead sheaths are malleable and resistant to damage, but fibers, DTS and dynamic rating models are combined.
are not as robust in resisting installation and third-party
damage as other thick-wall extruded and welded sheath The typical type of fiber used for this application is
types. common telecommunication-grade 50/125 μm diameter
multimode fibers. In combination with the temperature-
4.17.7 Other Sheath Considerations for Extruded monitoring devices available today, these fibers offer
Dielectric Cables good temperature accuracy and spatial resolution but
Extruded dielectric cables are subject to substantial only for limited lengths. For longer-length cable circuits,
radial expansion and contraction during heat cycling in single-mode fibers may become necessary, which in turn
service compared to self-contained fluid-filled cables. reduces temperature accuracy and spatial resolution.
The sheath design has to take this expansion and con-
traction into account. When concentric neutral wires The cladding and the coating of the glass fiber deter-
are used, alone or in combination with laminated or mine its temperature tolerance. Standard telecommuni-
lead sheaths, adequate bedding is required to limit the c at i o n - g r a d e fi b e r s a re l i m i t e d t o a m a xi mu m
indentation of the wires into the cable insulation shield. temperature of 70°C. This, however, is sufficient for the
For cables with rigid tubes, such as the extruded alumi- typical operating temperature in the sheath area. Higher
num and the welded aluminum and copper sheaths, an temperature ratings are available at increased cost.
adequate expansion gap has to be provided for the cable
core expansion. Glued laminated sheaths, and in partic- For most designs, one or more single and or multimode
ular lead sheaths, have to provide a certain elasticity, fibers are contained in a small, stainless steel tube. The
because they expand and contract with the cable core. diameter of the stainless steel tube varies, but is in the
The jacket composition, thickness, and application range of 1 to 2 mm (40 to 80 mils).
should be designed accordingly.
The ideal place for the fiber in any cable would be at or
All designs have to provide adequate electrical contact inside the conductor, which is both the hottest part of
between the cable core insulation shield and the metallic the cable and, at the same time, the cable component
sheath. Semiconductive, or sometimes conductive with the fastest temperature response to operational
(metal woven), tapes are selected specifically for each changes. Placement of the fiber in the conductor is,
sheath type. however, not very practical, considering the conductor
voltage, exposure to high operating temperature, and
Cables with rigid tube sheaths, in particular corrugated complexity of splicing. Therefore, the next best place is
sheaths, permit higher clamping forces in the installa- the incorporation of the fiber tube underneath the
tion design, while other sheath types limit the clamping sheath, as shown in Figure 4-44.
force, requiring the use of elastomeric liners or springs.
For cables with concentric neutral wires, one or more
The allowable short-circuit current for a specific sheath stainless steel tubes typically replace one or more of the
type is a function not only of the sheath material and
cross section, but also of the maximum allowable tem-
perature as a result of short-circuit duty. This tempera-
ture depends on the nonadiabatic effects, if considered,
as well as on the temperature tolerance of the adjacent
cable components. There are different specifications,
standards, and regional practices, and manufacturers
may permit different maximum temperatures for the
same sheath material.
Addition of DTS Fibers into the Cable
The operating temperature of a cable is mostly calcu-
lated for the cable and its operating environment as
described in Chapter 11. However, variations of the Figure 4-44 Distributed temperature sensing
thermal environment and assumptions and inaccuracies optical fibers incorporated into bedding
related to the thermal environment and the calculation tapes (courtesy of Southwire).
4-46
neutral wires and are applied together. The two adjacent 4.18 CONSTRUCTION: JACKETS
neutral wires may be oversized relative to each stainless The integrity of the jacket is internationally recognized
steel tube in order to provide crush protection for the as the key factor in ensuring the longevity of the cable.
tubes and the fibers contained therein. The jacket usually deteriorates at a far greater rate, due
to external damage during installation and in service,
For other sheath types, the fiber tubes will be sand- than the primary XLPE or EPR insulation, due to elec-
wiched in the bedding tape layers between the cable trical and thermal aging. Money should never be saved
insulation shield and the metallic sheath, as shown in by skimping on the quality or the thickness of the
4-44. Additional strength members, such as wires or flat jacket. Thicknesses specified by ICEA S-108-720-2004
cable assemblies or semiconductive composites may be are given in Table 4-13. The table also gives the magni-
used for both crush protection and to limit the indenta- tude of the factory in-line spark test that is applied after
tion of the fiber tube into the insulation shield. extrusion. The jackets on specially bonded cables are
supplied with a conducting layer. If this is a graphite lac-
Another possibility is to incorporate the stainless steel quer, it is applied after the spark test. If a layer of
fiber tube into the outer jacket encapsulating the metal- extruded semiconducting polymer is applied in the form
lic sheath, which, while thermally removing the fiber of a bonded coextrusion, then its thickness may be
further from the conductor, avoids any interference in included in the values given in Table 4-13 up to a maxi-
the interface between the insulation shield and sheath. mum of 20%. It is no longer possible to perform an in-
line spark test, and so a dc voltage test is specified to be
In many applications in North America, cables are performed on the completed cable shipping length of
installed in duct-manhole systems, which require that 200 V/mil (8 kV/mm) on the average thickness in Table
the fiber tube be incorporated into the cable construc- 4-13 up to a maximum of 25 kV for one minute.
tion, as described above. Alternatives that have been
used are to strap a fibre cable or an empty auxiliary tube The jacket is required to protect the inner layers of the
(sometimes called a sub-duct) to the power cable, in the cable from the ingress of water. The main effects of
latter case the fibre is blown in later. Another solution is water are corrosion of metals, loss of contact conductiv-
to place a small auxiliary duct in the vicinity of the hot- ity between metals, and water treeing in PE, XLPE, and
test duct within the ductbank. In tunnel and buried EPR insulations. The jacket is also required to give a
applications, the fiber is often fixed to the outer surface degree of protection from the diffusion of aqueous solu-
of the jacket after the cable has been laid and jointed. A tions of natural ionic salts and industrial and agricul-
variant of this is to attach an empty tube to the cable tural chemicals. Ionic solutions set up an electrolytic cell
and later blow the fibers into position, thus permitting between dissimilar conducting materials, such as the
replacement of the fibers, should they deteriorate with graphites and the different metals, which accelerate cor-
time. The advantages of external fiber placement are rosion. The rate of corrosion is also increased by the 60-
that the risk of fiber damage during cable manufacture Hz voltage in specially bonded systems, induced in the
and installation is eliminated, and the length of fiber underlying metallic sheath, foil, or shield wires. In these
installed without joints is increased thereby reducing the systems, it is good practice to break the electrolytic cell
number of fiber joints and the total signal attenuation. by insulating the metal surface by, for example, a bitu-
The disadvantage in external fibre placement is that the minous coating.
sensitivity of temperature detection is reduced, both in
reduced temperature rise and in increased time response The diffusion rates for water through different combina-
to rapid load changes. For more information on DTS, tions of metallic sheaths and polymeric jacketing mate-
see Sections 12.2.8 and 11.5.1.
Table 4-13 Jacket Thicknesses for Extruded Cables (ICEA S-108-720-2004)

Diameter over Shield
or Sheath Thickness
Minimum Maximum Spark Test
Shield Type in. mm mils mm mils mm kV
0-2.500 0-63.50 100 2.54 150 3.81 10
Tape or Wire Shield
>2.500 >63.50 125 3.18 185 4.70 12.5
0-2.250 0-57.15 70 1.78 105 2.67 7
Metal Sheath Shield 2.251-3.000 57.18-76.20 85 2.16 135 3.43 7.5
>3.000 >76.20 100 2.54 160 4.06 10
4-47
shown in Figure 4-45. Polyethylene, and in particular

Table 4-14 Moisture Diffusion Rates through
Sheath/Jacket Combinations
HDPE, is significantly more robust than PVC at the
higher temperatures that occur due to solar heating dur-
Diffusion rate
ing installation and due to operation at emergency tem-
Sheath Jacket [μg.m/m2.day]
peratures. A temperature difference of typically 10°C
Corrugated Aluminum PE 0
exists between the conductor and jacket. Thus when the
Laminated Foil Copper or Aluminum PE <0.005
conductor is heated to a prolonged emergency tempera-
None HDPE 20
ture of 105°C, the jacket experiences a temperature of
None PE 80
95°C. Caution should be exercised in the selection of the
None PVC 350
grade of PE jacket because an HDPE material at the
lower density end of the grade range will be softened at
rials are given in Table 4-14 (Ritter et al. 1994). The only 95°C and will have melted at 115°C, as indicated in Fig-
completely watertight material is an extruded or welded ure 4-7.
metallic sheath. The very small diffusion rate reported
for the metallic foil is attributed to the foil overlap. PVC jacketing materials are softened by the addition of a
None of the polymeric jackets is watertight, but they plasticizer which gives advantages of high flexibility and
have waterproofing properties. The diffusion rate minimal shrink-back. Zero-halogen fire performance
through the HDPE jacket (high-density polyethylene) is grades, using highly filled PE compounds, have low flam-
18 times less than that through the PVC jacket. The mability and have reduced smoke propagation. PVC
HDPE jacket is thus better suited to a wet design of jackets become brittle at low temperatures and should
cable without a metallic sheath, especially where there is not be installed at sub-zero centigrade temperatures.
risk of external water.
PE jackets are more robust, but are stiffer than PVC. PE
In addition to the 60-Hz induced voltage, the jackets on jackets can be installed at -10°C. The higher-density
specially bonded sheath systems experience a significant grades of PE jackets are more susceptible to stress
transient voltage to ground each time the circuit is cracking, which is the mechanism in which the combina-
switched or when it transmits a lightning surge. The tion of locked-in stress and a surface notch (scratch,
presence of sheath voltage limiters typically reduces this abrasion, or sharp indentation) generates spontaneous
to 5% of the surge voltage. The jackets on specially splits with time. ICEA S-108-720-2004 specifies that
bonded systems are usually specified to withstand both each grade passes an environmental stress cracking test
a commissioning HV test and a routine maintenance to ASTM D 1693, in which a sample of jacket material
test to demonstrate that they have not suffered damage is molded, slit with a razor blade to 20-25 mils
during installation. The jacketing materials must, there- (0.51-0.64mm), bent, and then immersed in the cracking
fore, be good electrical insulators with a high electrical liquid Igepal CO-630 at 50°C for 48 hours. The density
resistivity, or it will not be possible to apply a suffi- ranges of the base PE resin, less additives, in kg/m3 are:
ciently high test voltage with minimal leakage current
necessary to discriminate the fault current at sites of LDPE MDPE HDPE
jacket damage. Caution should be taken in the selection
of the jacketing material for water-immersed applica- 910-925 926-940 941-965
tion, such as direct burial and in ducts liable to flood,
because fire performance jackets usually contain fillers
that reduce their waterproofing and insulating proper-
ties when saturated.
The jacket is required to withstand the abrasion and

penetration that occurs when it is pulled into a duct-
manhole system or into a trench on rollers or skid plates.
In service in a duct system, the jacket experiences contin-
uous abrasion due to thermomechanical cable move-
ments, and high sidewall pressure at bends and indeed
wherever the cable is curved. In a buried installation, the
cable experiences vertical soil and vehicular loading and
at bends sidewall loading, which may cause pebbles to Figure 4-45 Comparative indentation depth of HDPE and
penetrate the jacket. The penetration properties of poly- PVC jackets at different temperatures (courtesy of CCI).
meric jackets are highly temperature dependent, as
4-48
PE jackets exhibit significant locked-in retraction loads, was mainly for MV and HV cables, this required the
which exert high tensile forces on a large cable of 0.5- addition of a higher level of peroxide to achieve high
1.5 t at the accessories. PE jacket retraction has been output line speeds for thin-wall insulations. However,
recorded to break plumbs on lead and aluminum this produced an increased level of scorch contaminants
sheaths, the foil on laminate sheaths, and bonding con- (ambers) that reduced the length of the extrusion run
nections to shield wires. Thus accessories are required to and an unnecessary level of by-products in thicker-insu-
be designed to either withstand the forces or to provide lation EHV cables that increased the time of the degas-
a tolerance for movement. sing process. The start of volume manufacture of EHV
XLPE cables (230-500 kV) in 2000-2005 initiated the
development of compounds with reduced peroxide lev-
4.19 MANUFACTURE OF COMPOUNDS
els and a base PE resin with additional functional side
group molecules that increase the crosslinking efficiency
4.19.1 Manufacture of XLPE Insulating Compound
and so reduce the volume of dicumyl peroxide and its
The supplier of the base PE resin (i.e., PE before it is by-products. The PE resin with functional side groups
compounded with additives) manufactures very high has also been developed for dc transmission with the
volumes of low cost “film”-grade PE material for the objective of reducing the loss of dc and impulse voltage
general nonelectrical market. The central part of the strength due to the space charge accumulation of
manufacturing batch of PE produced in the high- charged by-products.
pressure reactor process is very clean and is the most
suited for high stressed EHV insulation, whereas the PE A low-density PE homo-polymer resin is preferred for its
at the start and end is suitable for either HV and MV high-stress EHV performance, because it has low viscos-
cable or nonelectrical applications. For EHV applica- ity and can be super-filtered to the highest levels of
tions, the center of the run is cut and segregated, and the purity. Viscosity is characterized by the MFI (melt flow
rate of sample testing is increased to verify that the index), in which the weight of compound flowing
center cut is clean. These additional processes and the through a vertical small bore tube in 10 minutes at 190°C
reduced volume of high purity material significantly is measured. The downside is that after extrusion it
increase the price of the compound. exhibits a significant rate of droop off the conductor and
hence loss of circularity and concentricity in some types
Additives are mixed into the PE resin to provide process of CV lines. Droop is more pronounced on cables with a
capability. The pellets of pure PE resin are first “com- high ratio of core diameters and on EHV cables because
pounded” (mixed) with antioxidants and process-aids. of the increased time to raise the thick insulation and
The antioxidants protect the PE from the combined large conductor thermal mass to the crosslinking tem-
effects of residual oxygen (air) entrained within the perature. Table 4-15 shows that grades of PE are avail-
extrudate melt in the factory curing process time of ~10 able with a 250% increase in viscosity to permit the
hours at 150-200°C and within the cable construction in manufacture of improved geometry cables for EHV
the service life of around 25-40 years at 90°C. The pro- application. The downsides of low-droop compounds
cess-aids lubricate the PE through the compounding are that the compounds are more difficult to filter to
and extrusion processes when in contact with the hot high purity and that the extrusion torque and power are
metal surfaces of the extruder, strainer packs, and extru-
sion dies. The penultimate process is to super-filter the Table 4-15 Examples of XLPE Insulation Grades for
compound melt through 5-10 layers of stainless steel Different CV Lines
mesh, with the finest layer being in the range of 700- CV Line CV Line
1000 mesh, to remove any particles and agglomerations. Type 1 pType 2
To prevent initiation of premature crosslinking during Super-
filtration Low Droop
the compounding and filtering processes, the peroxide is PE Compound Property Material Material
added last—for example, by melting and adsorbing Melt flow rate
through the pellet surface—to achieve a typical level of g/10min 2 0.8
(2.16kg at 190oC)
2-2.5% by weight. Rheometry (Gottfert
Elastograft Nm 0.59-0.74 0.64-0.80
The compound formulations and processes are propri- torque value)
etary to the compound manufacturer. Suppliers of pre- Density kg/m3 922 923
compounded materials produce utilitarian formulations CV Lines Type 1: Inherently low droop lines: VCV, CCV silicone
lines, and CCV nitrogen lines with special gas circulation treatment.
suitable for application in a range of types of extruders,
CV Lines Type 2: Other types: MDCV and nitrogen CCV lines
dies, and CV lines, with a range of different insulation without special processes.
thicknesses and line speeds. When the market demand
4-49
increased by around 10%. The viscosity of the PE can be face, expressed by uniform distribution of particle sizes,
varied by altering the molecular weight (chain length) of minimum particulate contamination, and high chemical
a homo-polymer, or by using a co-polymer. Table 4-15 purity. Acetylene black is obtained by burning acetylene
shows that the density is little changed at 922-923 kg/m3. gas. The insulation and semiconducting compounding,
transporting, storing, and handling plants are com-
Some manufacturers of high-stressed EHV cables, par- pletely segregated within separate sealed buildings to
ticularly those in Japan, compound the insulation in- avoid the risk of cross contamination. The fine black
house to achieve the optimum combination of purity powder is difficult and hazardous to handle, and so is
and processing characteristics best suited to their CV loaded into the polymer to form a “master batch,” which
line and their cable designs. The risk of material con- is then pelletized. The semi-con compound is melted and
tamination by foreign particles during transportation is filtered to break up agglomerates and so minimize the
not changed, because both the base PE resin and the formation of shield protrusions (Figure 4-47). The
XLPE compound have to be transported to the cable dicumyl peroxide crosslinking agent is added last.
factory. In both cases, extreme care is required by the
use of dedicated, large-volume, van boxes, road tankers
(20 ton), or octabins (1 ton). The granules of PE or
XLPE are transferred from the road tanker or bin in
either a clean garage or a clean room. In-house com-
pounding has the advantages that: (a) immediately prior
to the final process, the addition of the peroxide, the
compound is super-filtered; (b) it is handled entirely
within enclosed factory pipe-work handling systems;
and (c) it is stored for the minimum period under con-
trolled temperature conditions in dedicated silos. In
comparison, precompounded material is transported
and stored for longer periods in a range of climatic con-
ditions, which risk sweat-out of the peroxide at low tem-
perature (<13°C), melting at >28°C, and premature
peroxide decomposition reactions at higher solar heat-
ing temperatures. If the temperature is too low, dicumyl-
Figure 4-46 Left: VCV tower. Center: XLPE compounding
peroxide may sweat-out of the pellets in the form of a plant. Right: CCV tower (courtesy of CCI).
surface dust to later form concentrated accumulations
in the extruder feed lines.
The building at the base of the CV tower in Figure 4-46

houses the clean rooms for transfer of PE from road
tankers, the insulation and semiconducting compound-
ing plants, and the storage silos that are directly con-
nected to the extruders in the top of the VCV tower (left)
and in the CCV stub tower (right).
4.19.2 Manufacture of Semiconducting

Compound
Approximately 40 % by weight of carbon black is loaded
into an EBA or EVA elastomeric resin base, which is
used to retain bending flexibility and match it to the
XLPE insulation. Acetylene black is the preferred type
Figure 4-47 Small protrusion from the semi-con
for EHV cables and is specified by AEIC CS9-06. It gives conductor shield into the insulation (courtesy of CCI).
the smoothest and most consistent shield electrical sur-
4-50
4.20 MANUFACTURE OF EXTRUDED CABLE The pellets of XLPE insulation and shielding com-
CORE pound are either lifted from the ground floor in sealed
stainless steel air lift pipes, or are lifted in their trans-
For reasons of economy and consistent extrusion qual-
port boxes, which are then located adjacent to the
ity, it is usual to run an extruder in a continuous opera-
extruder or one or two floors below it. The pellets are
tion lasting 10-20 days, termed an “extrusion campaign”
usually, but not always, dried before being fed into the
or “extrusion run.” The conductor is stranded in the
extruder hoppers.
maximum lengths possible that can be wound onto a
process reel and lifted for transportation within the fac-
Figure 4-49 shows a bag of pellets in a Class NASA
tory. This is usually less than the length of a core that
100,000 clean room being fed directly into an extruder
can be extruded in one campaign, and so it is necessary
hopper on the floor below, a method that is popular in
to join lengths of conductor together without stopping
North America for HV cables.
the extruder. The conductors are joined together when
they enter the extrusion line to form a continuous length
It is essential that the pellets are metered into the hop-
by means of temporary flush connections, usually com-
per and into the throat of the extruder at a constant
pression ferrules, but sometimes welded connectors. The
rate, or the phenomenon of “surging” can occur in
conductor is dynamically stored in a number of turns on
which the extruder delivers a time-varying output, as a
an accumulator consisting of two grooved reels. During
consequence of which the core diameter and in particu-
normal operation, the reels are positioned far apart.
lar the shield thicknesses fluctuate. One solution is to
During the conductor connection process, the reels are
use a loss-in-weight feeder.
moved together to release and feed conductor at a con-
stant rate into the extruder. This permits the trailing end
The extruder includes an annular helical gap between a
of the conductor to remain static for sufficient time to
long screw (Figure 4-50) with regions of different pitch
complete the conductor connection.
inside a cylindrical container, both of which are heated in
zones to precise temperatures. The extruders are referred
The speed of extrusion and crosslinking is low—for
to by the diameter of the screw and the maximum output
example, the typical throughput speed of a large 400-kV,
rate. Exam ple screw siz es used to manufacture
4000-kcmil (2000-mm 2 ) XLPE core is 0.15-0.3 m per
400/500-kV cables are 6 in. (150 mm), 7 in. (175 mm),
minute, thus the maximum length that can be extruded
and 8 in. (200 mm). The length-to-diameter ratio is typi-
in a continuous 10-20 day campaign is in the range of
1.3 to 5.4 miles (2.2 to 8.6 km). Taking a typical ship-
ping reel length of 2600 ft (800 m), the extrusion cam-
paign would produce 3-10 shipping reel lengths.
4.20.1 Extruders
Figure 4-48 shows the insulation extruder at the top of a
VCV (vertical continuous vulcanizing tube). The con-
ductor shield extruded is aligned facing the camera. The
insulation shield extruder is out of view on the left side.
Figure 4-48 VCV extruder floor. Left background: Figure 4-49 Insulation pellets loading from Clean
conductor shield extruder. Left foreground: triple head die Room into extruder throat (below floor) (courtesy
set. Right: insulation extruder (courtesy of CCI). of Southwire).
4-51
Figure 4-50 Extruder insulation and semiconducting shield screws (courtesy of CCI).
cally 20:1-22:1; thus an 8-in. (200-mm) screw would be 120°C. For EPR extrusion, a strainer pack typically
up to 14.5 ft (4.4 m) long. contains five layers of stainless steel mesh, with sizes
from 7 down to 82 mesh, with a typical hole/light size of
The slowly rotating screw has flights of variable pitch 7 mils (170 μm). For XLPE insulation, the mesh size is
and depth such that the first flights assist pellet mixing typically from 20 down to 100/200 mesh.
and consolidation, and the final flights compress the
melt and drive it out through the neck of the extruder Pressure-tight optical windows are provided in some
(Figure 4-48) as a clear, void-free liquid under pressures extruder necks to permit automatic “melt flow inspec-
as high as 700-1000 psi (50-70 bar). The pressure is tion” equipment to monitor the melt. Beams of visible
required to dissolve free air, force the melt through the light count and size the level of residual particulate con-
high hydraulic impedance formed by the strainer pack tamination and in some equipment identify the type of
and the die head, and resist the back pressure of 145 psi contaminant.
(10 bar) from the CV tube.
A similar extrusion process is followed for the semi-con
The neck connects directly to a strainer pack, consisting conductor shield and insulation shield; however, the
of a number of stainless steel meshes that are supported volumes of extrudate per unit length of cable are signifi-
from bursting by a thick steel breaker plate perforated cantly smaller than the insulation, requiring smaller
with an array of longitudinal holes, as shown in Figure screw diameters and lengths, as shown in Figure 4-50.
4-51. The melt passes through layers of progressively
finer woven, stainless steel mesh. The purpose of the The close arrangement of the three extruders necessary
strainer pack is to remove particles and agglomerations, for them to be connected to the triple extrusion die is
and to break up ambers and gels that may have been called the “extruder group,” as shown in Figure 4-52.
formed during material handling and extrusion. The
temperature of the melt is held at less than 140°C to pre- The semiconducting shield melt is injected first into the
vent premature crosslinking occurring in the strainer upstream end of the die body and then onto the conduc-
pack and dies. The typical extrusion temperature is tor. Figure 4-53 shows the extrusion channels in different
cross-head dies, which are so named because the melt is
injected into the body of the die at approximately 90° to
the axis in the cross direction of the conductor. The
semi-con melt in Figure 4-53 (left) is divided by a knife
edge gate into two semihelical channels, which change
the direction of the melt and which meet again further
downstream at the opposite side of the die. The down-
stream edges of the channels are of lower level to permit
the melt to overflow them and to form a concentric
annular flow over the cylindrical surface of the die. The
end of the die is formed by the “nipple,” which is a coni-
cal component machined to fit the diameter of the par-
Figure 4-51 Strainer housing and breaker plate ticular cable. The semi-con melt flows down the nipple
(courtesy of CCI). and then along a long cylindrical land until it flows onto
4-52
the conductor. The semi-con nipple and nipple holder fit tinue onto the cable. The radial knit lines can be seen in
inside the sleeve next to it, the outer surface of which the finished cable when thin slices are cut from the cable
holds the helical channels and nipple for the main insu- and viewed under polarized light. The high extrusion
lation layer. The insulation nipple holder and nipple fit pressures compress the insulation melt, which swells in
inside the sleeve of the outer semicon layer, not shown. diameter when it emerges from the die. This expansion
is guided by a trumpet-shaped flare. The flow of the
The knife edge gate, which divides the flow into the two melt is not completely parallel and streamlined, proba-
channels, and the arrow head on the opposite side, bly due to an inevitable slight eccentricity in the setting
which recombines the flow, produce two diametrically up of the dies at the start of the extrusion run and to the
opposite “knit lines” in the insulation melt, which con- significant swell of the melt in diameter. In some cables,
Figure 4-52 Extruder group (courtesy of General Cable Corp.).
Figure 4-53 Triple extrusion cross-head dies. Left: inner conductor

shield die and insulation die. Right: Insulation die, front and back
(courtesy of CCI).
4-53
this results in taking a thin wisp of semiconducting com- sion cost. Each line type is also capable of operating
pound radially inwards into the insulation along the outside its optimal range. Process development contin-
length of the cable, thus forming a deadly electrical ues to be directed at widening the optimal range of both
stress raiser, as shown in Figure 4-54. existing and new lines to address new cable opportuni-
ties. Specifications such as AEIC CS9 (AEIC 2006),
4.20.2 Types of CV Lines ICEA S-108-720 (ICEA 2004), IEC 62067 (IEC 2001),
The three layers emerge from the die as viscous liquids at and IEC 60840 (IEC 2004) do not mention types of CV
an extrusion temperature carefully controlled at ~120°C, line, but do give sufficient dimensional tolerance such
which is below the temperature at which the dicumyl that a cable from any of the CV line types, if carefully
peroxide crosslinking agent becomes reactive. To be able operated, would not be precluded. Examples of factors
to crosslink the XLPE insulation and the shield adjacent in CV line specification are:
to the conductor in a reasonable time, it is usual for the • Voltage range: comprising insulation and shield vol-
larger conductors used in HV and EHV cables to be pre- ume, thickness, and overall diameter. Insulation
heated prior to entering the extrusion head. thicknesses are given in Table 4-8—e.g., 303 mils
(7.7 mm) at 69 kV to 1002 mils (25.5 mm) at 345 kV.
It is then necessary to start the reaction as quickly as Note that some 400- and 500-kV cables have been
possible to maintain the circular shape and to limit the supplied with insulation thicknesses of 1260 mils
time for the melt to ‘sag’ from the conductor and spoil (32 mm). The range in thickness is ~4:1.
the circular shape of the core. The molten cable core
• Design stress range: comprising material types and
passes directly into a CV tube (continuous vulcanizing
handling, insulation and shield purity, and geometric
tube), typically containing a heat transfer medium of
precision. Design stresses are given in Table 4-8—e.g.,
nitrogen gas or silicone oil at 145 psi (10 bar) pressure.
100 V/mil (4 kV/mm) at 69 kV to 350 V/mil
The temperature of the insulation and shield melts is
(14 kV/mm) at 345 kV. Note that some 500-kV cable
raised to obtain an inner insulation temperature of
have stress levels of 16 kV/mm. The range in design
around 160°C by a combination of radiant heating from
stress is ~4:1.
the tube at 200°C and by circulation of heated gas or sil-
icone fluid. Precise and uniform control of the melt tem- • Conductor range: weight per unit length and reel
perature after it has left the die is critical in maintaining weight and conductor type—e.g., concentric, seg-
th e cylindrical in sulation shape. As previously mental milliken, copper, and aluminum. Dimensions
described, special insulation compound with a low MFI of conductor sizes can be derived from Chapter 3,
is available to reduce droop, but at the penalty of Tables 3-4 to 3-8—e.g., a 250-kcmil compressed cop-
increased difficulty in straining. per conductor has a diameter of 0.56 in. (14.2 mm)
and a weight of 0.77 lb/ft (1.15 kg/m) and a
Different types of CV line have evolved to successfully 5000-kcmil copper conductor has a diameter of ~2.5
meet the common objectives of extruding cable in the in. (63.5 mm) and weight of 15.9 lb/ft (23.6 kg/m).
transmission-class range, and this evolution will con- The range in weight is 20:1. The highest conductor
tinue as the demand for XLPE and EP R cables weight without insulation, suspended in a 200-m tube
increases. When first purchased, each CV line is speci- is approximately 4.5 tons. The highest conductor
fied to be optimized to address a certain product range, weight to be withstood and controlled by the input
a production throughput rate, and a competitive extru- capstan is typically twice this figure.
• Output: the output of large conductor EHV cable
from a CV line is approximately half that of a tape
lapping and impregnating line; thus a manufacturer
requires twice the number of CV lines to supply the
same market with XLPE cable. This applies strong
commercial drive to the development of higher-speed
CV line types.
There has been a significant move towards VCV lines in

Japan and Europe for EHV cables with large conduc-
tors. Other types of CV lines are in use with a good
track record in the manufacture of 400-kV XLPE cables
with large conductors, and appear likely to continue in
Figure 4-54 Cross-section of cable slice: penetration of the future. These lines are CCV-Silicone, CCV-Nitrogen,
semi-con from the insulation shield into the insulation and MDCV. They are each described below, following
knit line (courtesy of CCI).
4-54
the detailed description of a VCV line, because this is 1. The conductor is guided off the input reel into the
the simplest in concept. base of the tower.
2. Some manufacturers apply a temporary tape after
4.20.3 VCV (Vertical Continuous Vulcanization) stranding to protect the surface of the cable and
VCV lines produce excellent core geometry. The down- thence the shield and insulation from damage or
ward movement of the conductor and the gravitational contamination. The protective tape is removed at
droop of the PE melt after extrusion are in axial align- the base of the tower.
ment, thereby naturally maintaining conductor concen- 3. The conductor passes three or more turns around a
tricity and circularity. This also permits low viscosity dancer accumulator. In normal operation, the mov-
(high MFI) XLPE compound to be extruded with the ing wheel is at the extreme of its travel such that the
advantage that it can be pre-strained to the highest lev- maximum length of strand is stored. When addi-
els of purity. tional conductor is required during a continuous
extrusion run, the moving wheel is slowly moved
The length of the VCV tube necessary to heat and cool inwards so that the stored conductor is released at
the core determines the throughput speed and the the normal throughput rate. This gives time for the
height of the tower. The length of the VCV tube has to new leading end of the conductor to be connected
match the maximum output of the extruder. A 150-175 to the trailing end without stopping the CV line.
mm diameter extruder requires a tower of approxi-
4. At the top of the tower, the conductor passes over a
mately 80 m in height, whereas a 200 mm extruder
power-driven capstan wheel. The speed of the cap-
requires a tower of 100-120 m height.
stan is the key controllable process variable in the
CV line, and is used to achieve and maintain a con-
The disadvantages of VCV lines are that they are expen-
stant core diameter. The diameter of the hot core is
sive to construct and local government planning rules
measured automatically at the output of the
may prohibit or limit the height of the tower.
extruder and at the output of the CV tube at the
bottom of the tower. If the diameter of the cable
The key equipment in the VCV tower is generally com-
starts to reduce, the feedback control automatically
mon to each type of CV line. A typical description is
slows down the capstan, so that a greater insulation
given below in relation to Figure 4-55:
thickness is achieved from the constant-output
extruder group.
Figure 4-55 VCV extrusion and curing line (courtesy of EPRI).
4-55
5. The conductor is preheated by induction heaters. and the pressure prospectively lost, such that the
6. The extruder group applies the shields and insula- end 100-m length of the core and conductor is also
tion, operating at constant, synchronized screw process scrap. Thus, although a short CV tube may
rotational speeds. Note that the three extruders have have a lower output rate, its percentage yield is
been rotated from plan to elevational view in the higher than a longer tube.
diagram. 12. The core passes through a diameter gauge, as shown
7. An X-ray gauge automatically measures the diame- in Figure 4-56, which measures the final “cold diam-
ter and eccentricity of the hot core. This permits the eter” of the core at ambient temperature. This is the
concentricity of the dies to be adjusted manually at diameter that is checked against the specified value
the start of the run, and the mean diameter of the agreed with the conductor. Because of the uncer-
core to be automatically controlled during the run. tainty of the magnitude of the thermal contraction
between the hot and cold diameter measurements, it
8. In the heating zone, the cable is heated to crosslink-
is usual to apply additional insulation thickness.
ing temperature by a combination of radiant heat-
ing from the induction heated tube and by 13. Optionally the core may pass through an ultrasonic
circulation of heated nitrogen or silicone oil. The monitor, as shown in Figure 4-57, to detect the pres-
CV tube is pressurized to typically 10 bar to keep ence of harmful defects such as shield hang-ups,
the gaseous by-products of crosslinking in solution. particles, and voids. Figure 4-58 shows the cross sec-
The nitrogen gas or silicone oil has to be cleaned tion of a shield hang-up; for clarity, the shield and
and filtered of the crosslinking by-products and of conductor are shown as white in color. The use of
any thermal degradation, to prevent the insulation
from being contaminated by the circulation,
because the heat transfer media to some degree
impregnates the insulation.
9. The cooling zone contains recirculated water, which
is an efficient heat transfer medium. The water has
to be cleaned and filtered to protect the cable insula-
tion and the CV tube.
10. The bottom turn-round wheel alters the core direc-
tion from vertical to horizontal. Some VCV towers
have a vertical shaft that continues underground.
Some VCV lines are completely underground. The
turn-round wheel then turns the cable through 180o
so that it re-ascends the shaft back to the ground
surface. The XLPE insulation has to be sufficiently Figure 4-56 Diameter and eccentricity laser gauge at
outlet of line (courtesy of CCI).
cool that it will not be deformed by the wheel.
11. The cool cable core emerges through a compressed
rubber water seal at the base of the CV line, as
shown in Figure 4-59. The water seal has to with-
stand 10 bar water pressure, while allowing the
extruded insulation shield to slide through without
damage. The CV tube cannot be properly pressur-
ized, and the core consistently extruded, until the
leading end of the core emerges through the seal.
The small vibration of the emergence causes a small
disturbance at the die head and imprints an
unwanted annular “water seal mark” in the core,
which, if significant, produces a secondary mark
when this too emerges through the water seal. Thus
the setting-up process produces a significant in-pro-
cess scrap of one to three times the length of the CV
line—a length of 330-1000 ft (100-300 m). Similarly,
at the end of the extrusion run, when the trailing
end of the conductor and the connection to the Figure 4-57 Ultrasonic in-line monitor at outlet
hawser passes through the die, the core is disturbed of CV line (courtesy of Jicable).
4-56
4.20.4 CCV (Catenary Continuous Vulcanization)-

Nitrogen
CCV-nitrogen lines have experienced the widest use
worldwide, particularly for HV cables, because the high-
est output rates are achieved at the lowest cost by build-
ing long CV tubes at near ground level so that the rate
of cross-linking is matched to the output rate of the
largest extruders. The extruder group is mounted in a
stub tower, which is significantly shorter and lower in
cost than a VCV tower. Figure 4-46 compares the height
of a VCV tower containing two extruder groups side by
side and a CCV stub tower containing two CCV groups
one above the other.
The components of a CCV line are shown in Figure 4-60,

and are essentially the same as those in a VCV line. The
Figure 4-58 Shield hang-ups (courtesy of CIGRE). disadvantages of a CCV-Nitrogen line are gravitational
droop of the melt, which, if not properly managed, can
ultrasonic detectors is comparatively recent and has produce an eccentric and noncircular core, and the core
been applied to some EHV and MV submarine touches down onto the floor of the cooling zone. CCV-
cable applications (Smith et al. 2003). nitrogen lines are presently being developed to handle
14. The cable is driven by a caterpillar cable engine onto large conductor EHV cable, using high output extruders
a powered take-up reel (drum). The reel usually with large screw sizes of 7.87-11.8 in. (200-300 mm)
accommodates three or four spans of dispatch diameter.
lengths as one continuous length, which is dictated
by the lifting capacity and the floor strength of the The differences in design, as compared to a VCV line,
factory. When the process reel is full, the cable is cut and methods of reducing droop are described in relation
and is taken up onto a second reel. Samples are to the legend in Figure 4-60:
taken from the leading and trailing ends for analysis
in the routine material laboratory. Further samples 6, 7, 8, and 9: The extruder group and die head are
are taken from each successive length. angled downwards in line with the catenary angle.
10, 11, and 12: The CCV tube is formed in the shape
of an asymmetric catenary.
14: The core has to be tensioned by the turn-around

wheel to keep the core in the heating zone from
touching the side of the CV tube and being partially
flattened. An induction sensor feeds back the posi-
tion of the core to the turn-around wheel. The CCV
tube angle and shape are optimized for the optimum
range of cable dimensions to be manufactured. The
core usually touches down and runs along the side
of the tube in the cooling zone.
11 and 12: Condensed by-products of crosslinking

accumulate at the lowest point of the catenary, and
Figure 4-59 Outlet of extruded core from a VCV line so a drain is provided in a larger diameter tube posi-
(courtesy of CCI). tion, called the “fat boy.”
4-57
Figure 4-60 CCV extrusion and curing line (courtesy of EPRI).
Figure 4-61 CCV line: Heating zones MV (shallow) and HV/EHV (steep) curing tubes
Methods of reducing insulation droop in CCV-Nitrogen • Circulate gas locally at the die head, sometimes
lines are: heated and sometimes cooled, to achieve more uni-
• Steepen the angle of the catenary tube for higher- form temperature and dimensional conformity.
voltage HV and EHV cables, as shown in Figure • Rotate the core at the output end by locating an out-
4-61. put core twister at 13 in Figure 4-60. The core twister
• Use an increased viscosity (low MFI) XLPE com- is a caterpillar drive that is inclined to the axis of the
pound. core, such that a rotational movement is imparted to
the emerging core. The core rotation runs back up the
core through the CCV tube to the die head. The
4-58
droop is thus reduced in depth by being rotated heli- annulus of the die adjacent to the “trumpet,” which usu-
cally around the core. ally has a short cylindrical part (the land). The func-
• Rotate the conductor at the input end by locating an tions of the die and the CV tube are combined by
input core twister between 5 and 6 in Figure 4-60. bolting a tight-fitting tube to the horizontal die assem-
The twister consists of a caterpillar that is rotated bly, such that the tube becomes a “long land die,” and in
about its axis. The advantage is that the rotational effect acts as a long mold tube with an open end into
torque does not have to overcome the rotational drag which the XLPE melt is cast. The die tube is approxi-
of the core in the touch-down length in the cooling mately 40-60 ft in length (12-18 m) and has to be
zone. A possible downside is that the caterpillar has machined to suit the specified cable outer diameter. The
to be able to grip the conductor without disturbing length is required to provide sufficient heat transfer and
the outer layers. to generate back pressure.
4.20.5 CCV (Catenary Continuous Vulcanization)- The XLPE melt is in direct, sliding contact with the
Silicone Oil tube, with the frictional constraint and hydraulic imped-
ance generating sufficient XLPE pressure to keep the
CCV-silicone lines produce excellent core geometry and gaseous by-products in solution in the insulation. The
can use low-viscosity (high-MFI) insulating material. A tube temperature is heated to 270-350°C to directly heat
few CCV lines have been designed to be suitable for the the shield and insulation-melt to the crosslinking tem-
manufacture of large-conductor transmission-class perature by conduction. To prevent the liquid-melt from
EHV cables by steepening the angle of the catenary tube sticking, the inside of the tube is lubricated by injecting
and by filling the CCV tube with silicone oil in place of a layer of silicone oil onto the semi-con insulation shield
nitrogen and water. Silicone oil has a similar density to through a porous section of the tube wall. The high sur-
XLPE, such that the insulation has a near neutral buoy- face temperature and the presence of diffused dicumyl
ancy, thereby eliminating gravitational droop. Silicone peroxide can crosslink the silicone oil and so degrade it.
oil is used as the heat transfer medium to heat the core The solution adopted in the MDCV line is to extrude a
to crosslinking temperature and to cool it afterwards. fourth layer over the semi-con insulation shield, com-
The advantages of a CCV-silicone line are the same as prising a buffer layer of semi-con compound not con-
the CCV-nitrogen lines—i.e., lower installation cost and taining the peroxide additive. A small volume of silicone
higher output rate. Disadvantages are the cost of the sil- fluid can be expected to be absorbed by the shields and
icone fluid and the presence of a small residual volume XLPE insulation.
of silicone fluid in the insulation.
The advantage of the MDCV line is that the installation
4.20.6 MDCV (Mitsubishi Dainichi Continuous cost is low, being at ground level, with no need of a
Vulcanization) tower, and the resultant core is cylindrical. The disad-
MDCV lines extrude at ground level, thus facilitating vantages are the conductor tends to be eccentric within
the handling and inspection of the conductor, the han- the insulation, because, in a shallow catenary shape, it is
dling of the insulation and shield materials, the extruder not possible to apply sufficient conductor tension to
and die tooling, and the maintenance of the equipment. keep the conductor absolutely straight and horizontal;
The MDCV line can be fitted into an existing factory the dies are costly to manufacture, clean, and maintain;
building without the need to construct either a VCV
tower or a CCV stub tower.
A significant proportion of transmission-class cables

have been extruded using horizontal “long land” die
lines. Several of the key 400-kV 1600-2500 mm2 cable
projects in Europe were extruded in a long land die CV
line. The name MDCV is derived from the Mitsubishi-
Dainichi Nippon Company, which improved the exist-
ing HCV (horizontal continuous vulcanization) process
to improve concentricity (Tanaka and Greenwood 1983,
page 47).
The layout of the MDCV line is given in Figure 4-62.

Figure 4-62 MDCV line (courtesy of Blackwell
The terminology “land” is the name given to the outer Publishers).
4-59
the surface temperature of the cable is high, risking pressure is created. The pressure can prospectively dis-
accelerated thermal aging; and the CV tube is compara- place molded insulation in joints and initiate leaks at
tively short. One solution that is used to improve con- terminations.
centricity is to extrude with a higher-viscosity (low
MFI) XLPE insulating compound. A downside is that Specification AEIC CS7-93 required a 20-day waiting
the compound is more difficult to strain to high levels of period between extrusion and electrical testing in the
purity. Another solution is to rotate the conductor and absence of accelerated degassing. It should be noted,
core using caterpillar twisters. however, that no transmission-class cable will adequately
degas at ambient temperature within 20 days without the
4.20.7 Degassing Process addition of temperature.
XLPE and EPR cables use the same dicumyl peroxide
crosslinking agent, which produces similar volumes of Therefore, all manufacturers accelerate and control the
gas in both cables. Traditionally, degassing processes degassing of cables by heating the cable, either in a
were not performed on MV cable designs of less than degassing chamber, through electrical heating by the
46 kV. The gaseous by-products were low and able to means of conductor current, or a combination of heat-
diffuse comparatively quickly through the thin insula- ing and the application of a vacuum on the conductor.
tion wall, both in the factory and as installed. However,
the thick insulation walls at 138 kV produced sufficient Materials and processes used in cable extrusion and cur-
gas volume to create a pressure of 1.5 bar in the accesso- ing vary as well as the cable geometry, wall thickness,
ries on metal sheathed cables and in some cases was elected degassing temperatures, and durations. Typi-
noted to inflate the jacket in wet cables. It is now normal cally, degassing temperatures for XLPE cables range
to perform a degassing process on higher stressed cables from 60°C to 75°C. Too high a degassing temperature
at 138 kV and above with metallic sheaths containing distorts the circular shape of the cable, resulting in loos-
either XLPE or EPR insulation. ening of the turns of cable and damage of the insulation
shield in the next process.
The degassing process is a comparatively low technol-
ogy, which uses simple ovens, (Figure 4-63) to gently The duration varies between manufacturers and
heat the cable. The process times of 3-5 weeks for 138- depends on the insulation thickness and cable geometry
to 400-kV cables are generally longer than the extrusion and ranges from 2 (46 kV) days to approximately 20
times, and so degassing can become the critical path (115-220 kV) days, and more with some very large
that controls the output of shipping reels from the fac- 400-500 kV cables being degassed for up to 35 days.
tory. A large oven capacity is required to keep pace. Degassing EPR Cables
Degassing XLPE Cables The crosslinking process for EPR cables is similar to
Degassing of the insulation is necessary for two reasons. that in XLPE cables. Accelerated degassing, though, is
First, the pressure of the trapped gas in the insulation not a common practice for EPR cables. The reasons for
can mask insulation defects, in particular voids, from the different approach include the high filler content in
being detected during the ac withstand and partial dis- EPR, the lower peroxide levels, the use of EPR cables at
charge factory acceptance tests. Second, metallic lower voltages, the curing technique, and the use of EPR
sheaths applied over the core of a cable that has not in predominantly wet designs. However, for EPR cables
been degassed contains the gas such that an equilibrium that receive a metallic sheath, degassing procedures sim-
ilar to those for XLPE cables are in use in Switzerland
and Italy.
4.21 QUALITY CONTROL

An extruded cable is inaccessible for visual inspection
during the extrusion, crosslinking, and subsequent fin-
ishing processes, and so an effective quality system is
vital in achieving consistent reliability in service.
The electrical strength and service life of an extruded

cable are determined by both macro defects and the lev-
els of micro impurities in the extruded insulation. Inten-
Figure 4-63 Bank of four degassing ovens (courtesy of sive quality control is, therefore, essential at each stage
CCI). in the manufacture of transmission-class cables.
4-60
The measures of a good cable manufacturer are: In most cases, cable manufacturers require the com-
• Routine testing of incoming materials and sampling pound supplier to perform the 2% sampling test to meet
the AEIC requirement. The compound supplier certifies
of certified supply materials
the compound to conform to specifications based on the
• Trained factory operatives with knowledge of quality sampling data. Some manufacturers perform their own
faults and their causes insulation compound inspection.
• An immediate fault-reporting system
• A no-blame culture There are two different inspection methods, both being
off-line. One method is extruded film inspection, and
• Sufficient numbers of laboratory technicians to keep the other is pellet inspection. More details may be found
pace with tests on in-process samples in the ICEA standard for distribution cables, ICEA S-
• A “black museum” of defects, causes, and solutions 94-649. Figure 4-64 shows an extruded tape inspection
• A statistical process control (SPC) system cleanliness checker in an in-house compounding plant.
The instrument uses a light beam to automatically
• Quality reporting (fault and SPC) direct to manage- count, size and label particulate contaminants. For
ment transmission cables, depending upon the material sup-
pliers’ specifications, the insulation material should not
The typical manufacturing time for an extruded 400-kV include any contaminants larger than either 4 mils
XLPE cable is 12 weeks. Thus, if a defect is detected by (102 μm), or 3 mils (76 μm), as shown in Table 4-16.
electrical failure at routine HV test, the following four to
six extrusion campaigns will also be in doubt, with a Semiconducting Compound
cumulative in-process cable length of 5 to 32 miles (8 to AEIC CS9-06 requires that the sulfur and ash contents
51 km) to be placed in embargo until the cause of the be less than 0.005% and 0.01%, respectively, and that
failure is isolated. The routine HV test is a coarse test acetylene-black-filled compound be used for conductor
that only detects gross defects that cause early failure in shield stresses of greater than 200 V/mil (8 kV/mm).
service; it does not improve quality or ensure service life.
Insulation quality is determined at the front end of man-
ufacture during materials compounding and extrusion.
The manufacturing and quality control measures are

described for XLPE cables, but they are closely similar
for EPR cables. Where appropriate, the test levels are
given for both XLPE and EPR cables, noting that AEIC
CS9-06 and ICEA S-108-720-2004 specify requirements
for both cable types together. Section 4.21.2 describes
additional quality measures specific to EPR material
and cable.
4.21.1 Materials Inspection (XLPE and EPR)

The AEIC raw material and inspection requirement is
not typically applied in Europe, but is agreed between
purchaser and supplier. Therefore, materials for cables
in North America are specifically tested according to
AEIC CS9-06 or better.
Figure 4-64 Cleanlness of insulation compound
Insulating Compound being checking on extruded tape in in-house
compounding plant (courtesy of CCI).
Specification AEIC CS9-06 requires the inspection of
2% of the insulation material volume for contaminants;
material not inspected by the supplier must be inspected Table 4-16 Allowable Numbers of Insulation Contaminants
by the cable manufacturer. AEIC does not specify a Insulation Insulation Contaminant Size [μm]
Applica- Contami-
maximum number or size of any contaminants, but tion nants 50 – 70 70 –100 100 – 200
requires reporting of the size and number of contami-
EHV Number/kg 10 0 0
nants for information and a requirement for statistical
analysis of the data. HV Number/kg - 10 0
4-61
Table 4-17 Allowable Number of Shield Protrusions • Moisture content

Materials Inspection Specific to EPR • Silica content
Width of Shield Protrusions
at 50% Height: [μm]
• Particle size distribution
Shield • Grit content
Application Protrusions >150 >200 >500 1500-2500
EHV II Number/m2 20 10 1 0 • Residue on a 325 mesh size screen
EHV I Number/m2 50 20 1 0 • Wettability
HV Number/m2 - 40 2 1 • Maximum allowable percentage of metal salts
• Iron content
Typical smoothness specifications agreed with suppliers • Type and degree of treatment (if pretreated)
for semiconducting material used for both conductor
and insulation shields are given in Table 4-17. The table The specification for the antioxidant should specify the
gives the maximum allowable numbers of shield protru- chemical designation, melting point sharpness, and per-
sions per square meter of surface segregated by the pro- centage of allowable insoluble particles. The specifica-
trusion width at 50% height (W50). An example of the tion for the red-lead stabilizer should indicate the red-
shape of a protrusion is shown in Figure 4-47. lead content in the master batch. The degree of purity
should be specified in all raw material specifications
4.21.2 Materials Inspection Specific to EPR defining the type and amount of contaminants.
A high-voltage EPR-insulation compound typically
consists of seven to nine ingredients, of which all but If a raw material cannot be procured commercially in
two or three (copolymer or terpolymer, mineral filler, the purity and form deemed necessary to achieve the
and polyethylene, if used) are present in the compound required end results, then it is necessary to treat the
in less than 5% by weight. Hence, in addition to the for- material to achieve the required form and purity. In
mulation of the compound, quality-control procedures addition to incoming raw material inspections, certified
are essential to ensure the quality uniformity and clean- test reports covering each batch of each ingredient,
liness of all ingredients. Quality-control procedures showing test results and demonstrating compliance with
must also ensure uniform dispersion of the ingredients the applicable material specification, should be submit-
in the mixing operation to achieve a homogeneous com- ted by the vendor prior to shipping the ingredient. No
pound free of significant contamination. These steps are changes in the ingredients for chemical composition,
essential to achieve a high-quality, high-voltage trans- method of processing, or approved supplier should be
mission cable insulation. permitted without prior approval of the purchaser.
Raw Material Specifications EPR Compound Specification and Requirements

Raw material specifications are of paramount impor- To optimize ingredient dispersion and ensure a homoge-
tance. They should clearly define the minimum require- neous compound free of significant contamination, mix-
ments to achieve the desired performance. For example, ing should be done in either a dedicated intensive
the specification for the polymer should include the fol- internal mixer or a thoroughly cleaned continuous
lowing: mixer. Further, mixing should be done in master batches
with fine screening (using a no. 200 or smaller mesh
• Copolymer or terpolymer
screen) of each batch and rescreening of the final com-
• Ethylene content or degree of crystallinity pound prior to pelletizing.
• Mooney viscosity Quality-Control Tests for EPR Insulation
• Volatile matter Quality-control tests for a high-voltage EPR-insulation
• Moisture content compound should include the following:
• Gel content • Specific gravity serves to verify that the compound

has been mixed in accordance with the formulation.
• Purity
• Appearance of the compound serves to indicate the
• Form
degree of dispersion of the ingredients and the clean-
• Packaging liness of the compound. An industry standard is
needed to provide a method to measure dispersion
The specification for the mineral filler should specify and to set limits for various voltage ranges.
such items as: • Unaged physical properties indicate that the com-
• Bulk density pound formulation has been followed in the mixing
4-62
operation, particularly when the results are compared compliance and consistency with the test results
with data from other batches cured under the same obtained on prior batches of the same compound
conditions. These properties show that the insulation should be considered.
from the completed cable will meet specification
requirements. 4.21.3 Quality Control of CV line
• Aged physical properties also indicate that the com- An example of typical quality monitoring positions in a
pound formulation has been followed in the mixing VCV line is shown in Figure 4-65. Other similar, or
operation, particularly with regard to the antioxi- alternative, methods may be applied to CCV and
dant, and that the aged insulation from the com- MCDV lines. The monitoring positions are described
pleted cable will meet specification requirements. below in relation to the numbered legend:
• Specific gravity and unaged and aged physical prop- 1. Pellets of uncrosslinked PE insulation are sampled
erty tests are typically performed on molded slabs and extruded into tape, which is automatically light
cured for five minutes at 177°C, prepared in accor- scanned and particle counted. This is usually under-
dance with ASTM D 3182 and following the proce- taken before the pellets are admitted to a sealed
dure of ASTM D 412 using Die C. storage silo at the base of the VCV tower. A batch of
• Viscosity measurements—i.e., Mooney Scorch Test nonconforming material should not be admitted to
the silo.
and Monsanto Oscillating Disc Rheometer Curves—
determine the time to achieve optimum crosslinked 2. The melt of uncrosslinked PE insulation leaving the
density of the material. Specifically, the Mooney extruder is filtered by a stainless steel strainer pack.
Scorch Test provides information on resistance to The pressure rise across the strainer pack is moni-
precuring, and the Monsanto Rheometer measures tored to detect when the increasing number of
the viscosity before cure and the hot modulus of the ambers, gels, and other contaminants near the end
compound. of the extrusion campaign are starting to block the
strainer. The campaign must be aborted before this
• Electrical properties—i.e., dielectric constant, dissipa-
occurs to prevent the pack from suddenly bursting
tion factor, ac dielectric strength, and impulse
and releasing pieces of stainless steel wire into the
strength—are critical to the performance of the cable.
insulation.
They are best measured on model cables extruded and
cured in the laboratory for each batch of compound. 3. In some CV lines, a light scan automatically counts
and identifies types of particles within the transpar-
All of the above tests should be completed with satisfac- ent melt as it passes through the neck connecting
tory results before incorporating the compound into the the extruder to the die head. Nonconforming insu-
plant delivery system. Both the material specification lation would require either its position in the cable
Figure 4-65 Example quality monitoring positions on a VCV line.
4-63
to be noted or the extrusion campaign to be management staff to accept the defect and thereby
aborted. A benefit of a melt scanner is that the pro- avoid the high cost of scrap and program delays.
cess can be monitored near the end of the extrusion 8. Finally each reel length is fitted with test termina-
campaign (run) when the number of ambers and tions, and the HV acceptance tests are performed.
other contaminants increases to an unacceptable These tests include an HV voltage withstand test, a
level. partial discharge test, and a measurement of insula-
4. The conductor is centered within the extruded mol- tion capacitance. The resistance of the conductor is
ten insulation by an X-ray gauge. This is also used also measured. Samples of core are taken, and simi-
to measure the thickness of the shields and insula- lar tests to those described in item 6) are under-
tion. This geometric information is used to keep the taken.
diameter of the cable constant by controlling the
speed of the capstan wheel inputting the conductor 4.21.4 Inspection of Cable Insulation
into the extruder. It should be noted the measure- Samples of cable are cut from the ends of the process
ments are “hot” measurements, and allowances for reel “cut positions” during each extrusion campaign for
thermal contraction have to be made to calculate in-process tests and from the ends of the cable length on
dimensions at ambient temperature. the shipping reel for formal routine acceptance tests.
5. In some CV lines, the extruded core exits the CV Triplicate samples are produced.
tube and passes through a small water bath to per-
Shield Protrusions and Insulation Particle Counts
mit it to be continuously scanned by an annulus of
and Voids
ultrasonic transducers (Smith et al. 2003). The
The first two samples have the conductor removed to
transducers are set to detect shield and insulation
produce the cylindrical core samples in Figure 4-66 for
defects, particles, and impurities. The dimensions of
particle counting. One set is stripped of the insulation
the insulation and shields are also monitored and
shield and is heated to greater than 120°C for transpar-
recorded. The position of a nonconforming cable
entization in either an air-heated oven or in a bath of sil-
core is automatically recorded, and an alarm initi-
icone oil or glycerol. This permits a general qualitative
ated. A decision is taken to cut out the affected
visual examination for the presence of particulate con-
length or to abort the extrusion campaign. The ben-
tamination and “clouds” in the insulation and defects
efits are that the positions of unacceptable macro
on the conductor shield; however, defects at the outer
defects are located with accuracy in the extrusion
shield cannot be seen.
campaign, and the lengths of unaffected sound core
are identified, permitting them to be released into
The second core sample is then spiral cut on a lathe to
service. This kind of monitoring is optional and is
produce 20 to 50 turns. The spiral is then cut into circu-
supplementary to the final electrical routine test (8),
lar “slinkies,” as shown in Figure 4-67. The thickness
which is mandatory.
required in AEIC CS9-06 is about 25 mil (0.64 mm).
6. The diameter of the cable core at near ambient tem-
perature is monitored by a laser gauge and compared
to the specified nominal and minimum insulation
and shield thickness and eccentricity values.
7. The cable core is cut into the maximum lengths that
can be wound onto factory process reels. Samples of
core are cut out at the leading and trailing ends and
at the intermediate cut positions. Similar samples
will be taken after the degassing process. The sam-
ples are microscopically examined for the correct
geometry and the absence of defects. Samples of the
insulation and screens are removed for measure-
ment of physical properties. A length of core is
heated and made transparent, and a simple visual
internal examination is undertaken of the insulation
and shield interfaces. It is important that these tests
are performed quickly to permit remedial action to Figure 4-66 Samples of XLPE insulation cut from
be taken following the discovery of a nonconfor- extrusion runs; the conductor has been removed
mance, since late discovery at the end of manufac- (courtesy of CCI).
ture puts unnecessary pressure on the factory
4-64
A statistically significant number of slinkies are then The types of insulation inclusions can be subdivided
mounted in a microscope, and a general scan is per- under the three types listed in AEIC CS9-06 into the fol-
formed to identify large particles or voids of greater lowing categories:
than approximately 2 mil (50 μm) in the insulation and • Voids: Note that it is difficult to identify whether
protrusions greater than 2 mil (50 μm) present on the
voids are gas or liquid filled.
shield interfaces. If particles in the higher stressed cables
of greater than 5 mils (125 μm) are present and are iden- • Contaminants:
tified to be metals or blacks, the cable length and extru- —“Metals”: the type of metal can be identified by
sion run is placed in embargo, while a decision is taken scanning electron microscope (SEM).
on scrapping the length(s) and an investigation is per- —“Blacks”: usually burnt PE.
formed to locate the source of the problem. If no large
—“Gels”: transparent polymer.
particles are present, then the slinky is divided into a
number of squared graticules scratched on the surface, —“Others”: fibers, silicas, friable materials, nonme-
and a more detailed count is taken of the insulation and tals, etc.
of the shield interfaces at magnifications of (X50 to • Ambers: oxidized PE.
X250), as shown in Figure 4-68.
The maximum permitted AEIC CS9-06 insulation
inclusion sizes and shield (inner and outer) defects are
given in Tables 4-18 and 4-19 for a range of conductor
shield stresses.
Table 4-18 Maximum Permitted Dimensions of Insulation
Inclusions (AEIC CS9-06, with additions by author)
Conductor
Shield Stress Void Contaminant Amber
V/mil kV/mm mil μm mil μm mil μm
100 4 2 50 5 125 10 250
125 5 2 50 5 125 10 250
150 6 2 50 5 125 10 250
175 7 2 50 5 125 10 250
200 8 2 50 5 125 10 250
225 9 2 50 5 125 10 250
Figure 4-67 A circular ‘slinky’ cut from a spiral of XLPE 250 10 1.5 38 5 125 10 250
insulated core (courtesy of CCI).
275 11 1.5 38 5 125 10 250
300 12 1.5 38 5 125 10 250
325 13 1 25 4 100 8 200
350 14 1 25 4 100 8 200
Table 4-19 Maximum Permitted Dimensions of Shield

Defects (AEIC CS9-06, with additions by author)
Stress Void Protrusion
Shield V/mil kV/mm mil μm mil μm
100 4 2 50 3 75
125 5 2 50 3 75
150 6 2 50 3 75
175 7 2 50 3 75
200 8 2 50 3 75
Conductor 225 9 2 50 3 75
250 10 1.5 38 3 75
275 11 1.5 38 3 75
300 12 1.5 38 2.5 63
325 13 1 25 2 50
350 14 1 25 2 50
Figure 4-68 Routine particle counting on XLPE

Insulation All All 2 50 5 125
cable slinkies (courtesy of CIGRE).
4-65
The maximum permitted AEIC CS9-06 and ICEA S- hot set and tensile tests, as shown in Figure 4-13. AEIC
108-720-2004 insulation eccentricity limits are given in CS9-06 refers to ICEA S-108-720-2004. Generally the
Table 4-20. The percentage eccentricity is defined as: test results will be significantly within the limits, and so
it is good practice to plot an SPC chart of the actual
(Tmax − Tmin )
x 100 results with time. A sudden test divergence from one
Tmax sample, even though the result is still within specifica-
tion, may indicate either that the cable has been over-
The eccentricity percentages may be applied to the heated by controller failure as revealed by the aging test,
range of insulation thicknesses in Table 4-8, calculated or underheated perhaps by a heater burnout as revealed
from AEIC CS9 (AEIC 2006) stress limits, to determine by the hot creep test.
typical maximum permitted variations in thickness— Insulation Crosslinking
e.g., 12% on a 69-kV cable with 303 mils (7.7 mm) insu- “Hot creep” tensile tests (named “hot set” tests in
lation is 36.3 mils (0.92 mm), and 10% on a 345-kV Europe) are performed to check that the insulation has
cable with 1002 mils (25.5 mm) insulation is 100 mils sufficient crosslinking. It is good practice to take sam-
(2.6 mm). ples from the inner, mid-insulation, and outer insulation
Insulation Physical Properties if the insulation is sufficiently thick. If the elongation is
Figure 4-69 shows longitudinal slices approximately greater than 75% for XLPE, then the insulation has a
2 mm thick that are cut from the samples to permit high proportion of uncrosslinked material. Similarly if
“dumbbell” tensile test samples to be stamped out for the residual extension, after deloading and cooling back
to ambient temperature, is greater than 10% for XLPE,
Table 4-20 Insulation Eccentricity Limits (AEIC CS9-06, there is insufficient elasticity, again indicative of inade-
with additions by author) quate cross linking. The ICEA S-108-720-2004 test con-
Conductor Area Eccentricity ditions and requirements are given in Table 4-21 and
Rated Voltage
kV kcmil mm2 %
IEC 62067 in Table 4-22. The IEC test temperature is
500- 240-
higher, the limit for elongation under load is the same,
69 wet 12 and the allowance for tensile set is greater. In ICEA S-
4000 2000
108-720-2004, EPR Class I is for discharge-free and dis-
500- 240- charge-resistant designs, and Class II is for discharge-
69 dry 12
4000 2000 free designs only. IEC 62067 does not distinguish the
115 dry
750- 400-
12
type of EPR.
4000 2000
750- 400- Insulation Aging Antioxidant Content
138 dry 12
4000 2000 The ICEA S-108-720-2004 insulation requirements for
161 dry
750- 400-
10
resistance to aging are shown in Table 4-23 and IEC
4000 2000
1000- 500-
230 dry 10 Table 4-21 Requirement for Degree of Crosslinking (Hot
5000 2500
1000- 500-
Creep), ICEA S-108-720-2004
345 dry 10
5000 2500 Requirement XLPE EPR I and II
Time in air oven 15 minutes 15 minutes
Temperature 150oC 150oC
Applied stress 29 psi 0.2 MPa
Elongation ≤175% ≤50%
Tensile set ≤10% ≤5%
Table 4-22 Requirement for Degree of Crosslinking (Hot

Set), IEC 62067
Requirement XLPE EPR
Time in air oven 15 minutes 15 minutes
Temperature 200oC 250oC
Applied stress 29 psi 0.2 MPa
Elongation ≤175% ≤175%
Figure 4-69 Sample of XLPE cable taken from an Tensile set ≤15% ≤15%
extrusion campaign cut (courtesy of CCI).
4-66
62067 in Table 4-24. The purpose of the aging test is to permitted for some loss of shield material and separa-
show that sufficient antioxidant has survived the extru- tion of the bond with the insulation.
sion and crosslinking processes to protect the insulation
for its specified service life. The effect of aging is embrit- 4.21.5 Tests on Extruded Core
tlement and discoloration, as evinced by loss of elastic-
Insulation Shrink-back
ity (elastic extension) and reduction in tensile strength.
Insulation shrink-back, called insulation retraction in
Shield Crosslinking Europe, is the contraction that occurs when the cable is
The degree of crosslinking of the shields is checked by cut. Shrink-back has two parts. The first is the locked-in
the ICEA S-108-720-2004 “wafer boil” test. A complete thermal contraction stress produced when the cable
ring of the core at least 25 mil (0.64 mm) thickness, con- cools from 107°C to ambient temperature, which is a
taining the insulation and inner and outer shields, is cut function of the contraction that occurs when the semic-
from one of the core samples shown in Figure 4-66. The rystalline lamellae structure is formed. The second is
sample is immersed in boiling decahydronaphthalene or higher and occurs when the cable is heated during joint-
xylene with 1% by weight of Antioxidant 2246 for 5 ing or in service above the crystalline melting point of
hours. A complete ring of each of the shields shall 107°C, thus releasing the locked-in memory formed by
remain without interruption or cracking, although it is contraction from the insulation mean crosslinking tem-
perature of 160-180°C. The latter dictates that accesso-
ries be specially designed to resist and contain shrink-
back if a cable system is required to operate above 90°C.
Table 4-23 Requirement for Insulation Resistance to Aging, ICEA S-108-720-2004 specifies a limiting shrink-back
ICEA S-108-720-2004 value to provide a degree of protection for accessories in
Requirement XLPE EPR I EPR II the normal operating range of less than 90°C. Shrink-
Unaged tensile ≥1800 ≥12.5 ≥700 ≥4.8 ≥1200 ≥8.2 back of the insulation cut prospectively opens a large
strength psi MPa psi MPa psi MPa void at the center of a straight joint, resulting in imme-
Unaged elonga- diate failure. It is recommended that accessories are also
250% 250%
tion at break
designed to limit the retraction by positioning the insu-
lation cut in an electrically nonstressed area and provid-
Ageing time in air
168 hours 168 hours ing an axial tolerance before the insulation cut retracts
oven
Ageing into the stressed region.
121oC 121oC
temperature
Ratio of aged to The degree of insulation shrink-back is a variable
unaged tensile ≥75% ≥75% ≥80% between different cable designs and different CV pro-
strength
cesses. Cables with binder tapes over the conductor
Ratio of aged
to unaged ≥75% ≥75% ≥80% exhibit a greater shrink-back because this reduces fric-
elongation tional grip. Extruded cores that experience too great a
thermal shock in the CV line when being cooled below
Table 4-24 Requirement for Insulation Resistance to crosslinking temperature also produce increased shrink-
Aging, IEC 62067 back.
Requirement XLPE EPR
Unaged tensile
≥1800 psi ≥12.5 MPa ≥700 psi ≥4.2 MPa ICEA S-108-720-2004 requires that five 1.5 ft (0.45 m)
strength cable samples be heated at 50°C for 2 hours and then be
Unaged elonga- allowed to cool for two hours at room temperature. This
200% 200%
tion at break
cycle shall be repeated three times. The distance between
the edge of the insulation and the end of the conductor
Aging time in air
168 hours 168 hours (total of both ends) shall be less than 0.3 in. (7.62 mm).
oven
Aging tempera-
135°C 135°C
ture IEC and BS 7912 measure the total prospective shrink-
Tensile strength: age by heating to above crystalline melting point. The
Maximum varia- ±25% ±30% 200-mm samples are heated to 130°C for 6 hours. The
tion after aging
distance between measuring cut marks is 200 mm. The
Elongation at
break
±25% ±30% maximum permissible shrinkage of the total sample
length is 4%.
4-67
Crosslinking By-product Concentration CS9-06, with the voltages given in Table 4-25. It can be
At present, tests to ensure that a satisfactory degree of seen that the withstand test voltages are in the ratio of
degassing has been performed are not included in North 2:1 to 3:1 to the phase-to-ground voltage. The 2:1 test
American or in IEC specifications, and so a formal voltage for the 345-kV cable produces the limiting stress
request has been made to CIGRE to investigate. at the conductor shield of 700 V/mil (28 kV/mm). It is
potentially damaging to apply a constant ratio test volt-
Measurements were made in the United Kingdom on a age irrespective of the design stress, because of the risk
range of production cables with up to 400-kV insulation of damage to an otherwise sound cable. It is known that
thicknesses. These showed that a normal degassing pro- there will always be a population of inclusions present in
cess should remove methane and water vapor by more extruded insulation, and in the absence of macro defects
than 1% by weight of the insulation, leaving typically 1 this will dictate the breakdown strength and service life.
to 1.5% of the low volatile by-products permanently and Jocteur showed that the endurance of PE insulation
harmlessly in the insulation. UK specification BS 7912 samples containing deliberately added metallic particles
for the manufacture of XLPE cables up to 132kV was reduced when the applied stress exceeded 760 V/mil
includes a test to check degassing and is described below (30 kV/mm) (Jocteur et al. 1972) For full-sized cables, it
for information: is accepted practice to apply a safety margin and limit
the maximum test stress to 700 V/mil (28 kV/mm).
Five small samples of insulation are taken from
extruded core that has been submitted to a factory It is good practice to perform a partial discharge at low
degassing process, each 20 mg in weight, from the inner, voltage upon first energizing to ensure that there is an
middle or outer part of the insulation. A thermo-gravi- absence of voltage-induced PD in the insulation and
metric test is then performed (apparatus which heats contact-induced PD between the shields and the shield
and measures weight change). The samples are heated conductors. The PD equipment is then usually discon-
from 30°C to 175°C at 50°C /min (i.e. in 2.9 minutes) nected to protect it in the event that failure occurs dur-
and held at 175°C for 30 mins. The elevated test temper- ing the main withstand test. The PD equipment should
ature and small sample size produce a more rapid diffu-
sion than the low factory process temperature. The test
temperature is above the boiling point of water, methane
and alpha-methyl styrene (160-165 °C) and is below the
boiling points of acetaphenone of 202°C and cumyl
alcohol of 215-220 °C. The total weight loss during the
30 minute test period is required to be less than 1.6% of
the original insulation weight. Additionally the weight
loss in the first 5 minutes is required to be less than
0.9%: the weight loss in the next 10 minutes is to be less
than 0.475% (record for information only) and the
weight loss in the last 15 minutes is to be less than
0.225%.
4.21.6 Electrical Tests on Main Insulation Figure 4-70 Screened room routine electrical
Each shipping reel is subjected to the voltage withstand production test on XLPE cable (courtesy of
test and the partial discharge test required in AEIC Southwire).
Table 4-25 Electrical Tests on Main Insulation, AEIC CS9-06

Max
Voltage Stress Withstand Test Partial Discharge Voltages
V Vg BIL 2.0 Vg 2.5 Vg 3.0 Vg Vt/Vg Vt/Vg Vt/Vg
kV kV kVp V/mil kV/mm 1.0 hr 0.5 hr 0.25 hr 1.0 1.5 2.0
69 40 350 150 6 80 100 120 40 60 80
115 70 550 200 8 135 160 200 65 100 135
138 80 650 200 8 160 200 240 80 120 160
161 92 750 225 9 185 230 280 95 140 185
230 132 1050 275 11 265 330 N/A 135 200 265
345 200 1300 350 14 400 N/A N/A 200 300 400
4-68
then be reconnected at the end of the withstand test and nal internal pressure of 3-5 psi of dry air or nitrogen
the formal PD test performed. AEIC CS9-06 and ICEA within the extruded metal-sheath. A pressure gauge has
S-108-720 specify that no discernable PD should exist been fitted, such that the integrity of the end caps could
with a minimum detection sensitivity of less than 5 pC. be checked in transit and installation.
Allowing a conservative signal-to-noise ratio of 2:1 pro-
duces a prospective PD pass level of less than 10 pC. Residual methane gas will continue to diffuse from the
The latter level is too high for modern high-stressed insulation even though the cable core has been subjected
transmission-class cables. to a factory degassing process. Care should be taken to
vent the gasses prior to cutting the cable and prior to
Manufacturers of EHV (230-500 kV) cable worldwide installing it in an enclosed space, with the objective of
generally work to higher standards by testing in electri- reducing the risk of fire and suffocation. The jointing
cally screened rooms, as shown in Figure 4-70, with area should be well ventilated and checked with a meth-
screened power supplies and with filter detection sys- ane gas detector prior to commencement of work.
tems that permit quiet detection frequencies to be
selected in between the spectrum of frequencies gener- 4.23 DESIGN STRESS AND SERVICE LIFE
ated by external radio frequency and atmospheric noise.
The selection of safe design stresses at the conductor
Manufacturers have regarded the investment in screened
shield and insulation shield is a key choice in ensuring
rooms as equally important as the clean-room materi-
that the specified service life will be achieved. Figure
als-handling equipment and the CV line. It is normal to
4-17 gives the variation of design stress against transmis-
be able to detect PD in the range of 1-2 pC or less, with
sion voltage from an EPRI worldwide survey showing
a background noise level of approximately 0.3 pC.
how the conductor shield design stress has approxi-
Should a 1-2 pC persistent voltage-induced PD pulse be
mately been doubled from 8 kV/mm at 220/230 kV to
detected and confirmed to be within the cable, the cable
16 kV/mm at 500 kV (EPRI 2002). Figure 4-18 and
should not be released into service, even though this is
Table 4-2 show that for large-conductor cable systems
below the minimum 5pC specification detection limit.
containing joints, the design stress has been limited to
14.6 kV/mm. The conductor stress limit has primarily
For more details on testing, refer to Chapter 14.
been driven by the need to reduce the cable diameters to
obtain a worthwhile length on the cable shipping reel.
4.22 SHIPPING These lengths are then competitive for the applications
The length of cable that can be shipped on a reel is met by the smaller-diameter fluid-filled SCFF cables,
determined by the dimensions of the reel: the minimum which operate at between 15 and 18 kV/mm. Equation 3-
diameter DH of the drum (also named hub), the internal 71 in Chapter 3 shows that the reel length is approxi-
width (traverse)W, and the outer diameter DO. The mately inversely proportional to the cable outer diame-
equation for the cable length is given in Chapter 3, Sec- ter, hence the drive to reduce the thickness of the XLPE
tion 3.6.7, Equations 3-71, 3-72, and 3-73. In AEIC insulation. The limit in conductor shield stress for large-
CS9-06 (AEIC 2006), the minimum drum diameter DH conductor cables at 500 kV is set to limit the stress expe-
is expressed by multiples k of the cable outer diameter
DC for different sheath types and different insulation
thicknesses, as given in Table 4-26. Table 4-26 Minimum Diameter of Reel Drum for Extruded
Cables (AEIC CS9-06, Table 10-1-1, with editorial changes
by author)
Having produced high-quality dry cable insulation, it is
important that the ends of the cable are sealed to be Insulation Reel Drum
Thickness Diameter
completely watertight. The end caps and pulling bolts Ratio
should be capable of withstanding submersion in water- Sheath mils mm k=DH/DC
filled ducts and trenches, and internal gas pressure <500 <12.7 14
expansion and contraction due to changes in ambient Lead alloy 500-800 12.7-20.3 18
temperature. The cable caps should also be reinforced >800 >20.3 22
onto the jacket and metallic sheath, if present, to pre- Aluminum tubular all all 30
vent the retraction forces present in PE jackets from Aluminum tubular bonded to
all all 18
breaking the watertight seal. jacket
Corrugated copper or <800 <20.3 18
Some large-conductor, high-value 230-kV, 400-kV, and aluminum >800 >20.3 22
500-kV XLPE cables in the United Kingdom and Japan
have been supplied with the added protection of a nomi-
4-69
rienced by the accessories to 8 kV/mm operating stress, step in design is to apply a performance margin of say
as shown in Figure 4-18 and in Table 4-2. It should be 10% to the breakdown stress to ensure that the cable will
noted that the majority of 500-kV accessories previously predictably pass the specified qualification BIL level.
used in Japan are field-molded designs and not factory The outer diameter of the insulation and its thickness
premolded one-piece types, as now used in other coun- are then calculated according to Chapter 3, Equations
tries, which generally have lower design stresses. It can be 3-40 and 3-47. The stress (E) at normal working voltage
seen that, at 400 kV, lower conductor shield stresses of is calculated simply by dividing the impulse design stress
11.5-12.7 kV/mm has been used, and this has limited the by the ratio of lightning impulse voltage to working
accessory/insulation shield stresses to 5-6 kV/mm. The voltage. This ratio is derived from the insulation coordi-
stress limits given in AEIC CS9 (2006) for 345-kV cable nation levels for the particular transmission system, as
of 14 kV/mm at the conductor shield and 6 kV/mm is in given in Chapter 3, Table 3-1 for North America. For a
accord with the highest international experience at the large-conductor 400-kV cable, this typically results in a
400-500 kV system voltage levels. design stress of E = 15 kV/mm at ac working voltage.
The second fortuitous occurance is that, although the ac
The question then arises: What tests and calculations strength of the insulation reduces with time t according
have been performed to give reasonable assurance that to a E n t = k life relationship, as given in Chapter 3,
345-500 kV cable systems at these high design stresses Equations 3-60, 3-61, and 3-62, the long-term break-
will achieve a 40-year life? Service applications are still down strength of the cable insulation is high, at approx-
new, with the oldest circuits with large conductors imately E = 30 kV/mm. The latter provides an inherent
(greater than 1600 mm 2 ) and joints having less than design margin of 2:1—i.e., a 100% factor of safety. How-
10-15 years experience (Table 4-1). ever, it took some 40 years of development before the
knowledge and means were gained to achieve the insula-
The following sections describe the range of test and cal- tion quality to predictably pass qualification tests and
culation methods that have been performed to give a operate at the highest system voltages. The limitation
40-year-life prediction. was usually the accessories; however, the cable design
margin also provided a margin for them, although
4.23.1 Factors Affecting Design Stress somewhat reduced.
The calculation of design stress for XLPE and EPR
cables has been hampered from the start of the technol- The breakdown strength of extruded cables is now also
ogy by the lack of knowledge of the insulation defects well known to be dictated by the presence of “defects,”
that initiate breakdown and therefore the inability to not by the inherent strength of the polymer. The defects
predict failure from first principles. However, cable engi- in this case are contaminants (particles and voids) in the
neers have long experience in the inability to scale up crosslinked insulation, protrusions on the semiconduct-
stress levels from bench-top experiments to full-size ing shields, and variations in the extrusion morphology.
cables, although this experimental data is useful in Unlike a taped insulation, the defects are not inherently
explaining trends in full-size cable phenomena. The insu- controlled in geometry, location, or numbers. Thus the
lation design data is always derived from testing many materials and manufacturing technologies have had to
full-size cables, usually during development and qualifi- evolve with the objective of limiting the size and num-
cation tests, and then extrapolating the design stress to bers of the contaminants and protrusions to give the
other system voltages and conductor sizes. A design fac- degree of predictability necessary for design at the high
tor is then applied to ensure predictable success both in stresses for EHV operation.
qualification testing and in service operation.
It has been suspected that a threshold stress exists under
This was the case in the low-pressure fluid-filled paper which XLPE insulation does not age. Much research is
cable. These cables fortuitously contained a geometri- directed to proving this on both bench-top samples and
cally repeatable “defect,” that reduced the variable full-size cables. One attractive theory is that ac electrical
strength of the fluid-impregnated paper to a completely stress induces an alternating mechanical stress at the
predictable stress level. The defect is the fluid-filled butt insulation/defect interface, and that aging is one of
gap between adjacent tapes. The thickness of the butt mechanical fatigue, and then crack growth followed by
gap is precisely known and controllable, being the thick- electrical tree initiation. If this is the case, then insula-
ness of the paper tape, and this ensures that the fluid tion polymers and compounds could be modified to give
does, or does not, break down at a predictable stress greater mechanical fatigue strength permitting further
under lightning impulse voltage conditions. The next increases in design stress, accessories permitting.
4-70
4.23.2 Comparison with Other Manufacturers’ nants is given in Table 4-18 and the shield protrusions in
Experience Table 4-19. The “metallic” and “black” contaminants
As an alternative to carrying out their own research, for the purposes of stress calculation are taken to be
development, and experimental work, manufacturers conducting particles. The ambers are taken to be high-
may elect to draw on the common body of experience permittivity dielectric particles with a permissible length
within the public domain and use the same electrical simply twice that of the metallic particle. The design for-
stresses being used on similar installations—for exam- mulae for the contaminants and protrusions include fac-
ple, from Figures 4-17 to 4-21 and Tables 4-1 and 4-2. tors based on the following:
The choice of stress will be supported by the completion • An ac breakdown stress in the XLPE adjacent to the
of qualification and long-term prequalification tests. tip of the contaminant of 300 kV/mm at 20°C in a 15-
minute ramp test to failure.
This is the lowest-cost route for the manufacturer,
• A service life of 40 years.
because it follows other manufacturers, thereby learning
from their previous mistakes and avoiding the costs of • A deterioration coefficient (ratio of stresses E1/E2)
extensive test/verification programs. based on the (Ent = k) aging law with an aging expo-
nent of n = 15 for the XLPE adjacent to the contami-
It would be prudent for the user to insist that qualifica- nant.
tion/prequalification tests include the following sup- • A stress raiser calculated at the tip of the idealized
porting information for the design stress levels for the conducting contaminant with an assumed tip radius
particular application: of 10 μm.
• The safety margin used in the design. Note that the • A breakdown stress reduction factor of 1.2 (20%) for
safety margin within the design will not have been operation at 90°C.
validated; the only fact known is the compliance with
• A design factor of safety of 1.1 (10%).
the qualification test regime.
• Although competitors’ operating stresses will be The factors of safety in the formulae in AEIC CS9
known, details of competitors’ grades of insulating (2004) Appendix A were based on design methods from
and semiconducting compound, material-handling Japan that had been proven to be suitable by many test
and quality-checking procedures, and extrusion results and service experience on full-size cable systems
know-how will not be known. in the laboratory. In particular, the Japanese manufac-
• Other key elements such as conductor design, dimen- turers had developed impressive test technology that
sional tolerances, degassing procedures, metal sheath detected PD from incipient pre-failure and, at high
design, and routine test regimes are difficult to deter- speed, to automatically remove test voltage, thereby pre-
mine and emulate. venting disruptive failure of the insulation. They were
then, by painstaking dissection of the cable loop, able to
• Long-term prequalification tests must still be per-
find the site of the incipient activity to prove that it had
formed on the cable and its accessories even if the
been initiated by a contaminant of measurable size. The
designs are identical to those of other manufacturers.
full-size test results were then correlated with bench-top
tests on insulation samples containing artificial stress
4.23.3 Mechanistic Approach to Design
raisers such as needles of semiconducting compound.
The new AEIC CS9 (2004) specification is an excellent However, the success of the mechanistic design approach
example of the mechanistic engineering approach to is largely dependent on the size of the selected factors of
cable design based on limiting the defects in XLPE cable safety and in particular on the achievement of very high
to size limits specified in factory routine acceptance quality standards in material and manufacturing.
tests. Appendix A of AEIC CS9 (2004) gives two design
formulae. One formula calculates the maximum permis- A similar approach to the above considered the effect of
sible void-design stress relationship to yield the void size larger sporadic shield defects that either caused short-
limits given in Table 4-19. The second formula calcu- term failure of HV and EHV cables at factory accep-
lates the maximum permissible contaminant-design tance test or were discovered during routine examina-
stress relationship for the insulation. The same relation- tion of insulation samples (Smith et al. 2003). The
ship is used to determine the maximum permissible size geometry of the defects was field plotted, and the break-
of a shield protrusion on the assumption that its shape down stress in the short times to failure of less than 5
is the same as that of the insulation contaminant (an minutes was determined to be typically 100 kV/mm. An
ellipsoid), but that it is cut in half on its minor axis. The endurance line was then extrapolated according to the
maximum permissible size of the insulation contami-
4-71
(Ent = k) law, and the times-to-failure of different mag- The key parameter for material quality and process con-
nitude shield defects were calculated at service voltage. trol is the slope of the line; theoretically it should be ver-
The sizes of defects that could pass the HV withstand tical such that each sample fails at exactly the same
test and pass unfailed into service were calculated, predictable stress.
thereby prospectively risking premature failure. This
work emphasized the benefits of stringent quality con- The key value for cable design is the lowest breakdown
trol and in-line monitoring measures at the start of stress, 20 kV/mm, which is coincidental in this distribu-
manufacture and during extrusion. tion with a 1% probability of failure. It is important to
dissect the samples to find the “gross” defect that initi-
4.23.4 Statistical Breakdown Testing ated failure at this low stress. This minimum breakdown
Samples of insulation are taken from the shipping reel stress is extrapolated by the (E n t = k) calculation to
as part of the factory acceptance test, permitting the obtain its residual strength at the end of the 40-year life.
contaminant and protrusion levels to be measured, as The residual 40-year strength must be greater than the
specified in AEIC CS9 (2006). However, the sampling original cable design stress level, as given in AEIC CS9
volume is necessarily very low compared to the volume (2006).
of insulation in the long length of cable on the reel. A
significant variation in contaminant shapes, sizes, and The maximum ac breakdown stress level in Figure 4-71
numbers is expected along the reel length. Thus, if short is 100 kV/mm. This illustrates that if the model results
10-m lengths of cable are cut from production runs were extrapolated to samples of 345-kV production
either for statistical quality breakdown tests, or experi- cable having a design stress of 14 kV/mm at 199 kV
mental tests, a wide variation in breakdown strength is phase voltage, a maximum ac test voltage of 1.42 MV
experienced. would be required. Similarly a 500-kV cable with a
design stress of 16 kV/mm would require a test voltage
Figure 4-71 shows the results of experimental ac ramp of 1.8 MV. For this reason, some manufacturers, partic-
breakdown tests, which compare two grades of insula- ularly in Japan, have equipped their laboratories with
tion on model cables of small conductor size and several test sets capable of testing in excess of 3 MV to permit
millimeters of insulation. The results are plotted as a full-sized EHV cables to be tested to failure.
Weibull probability distribution. It can be seen that 1%
(i.e., 1 in 100) of the samples of the Y grade material The compromise test method, when a high-voltage test
failed at a stress of 20 kV/mm, and 63% of the samples set is not available, is to extrude and test a model cable
at 70 kV/mm, a ratio in breakdown stress of 3.5:1. The with a reduced insulation thickness—for example, a 15-
63% stress (an arbitary value at which a log function in or 35-kV cable. One practical disadvantage is that com-
the formulae is equal to unity) is useful for comparing mercial production of EHV cable is significantly delayed
the performance of materials Y and X; material X when the model cable is extruded. Other disadvantages
clearly being superior. are that the conductor size and manufacturing conditions
are significantly different. As the insulation is thin com-
pared to the conductor diameter, the insulation and con-
ductor shield stresses are likely to be of similar
magnitude, resulting in a significant proportion of the
failures occurring where the insulation shield had been
removed within the test terminations and not in the cable.
Figure 4-72 shows the results of statistical process con-

trol tests performed on production cables using impulse
voltage instead of ac voltage. It has been found that
impulse voltage is also able to initiate failure at defects
(Peschke et al. 1995). The maximum breakdown stress
in Figure 4-72 is seen to be approximately 200 kVp/mm,
this being 2.3 times the BIL design level of 87 kVp/mm
for an AEIC CS9 230-kV, 1050-kVp BIL cable. Thus it
requires an impulse breakdown voltage of 2.4 MVp to
initiate failure, which is comparatively readily available
Figure 4-71 AC voltage breakdown results on thin wall in HV laboratories. A further advantage of statistical
model cables (courtesy of CIGRE). impulse testing is that the test loop can be assembled
4-72
faster than an ac test loop, often without need to remove 4.23.5 Endurance Tests on Cables
the outer shield, since in some cables the shield can use- These tests are intended to establish the stress/time rela-
fully act as a resistive potential divider under transient tionship, the residiual strength, and the value of the
voltage. If experimental correlation tests have been per- aging exponent n. Figure 4-73 shows stress/time plots
formed, it is also possible to obtain an indication of the for 100-m loops, different size cables at ambient temper-
equivalent ac breakdown stress. ature, and n (the reciprocal of the slope). The tests are
shown to have been performed for up to 50,000 hours
(5.7 years). In the top graph for EHV cable samples with
31-mm insulation thickness and HV samples with
12-mm thickness, some cable samples failed at stresses
between 30 and 50 kV/mm within 30 hours, whereas
samples with stresses in the range of 12-20 kV/mm con-
tinued without failure up to 15,000 hours (1.7 years).
The aging exponent n is greater than 12. In the lower
graph, samples of 20-kV MV cable with an insulation
thickness of 5.5 mm had withstood a stress of
10.9 kV/mm with 95°C load cycles under dry conditions
and had then been subjected to short-term ramp tests at
1 hour, 10,000 hours, and 50,000 hours, with short-term
failure occurring at a similar stress in the range of 40-
70 kV/mm, showing that no obvious aging had occurred
to the surviving samples.
4.23.6 Endurance Estimation on Full-Sized Cable

Systems, CSEM Method
The performance of a complete cable system is limited
by other factors than the cable insulation, principally by
the accessories and by thermomechanical effects. The
Figure 4-72 Impulse breakdown results on complete system performance can be modeled by ana-
production cable (courtesy of Jicable) (Peschke et lyzing the endurance time of both test loops and service
al. 1995). installations containing cables and accessories. This
Figure 4-73 Time-to-failure tests on different size XLPE cables (courtesy

of Jicable) (Peschke et al. 1995).
4-73
approach is termed the Cable System Endurance Model To enable the extrapolation of the endurance curve to
(CSEM) (Attwood 1998). the design life, an aging model is required. The extrapo-
lations included in Figure 4-74 are based on the inverse
Figure 4-74 shows endurance data in the form of load power approximation. Extrapolation of this line to the
cycles survived at specified stresses for XLPE cable sys- prescribed design life (in this case, a 40-year life—i.e.,
tems up to 1998. The time in service is plotted for com- 1.46 x 104 daily load cycles) gives the limiting value of
mercial circuits. The high-stress data represent cable electrical design stress at the conductor shield.
breakdown tests at elevated temperatures under daily
load cycles. Thus, any endurance curve must lie towards The conductor shield design stress identified by this
the right-hand edge of the installed cable system sur- approach was 11 kV/mm for XLPE insulation based on
vival data and to the left of the cable-only failure data. early service and test experience up to 1998. The conser-
vative cable dimensions predicted by this model are
The position of the endurance line may be further shown in Table 4-27. It is noted that 500-kV cable sys-
refined by the estimation of the scale parameter for the tems with design stresses up to 14.6 kV/mm with an insu-
surviving cables using th e m ethod of Weibayes lation thickness of 27 mm have since been placed in
(Attwood et al. 1998). This method assumes that the service (Table 4-2).
cable systems will fail, under a voltage ramp, in the same
manner as cables at elevated temperatures. The assumed The CSEM model has some significant engineering
shape parameter (β in Figure 4-75) and the estimated attractions, namely:
scale enable the position of the endurance curve to be • It tracks real cable system performance.
estimated at any desired level of probability of failure.
• The cable systems have been subject to either the rig-
ors of transmission network operation or the acceler-
ated test protocol.
• The analysis is based on large sample numbers of
cables and accessories.
• The low probability estimate (Weibayes estimates) of
the endurance curve is inherently conservative.
• It is a continually evolving and responsive model as
service experience (years in service), new type
approval, and prequalification test data are added
year on year.
Table 4-27 Insulation Dimensions Predicted by CSEM for

Selected Voltages and Conductor Sizes at a 1% Probability
of Failure
Voltage Conductor Size Insulation Thickness
(kV) (kcmil) (mm)
230 1000 19
230 2500 16
345 2500 25
Figure 4-74 Endurance data for XLPE cable systems 400 4000 31
(courtesy of Jicable). 400 5000 30
4-74
REFERENCES Atkinson, R. W. 1924. “Dielectric Field in an Electric

Power Cable.” AIEE Journal.
AEIC. 1993. AEIC CS7-93. Specifications for Cross-
Linked Polyethylene Insulated Shielded Power Cables Attwood, J. R., B. Gregory, M. Dickinson, R. N.
Rated 46 Through 138 kV. 3rd Ed. Association of Edi- Hampton, and R. Svoma. 1998. “Development of High
son Illuminating Companies. New York. Stress HV and EHV XLPE Cable Systems.” Session
Paper No. 21-108. CIGRE.
AEIC. 1996. AEIC CS6-96. Specifications for Ethylene
Propylene Rubber Insulated Shielded Power Cables Field Research on 345-kV Underground Cable Sys-
Rated 5 Through 69 kV. 6th Ed. Association of Edison tems.” Joint AEIC-EEI-Manufacturers Steering Com-
Illuminating Companies. New York. mittee, IEEE Transactions on Power Apparatus and
Systems. Vol. 85. No. 4. April 1966.
AEIC. 2000. AEIC CS8-00. Specifications for Extruded
Dielectric Shielded Power Cables Rated 5 Through 46 Bahder, G., T. W. Dakin, and J. H. Lawson. 1974.
kV. 1st Ed. Association of Edison Illuminating Compa- “Analysis of Treeing Type Breakdown.” CIGRE Paper
nies. New York. 15-05. Paris.
AEIC. 2006. AEIC CS9-06. Specification for Extruded Bahder, G. et al. 1976. “Criteria for Determining Perfor-
Insulation Power Cables and their Accessories Rated mance in Service of Crosslinked Polyethylene Insulated
Above 46 kV Through 345 kVac. Association of Edison Power Cables.” IEEE Transactions on Power Apparatus
Illuminating Companies. and Systems. Vol. 95. No. 5. September/October.
Andersen, P., M. Dam-Andersen, L. Lorensen, T. Ball, E. H. and G. Maschio. 1968. “The AC Resistance
Tanabe, S. Suzuki, O. Kjær Nielsen, S. H. Poulsen, and of Segmental Conductors as Used in Power Cables.”
B. S. Varrer Hansen. 1996. “Development of a 420 kV IEEE Transactions on Power Apparatus and Systems.
XLPE Cable System for The Metropolitan Power Vol. 87.
Project in Copenhagen.” Session Paper No. 21-201.
CIGRE. Ball, E. H., H.W. Holdup, et al. 1984. “Development of
Crosslinked Polyethylene Insulation for High Voltage
ANSI/ICEA. 2000a. ANSI/ICEA S-94-649-2000. Con- Cables.” CIGRE 1984 Session. Paris. August/Septem-
centric Neutral Cables Rated 5 Through 46 kV. ber.
ANSI/ICEA. 2000b. ANSI/ICEA S-97-682-2000. Util- Bartnikas, R. and K. D. Srivastava, eds. 1987. Power
ity Shielded Power Cables Rated 5 Through 46 kV. Cable Engineering. Waterloo, Ontario: Sanford Educa-
tional Press.
ANSI/ICEA. 2003. ANSI/ICEA T-28-562-2003. Test
Method for Measurement of Hot Creep of Polymeric Bartnikas, R. and K. D. Srivastava. 2000. Power and
Insulations. Insulated Cable Engineers Association. Communication Cables, Theory and Applications. IEEE
Press Series, McGraw Hill, 2000, ISBN 0-7803-1196-5
ANSI/ICEA. 2004. ANSI/ICEA S-108-720-2004. and 2000, ISBN 0-07-135385-2.
Extruded Insulation Power Cables Rated Above 46
Through 345 kV. BICC Cables. 1997. Electric Cables Handbook. Third
Edition. Blackwell Science. ISBN 0-632-04075-0.
Argaut, P., H. Auclair, and E. Favrie. 1993. “Develop-
ment of 500 kV Low Density Polyethylene Insulated British Standards Institution. 2005. BS 7970. “Tests and
Cable.” Third IEE International Conference on Power Requirements for Metallic Wire and Foil Sheath Con-
Cables and Accessories 10kV-500kV Conference Publi- structions of Power Cables having XLPE Insulation for
cation 382. Rated Voltages from 66 kV to 132 kV.”
ASTM. 1980. ASTM C-177-76, Part 18 of Annual Stan- British Standards Institution 2001. BS 7912 Power
dards. American Society of Testing and Materials. Phil- cables with XLPE insulation and metallic sheath, and
adelphia. their accessories, for rated voltages from 66 kV (=72.5
kV) to 132kV (=145kV). Requirements and test meth-
ods.
4-75
Champion, T. C., F. Marciano Agostinelli, and R. D. Dynatech. 1975. Measurements by Thermatest Depart-
Rosevear. 1993. “Short Term and Long Term Testing ment. Dynatech Corporation, Cambridge, Massachu-
Programme on 230 kV XLPE Cable System.” Page 108. setts.
IEE Third International Conference. Power Cables and
Accessories. IEE Conference Publication 382. Eich, E., E. M. Allam, and W. H. Cortelyou. 1976.
Some Theoretical Considerations Affecting Design of
CIGRE. 1974 ELECTRA 32. “Breakdown of Gasses in Lapped Plastic Insulation for Superconducting Power
Uniform Fields, Paschen Curves for Nitrogen, Air and Transmission Cables. Appendixes to December 1976.
Sulfur Hexafluoride.” CIGRE WG 15.03. EPRI Report No. EL-269.
CIGRE. 1992. ELECTRA 141. “Guideline for Test on Emanueli, L. 1930. High Voltage Cables. New York:
High Voltage Cables with Extruded Insulation and John Wiley & Sons, Inc.
Laminated Protective Coverings.” Pp. 53-65. CIGRE
WG21.14. Energy Networks Association. 2004. Technical Specifi-
cation 09-16. Issue 2. “Tests on Power Cables with
CIGRE. 2003. ELECTRA 151. “Recommendations for XLPE Insulation and Metallic Sheaths and their Acces-
Electrical Tests—Prequalification and Development, on sories for Rated Voltages of 66 kV, 110 kV and 132 kV.”
Extruded Cables and Accessories at Voltages of Greater London.
than 150 kV and Less than Equal to 400 kV.” CIGRE
WG 21.03. December. EPRI. 1978. Research to Determine the Acceptable
Emergency Operating Temperatures for Extruded Dielec-
CIGRE. 2006. TB 303. Revision of Qualification Proce- tric Cables. November. pp. 6-3 and 6-5. EPRI Report
dures for High Voltage and Extra High Voltage AC EL-938. Project 933-1. Final Report. Institut de Recher-
Extruded Underground Cable Systems. CIGRE WG che de l’Hydro-Quebec.
B1.06.
EPRI 2002. Cable System Technology Review of XLPE
Cinquemani, P. L., F. L. Kuchta, M. M. Rahman, F. EHV Cables, 220kV to 500kV, EPRI Palo Alto, CA
Ruffinazzi, and A. Zaopo. 1996. “105°C/140°C Rated :2002.1001846
EPR Insulated Power Cables.” IEEE Transactions on
Power Delivery. Vol. 11. Issue 1. January. Fukawa, M., T. Kawai, Y. Okano, S. Sakuma, S. Asai,
M. Kanaoka, and H.Yamanouchi. 1996, “Development
Clavreul, R. and P. Reignier. 1993. “Development of of 500-kV XLPE Cables and Accessories for Long Dis-
Moisture Analysis for High and Very High Voltage tance Underground Transmission Line. III. Electrical
XLPE and LDPE Cable Insulations.” Page 26. IEE Properties of 500-kV Cables.” Pp. 627– 634. Power
Third International Conference. Power Cables and Delivery. IEEE.
Accessories. IEE Conference Publication 382.
Graneau, P. 1979. Underground Power Transmission, the
Constantjnov, T. E. 1964. “Dielectric Materials Mea- Science, Technology & Economics of High Voltage
surement, and Applications.” IEE Conference. London. Cables. John Wiley and Sons. ISBN 0-471-05757-6.
Crine, J. P., S. Pelissou, and H. St. Onge. 1987. “Elemen- Gregory, B., J. Monteys, S. Barris, and J. M. Mendez.
tal and Ionic Impurities in Cable Insulation and 1999. “Installation of 220 kV XLPE Cables in a Tunnel
Shields.” JICABLE. Paris. September. pp. 206-213. whilst Minimising Electromagnetic Induction in Com-
munication Cables.” Pp. 148–153. Jicable.
Demaria, G., B. Gregory, C. L. Griffiths, R. N. Hamp-
ton, and S. P. Mainwaring. 1999. “The Design of Power Groeger, J. H., J. L. Henry, and A. Garton. 1988.
Cables to Operate at High Electrical Stress.” Pp. 37–43. “Location and Concentration of Ionic Impurities in
Jicable. Polymeric Cable Insulation.” Conference Record. IEEE
International Symposium on Electrical Insulation. Bos-
Dissado, L. A. and J. C. Fothergill. 1992. Electrical Deg- ton. June. pp. 300-305.
radation and Breakdown in Polymers. IEE Monograph.
Peter Peregrinus. ISBN 0 86341 196 7.
4-76
Hagen, S. T. and E. Ilstad. 1993. “Reduction of AC- Kang, B. P. 1963. U.S. Patent No. 3,077,514. February
Breakdown Strength Due to Particle Inclusions in 12.
XLPE Cable Insulation.” Page 165. IEE Third Interna-
tional Conference. Power Cables and Accessories. IEE Kitchin, D. W. and O. S. Pratt. 1958. “Treeing in Poly-
Conference Publication 382. ethylene as a Prelude to Breakdown.” AIEE Transac-
tions. Vol. 77. Pt. III. June. Pp. 180-186.
Henningsen, C. H., K. B. Muller, K. Polster, R. G.
Schroth. 1998. “New 400 kV XLPE Long Distance Knowlton, A. E. 1957. Standard Handbook for Electri-
Cable Systems, Their First Application for the Power cal Engineers. 9th Ed. New York. McGraw Hill.
Supply of Berlin.” Session Paper No. 21-109. CIGRE.
Kreuger, F. H. 1964. Discharge Detection in High Volt-
ICEA. 1973. ICEA Publication No. S-61-402. NEMA age Equipment. Heywood.
Standards Publication No. WC5-1973. Thermoplastic
Insulated Wire and Cable for the Transmission and Dis- Kreuger, F. H. 1989. Partial Discharge Detection in High
tribution of Electrical Energy. Appendix G. National Voltage Equipment. Butterworth & Co (Publishers) Ltd.
Electrical Manufacturers Association. Washington. ISBN 0-408-02063-6.
ICEA. 1984. ICEA S-68-516. Ethylene-Propylene-Rub- Lemainque, H. and G. Terramorsi. 1972. “Research and
ber-Insulated Wire and Cable for the Transmission and Development in France in the Field of Extruded Poly-
Distribution of Electrical Energy. National Electrical ethylene Insulated High Voltage Cables, Second Part,
Manufacturers Association. Washington. The Use of High Density Polyethylene in the Manufac-
ture of High Voltage Cables.” CIGRE. Session 28.
ICEA. 2004. ICEA S-108-720-2004. Standard for Paris.
Extruded Insulation Power Cables Rated above 46 kV
Through 345 kV. Insulated Cable Engineers Associa- Loeb, L. B. 1965. Electrical Coronas. Berkeley, Califor-
tion. nia: University Press. p. 348.
IEC. 2001. IEC 62067. “International Electrotechnical Marshuis, P. H. F., E. H. Kreuger, and P. P. Leufkens.
Commission. “Power Cables with Extruded Insulation 1988. “The Effect of Different Types of Inclusions on
and Their Accessories for Rated Voltages above 150 kV P.E. Cable Life.” IEEE Transactions on Electrical Insu-
(Um = 170 kV) up to 500 kV (Um = 550 kV) - Test lation. Vol. 23. No. 6. December. pp. 1051-1055.
Methods and Requirements.”
Mashikian, M. S. and J. H. Groeger. 1987. “Ionic Impu-
IEC. 2004. IEC 60840. International Electrotechnical rities in Extruded Cable Insulation: Analytical Detec-
Commission. “Power Cables with Extruded Insulation tion Techniques, Sources, Nature and Effects.”
and their Accessories for Rated Voltages Above JICABLE. Paris. September. pp. 199-205.
30kV(Um=36kV) up to 150kV (Um=170kV)-Test
Methods and Requirements.” Mason, J. H. 1959. “Dielectric Breakdown in Solids.” J.
J. Birks and J. H. Schulinan, ed. Progress in Dielectrics.
Illdstad, E. et al. 1990. “Influence of Mechanical Stress Vol.1. New York: John Wiley & Sons. Pp. 1-58.
and Frequency on Water Treeing in XLPE Cable Insula-
tion.” IEEE International Symposium. Toronto. June. Matsubara, H. and H. Miyauchi. 1970. “Heat Aging
Properties Near Melting Point Temperature of Polyeth-
Ishikawa, I., S. Nakamura, S. Utsunomiya, S. Yama- ylene and Chemically Cross-Linked Polyethylene.”
moto, and T. Niwa. 1994. “The Research of Peroxide Sumitomo Electrical Technical Review. No. 14. Novem-
Decomposition in XLPE Cables.” Pp. 506. Electrical ber.
Insulation and Dielectric Phenomena. IEEE.
Maxwell, J. C. 1891. A Treatise on Electricity and Mag-
Jocteur, R., M. Osty, H. Lemainque and G. Terramorsi. netism. London: Constable & Co. Ltd.
1972.”Research and Development in France in the Field
of Extruded Polyethylene Insulated High Voltage McMahon, E. J. and J. R. Perkins. 1963. “Volume Dis-
Cables.” Session paper CIGRE 21-07 1972. charges and Treeing: a Primer.” Proceedings of the 5th
Electrical Insulation Conference. Chicago. Pp. 98-100.
4-77
Melson and Fawsett. 1923. “Second Report on the Smith, C., S. Galloway, B.Gregory, S. Lloyd and D.
Research on the Heating of Buried Cables.” Journal of Notman, 2003 “The Development of an Ultrasonic
IEE. Vol. 61. Quality Monitoring Process for the Manufacture of
Enhanced Reliability HV And EHV XLPE Cables”,
Natta, G. 1963. “From the Stereospecific Polymeriza- Paper A.2.6, Jicable 2003 Conference.
tion to the Asymmetric Autocatalyc Synthesis of Mac-
romolecules.” Nobel Lecture. December 12. Steenis, E. F. 1989. Water Treeing—The Behavior of
Water Trees in Extruded Cable Insulation. N.V. Kema.
Peschke, E., R. Schroth, and R. Olshausen. 1995.
“Extension of XLPE Cables to 500 kV Based on St. Onge, H., R. Bartnikas, et al. 1978. Research to
Progress in Technology. Jicable 1995 paper A.1.2. Determine the Acceptable Emergency Operating Temper-
atures for Extruded Dielectric Cables. November. EPRI
Rath, R. A. V. and J. B. Allison. 1956. Polyethylene. Report No. EL-938. Project 933-1.
New York: Interscience. Figure 5-8.
Tanaka, T. and A. Greenwood. 1983. Advanced Power
Ritter, L. R. Scherer, M. Laurent, R. Bautz, and H. Cable Technology. Boca Raton, Florida.CRC Press Inc.
Zimmermann. 1994. “HV Cable Experience in Switzer- ISBN 0-8493-5165-0 (Vol. I) and 0-8493-5166-9 (Vol.
land—Particularities and Experience.” Paper 21-04. II).
CIGRE Conference. 1994 Session.
Toriyama, Y. et al. 1967. “Degradation of Polyethylene
Sato, T., M. Muraki, N. Sato, and Y. Sekii. 1993. by Partial Discharge.” IEEE Transactions on Electrical
“Recent Technical Trends of 500 kV XLPE Cable.” Insulation. Vol. El-1. August. Pp. 83-92.
Page 59. IEE Third International Conference. Power
Cables and Accessories. IEE Conference Publication Union Carbide Company. n.d. “A Critical Comparison
382. of XLPE and EPR for Use as Electrical Insulation on
Underground Power Cables.” Kabelitems. Wire and
Schiweck, L. 1969. “Investigations into the Breakdown Cable. No. 157
Process in Epoxy Resin Mouldings at High Voltages.”
Elektrotechnische Zeitschrigt (A). Vol. 90. May. Pp. 675- Union Carbide Company. n.d. “Treeing Update Part I:
678. History.” Kabelitems. Wire and Cable. No. 150.
Sheldon, R. P. and K. Lane. 1965. “Thermal Conductiv- Union Carbide Company. n.d. “Treeing Update Part
ities of Polymers II—Polyethylene.” Polymer. Vol. 6. p. III: Water Trees.” Kabelitems. Wire and Cable No. 152.
205.
Wartusch, J. and H. Wagner. 1977. “About the Signifi-
Silver, D. A., R.G. Lukac, and R.A. Hartlein. 1983. cance of Peroxide Decomposition Products in XLPE
Optimization of the Design of Metallic Shield—Concen- Cable Insulations.” IEEE Trans on Electrical Insulation.
tric Conductors of Extruded Dielectric Cables Under Vol. 12. pp. 395-401.
Fault Conditions. April. EPRI Report No. EL-3014.
Project 1286-2. Whitehead, S. 1951. Dielectric Breakdown of Solids.
Oxford: Clarendon Press.
Simmons, D. M. 1923. “Cable Geometry and the Calcu-
lation of Current-Carrying Capacity.” AIEE Transac- Woodruff, L. F. 1938. Principles of Electric Power Trans-
tions.. mission. 2nd Ed. New York. John Wiley & Sons.
Simmons, D. M. 1932. “Calculation of the Electrical

Problems of Underground Cables.”
The Electric Journal. East Pittsburgh, Pennsylvania.
May/ November.
4-78
ADDITIONAL BIBLIOGRAPHY—WATER TREES McMahon, E. J. and J. R. Perkins. 1963. “Volume Dis-

charges and Treeing: a Primer.” Proceedings of the 5th
Bahder, G., T. W. Dakin, and J. H. Lawson. 1974. Electrical Insulation Conference. Chicago. pp. 98-100.
“Analysis of Treeing Type Breakdown.” CIGRE. Paper
15-05. Paris. Mashikian, M. S. and J. H. Groeger. 1987. “Ionic Impu-
rities in Extruded Cable Insulation: Analytical Detec-
Crine, P., S. Pelissou, and H. St. Onge. 1987. “Elemental tion Techniques, Sources, Nature and Effects.”
and Ionic Impurities in Cable Insulation and Shields.” JICABLE. Paris. September. pp. 199-205.
JICABLE. Paris. September. pp. 206-213.
Steenis, E. F. 1989. Water Treeing—The Behavior of
Groeger, J. H., J. L. Henry, and A. Garton. 1988. Water Trees in Extruded Cable Insulation. N. V. Kema.
“Location and Concentration of Ionic Impurities in
Polymeric Cable Insulation.” Conference Record. IEEE Tanaka, T. and A. Greenwood. 1983. Advanced Power
International Symposium on Electrical Insulation. Bos- Cable Technology. Vols. I & II. Boca Raton, FL: CRC
ton. June. pp. 300-305. Press.
Illdstad, E., et al. 1990. “Influence of Mechanical Stress Union Carbide Company. n.d. “Treeing Update Part I:
and Frequency on Water Treeing in XLPE Cable Insula- History.” Kabelitems. Wire and Cable No. 150.
tion.” IEEE International Symposium. Toronto. June.
Union Carbide Company. n.d. “Treeing Update Part
Kitchin, D. W. and O. S. Pratt. 1958. “Treeing in Poly- III: Water Trees.” Kabelitems. Wire and Cable No. 152.
ethylene as a Prelude to Breakdown.” AIEE Transac-
tions. Vol. 77, pt. III. June. pp. 180-186.
4-79

Pipe-Type
Authors: Frank L. Kuchta, Prysmian Cables and Systems
Jay A. Williams, Power Delivery Consultants, Inc.
Reviewer: Joseph T. Zimnoch, Consultant
Pipe-type cables have been the preferred cable type in North America from the earliest
transmission cable installations in the 1930s through the start of the twenty-first century,
and they still account for the largest percentage of installed length. Although extruded-
dielectric cables are becoming more common—even at 345 kV—there is still a great deal
of activity in the pipe-type cable industry, in uprating, relocating, and modifying existing
systems as well as installing new circuits both on land and for significant water crossings.
This chapter provides details of cable construction for each component of the cable sys-
tem, from conductor through pipe coating, and provides specific parameters that may be
referenced when designing a pipe-type cable system. The material provides a historical
perspective on insulations and on dielectric fluids used for impregnating the insulation
and filling the cable pipe, and addresses the innovations in these components in the last
several decades. The chapter also includes a discussion of pipe-type cable manufacturing
and quality control.
Related information may be found in other chapters. Chapter 2 describes the pipe-type
cable as well as other cable system types and their common applications. Chapter 3 pro-
vides details on components such as conductor constructions and sizes, etc. Chapter 8
describes accessories for all cable types. Chapter 9 provides important information on the
hydraulic design and cooling of pipe-type cables. Chapter 10 addresses the corrosion pro-
tection system that is vital for a long, trouble-free life of the cable pipe. Chapters 12 and
13 cover designing and installing the cable system, Chapter 14 describes testing, and
Chapter 15 addresses Operation and Maintenance. Chapter 16 reviews power system
considerations for all cable types.
Frank L. Kuchta received his BSEE degree from Fairleigh Dickinson

University in 1982. Upon graduation, he joined Prysmian Power
Cables and Systems (formerly Pirelli Cable Corporation), and has held
various positions in design and application engineering. He is currently
Assistant Chief Engineer Manager of High Voltage Engineering in the
Research, Development and Engineering Department. His duties
include development of new products, high-voltage cable system
design, and industry standards for the North American operations. Mr.
Kuchta is currently President of the Insulated Cable Engineers Association (ICEA), a
Member of the IEEE Power Engineering Society and Standards Association, a Voting
Member of the Insulated Conductors Committee (ICC), a Member of CSA, and a
Member of CIGRE. He is the author of multiple technical papers and publications and
the holder of two patents.
5-1
Chapter 5: Cable Construction: Pipe-Type EPRI Underground Transmission Systems Reference Book
Jay Williams has degrees from Brown engineering group. He co-founded Power Delivery Con-
University and New York University. sultants, Inc. in 1992, and has continued to specialize in
He worked in the cable section at Con underground transmission cable systems. He was
Edison in the 1960s and early 1970s technical editor and principal author of the 1992
when the utility was installing major edition of the Underground Transmission Reference
lengths of 345-kV cable. He was part Book, and was project engineer for several other major
of the design team for the Waltz Mill EPRI projects including the original Underground
Test Facility, and was in charge of the Transmission Workstation. Mr. Williams is a Fellow of
utility advisory committee for early EPRI supercon- the IEEE and a registered Professional Engineer in New
ducting cable projects. Mr. Williams spent 20 years at York and Ohio.
Power Technologies, Inc., where he headed the cable
5-2
EPRI Underground Transmission Systems Reference Book Chapter 5: Cable Construction: Pipe-Type
5.1 INTRODUCTION The three phases making up a line are pulled together
Impregnated-paper insulation has been the most widely into a previously installed, coated and cathodically pro-
used dielectric material for power transmission cables tected steel pipe, with distances commonly 3000 ft
through the start of the twenty-first century. Pipe-type (980 m) between splices for 345-kV cables. After the
cables using impregnated paper or laminated paper entire line is installed, including splices and termina-
polypropylene have predominated in North America, tions, and the pipe welding is completed, the line is evac-
and there are significant pipe-type cable installations in uated, then filled with a dielectric liquid (or possibly
Ireland, South America, the Middle East, and Japan. nitrogen gas for HPGF cables). The liquid is pressurized
The first pipe-type cable was installed in 1932, and to a nominal 200 psig (1380 kPa) using a pressurizing
there are many new pipe-type cable projects under way plant that also has a reservoir tank to accept volume
in 2006. changes due to system thermal expansion and contrac-
tion, as well as pumps, controls, and alarm systems. The
Pipe-type cables have been the most commonly used presence of the dielectric liquid and need for a pressuriz-
transmission cable type in North America for several ing plant must be taken into account for the HPFF
major reasons: cable system, but those considerations do not exist with
a high-pressure gas-filled cable system. Figure 5-1 shows
• The pipe itself is very rugged. a 345-kV HPFF cable.
• The system is reliable.
• The pipe can be installed in relatively short street This chapter describes the construction of pipe-type
openings, minimizing traffic disruption in crowded cables. Section 5.2 reviews the main conductor designs.
urban areas, and it can generally be installed in Section 5.3 discusses conductor and insulation shields.
longer lengths with fewer manholes and more quickly Section 5.4 covers the insulation of pipe-type cables,
than a concrete-encased duct bank. including the history of its development, impregnated
paper, laminated paper polypropylene, applications,
• Presence of the dielectric liquid provides several electrical and physical properties, breakdown mecha-
options for cooling to improve the circuit rating. nisms, and impregnating liquid. Section 5.5 describes
the outer layers and filling medium. Section 5.6 identi-
Since the development of pipe–type cables in the mid fies quality control procedures, and Section 5.7 identi-
1930s, more than 80% of the 4200 approximate circuit fies specifications relevant to pipe-type cables.
miles of underground high-voltage transmission cables
installed in North America, at voltages up to 345 kV,
5.2 CONDUCTORS
have been pipe-type cables.
The main conductor designs utilized in pipe-type cables
The first commercial usage of laminated paper polypro- in the United States are round (concentric, compressed,
pylene (LPP) insulation instead of conventional paper or compact) and compact segmental. The conductor
insulation occurred in 1987 at the 345-kV voltage level. construction used in a particular cable application
Since that date, over 2 million conductor feet of this depends on the conductor size, cable type, conductor
LPP insulated pipe-type cable have been installed in new material, cable rating, and installation conditions. Cop-
345/230-kV circuits and for recabling the pipes of older per compact segmental conductors are used for a major-
existing 138/115-kV circuits.
5.1.1 General Description

Pipe-type cables have the three cables, insulated with
tapes of kraft paper or laminated paper polypropylene,
installed in a common steel pipe. The pipe is pressur-
ized, usually with a dielectric liquid, and then is called a
high-pressure fluid-filled (HPFF) or high-pressure liq-
uid-filled (HPLF) cable system. This cable system was
called high-pressure oil-filled (HPOF) until the 1970s,
when synthetic dielectric liquids began to replace the
mineral oils that had been used earlier. Pipe-type cables
rated up to 138 kV (and possibly 230 kV in the future)
can also be pressurized with nitrogen, and are then
Figure 5-1 345-kV HPFF pipe-type cable (courtesy of
called high-pressure gas-filled (HPGF) cables.
The Okonite Company).
5-3
ity of the high-current installations in North America. roundness on segmental conductors was to use four
For example, several utilities have standardized on equal-size segments with sharp corners and a tight-
2500 kcmil (1266 mm2) compact segmental copper con- fitting, solid copper rod as a center filler. The copper
ductors for 345-kV installations. Aluminum conductors rod stabilizes the conductor segments during bending,
have been used, but are much less common. Chapter 3 and the sharp corners and equal-size segments make it
provides more detail on conductors, including physical possible to obtain a very circular conductor with little
and electrical properties. The following section provides or no fillers. With this method, all segments are
comments specific to pipe-type cables. wrapped with paper to keep the segments insulated from
one another. Two of the segments are wrapped with
Copper conductor strands are either uncoated or coated insulating paper, while the other two are wrapped with a
with tin or lead alloy. Since earlier paper-cable impreg- carbon-black paper tape, thereby ensuring electrical
nating fluids had a tendency to increase the dissipation continuity between the conductor and the conductor
factor slightly when in contact with bare copper, tin and shield. It was found through measurements of ac/dc
lead-alloy-coated strands were employed when it was resistance ratio that all four segments could be wrapped
desirable to minimize this change. These practices are with the carbon-black paper without increasing the con-
considered unnecessary today. ductor losses. This is because the carbon-black paper
has a sufficiently high resistance to prevent circulating
Insulating enamel coatings have been applied to copper currents from flowing across the segments. Figure 5-2
conductor strands or the strands were intentionally oxi- shows an older conventional segmental conductor and
dized to increase the interstrand resistance and thereby an example of a typical improved design.
reduce conductor losses due to proximity effects. The
use of enamel-coated strands is quite rare, however, Another commonly used variation of this improved
because the marginal increase in ampacity does not jus- design utilizes two bare copper segments and two
tify the increase in cost and difficulty in joining the con- slightly smaller segments, which are spirally wrapped
ductor. Aluminum readily forms a tenacious, high- with a very tough special insulating paper tape for seg-
resistance oxide layer and no additional coatings are ment isolation, along with only one close-fitting nylon
needed. See Chapter 3. rod filler in the center interstice.
For pipe-type cables of siz es larger than about

1250 kcmil (633 mm2), up to which conventional round
conductors have been used, segmental conductors are
primarily utilized. The advantage in reduced ac resis-
tance is quite significant. This is directly related to the
close proximity of the cables in the pipe. Copper is used
in most installations. However, there are a smaller num-
ber of cases in which segmental aluminum conductors
have been used (Roughley 1974). Copper conductors are
usually the preferred economic selection for pipe-type
cables, despite their high transverse conductivity relative
to aluminum.
In recent years, with the advent of reduced-wall cable

designs (Prime 1981, 1986) and hot-impulse withstand
Figure 5-2 Conventional segmental conductor and
requirements for pipe-type cables, it was found that a improved design.
very smooth and round conductor helped to ensure sat-
isfactory performance. One method of improving
5-4
5.3 SHIELDS referred to as a foil-backed paper tape. Some manufac-

turers employ a perforated metal foil in this tape to
5.3.1 Conductor Shield facilitate drying and impregnating the paper insulation,
The conductor shield of an impregnated-paper-insu- while others have used a no perforated foil with satisfac-
lated cable consists of two or more carbon-black paper tory results.
tapes (or occasionally metallized paper tapes) applied
with an open butt, with each overlying tape covering In higher-voltage cables, the insulation shield is a mirror
the butt gaps in the underlying layer. In some designs image of the conductor shield, including the optional
of extra-high-voltage (EHV) cables, a duplex carbon- duplex paper tape (carbon side out) and carbon-black
black tape is applied over the carbon-black paper tapes. paper tapes beneath the perforated metallized paper or
The duplex tape consists of an insulating tape and a metallized carbon-black paper tapes.
carbon-black paper tape, which are bonded together.
The duplex tape is applied with the carbon-black side 5.4 INSULATION AND IMPREGNANT
against the conductor shielding. Such a tape provides a
more uniform transition from the conductor shield to 5.4.1 Background—Paper and Laminated Paper
the insulation wall. The use of the duplex tape in EHV Polypropylene Insulation
cable is often necessary to meet very low dissipation-
Paper Insulation
factor requirements. It has been theorized that, when
The first use of paper tape applied helically as a wire
the butt gaps at the conductor have one surface consist-
insulation is thought to have been in 1872 by
ing of a carbon-black tape, movement of the carbon
Wortendyke, a Virginia paper manufacturer, for a bell
particles in the applied electric field causes an apparent
system in his home. McCracken improved on Wor-
increase in the measured cable dissipation factor
tendyke’s crude taping machine and patented the
(Gooding and Blodgett 1964; Gazzana-Priaroggia et al.
concept of paper-insulated wire in 1884. Vacuum drying
1964; Kojima et al. 1965). The duplex tape minimizes
and hot, pressurized impregnation were patented in
such butt gaps. Figure 5-3 illustrates conductor shield-
1885 by Jacques. The Norwich Wire Company, founded
ing for impregnated-paper-insulated cables.
by McCracken, began manufacturing paper cables in
1892. In the late 1890s, rubber and paper, both covered
5.3.2 Insulation Shield
with a lead sheath, competed for supremacy as the
The insulation shield design varies with pipe-type cable preferred power cable insulation, but by 1902, paper was
voltage class and dissipation-factor requirements. The the clear winner. Paper-insulated cables impregnated
most basic insulation shield, used in lower-voltage with rosin oils or compounds and covered with lead
classes, consists of a metallized paper tape that is sheaths provided the means for the early evolution of
applied with the metallized side facing away from the modern power systems in the first two decades of the
insulation. The metallized paper tape consists of a paper twentieth century.
tape bonded to a thin aluminum foil. This is also
In 1917, Luigi Emanueli developed the fundamental
fluid-filled, pressurized cable system, which permitted
higher voltages and the use of lighter mineral fluids in
place of the lossy rosin compounds. The 1920s saw
improvements in fluids and paper selection that raised
the performance levels of both solid-type and fluid-filled
paper-insulated cables. A 66-kV solid-type system with
fluid reservoirs on each joint was installed in Cleveland
in 1923. In 1926, a 66-kV fluid-filled cable was installed
in England, and 132-kV fluid-filled cables were installed
in Chicago and New York. In 1932, Okonite provided
the first experimental “Oilostatic” fluid-filled pipe-type
cable operating at 200 psi (1380 kPa) for a 66-kV line at
Plymouth Meeting near Philadelphia, and the British
installed a gas-pressurized pipe-type cable in London.
Figure 5-3 Conductor shielding for paper cables In 1934, the first commercial 138-kV pipe-type cables
(courtesy of Prysmian Cables & Systems). were installed by the Pennsylvania Railroad Company
in Baltimore tunnels.
5-5
From the 1930s through the end of the twentieth cen- domestic paper producers, and the door opened to
tury, pressurized paper-insulated cables dominated the foreign suppliers. Medium-and high-density Fourdrinier
underground transmission industry with no significant papers, in use for some time in Europe in self-contained
changes in basic concepts. There has been, however, a fluid-filled (SCFF) cables, began to be imported as a
continuing evolution of this technology in response to second source of paper, putting further pressures on
increasing performance levels and changing market con- domestic producers.
ditions. In the mid-1950s, four cable companies in the
Laminated Paper Polypropylene Insulation
United States, working with their paper suppliers,
Impregnated-paper laminar insulation provided efficient
achieved cable characteristics suitable for 345-kV sys-
and reliable cable insulation for distribution and trans-
tems, and each supplied cables and accessories to a dem-
mission cable applications throughout most of the twen-
onstration project at Cornell University sponsored by
tieth century. However, as power demands in the United
the Association of Edison Illuminating Companies
States increased steadily, with projections of continuing
(AEIC), Edison Electric Institute (EEI), and the cable
rise, research seeking lower-loss synthetics for laminar
and accessory manufacturers (AEIC 1966). That pro-
insulation was intensified worldwide.
gram was successfully completed in 1964. Consolidated
Edison Co. of New York had already installed the first
It was clear that for the higher-voltage 800- and
of many 345-kV cable systems. Paper-insulated pipe-
1100-kV voltage systems, transmission cables would
type cables reached their highest level of development in
require a new, low-loss laminar insulation superior to
a second industry-sponsored program to establish
impregnated-cellulose paper. This basic concept of
500-kV cable systems. In 1972, cable systems provided
combining layers of paper and polymer tapes was
by the same four U.S. manufacturers successfully com-
originally cited in a 1963 U.S. patent describing a
pleted a four-year test program at the Waltz Mill Test
modified high-voltage cable, which employs a bonded
Site in Pennsylvania. Each of these systems demon-
laminate insulation of synthetic film and synthetic paper
strated complete stability and the lowest dielectric losses
(Kang 1963). However, synthetic paper never proved
ever obtained with large-scale paper dielectrics.
practical or successful. Polypropylene film replaced
polyethylene in subsequent industry developments,
The 1970s witnessed turmoil in the transmission cable
principally to attain higher temperature capabilities that
industry. Expenditures for new circuits fell off rapidly,
at least matched the thermal stability of conventional
creating competitive pressures on all of the manufactur-
cellulose-paper insulation.
ers. Paper costs came under close scrutiny; this led to
compromises in controllable properties that had taken
In the United States, the Electric Power Research Insti-
years to establish. During this time, the utility industry
tute sponsored R&D projects that resulted in the suc-
imported the high-temperature impulse withstand
cessful laboratory development of a 765/800-kV
requirements used in Europe. It was quickly found that
laminated paper polypropylene (LPP)-insulated pipe-
papers so carefully developed to optimize dielectric loss
type cable system, which subsequently completed the
characteristics could not easily meet the impulse criteria
two-year Waltz Mill field test programs and earned
at elevated temperatures. Higher-density papers were
industry approval for commercial applications in 1985
required, resulting in increased dielectric losses.
(Allam 1982; Allam et al.1986). Further developments
and refinements by the mid-1980s produced a 345-kV
There was also a restructuring within the paper industry.
LPP-insulated pipe-type cable system, consolidating the
The energy crisis and environmental pressures caused
new technology for LPP cable applications in the
irreversible changes in the availability of many high-
138-500 kV, U.S. commercial range (Allam et al. 1988;
performance papers. The declining production of cable
Eich et al. 1976). (The AEIC designation for this insula-
papers coupled with the general demand for paper and
tion is laminated paper polypropylene. A variety of
paper products contributed to the problem. The early
abbreviations are used worldwide. LPP is used in the
1970s saw a worldwide shortage of pulp, which limited
2007 edition of this book, replacing the trademarked
options open to paper makers. One supplier of a
term PPP that was used in the 1992 edition.)
premium high-alpha pulp, which yielded papers with
exceptionally good loss characteristics, discontinued
Transmission systems insulated with LPP can signifi-
production in favor of high-volume, standard pulp
cantly reduce capital and operating costs and provide an
grades. Cylinder paper machines, which had produced
attractive new alternative for utilities that are planning
most of the cable papers in service, were unable to
system expansion. Although somewhat higher in cost
compete with the faster Fourdrinier machines. As a
than conventional paper, LPP insulation is cost-effective
result, cable manufacturers had to qualify new types
in view of cumulative system benefits, including smaller
and sources of paper. There was a rearrangement of
cable diameter, smaller pipe size for pipe-type cable,
5-6
smaller and fewer joints and manholes, longer pulling mal characteristics of the LPP-insulated cable system
lengths, and lower installation and operating costs. (Allam et al.1988).
Additionally, the significant savings in cost-of-losses is a
major factor. 5.4.2 Impregnated Paper
The term “paper” is an extremely broad term. Paper is
LPP is applicable both for new installations, and for remade by suspending fibers in water to form pulp, then
cabling existing systems (retrofits for voltage and/or con- depositing the pulp onto a wire screen or felt and
ductor-size upgrade). LPP is equally applicable to removing the water to leave the fibrous stock in sheet
medium- or high-pressure self-contained fluid-filled form. The raw material is usually derived from plants
cables, which are employed extensively for underground containing cellulose, but papers can be made from syn-
transmission in most countries. LPP laminate is avail- thetic fibers such as nylon or polypropylene. Synthetic
able from overseas suppliers. papers are differentiated from cast or extruded films of
the same polymer by having a finite air resistance, or the
MVA ratings for paper-insulated versus LPP-insulated ability to pass some amount of air through the sheet.
345-kV pipe-type cables are compared in Figure 5-4 for Films are said to have infinite air resistance.
a range of copper conductor sizes. A physical-size com-
parison, shown in Figure 5-5, of equally rated systems The properties of a paper depend on the choice of raw
illustrates the reduced size of the LPP-insulated cable material, its preparation, the use of additives, the nature
assembly, leading to reduced costs and improved ther- of the paper-forming process, and treatment after for-
mation. The process is so complex that it is virtually
impossible to produce the exact same paper in two dif-
ferent mills. The characterization of paper at each step
in the process of formation and in its final form is done
in accordance with test methods set up by TAPPI, the
Technical Association of the Pulp and Paper Industry.
The American Society for Testing and Materials
(ASTM) has also established many test methods that
are useful in the relation of paper characteristics to
functional requirements. ASTM-D 202 is the general
characterization and testing specification for electrical-
grade papers (ASTM 2002a).
Paper classification begins with the raw material that is

used to make pulp, and the process of reducing the raw
Figure 5-4 Comparative MVA ratings for 345-kV pipe- material to pulp. Prior to 1850, the only practical source
type cables. of cellulose was used rags of cotton and linen. The
Figure 5-5 Comparable cable cross sections of paper-insulated and LPP-insulated 345-kV cables (courtesy of
Cablec Utility Cable Company). (Note that this figure uses the term PPP, which is Cablec's trademark for the
laminate that is now commonly called LPP.)
5-7
chemical reduction of wood to pulp by the soda process with cellulose of long chain lengths, or a high DP. This
was introduced in 1855. In 1867, the Keller process for preferred structure is the alpha cellulose, which exhibits
the mechanical reduction of wood to pulp reached the a DP that may be as high as 10,000. Alpha cellulose is
United States from Germany. The sulfite process arrived not soluble in water or alkali solution. Beta cellulose
in 1882, and the sulfate process, which was predomi- and gamma cellulose are shorter chain structures that
nantly used through the 1980s, was first applied in 1909. will dissolve in an alkali solution.
Semichemical pulping, which combined mechanical
reduction with various chemical processes, was intro- Hemicellulose is a nonfibrous polysaccharide material,
duced in 1925, opening the door to the use of hard- with a shorter chain structure than cellulose. The DP of
woods, which were plentiful in the United States. The hemicellulose is in the range of 150. The hemicelluloses
sulfite process is being used again now because of the can contribute to interfiber strength, but they are also
lower level of residual acidity in the paper and the quite lossy. Papers suitable for EHV cable are made with
improved archival qualities. high-alpha pulp which has low levels of hemicellulose.
The first papers used for electrical insulation were of the The transformation of pulp into paper is accomplished
hemp family, using fiber reclaimed from used cordage as by spreading the pulp uniformly over a flat surface and
the source of cellulose. The increase in electrical perfor- removing the water. The individual fibers are strongly
mance required by the move to higher voltages in the attracted to each other by molecular forces. As water is
1920s, together with cost considerations, caused a shift removed, the fibers settle in to form a coherent sheet.
to kraft papers made with pulp produced by the newly The preparation of the fibrous stock is crucial to the
implemented sulfate process. Kraft papers were more successful formation of the paper sheet. The pulp is put
uniform and easily tailored to meet specific structural through a mechanical beating process, which has the
requirements, one of which was the need to obtain more effect of breaking up large groups or bundles of fiber
thorough impregnation. Rope papers, while mechani- into a smoother suspension of smaller fiber groups. In
cally tough, had poor dimensional control and relatively so doing, the attractive forces between fibers are
high dielectric losses—factors that kraft papers dramati- increased. In fact, the pulp becomes noticeably sticky
cally improved. Since the 1950s, virtually all of the paper after beating, even when no additional sizing has been
used for 69-kV and higher-voltage cables in the United added. In general, more beating produces a stronger
States has been kraft paper. paper up to a point. Since beating is an intensive activ-
ity, some damage to individual fibers is unavoidable.
Kraft paper is made from coniferous wood such as pine, Too much beating causes more loss in fiber structure
spruce, and cedar. Wood consists of about 40% cellu- than is gained in interfiber attraction, and the paper
lose, 30% hemicellulose, and 30% lignin. Lignin is an strength falls off.
amorphous polymer that binds the cellulose fibers and
hemicellulose into the strong textured wood structure. The characteristics of the paper sheet are determined by
Sulfate or sulfite processes are used to chemically sepa- the manner in which the fibers orient themselves as they
rate the cellulose from lignin and hemicellulose mate- settle into the sheet. If all of the fibers were to lay down
rial. As in most separation processes, the more complete parallel to the surface and in the same direction, the
the separation, the more time-consuming and costly is paper would be very dense and have high tensile
the process. Also, the sought-after cellulose is damaged strength in the direction of the fibers (grain-long direc-
as the process progresses, so the process must be opti- tion), but very low cross-grain strength (e.g., newsprint).
mized to obtain the desired pulp/paper characteristics at If the fibers were made to lie randomly in all directions,
reasonable cost and yield. the paper would be less dense and have a moderate ten-
sile strength in any direction. This sheet would be called
The main component of paper is the cellulose fiber. Cel- “square,” because it would have isotropic mechanical
lulose is a carbohydrate, specifically a polysaccharide properties (e.g., bond paper). If some of the fibers were
whose base sugar unit is C6H10O5. The number of units oriented obliquely to the plane of the sheet, and the
in the chain-type molecule is called its degree of poly- sheet thickness was large relative to fiber length, the
merization (DP). The individual molecular chains tend paper would be soft and have a low compressive modu-
to align themselves into bundles known as microfibrils, lus (e.g., tissue paper). Beating establishes the capabili-
which consist of crystalline regions separated by amor- ties of the pulp; the wet end of the paper machine
phous regions. Thousands of microfibrils combine into controls the actual formation of the paper sheet. Refer-
a long tubular structure called a fibril, or fiber. Many ences (Casey 1961; Grant 1957; Kelk and Wilson 1965;
properties of paper are improved by making the paper
5-8
Bartnikas and Srivastava 1987) provide good back- The second type of machine is the cylinder machine that
ground information on paper for high-voltage cable use. was invented by John Dickinson in Germantown, Penn-
sylvania in 1809, and put to use before the Fourdrinier
As noted previously, two fundamental machines are machine reached America. The cylinder machine forms
used for the manufacturing of essentially all papers used the paper sheet on a wire mat, which covers a large cyl-
in modern times. The first was the Donkin machine inder. The cylinder is mounted in the horizontal plane
invented by Louis Robert in France in 1799 and devel- and sits in a vat of the pulp/water suspension with about
oped into a commercial production machine by Gryan one quarter of the cylinder above the liquid surface. The
Donkin in England in the early 1800s. Since the Donkin cylinder wall is open and the ends of the cylinder are
machine development had been financed by the Four- hydraulically isolated from the vat, allowing the water
drinier Brothers, this type of machine assumed their level inside the cylinder to be maintained at a lower level
name. The first Fourdrinier machine to be constructed than the water in the vat. As the cylinder rotates, water
in the United States was set up at Saugerties, New York flows inward through the wire and the cylinder wall and
in 1836. deposits the fiber on the wire. A continuous felt belt
contacts the cylinder at the top and transfers the sheet
The Fourdrinier machine uses a continuously moving from the cylinder to the felt.
belt made of a wire mesh. Pulp suspended in water is
uniformly poured onto the wire at the start of its hori- The characteristics of the paper are established by the
zontal run. The wire passes at high speed over several fiber orientation on the wire, which is affected by the
rolls that support the wire and, with the help of gravity, radial flow of the pulp suspension through the wire into
much of the water is drained off, leaving the pulp the cylinder, the tangential flow of the wire surface rela-
behind. The wire then passes over a series of suction dri- tive to the pulp suspension in the vat, and the centrifu-
ers, which pull off the bulk of the remaining water gal forces caused by the cylinder rotation. Unlike the
through the wire. As the fiber sheet dries, it picks up Fourdrinier machine, these controlling variables act on
strength. By the time the fiber sheet passes the suction the formation process for almost the entire time that the
boxes and one last suction roll, it can support its own sheet is in contact with the wire. Flow through the wire
weight. The sheet is then lifted off the wire and trans- and centrifugal force tend to orient the fiber normal to
ferred to a continuous felt belt. On the felt, the sheet the surface of the sheet which helps build a thick soft
passes through roller presses, steam-heated drier rolls, sheet. Tangential drag forces cause the fibers to lay
and finishing rolls to emerge from the machine in its down on the wire in the machine direction. Flow in the
final dry form. vat can be either in the direction of rotation, to mini-
mize fiber laydown, or opposite to rotation to maximize
One of the most critical points in the machine is the flow machine-direction orientation. The former is used for
spreader, where the pulp suspension first contacts the paperboard production, and the latter is the configura-
wire. The properties of the paper are controlled by the tion for very thin but strong (in the machine direction)
manner in which the individual fibers in the pulp settle papers such as cable paper. Cylinder papers are inher-
onto the wire. The flow tends to provide some alignment ently anisotropic, either in the machine direction or in
of fiber in the direction of motion, known as the the direction normal to the sheet. It is almost impossible
machine direction, which imparts anisotropic properties to make isotropic paper of any strength and quality on
to the paper. If the wire is made to travel slightly faster this type of machine.
than the flow, the pulling action of the wire on the pulp
fibers results in more orientation of the fibers. In gen- The initial and the operating costs of cylinder machines
eral, however, Fourdrinier paper tends to be only are low compared to the Fourdrinier machine, but their
slightly anisotropic because of the lack of a positive running speed is limited predominantly by the centrifu-
means to orient fiber in the machine direction. This is gal forces tending to throw the fiber and sheet off the
an advantage for most paper applications. The Four- cylinder. The Fourdrinier machine has about a 10:1 run-
drinier machine is capable of very high running speeds, ning speed advantage, which explains why most paper is
limited only by the rate that water can be withdrawn made on them. Paperboard and thin, highly anisotropic
from the sheet. Fourdrinier machines dominate the papers serving niche markets occupy most cylinder
paper industry because of their high running speed and machines. In the 1980s, these niche markets deteriorated
low production costs. due to the increase in the use of synthetic materials and
competition from Fourdrinier machines.
5-9
Paper Characteristics laminating an extruded polypropylene film that is sand-

The paper characteristics that are critical to the manu- wiched, while hot, between two cellulose-paper layers.
facture of high-quality cable are: The three layers are bonded using two pressure rollers.
• Tensile strengths in both machine and cross-machine No adhesive or copolymer materials are used in making
the LPP laminate.
directions
• Compressive modulus normal to the surface (soft- The LPP-insulated cable approach was successfully
ness) and in the cross-machine direction (stiffness) developed in Japan, the United Kingdom, and the
• Tensile modulus of elasticity United States, because extensive worldwide research had
• Torsional modulus shown that synthetic and semisynthetic dielectric mate-
rials alone, in general, could not meet some of the fol-
• Surface friction in the cross-machine direction lowing critical basic requirements:
• Hygroexpansivity • Low dielectric loss
• Air resistance • High dielectric strength
• Density • Good resistance to swelling in oil
• Dielectric loss • Adequate ratio of tensile modulus to compressive
• Dielectric strength modulus
• Cost-effectiveness
All of these characteristics are under the control of the
paper manufacturer. Needless to say, a cable manufac- The major advantages of LPP over conventional paper
turer has a very close relationship with his paper sup- in EHV cables are due to the polypropylene film and
plier. Slight changes in any of these characteristics can include the following:
upset a delicate balance.
• Lower dielectric loss
5.4.3 Laminated Paper Polypropylene • Higher ac and impulse strength
LPP insulation utilizes a composite construction involv- • Thinner insulation walls
ing a layer of low-loss, electrical-grade, homocast
• Higher ampacity ratings
polypropylene film (unoriented), bonded without adhe-
sive between two layers of high-purity kraft electrical-
Pipe-type, LPP-insulated cable designs up to 800 kV are
insulating paper, as shown in Figure 5-6. This laminate
now commercially available. The basic electrical, physi-
tape looks and handles like a paper-insulating tape of
cal, and thermal properties of LPP insulation and LPP
equal thickness. The LPP composite is manufactured by
cables are described in following sections.
5.4.4 Cable Application

General
The dielectric properties of impregnated paper have
made this material suitable for many insulation func-
tions ranging from motors and transformers to capaci-
tors and cables. However, it is the mechanical attributes
of paper that make this the material of choice for such a
broad range of applications. The dielectric strength of
paper is greatest when the electrical stress is normal to
the paper surface, which is consistent with the wrapping
of a high-voltage conductor with paper tape. The
mechanical strength of the tape becomes the limiting
factor for the wrapping operation, especially for cable
manufacture where high production speed is essential.
Figure 5-6 Laminated paper polypropylene insulation Early cables used paper made from rope fiber because
(courtesy of Cablec Utility Cable Company).
this type of paper has the highest strength obtainable
5-10
with natural fiber. The change to dielectrically superior This butt space is required to allow the core to bend
kraft papers was made possible by improvements in the without causing the adjacent tape turns to collide and
paper industry and the cable manufacturing equipment. wrinkle. Smooth bending is possible only if each tape
Manufacturing requirements are very similar for LPP can slide between the tapes below and above it. Sliding
insulation except for the drying and impregnating cycles is dependent on the coefficient of friction between two
as discussed later. layers of paper and the compressive force in effect. The
position of each tape must be carefully controlled rela-
Manufacturing Requirements tive to the tape that it covers to ensure the optimum dis-
Cables are made by the helical application of many lay- tribution of butt spaces. The paper must have sufficient
ers of tape to a conductor in a horizontal taping stiffness to resist collapsing into a butt space of either of
machine. Conductor is supplied on a reel that is set up its neighbors. It also must have enough strength in the
behind the taping machine. The conductor comes off its cross-machine direction to withstand the tensile force on
reel, passes through the taping machine where it is insu- the outside of a bend that is trying to split the paper.
lated and is taken up on the larger, impregnating reel.
When full, this reel is transported to the impregnating As the number of tapes applied under tension increases,
vessel, in which the paper is dried and impregnated with the compressive force between the lower layers increases
insulating liquid at high temperature and vacuum. and causes the diameter of each layer to be reduced.
When impregnation appears to be complete, as deter- This, in turn, relaxes the tension in the lower tapes in
mined by monitoring the cable’s capacitance, the pres- accordance with their modulus of elasticity and prevents
sure in the vessel is raised to several atmospheres, to the compressive forces from building to infinity. A suc-
ensure complete impregnation, and the tank is slowly cessfully taped cable will have a positive interlayer pres-
cooled to room temperature. Once cool, the impregnat- sure between all tapes and will bend without wrinkling
ing reel is removed from the tank and transported to a or tearing any tapes, or collapsing any tapes into adja-
machine that applies the skid wire to the cable. The fin- cent butt spaces. This requires a careful balance of
ished cable is taken up on a shipping reel and moved to paper characteristics and taping tensions. Each time a
the test department for full-reel tests. taping line is set up for a new production run, a dry
bend test must be conducted before starting production.
The process described above is called “mass impregna- This test consists of bending the cable to the impregnat-
tion,” because the cable is dried and impregnated in a ing reel diameter and carefully dissecting a few feet of
tank before its outer covering is applied. This is the pre- insulation to check for proper tape sliding and butt
ferred process when the impregnant is a high-viscosity space distribution. Fine-tuning the tape tensions and
fluid, because these fluids must be at elevated tempera- positions is done at this time. The setup and dry bend
ture to be pumped from a storage vessel to the impreg- tests consume a few hundred feet (about 100 m) of con-
nating tank and back to the storage vessel. At or near ductor and require considerable time.
room temperature, the impregnant is too thick to drain
out of the paper before the cable is covered. High-voltage pipe-type cables require from 50 to 200
tapes to build the insulation wall. This requires many
High-voltage cables are made in taping machines that taping heads in tandem. It is customary to reverse the
contain many taping heads installed in tandem. Each direction of rotation of each successive head to balance
head is a rotating frame designed to apply a fixed num- the torque being applied to the core. Reversing direction
ber of tapes to the core in the same direction. The paper also eliminates the necessity of direct coupling between
tape rolls are mounted in the head with a series of guide successive heads to ensure proper tape overlay. With
rollers that transfer each tape to the core at a controlled opposite rotation, the speed of each head (relative to
angle and position, and with a controlled tension as the core movement) can be different. This allows the use of
head rotates around the core. Figure 5-7 shows a cable-
taping line and the details of an individual taping head.
The head design fixes the number of rolls per head, and
the maximum roll diameter that can be mounted. Com-
mon machines use 10-, 12-, or 16-roll heads, but heads
holding up to 40 rolls have been used successfully. As
the head rotates, each tape is wrapped onto the core.
The core must move in synchronism with the head rota-
tion to ensure a uniform application of the tape with an
open butt space between successive turns. Figure 5-7 Cable-taping line and taping head (courtesy of
Prysmian Cables & Systems).
5-11
different tape widths, increasing as the diameter Priaroggia 1960a). When the paper characteristics are
increases, which has become standard practice. It is also known, the interlayer pressure profile can be calculated
customary to increase tape thickness as the diameter for an assumed distribution of taping tensions. Upper
increases to reduce the number of tapes required to and lower limits for successful sliding can also be
build the wall thickness. Tape thickness establishes the calculated. Taping tensions are selected that result in an
butt gap thickness, which can limit the dielectric interlayer pressure profile that lies within the limits.
strength of the cable. Increasing butt gap thickness as Each manufacturer must apply considerable experience
the diameter increases takes advantage of the decreasing to the problem of establishing the best taping tensions.
radial voltage stress in the outer portion of the wall. The capabilities of the machine, the properties of the
Most cable engineers believe that graded constructions, paper, the changes that occur on impregnation, and the
where both the tape thickness and width are increased intended service of the cable must be considered.
with diameter, result in the best mechanical and electri- S e r i o u s p ro b l e m s c a n d e v e l o p d u r i n g t h e
cal performance. manufacturing process. Cables that develop visible
flaws on the impregnating reel or in the skid wire
An added complication is the fact that the paper’s machine bring the entire manufacturing line to a halt.
mechanical characteristics change when the paper is Although single-point damage can usually be repaired,
dried. Most notably, the tapes shrink, and their fric- visible buckling, creasing, or other indications of
tional interaction also changes when impregnated. improper taping usually result in scrapping the entire
Thus, the cable structure changes considerably in the length, and may involve redesigning and requalifying
impregnating tank. In its final form, there must still be the taping structure.
positive interlayer pressure throughout the insulation
wall, and the tapes must still slide to avoid collapse, 5.4.5 Insulation Electrical Properties
tearing, or wrinkling. In general, impregnation The electrical behavior of the dielectric insulation of a
improves the mechanical stability of the structure, which metallic conductor can be fully characterized by the fol-
is why shipping reels can be reasonable in size and lowing four parameters:
installation bending can be more severe than on cable in
the dry state. • Electrical capacitance
• Dielectric loss
The reduction in paper dimensions upon drying causes • Insulation resistance
a major problem on cables with insulation walls greater
than 0.5 in. (1.3 cm). The solution is to predry the paper • Dielectric strength
before it is applied to the cable. Paper in equilibrium General Comments: Kraft Paper
with normal atmospheric conditions, where the relative
humidity is in the 60% range, contains about 5% water Capacitance
by weight. The cable manufacturer receives paper in The capacitance of a cable is proportional to the relative
jumbo rolls, which may weigh up to 1000 lb (450 kg) permittivity, sometimes called SIC (specific inductive
and are about 30 in. (76 cm) wide. The paper is pro- capacity) or dielectric constant, of the dielectric, and
cessed by a slitting machine to make rolls of the correct depends on the geometry of the conductor and shield
width and diameter. By drying the paper at the slitting that defines the boundaries of the dielectric. Paper den-
operation and maintaining the taping machine in a con- sity exerts the most influence on relative permittivity,
trolled, low-humidity environment, it is possible to which increases with increasing density. The values in
lower the moisture content of the paper to about 1.5%. common use include 3.4 for low-density, impregnated-
This improvement, though expensive, is necessary to paper insulations, such as the early 345-kV cables pro-
achieve the performance levels required by kraft-paper duced in the 1960s; 3.5 for medium-density cables in the
EHV cables. LPP tapes are not dried to such low mois- 69 to 161-kV class; and 3.5-3.6 for cables made with
ture levels. higher-density papers, which include HPGF pipe-type
cable to 138 kV and the newer EHV cables made to
The relationships between taping tensions and bending withstand hot-impulse requirements.
characteristics can be analyzed mathematically. The
classic work on this subject is a small textbook by The capacitance of a paper cable presents a convenient
Gazzana-Priaroggia, Occhini, and Palmieri of Prysmian means for monitoring the impregnation process. The
(for merly Pi rel li) published in 1960 (Gazzana- cable’s capacitance is proportional to the relative per-
mittivity of the materials that comprise the dielectric.
5-12
The cellulose fiber may have a relative permittivity in the When the empty space in the paper is filled with the
range of 5 or 6 but only occupies about 60% of the impregnant, the loss is increased due primarily to con-
paper volume. Impregnants have a relative permittivity duction mechanisms in the fluid. Mineral oils are more
between 2.0 and 2.3. When the impregnant fills the active in this respect than synthetic fluids such as poly-
empty space around the cellulose fiber and the open butene and alkylbenzene. Conduction loss can show an
butt spaces, the capacitance of the cable increases by up increase with stress because of the Wien effect, which
to 10%. Continuous monitoring of the cable capacitance predicts that there will be more dissociation at higher
can indicate when the empty space in the cable has been fields. It can also exhibit a decrease with increasing
essentially filled with impregnant. Only then is it safe to stress due to barrier effects caused by the paper fiber,
raise the pressure on the cable to ensure complete which tend to limit the travel of ionic species, a mecha-
impregnation. If the pressure is raised before sufficient nism suggested by Garton (Garton 1941). Improve-
fluid has entered the cable, the insulation wall may be ments in the production and treatment of dielectric
crushed by the imbalance of pressures. The crushing fluids have almost eliminated the effects of impurities in
results in a pressure ridge or crease running axially in cable fluids. The dielectric loss of modern cables is
the paper tapes. When the crushing occurs at the top of essentially that of the dry paper. It shows almost no
the tank, it is due to premature pressurization; when it change with stress and has a characteristic, slightly U-
occurs at the bottom of the tank, the insulation wall is shaped curve with temperature as shown in Figure 5-8.
most likely too soft to withstand the initial pressure
caused by the head in the tank. Water is a major contributor to dielectric losses. In
bound form, it is a polar molecule that causes rotational
The ability to monitor cable capacitance during impreg- loss in an ac field. Unbound water that is free to enter
nation requires that the cable have a conducting outer the impregnant readily dissociates to contribute conduc-
shield. This is provided by one or two carbon-black tion loss. Drying the paper insulation at high tempera-
and/or metallized paper tapes. In the early days of mass ture and vacuum is mandatory for successful operation
impregnation, the metallized tape was perforated to of transmission cables. The actual degree of drying usu-
enhance the drying of the paper and the flow of impreg- ally determines how close the final loss characteristic of
nant radially into the insulation wall. In the 1970s, it a particular cable comes to the theoretical dry paper
was determined that perforated metallized tapes may be level achieved in a lab cell. Drying at temperatures in the
an unnecessary complication for standard impregnants. range of 120-130°C with vacuum levels in the tank
Modern cables may or may not use perforated metal- under 100 microns will produce cable with a final mois-
lized tapes for this function. ture content of its paper on the order of 100 ppm. When
the moisture content is increased to 500 ppm, the cable
Dielectric Loss
dissipation factor begins to show an upturn at room
Dielectric loss in a cable insulation is undesirable,
t e m p e r at u re a n d t e m p e r at u re s ove r 1 0 0 ° C. A t
because it increases the cost of losses and results in a
1000 ppm, these upturns are enhanced, and the
temperature rise of the conductor, which reduces the
midrange level is slightly raised. Higher moisture con-
cable’s ability to carry current and transfer power.
tent causes the U-shape to disappear, and the loss
Dielectric loss is the result of several mechanisms within
the dielectric. In general, three main categories of loss
mechanism exist: ionic conduction, polar rotation, and
gaseous ionization. Pressurized, fluid-filled cables are
expected to have no gaseous ionization at normal oper-
ating stress, so that this mechanism, which limits the
performance of solid-type and gas-filled cables, is not a
factor for fluid-filled cables. The loss inherent to the
dried paper structure is the lower limit of loss achievable
in a completed cable. This loss is the result of the ionic
conduction and polar rotation mechanisms. Ionic con-
duction loss results from the dissociation of residual
contaminants and the small amount of water not
removed by drying, while the molecular structure of the
alpha cellulose, hemicellulose, and lignin contribute
varying degrees of polar loss mechanisms, mostly in the Figure 5-8 Dissipation factor vs. temperature for typical
hemicellulose and lignin. impregnated paper.
5-13
increases with temperature over the entire range of Paper density and purity establish the basic level of loss
interest. Figure 5-9 shows dissipation factor vs. moisture obtainable in the finished cable. Other things being
content for a typical impregnated paper. equal, loss increases with increasing density. The quality
of the paper also plays an important role. The introduc-
tion of deionized water in the paper mill was an impor-
tant step in improving paper purity. This was done by
running the paper machine with a closed water system
that was equipped with a deionizing electrostatic filter.
The evolution of cable technology, which includes

improvements in taping, drying, and impregnation, and
the papers and fluids used by the industry, is illustrated
in Figure 5-10, which plots the maximum cable dissipa-
tion factor allowed by the AEIC cable specifications for
paper-insulated, lead-sheathed (PILC) cables, SCLF
cables, and HPFF pipe-type cables against time. A log
scale is used for dissipation factor to improve readabil-
ity. This graph clearly shows the distinction between
Figure 5-9 Dissipation factor vs. moisture content for
typical impregnated paper.
Figure 5-10 Comparison of dissipation factor between low-voltage PILC cables and high-voltage pressurized cables.
(Note: This figure uses the abbreviation SCLF [self-contained liquid-filled], which is termed SCFF [self-contained
fluid-filled] in this reference book.)
5-14
low-voltage PILC cables and high-voltage pressurized the butt space, may use very thin tapes near the conduc-
cables used for power transmission. tor, 2-2.5 mils (0.05-0.06 mm) thick. These inner tapes
are manufactured to have very high density (super-
Insulation Resistance
dense) and very high air resistance (supercalendered) to
Insulation resistance is a parameter of interest for dc
minimize the damage that can be caused by ionization.
applications and dc testing of ac cables. Ideal dielectrics
Table 5-1 summarizes the important electrical properties
would have infinite insulation resistance. This is
of impregnated-paper insulation. Several references
achieved in vacuum and gaseous dielectrics; very pure
(McAvoy and Waldron 1971; Constanitov 1964; Tanaka
extruded polymeric insulations come close to this ideal,
and Greenwood 1983; Salvage and Gibbons 1960; Gaz-
but impregnated-paper dielectrics have finite insulation
zana-Priaroggia and Palandri 1955; Hall and Kelk 1956;
resistance due to inherent conduction mechanisms asso-
Dieterle 1960) provide relevant background information
ciated with the cellulose chain structure. The fluid
on dielectric strength of cable papers.
impregnant provides the medium to support additional
electrolytic activity. Laminated Paper Polypropylene Electrical
Properties
Dielectric Strength
The electrical properties of LPP insulation are summa-
Dielectric strength is the parameter that determines the
rized in Table 5-2. The primary characteristics include
physical dimensions of a cable. The primary function of
dissipation factor, relative permittivity, ac breakdown,
a cable’s insulation is to support the voltage applied
and impulse breakdown. Although a single-tape mea-
between the conductor and the grounded shield. This
surement of dissipation factor is included for basic ref-
voltage is the nominal phase-to-ground operating volt-
erence, all properties have been evaluated on either full-
age augmented by occasional surges caused by switching
size or miniature cable constructions.
operations and faults on the power system, and over-
voltages due to lightning strikes on connected overhead
lines and other exposed electrical equipment. Table 5-1 Electrical Properties of Impregnated-Paper
Insulation
The dielectric strength of cables is affected by several Characteristic Typical Value Sample
parameters under the control of the paper and cable Dissipation factor – % (tan δ) 0.23 Cable
manufacturers. The controlling test is the basic insula- Ac step-breakdown volts/mil
1300 (50) Miniature cable
tion level (BIL) impulse test; ac tests seldom result in (kV/mm)
breakdown. Cables are designed to have a strength in Impulse step-breakdown
3300 (130) Miniature cable
excess of the required BIL. Density and air resistance are volts/mil (kV/mm)
two paper properties that are generally agreed to influ- Relative permittivity 3.5 Cable
ence surge strength, with higher values of either property Dielectric loss factor (tan δ x rel-
0.0081 Cable
ative permittivity)
increasing strength. Paper thickness also contributes to
strength for two reasons: thinner papers have higher
strength, other things being equal; and paper thickness Table 5-2 Electrical Properties of LPP Insulation
establishes butt gap thickness, which may inversely affect
Temp. Typical
surge strength. Increased density is accompanied by Characteristic
(°C) Value Sample
increased dielectric loss, so one must compromise on Dissipation Factor—% Single LPP
80 0.07
density. Air resistance of the paper is achieved by slower (tan δ) tape
running speeds on the paper machine, increased beating Dissipation factor—%
80 0.075
345-kV pipe-
of pulp, and more calendering in the finishing section of (tan δ) type cable
the paper machine, all of which raise the cost of the Ac step-breakdown
105 1300 (50) Model cable
volts/mil (kV/mm)
paper. Paper thickness establishes the number of tapes
Impulse step-breakdown
required to build the insulation wall. There are pressures 105 3800 (150) Model cable
volts/mil (kV/mm)
on the manufacturer to use as few tapes as possible and Impulse step-breakdown 345-kV pipe-
90 3300 (130)
as few thicknesses as possible to simplify material han- volts/mil (kV/mm) type cable
dling. High-voltage, pipe-type cables use only two thick- Relative permittivity 80 2.8
345-kV pipe-
nesses of paper, typically 5 mils (0.13 mm) and 6.5 mils type cable
(0.17 mm). EHV cables may use three or four thick- Dielectric loss factor
345-kV pipe-
(tan δ x relative permittiv- 80 0.0021
nesses, starting with 4 mils (0.10 mm) at the conductor type cable
ity)
and going to 8 mils (0.20 mm) on the outside of the
cable. HPGF cables, which have to operate with gas in
5-15
5.4.6 Physical Properties impregnant. For example, paper impregnated with min-
eral oil has a resistivity of about 500C°-cm/W, but the
Kraft Paper same paper impregnated with polybutene has a resistiv-
The specific design of a cable—which involves the selec- ity of 600C°-cm/W. Higher density and hardness result
tion of paper quality, tape thicknesses, tape widths, and in lower resistivity. Resistivity is best determined by
tape tensions—is the responsibility of the cable manu- direct measurement on a piece of completed cable. It is
facturer, who determines the balance of design parame- necessary to energize the cable with dc current to estab-
ters to satisfy the functional requirements of the cable lish a known heat source at the conductor. The resistiv-
based on experience. One aspect of a cable that usually ity is then proportional to the temperature rise of the
varies from one manufacturer to the next is the per- conductor relative to the shield. Resistivity is a trade-off
ceived hardness or firmness of the insulation. The ten- parameter because of its interaction with dielectric loss,
dency to correlate cable hardness with impulse strength relative permittivity, and dielectric strength.
is generally supported by test data. A hard cable has a
higher density than a soft cable, which gives it greater High-voltage HPFF cables benefit from low resistivity,
strength, but also higher loss. It is unwise, however, to because dielectric losses are low, and the temperature
presume that all cables should be as hard as possible, rise that they cause is minor. EHV cables, on the other
because there are many other factors to be considered. h a n d , h ave h i g h e r di e l e c t ri c l o s s a n d a h i g h e r
A cable’s hardness is determined by the interlayer pres- temperature rise due to dielectric loss. This temperature
sure throughout the insulation wall after drying and ri s e i s p ro p o r t i o n a l t o t h e p ro d u c t o f t h e r m a l
impregnation. It is increased by higher taping tensions resistivity, relative permittivity, and the dissipation
and decreased by the change in moisture content in the fac to r of the in sul ati on , wh ich give s ris e to
drying operation. It is also affected by the average butt optimization of these parameters.
space width, paper density, and elastic modulus. After
taping, the cable must be wound onto the impregnating Laminated Paper Polypropylene
reel without distress, deformation, or uneven sliding of Physical design characteristics of LPP insulation (as
tapes. This requires a careful balance of taping tensions with cellulose paper) are highly important in determin-
and humidity in the taping room. In general, harder ing essential taping parameters used for manufacturing
cables are more difficult to manufacture consistently, the insulated cable core. Since the outer surfaces of the
and have a greater tendency toward localized tape dis- LPP laminate are paper, the same EHV taping line facil-
turbance during handling and subsequent service, due in ities can be utilized in cable manufacturing.
part to the higher stored mechanical energy in the insu-
lation wall. Table 5-3 summarizes the important physi- The ultimate objective of the design procedure is to
cal properties of impregnated-kraft paper insulation. develop a proper combination of paper/polypropylene
laminate ratio and tape sizes, coordinated with appro-
The principal thermal characteristic of paper insulation priate taping tensions and angles, to produce a uniform
is its thermal resistivity. This parameter determines the and firm, insulated cable core, which is stable during
temperature rise of the conductor relative to the shield bending, particularly in the dry-core state.
for a given geometry. The thermal resistivity of HPFF
insulations lies between 500 and 600C°-cm/W. It is The basic and essential physical characteristics that
affected by paper density, cable hardness, and type of describe LPP insulation are presented in Table 5-4.
Basically, LPP insulation (as cellulose paper) is limited

Table 5-3 Physical Characteristics of Impregnated-Kraft
thermally to operating temperatures in the region below
Paper Insulation
120-130°C. Specified maximum continuous ratings are
Characteristic Typical Value
limited to the 80-90°C range. The paper and polypropy-
Tensile modulus a
lene in laminates are well matched thermally.
E1(MD) psi x 10-3(MPa) 1800 (12,500)
E2(CD) psi x 10-3 (MPa) 670 (4,600)
Compressive modulus—E3 psi x 10-3 (MPa) @ It has been well established that cellulose paper
3500 (24,000) insulations afford marginal stability at temperatures
1 kg/cm2
Thermal resistivity with polybutene ρi : C°- 600

above 120-130°C, since they exhibit increasing thermal
degradation in physical properties over time beyond
cm/W
these limits (Gazzana-Priaroggia 1960b). The industry
a. The physical moduli, as well as tensile strength, control the recommends limits of 100-105°C for long life. Since
physical stability of the taped cable core during manufac-
melting temperatures of 160-170°C for polypropylene
ture and throughout service life.
materials are cited, thermal ratings for LPP-insulated
5-16
cable systems are at least equivalent to those for paper- ing surge or operating voltage may then trigger ioniza-
insulated systems. tion discharge, creating an incipient flaw (carbonized
track), and leading to eventual failure.
A representative section of the Waltz Mill 345-kV LPP
cable was thermally evaluated in the laboratory (Allam Breakdown mechanisms for impregnated laminar insu-
et al.1988). An average thermal resistivity value, ρi of lations, such as LPP or paper, essentially involve a com-
575 C°-cm/W, was measured, and a working “design” mon process. However, the central polypropylene film
parameter for ρi of 600 C°-cm/W was established for element in LPP tends to enhance its dielectric strength
use in determining ampacities on LPP-insulated cables. compared to paper, thereby permitting a significant
reduction in insulation-wall thickness. In addition, the
5.4.7 Breakdown Mechanisms close match in relative permittivity between the film ele-
Under normal operating conditions, including proper ment and the fluid impregnant reduces stress concentra-
temperatures and pressures, HPFF pipe-type transmis- tion in butt-space channels.
sion cables (employing either LPP or paper insulation)
exhibit excellent dielectric strength under ac voltage and 5.4.8 Impregnating Liquid
impulse or switching-surge duty. As with all cable sys- The choice and application of impregnating fluid are
tems, surge arresters provide system protection from critical to the manufacturing and performance of pipe-
excessive transient-voltage excursions. Breakdown in type cables.
pipe-type cable is possible under abnormal operation,
such as loss of either system pressure or surge-arrester Mass-impregnated cables require a moderately viscous
protection, or possibly as a result of thermomechanical impregnant that will not drain excessively after leaving
aging and degradation in service. A lightning or switch- the impregnating tank. In general, the highest viscosity
fluid that can be readily filtered, degassed, and pumped
Table 5-4 Physical Characteristics of LPP Insulation between the impregnating tanks and the storage tanks
Temperature is chosen.
Characteristic (°C) Typical Value
Tensile modulusa Pressurized fluid-filled cables overcome the shortcoming
E1 (machine direction) psi x 10-3
25 1100 (7,600) of solid-type cables by eliminating voids in the insula-
(MPa) tion, which ionize under stress and cause damage to
E2 (cross-machine direction) psi both fluid and paper.
25 650 (4,480)
x 10-3 (MPa)
Compressive modulus—E3 psi x
10-3 (MPa) @ 1 kg/cm2
25 800 (5,515) Cables pressurized with gas are designed to tolerate
Polymer ratio—%b 25 45
voids by keeping the pressure in the voids high enough
Thermal resistivity with poly-
to prevent serious ionization. The requirements of the
impregnant for gas-filled cables are similar to nondrain-
butenec ρi : C°-cm/watt
85 600
ing solid-type cables. The impregnant must be highly
Adhesion—peel strength in poly-
25 0.40 (0.071) viscous to keep it in place during handling in the plant
butened lb/in. (kg/cm) width and throughout its service life. In self-contained cables
Cable freeze stabilitye –100 Stable with gas channels, the impregnant must not be allowed
Coefficient of friction (CD) (cross- to clog the channels and cut off the pressure source. In
25 0.35 (Dry)
machine direction)f
high-pressure gas-filled pipe-type systems, a thick
Tendency to swell in oilg 90 Minimal
impregnant is desired to minimize the tendency of the
a. The physical moduli, as well as tensile strength, control the impregnant to drain out of the insulation into the pipe,
physical stability of the taped cable core during manufac- a phenomenon that is accelerated by high operating
ture and throughout service life.
temperature and elevation differences.
b. Polymer ratio influences the electrical properties.
c. Thermal resistivity strongly influences ampacity ratings.
d. Adhesion of the laminate influences cable The primary function of the impregnant is to fill and
structural stability. wet the individual paper fibers with a dielectrically sta-
e. Freeze stability ensures cable safety in performing hard ble and chemically inert substance. The combination of
freeze system repairs. paper and a suitable impregnant has a dielectric
f. Coefficient of friction influences dry core cable stability in strength much higher than either component alone. The
bending and handling. improvement in the dielectric strength of impregnating
g. Stability to oil swelling is essential for uniform cable flexi-
bility, as well as to ensure efficient oil flow during impreg-
fluid when its volume is filled with paper fiber is
nation and subsequent cable system operation in service. undoubtedly caused by barrier effects, which limit the
5-17
mobility of ionic species and electronic charges exposed surface of the insulation) of pipe-type cables are referred
to the accelerating forces of electrical stress. The to as outer cable layers. These components are
improvement in paper strength when impregnated is employed in pipe-type cable designs to provide one or
likely due to several effects, including the containment more of the following:
of free charges associated with the paper fiber by molec- • Metallic drainage of capacitive charging current
ular interaction with the hydrocarbon, the reduction of
stress enhancement at the surface of the fibers caused by • Adequate short-circuit current-carrying capability
permittivity mismatch, and a quieting of ionic activity in • Protection against external mechanical damage
the space surrounding the fiber caused by the high den- • Protection from moisture ingress into the cable
sity of the impregnant.
• Reduction of impregnant drainage during shipment
The key point is that each paper fiber must be wetted by and storage
the impregnant to achieve the necessary improvement in • Mechanical strength during and after installation
dielectric strength. When properly impregnated, the • Pressurizing enclosure
next level of dielectric weakness of the paper-fluid com-
bination are voids, where charged particles can accumu- Moisture Seal/Outer Shielding Assembly
late sufficient kinetic energy to initiate the destruction The first component applied outside the insulation
of the material surrounding the void. This weakness is shield of a pipe-type cable is referred to as either the
countered by higher pressure (density) in the void, as in moisture seal or the outer shielding assembly.
gas-filled systems, or prevention of void formation, as in
fluid-filled systems. The next level of strength is limited The primary purposes of the moisture seal/outer shield-
by the unavoidable free-liquid volume associated with ing assembly are to provide a semipermeable seal over
butt spaces between successive paper turns. This weak- the insulation structure, so that impregnating fluid can-
ness is aided by keeping the butt gap width as small as not readily drain from the cable insulation, and to pre-
possible and controlling the radial thickness of the gap vent moisture from entering the cable insulation during
in proportion to the local radial stress. The butt space storage, transportation, and installation. The assembly
electrical strength is improved by higher viscosity is purposely made semipermeable to allow fluid to move
impregnant and higher pressure. between the cable insulation and the pipe-filling fluid
and vice versa during load cycling, thereby preventing
Mineral oils served as the source of impregnants and the creation of voids in the insulation due to fluid star-
filling fluids for decades, but have been displaced by vation during the cool-down portion of the load cycle.
synthetic fluids for a number of reasons. Polybutene flu- The secondary purpose of the outer shielding assembly
ids were introduced in the 1950s as a means to lower is to provide mechanical protection to the insulation
dielectric losses for EHV pipe cables. Their use as and insulation shield during handling and installation in
impregnants and pipe-filling fluids has since become the pipe.
common. In the 1970s, mineral oils suffered from short-
ages of crude and from pressures on the environmental The moisture seal/outer shielding assembly consists of
front, particularly in the disposal of waste from the acid one or more layers of metallized polyester tape applied
wash necessary to reduce the loss characteristic to toler- directly over the insulation shield. Polyester is impervi-
able levels. This resulted in the increased use of poly- ous to cable-impregnating fluids. The metallizing con-
butene and blends of polybutene and alkylbenzene. By sists of a thin layer of aluminum, which is bonded to the
the mid 1980s, all pipe-type cables made in North polyester. The tapes are helically intercalated with the
America—both kraft paper and laminated paper metallized side of one tape facing the insulation shield
polypropylene—used polybutene as their impregnant, and the metallized side of the other tape facing away
and pipe-filling fluids were low-viscosity polybutene or from the insulation shield. In this way, electrical conti-
alkylbenzene or blends of polybutene and alkylbenzene. nuity is maintained between the insulation shield and
the subsequently applied metallic coverings.
5.5 CABLE OUTER LAYERS AND FILLING
MEDIUM One factor affecting the drainage of pipe-type cable
impregnant is the diameter of the shipping reel. Drain-
5.5.1 Outer Layers age is a function of the head pressure developed
between the top and bottom sections of a layer of cable.
All cable components external to the insulation shield
For this reason, it is desirable to apply a second set of
layer (the basic semiconducting carbon-black papers or
metallized polyester tapes on large cables that must be
metallized-paper layers in direct contact with the outer
5-18
placed on large shipping reels. The second set is applied wire materials are zinc alloy, bronze, brass, or stainless
with a lay direction opposite to that of the first set. steel depending on the cable weight and installation
Cable for long-duration shipment, especially to tropical conditions. Stainless steel provides the highest resistance
areas, often has a second set of moisture barrier tapes, to abrasion and the lowest coefficient of friction. Chap-
as well. ter 12 describes the effect of the coefficient of friction on
cable-pulling forces.
The second component of the moisture seal/outer
Armoring (Optional)
shielding assembly is a nonmagnetic metal tape applied
In special applications, one or two other components
helically and intercalated with a plain or metallized
may be included in the outer layers of a pipe-type cable.
polyester tape (metallized side facing conductor),
Armoring is employed in cables intended for operation
commonly called the shielding tape. The metal tape is
in vertical configurations such as long riser cables. The
usually copper, tinned copper, zinc alloy, or stainless
armor is clamped inside the pipe and supports the cable
steel. The purpose of the intercalated polyester tape is to
weight. Armoring may also be required in installations
cover the exposed edges of the metal shielding tape and
in which the pipe has sections with a large sloping pro-
increase the effective impedance of the metal shielding
file to prevent the cables from shifting downhill during
layer and, thereby, reduce the cable shield losses due to
load cycling. This downhill shifting can result in the
circulating currents in these components. Zinc alloy is
cable accumulating in the lower regions of the pipe, and
used because of its low cost. Tinned copper has been
more seriously, it can cause damage to joints or adjacent
used in earlier cables rated 345 kV or above, having
cable in these regions by bending them within the joint
stringent dissipation-factor requirements, where the
casing and/or wedging them against the downhill reduc-
possibility of an increase in fluid dissipation factor
ers (EPRI 1988, 1990; Bankoske et al.1978; Chamberlin
could result from contact of bare copper with earlier
1980). Chapter 8 describes the anchor joints at the top
cable impregnants and pipe-filling fluids. Stainless steel
of the slope and the skid joints at the bottom.
has been used because of its high strength. In present-
day practice, most manufacturers use only stainless steel
Armoring generally consists of one or more layers of
and zinc alloy materials. Figure 5-11 shows the moisture
rectangular nonmagnetic stainless-steel strips with
seal/outer shielding assembly on HPFF cable.
rounded edges. The strips are applied helically over the
Skid Wire Assembly outer shield assembly and under the skid wires. The
The next component of the outer layers of a pipe-type width and number of armor strips are chosen so that
cable is the skid wire assembly. The purpose of the skid there is 100% coverage around the core. The thickness
wire assembly is to protect the outer shielding assembly of the armor strips is usually 0.02-0.035 in. (0.5-
and reduce the coefficient of friction between the cable 0.9 mm), and the width is about 0.315 in. (8 mm) for a
and pipe during installation. The skid wires have a typical 345-kV cable. These dimensions vary depending
“D”-shaped cross section. Two wires are generally on the cable’s overall diameter. By nature of its design,
applied with the flat face of the “D” against the cable the armoring and skid wire assembly has a much lower
surface. AEIC specifications require that the axial spac- impedance than the normal outer covering, and this
ing between adjacent turns must be a nominal 1.5 in. may adversely affect the ampacity due to increased
(3.8 cm). The standard dimensions of the skid wires are shield losses.
0.1 x 0.2 in. (2.5 x 5 mm) height and width. Typical skid
Lead Sheath (Optional)
Sometimes a temporary lead sheath is applied to cables
that are expected to be kept in storage for long periods
of time, such as spare cable lengths. The temporary
lead sheath is generally 0.1 in. (2.5 mm) thick, and one
or two fine steel rip wires installed underneath to facili-
tate stripping. It is necessary to pressurize the sealed
length with nitrogen gas at about 5 psig (34 kPa) to
monitor the sheath’s integrity continuously during
long-term storage.
Pipe
The final outer component for pipe-type cable is the
Figure 5-11 HPFF cable showing moisture seal and
outer shielding assembly (courtesy of Prysmian Cables pipe. Three individually insulated and shielded conduc-
& Systems). tors are installed in the pipe. The pipe provides the her-
metic seal and enclosure for keeping moisture out and
5-19
containing the pressurizing fluid (liquid or gas) under 5.5.2 Pipe-Filling Medium
which these cables must operate. The pipe provides
Applications in Power Cables
mechanical protection to the cable as well as a path for
The principal types of liquids historically used in cables
fault current.
in the United States are mineral oils, polybutenes, and
alkylbenzenes. Alkylbenzene liquid has also been
Carbon-steel line pipe is used for practically all multicon-
blended with mineral oil to reduce its tendency to gas
ductor (three cables per pipe) cable installations.
under voltage stress and with polybutene liquid to
Although pipes and pipe systems of various alternative
reduce viscosity. Both polybutene and alkylbenzene liq-
designs—such as fiberglass-reinforced aluminum, stain-
uids and blends of both are being used as make-up flu-
less steel, steel/epoxy laminates, and carbon steel with
ids in HPFF cables (i.e., to top-off the fluid in the
high-permeability liners—have been proposed and stud-
system in the event of leaks or feeder rearrangement).
ied (Eager and Silver 1982; EPRI 1982), none has been
used extensively. Although these proposed systems often Blending of Insulating Liquids
offer savings due to increases in MVA rating versus higher Mineral oils, polybutene, and alkylbenzene liquids of
installation costs, the carbon-steel pipe is most versatile the types commonly used in high-voltage cables can be
with regard to handling, joining, and bending in the field. blended with each other, generally without any unusual
Nonmagnetic stainless-steel pipe is used for cable riser effects.
sections. In rare single-cable-per-pipe installations, non-
magnetic stainless steel and fiberglass-reinforced alumi- Lower-viscosity liquids may be obtained by blending a
num have been used (Grosser 1973). Nonmagnetic very-low-viscosity liquid (often an alkylbenzene) with
stainless steel was also used for one trial installation with the existing pipe-filling fluid to achieve the desired vis-
three cables inside in the 1970s but not used again after cosity. The viscosity of the resultant blend of liquids
that because of cost and other installation difficulties. may be calculated using the following formula:
C = AxB1-x (cSt) 5-1
Carbon-steel pipe per ASTM A523 Specification
(ASTM 2001), which has been coated internally and ⎛C⎞
externally, is employed in pipe-type cable circuits. The log ⎜ ⎟
wall thickness of the pipe is generally 0.250 in. (6.4 mm), X= ⎝ B⎠ 5-2
except for submarine crossings where a 0.375 in. ⎛A⎞
log ⎜ ⎟
(9.5 mm) or even heavier wall thickness is used for ⎝ B⎠
greater ruggedness and long-term reliability. The coat- Where:
ing on the inside of the pipe is a rust-inhibiting epoxy, A = Viscosity of one liquid, cSt.
which protects the inner surface during storage, facili- B = Viscosity of second liquid, cSt.
tates cleaning after installation, and provides a smooth C = Desired or resultant viscosity, cSt.
surface for pulling in the cables. X = Proportion of liquid A, by volume.
1-X = Proportion of liquid B, by volume.
The outer surface of the pipe is covered primarily for
corrosion protection and to provide an insulating cover- The accuracy of this formula is highest when the specific
ing, so that impressed-current cathodic protection can gravities of the two liquids are the same. In practice, the
be applied. In early pipe-type cable installations, the specific gravity of liquids used as cable impregnants and
pipe was coated with Somastic (asphalt mastic) or bitu- filling materials does not differ enough to significantly
mastic coal tar saturated glass tape and felt. The coating affect this calculation.
layer was about 0.5 in. (12.7 mm) thick for Somastic and
about 0.18 in. (4.6 mm) thick for the coal tar coating. Blends can also increase gas absorption, reduce oxida-
Present-day practice is to use pipe coated with layers of tion during aging, improve voltage breakdown strength,
butyl-rubber and high-density polyethylene or adhesive- decrease moisture transmission, and reduce dissipation
bonded extruded polypropylene. These coatings are rel- factor before and after aging. Care must be taken in
atively thin (0.06-0.12 in. [1.5-3.0 mm]) and have excel- blending, because in some cases, adverse effects can
lent resistance to bacterial and chemical attack, result. For example, the mixture may have a higher dissi-
abrasion, and mechanical damage. The pipe for long pation factor than either of the two liquid components.
water crossings using horizontal directional drilling is The properties of mixtures should be evaluated prior to
typically coated with a fusion-bonded epoxy, followed use. Polybutene liquids with different viscosities can be
by a tough polymer-concrete coating. Chapter 10 pro- blended to secure intermediate viscosities.
vides further information on corrosion coatings.
5-20
Types—Mineral Oils, Polybutene, Alkylbenzene, Polybutene Insulating Liquids

and Mixtures Polybutene liquids were the first to replace mineral oils
in high-voltage cables in the United States. Beginning in
Mineral Oils
the 1930s, these liquids were used as impregnants for
Mineral oils were the liquids that were used predomi- HPGF pipe-type cables (Sticher et al. 1948). Subse-
nantly in the United States as impregnants and filling quently, they have been used as impregnants in solid-
oils in transmission cables through the late 1960s type paper-insulated riser cables, in HPFF pipe-type
(Whitehead 1935; Gemant 1933). They were obtained cables both as the impregnant and pipe-filling fluid, and
from crude petroleums, which are a naturally occurring as the filling fluid in SCFF cables. Polybutene liquids
mixture, essentially comprising hydrocarbons removed are synthetic hydrocarbons. They are produced from
from the earth in liquid state. Those suitable for trans- petroleum, natural gases, or byproducts of coal gasifica-
mission cables are often classified as paraffinic, naph- tion. Similar to mineral oils, the main constituents of
thenic, or aromatic, depending on the main constituent polybutene liquids are carbon and hydrogen.
of their distillation residue. The main components of
mineral oil are carbon and hydrogen, together with Generally, in the manufacture of polybutene liquids,
small amounts of sulfur, nitrogen, and oxygen, and trace butylene or iso-butylene gas is separated from other
amounts of vanadium, sodium, nickel, and iron. Car- gases at the refinery and is typically distilled using a cat-
bon and hydrogen are present in paraffinic, naphthenic, alyst and an acid activator. The molecular weight range
and aromatic hydrocarbons; sulfur, nitrogen, and oxy- of typical polybutene liquids is 300 to 3500 AMU. The
gen are present with carbon and hydrogen in the hetero- liquids later are purified by removal of the catalyst, fol-
aromatic compounds. Mineral oil thus is a complex lowed by neutralization of the activator, distillation,
mixture of hydrocarbons with generally a small amount drying, and filtering to remove moisture and impurities
of hetero-aromatic and metallic compounds. to produce a satisfactory insulating liquid.
Since transmission cables operate at high-voltage Typical electrical, chemical, and physical properties of
stresses, the crude oil must be highly refined, so that it polybutene liquids commercially used in cables in the
has good electrical properties, oxidation stability, and United States are given in Table 5-6. Polybutene liquids
low gas-evolving properties. This requires low sulfur may have slightly different properties in practice than
and nitrogen content and a proper aromatic content. those shown in Table 5-6 because the butylene gas used
This is accomplished by successive distillation, filtra- may differ in molecular structure, which can result in
tion, and chemical treatments. The principal aims of the different properties in the final product. The polybutene
refining process are to remove or reduce waxes, sulfur, liquid used as an impregnant for HPGF pipe-type
nitrogen, and oxygen-containing compounds, and to cables was developed expressly for this product.
control the aromatic hydrocarbons. Although mineral
oils show excellent insulating properties after the refin- The polybutene liquid used for HPFF pipe-filling fluid
ing processes, they still possess measurable electrical
has also been used as make-up fluid in existing cable
conductivity and dielectric loss and are subject to rapid
systems. The polybutene liquid used in HPGF pipe-type
deterioration in the presence of small quantities of air
cables is an important component, because it must not
and moisture, especially when catalyzed by contact with
significantly drain from the insulation structure during
cuprous materials.
service. Consequently, an impregnant that has a high
viscosity, even at high temperatures, is necessary.
Typical electrical, chemical, and physical properties of
mineral oils that have been commercially used in trans-
The properties of polybutene liquids inherently are simi-
mission cables in the United States are given in
lar to paraffinic petroleum oils. However, the poly-
Table 5-5. The properties of oils that are used differ in
butene liquids are produced from a substantially pure
practice in some cases from the values shown in
gas, and the molecular weight can be varied over a wide
Table 5-5 because of sources and refining processes. The
range. The final electrical properties depend on the
properties shown for high-pressure fluid-filled (HPFF)
maintenance of purity and the removal of contaminants
cable impregnants are similar to those shown for gas-
introduced by the catalyst and activator, and on the dis-
pressurized cable impregnants.
tillation process.
5-21
Alkylbenzene Liquids tent product than mineral oil with superior gassing
Alkylbenzene liquids were introduced in SCFF cables in properties under electrical stress. These liquids have fre-
the late 1960s as an alternative to mineral oil (IEEE quently been used as HPFF pipe-filling liquids, often
[Japan] 1968). Alkylbenzene liquids were a more consis- blended with polybutenes.
Table 5-5 Typical Properties of Mineral Oils Commonly Used in U.S. Transmission Cables
ASTM D Test HPFF Cable-
HPFF Pipe-Filling Oil
Method Impregnating Oil
Electrical Properties
Dissipation factor, % (100 tan δ)
100°C. 60 Hz
As received 924 0.3 0.6
Aged, max. 1934A 2.0 2.8
Dielectric strength, kV, 60 Hz, 0.1 in. (2.5 mm)
877
spacing at
25°C >30 >30
80°C >30 >30
Dc volume resistivity,
x 1012 ohm-cm, 100°C
As received 1169 60 20
Aged, min. 1934A 0.1 0.5
Chemical Properties
Corrosive sulfur 1275 Noncorrosive
Total sulfur, % 129 0.07 0.09
Neutralization no., mg KOH/g
Aged, max. 1934A 0.06 0.10
Water content, ppm, max. 1533 30 30
Gassing under electrical stress, µl /min., 80°C 2300B –2 –2
Physical Properties
Viscosity, SUS at 88
100°F (37.8°C) 3350 775
210°F (98.9°C) 115 60
Pour point, °C 97 9 –20
Flash point, °C 92 235 193
Color, ASTM number, max.
Aged 1934A 5.0 4.0
Interfacial tension, dyne/cm 971 – 40
Coef. of thermal expansion x 10-4 1903
°F 3.7 3.7
°C 6.7 6.7
Specific heat. 32°F (0°C), 2766
Btu/lb 0.40 0.40
J/gram 0.93 0.93
Thermal conductivity 2717
Btu/hr-ft-°F 0.072 0.072
W/cm-°C 0.00125 0.00125
Specific optical dispersion, x 10-4 1807 113 110
Molecular weight, AMU 2224 450 370
Specific gravity. 15.6°C 1298 0.93 0.93
5-22
Alkylbenzene liquids are synthetic hydrocarbons. They chloride or hydrofluoric acid. The product is then puri-
are produced by first making the appropriate aromatic fied by the removal of the catalyst and acid by-products.
and n-paraffin, or olefin, then reacting the polymer with This is followed by distillation, filtration, and drying to
benzene in the presence of a catalyst such as aluminum produce an acceptable electrical-insulating liquid.
Table 5-6 Typical Properties of Polybutene Dielectric Liquids Commonly Used in U.S. Transmission Cables
ASTM D Test HPFF Cable-Impregnating
HPFF Pipe-Filling Liquid
Method Liquid
100°C, 60 Hz
Aged, max. 1934A 0.20 0.20
Dielectric strength. kV, 60 Hz, 0.1 in. (2.5 mm)
877
spacing at
25°C >35 >35
80°C >35 >35
x 1012 ohm-cm, 100°C
Aged, min. 1934A - 100
Chemical Properties
Inorganic chlorides and sulfates 878 None None
Organically bound chloride, W% 2522 0.004 0.004
Aged. max. 1934A 0.10 0.10
Gassing under electrical stress, µl /min., 80°C 2300B -20 -30

Physical Properties
100°F (37.7°C) 3500 560
210°F (98.9°C) 164 65
Pour point, °C 97 -20 -30
Specific gravity, 15.6°C 1298 0.87 0.85
Color, ASTM number, max
Aged 1934A 1.0 1.0
Coef. of thermal exp.
x 10-4/°F 1903 4.0 4.0
x 10-4/°C 7.2 7.2
Btu/lb 0.48 0.46
J/gram 1.12 1.07
Btu/hr.-ft.-°F 0.062 0.062
W/cm-°C 0.00107 0.00107
Molecular weight. AMU 2224 550 400
5-23
Typical electrical, chemical, and physical properties of liquids. Table 5-7 presents data on two typical liquids,
alkylbenzene liquids commercially used in transmission one having a conventional viscosity for the type of cable
cables are given in Table 5-7. The alkylbenzenes possess represented, and the other a lower viscosity.
low viscosity compared to mineral oils or polybutene
Table 5-7 Typical Properties of Alkylbenzene Cable Liquids

ASTM D Test
HPFF Pipe-Filling Liquid
Method
Low
Regular Viscosity
100°C, 60 Hz
Aged, max. 1934A 0.20 0.20
Dielectric strength. kV, 60 Hz
877
0.1 in. (2.5 mm) spacing at
25°C >35 >35
80°C >35 >35
x 1012 ohm-cm, 100°C
Aged, min. 1934A - -
Chemical Properties
Aged. max. 1934A 0.10 0.10
Inorganic chlorides and sulfates, ppm 878 None None
Organically bound chloride, ppm, max. 2522 40 40
Gassing under electrical stress, µl /min., 80°C 2300B -30 -30

Physical Properties
100°F (37.8°C) 150 98
210°F (98.9°C) 41 37
Pour point, °C 97 -40 -50
Specific gravity, 15.6°C 1298 0.87 0.87
Color, ASTM number, max. 1500
As received 1.0 1.0
Coef. of thermal exp. x 10-4 1903
°F 4.2 4.4
°C 7.6 7.9
Btu/lb 0.45 0.45
J/gram 1.05 1.05
Btu/hr.-ft.-°F 0.06 0.06
W/cm-°C 0.00104 0.00104
5-24
Both branched and linear alkylbenzenes are used for The power dissipated in the liquid that impregnates the
high-voltage cables. In general, the branched type is pre- cable insulation and has an apparent capacitance (C) is
ferred because of higher autoignition temperature, given by the following expression:
greater resistance to oxidation, and biodegradability. W=IE cos θ, (watts) 5-3
The greater availability of linear alkylbenzenes, however, = (Ic +I’a) tan δ (see Figure 5-12)
has resulted in their more extensive use. = ωC E2 tan δ
Electrical Properties Where:
W = power dissipated in the cable-impregnating
Dissipation Factor liquid, watts.
The dissipation factor (commonly called the power fac- ω = 2πf angular frequency, radians.
tor) of impregnating and filling liquids in transmission f = frequency of the applied voltage, Hz.
cables is an extremely important parameter. It must be tan δ = dissipation factor, dimensionless.
low initially to limit dielectric loss, and remain reason-
ably low while the cable is in operation to avoid cable If the liquid had no pure conduction or absorption cur-
failure due to excessive local heating of the insulation. rents, both I’a and I’’a would be zero, and the specimen
While dissipation factor values are dominated by the would act like a gas. The pure conduction current is ionic,
cable paper, the dielectric liquid plays a major role as and the absorption current is due to the movement of
well. ions, dipoles, and space charges. Low-dissipation-factor
liquid is achieved through proper refining and removal of
The dissipation factor is a measure of the losses as a moisture and gas.
consequence of the component of current in the liquid
that is in phase with the applied voltage and, when asso- The dissipation factor of polybutene and alkylbenzene
ciated with cable-insulating liquids, is generally consid- liquids is markedly lower than mineral oils. This is
ered as tan d, as shown in Figure 5-12 (Emanueli 1930). because of their greater purity, which results in lower
In this figure, ionic content, and the lack of dipoles and space charges.
E = applied voltage, kV. The increase in dissipation factor of the synthetic liq-
I = total current, A. uids, when aged in oxygen, is significant but much lower
Ia = absorption current. than in mineral oils. Because the synthetic liquids are
I’a = component due to capacitance. made from pure gas, proper processing can achieve a
I’’a = component due to absorption. low dissipation factor even with low-viscosity liquids
Io = true conductance, current. and even after aging in the presence of oxygen.
Ic = capacitance current. Dielectric Strength
The voltage breakdown strength of liquids must be high
for use in transmission cables. High breakdown voltage
is achieved by proper purification, refinement, and
removal of even small quantities of water.
There are two types of voltage breakdown in liquids

used in power cables—pure electric and thermal. The
former type does not occur often in high-voltage cable
while in service and is independent of pressure and tem-
perature. The latter type of breakdown is more com-
mon. The final stages of thermal breakdown generally
are the liberation of gas, gaseous ionization, cumulative
heating, and, finally, electron avalanche. The original
heating may be due to the pure conduction and absorp-
tion currents. High pressure significantly increases the
breakdown strength, because it suppresses ionization
and reduces the liberation of gas. This is one reason that
HPFF pipe-type cable is successful.
Figure 5-12 Current in typical cable dielectric liquid due to All commercial-grade dielectric liquids, with physical
applied voltage. properties suitable for use in transmission cable possess
approximately the same voltage-breakdown strength
5-25
wh e n m e a s u re d u n d e r s i m i l a r c o n d i t i o n s. T h e liquid by filtration and removing the air and moisture

breakdown strength of unused dielectric liquid depends by degassing and dehydration, the liquid (particularly
principally on the purity achieved by the distillation mineral oil) possesses anomalous conductivity or
and filtering processes at the refinery. The breakdown absorption characteristics because of the existence of
strength is dramatically reduced by small amounts of dipoles and space charges. Many of these factors that
water or particulates and somewhat reduced by the affect conductivity correspondingly affect resistivity,
presence of gas. Aside from these factors, the voltage- both dc and, to a somewhat different degree, ac. The
breakdown stress depends on the spacing between test industry has responded to this wide influence of
electrodes. The unit-breakdown stress for a given variables by relying on the low-voltage-stress dc
insulating liquid increases significantly as the thickness resistivity parameter to characterize liquids for use in
of the film is decreased; this is why thinner paper is transmission cables.
often used near the cable conductor to give thinner butt
Relative Permittivity
gaps. Aside from water and gas, contaminants not
The relative permittivity of mineral oil is approximately
removed during refining or that have unavoidably been
the same for all types, regardless of the source of the
a d d e d t o t h e l i q u i d s u c h a s p ap e r fi b e r s, e t c .
crude. When the oil is degassed and dehydrated, the rel-
significantly reduce voltage-breakdown strength. As the
ative permittivity at 60 Hz is usually in the range of
temperature of the dielectric liquid is increased, its
2.15–2.35. Theoretically, the relative permittivity is
voltage-breakdown strength decreases. The only
equal to the square of the index of refraction. An empir-
physical property of unused mineral oil that has an
ical formula that relates specific gravity (sp gr) of min-
effect on its breakdown strength is its viscosity, and the
eral oil with relative permittivity is:
effect is minimal. The basic classifications of mineral
oil, paraffinic or naphthenic, have little if any effect on relative permittivity = 0.237 + 2.23 x sp gr 5-4
voltage-breakdown strength.
Since the typical specific gravity at 15.6°C of HPFF
The voltage breakdown strength of polybutene and cable impregnating fluid is 0.93, the relative permittivity
alkylbenzene liquids is considerably higher than mineral given by the above formula is equal to 2.31.
oils. The higher voltage-breakdown strength of these
liquids is most likely due to their higher inherent purity The relative permittivity of polybutene liquids at 60 Hz
and low level of contaminants, rather than the basic is approximately the same as that of mineral oils; the
properties of the liquids. range is 2.08-2.30. The slightly lower values observed in
polybutene liquids are due to their low content of
Resistivity dipoles and space charges.
The resistivity of an insulating liquid depends on its
composition and refining process, moisture and gas con- The relative permittivity of alkylbenzene liquids at
tent, metallic contaminants, and other ingredients such 60 Hz is slightly higher than that of mineral oils and
as acids, esters, phenols, and sulfonates. Mineral oils polybutene liquids. The range is 2.1-2.5 depending
have lower resistivity than polybutene and alkylben- upon the number of aromatic rings present. The
zenes. Resistivity decreases with increasing temperatures higher relative permittivity is associated with the mul-
and decreasing viscosity. tiple ring structures.
The resistivity of insulating liquids is subject to the The variations of relative permittivity of the various
Ohm’s law expression only at low-voltage gradients l i q u i d s a re n o t s i g n i fi c a n t i n d e t e r m i n i n g t h e
(Whitehead 1935). The resistivity decreases nonlinearly c ap a c i t a n c e o f t h e c abl e a n d di e l e c t ri c l o s s o f
above a voltage stress of about 50 V/mil (2 kV/mm), impregnated-paper insulation because the relative per-
principally because of ionization. This voltage gradient mittivity of dry paper is generally in the range of 4.5-5.5,
is less than gradients in the butt spaces and other liquid- and this has a greater influence on the impregnated-
filled spaces that exist in transmission cables. Further- paper relative permittivity. There is little change in the
more, the resistivity increases with time immediately relative permittivity of liquids over the temperature
after the application of voltage stress in a nonlinear range encountered by transmission cables.
fashion before reaching a steady value. The resistivity
decreases markedly with temperature so that in trans- Physical Properties
mission cables, the additional complication exists that Viscosity
the resistivity of the liquid near the conductor is lower Viscosity is the most important physical property for
than that of the liquid near the insulation shield. Even liquids used in cables (Gemant 1933; Emanueli 1930).
after removing the contaminants by highly purifying the The optimum viscosity to use depends on the type of
5-26
cable system. For HPFF cables, a relatively viscous liq- Where:

uid is required for the impregnant. The viscosity must υ = viscosity, cSt.
be low enough for the liquid to move through the paper T = absolute temperature, K.
insulation during temperature changes due to loading in A&B = constants for a given liquid.
service, but high enough not to drain from the cable k = a universal constant for values of υ greater
during certain steps in cable manufacture, storage, and than 1.5.
installation. On the other hand, the viscosity of the
pipe-filling liquid for HPFF cable must be low enough Polybutene liquids can be manufactured to a specific
so that pressure can be maintained throughout the viscosity. The commercial range is quite wide, varying
length of the pipe at all times and pressure drop can be from about 115 to 112,000 Saybolt Universal Seconds
limited during rapid circulation. Low-viscosity liquid (SUS) at 100°F (37.8°C). This makes polybutene liquids
reduces transient pressures and permits longer feeds. attractive to the cable industry because, as can be
For HPGF pipe-type cables, a viscous liquid is required observed in Table 5-6, polybutenes have applications in
to prevent excessive drainage from the cable insulation cables requiring a wide range of viscosities, not feasible
into the pipe during normal service. with mineral oils even with additives. Furthermore,
polybutene liquids have a smaller change in viscosity
The viscosity of liquids decreases to a considerable with temperature than mineral oils. They are character-
extent with increasing temperature. The viscosity- ized by tackiness, which prevents excessive draining
te mperat ure relatio nship, or vi scosity index, is from the insulation structure during shipment and
important in cable applications. Viscosities of typical installation of HPFF pipe-type cables and during ser-
cable liquids, as related to temperature or viscosity vice of HPGF pipe-type cables. The viscosities of poly-
index, are given in Figure 5-13. butene liquids render them suitable for impregnation in
cable accessories since they have less tendency to drain
The curves of Figure 5-13 can be obtained if the viscos- at operating temperatures.
ity of a liquid is known at two temperatures. This is true
generally for cable liquids from approximately 5°C Alkylbenzene liquids are characterized by low viscosity.
above the pour point to 5°C below the flash point. The The commercial range used for high-voltage cable is
viscosity scale is based on an empirical equation that approximately 35-550 SUS at 100°F (37.8°C).
relates viscosity to temperature, as follows:
Pour Point
log log (υ + k) = A +B log (T) 5-5 The “pour point” is the lowest temperature at which a
frozen liquid just begins to flow when thawed at a pre-
scribed rate of temperature change. The pour point is an
important parameter for cable liquids, particularly for
filling liquids. For mineral oils, the pour point of the
crude is quite high, particularly for the paraffinic crudes.
The pour point is lowered in temperature by the refining
process when the waxes of the crude are reduced or
removed and further reduced by blending the mineral
oil with other types of liquids such as polybutenes. The
pour point of polybutene and alkylbenzene liquids is
lower than that of mineral oils for the same viscosities.
Flammability
The ease with which a liquid ignites and burns is an
important parameter, particularly for filling liquids in
HPFF and SCFF cables. The degree of flammability of
a liquid can be determined by a number of tests; the one
commonly used in the United States is the flash point.
The “flash point” is the lowest temperature at which
sufficient vapor is produced to ignite momentarily when
Figure 5-13 Viscosities of typical liquids used in U.S. a flame is applied to the liquid under controlled condi-
cables. (Note: This figure uses the abbreviation SCLF tions of temperature, time, and flame size (ASTM D
[self-contained liquid-filled], which is termed SCFF [self-
contained fluid-filled] in this reference book.) 92-02, 2003a). When the temperature of a liquid during
handling or in operation is significantly below that of its
flash point, it is not considered to be a fire hazard.
5-27
Mineral oils are flammable, and their combustion is an that of mineral oils. However, the differences among the
exothermic reaction that promotes fire propagation. cable liquids in these parameters are not significant in
Vapors of petroleum oil explode when mixed in the right cable operation.
proportion with oxygen if ignited by a spark. Such mix-
Specific Optical Dispersion
tures can occur in tank cars, drums, and other contain-
The specific optical dispersion of a mineral oil is the dif-
ers where oil is handled or stored. The higher viscosity
ference between the refractive indices of light of two dif-
and molecular weight oils have higher flash points.
ferent wavelengths (ANSI 1978). It can be used to
determine the amount of unsaturated compounds
The flammability properties of polybutene and alkylben-
present in the oil. Because the dispersion values are
zene liquids vary somewhat from those of mineral oils
essentially the same for naphthenic and paraffinic oils,
used in power cables. The flammability of polybutene liq-
values bear a direct relationship to the amount of aro-
uids can be modified to some extent by narrowing the
matic compounds present in the oil. Typical values for
boiling range during the manufacturing process.
mineral oils are given in Table 5-5. The refractive index
Interfacial Tension varies with the amount of contaminants held in solu-
The tension at the interface between two liquids is tion. It is advantageous for the mineral oil to have a rea-
known as interfacial tension. The interfacial tension of sonably high aromatic component to impart adequate
mineral oils against water is the usual method employed gas-absorbing tendencies, especially if subjected to ionic
(ASTM D 971-99, 1999) to measure this parameter. bombardment in air spaces and voids. This is particu-
This method is very sensitive to polar and ionic contam- larly true in SCFF cable applications. The specific opti-
inants in the oil that increase dissipation factor, reduce cal dispersion characteristic is not important in
resistivity, decrease voltage-breakdown strength, and polybutene and alkylbenzene liquids, because they
decrease resistance to oxidation. absorb gas under conditions of ionic bombardment.
Chemical Properties
The interfacial tension of polybutene and alkylbenzene
liquids generally is not measured in new supplies Sulfur Content
because, in their manufacture, the polar and ionic com- Sulfur in a mineral oil can decompose on heated metal
ponents are not significant in commercial grades used for surfaces, become corrosive, and produce metal sulfides
cables. After being in service, however, these liquids can that can affect the electrical properties of the oil and
pick up contaminants from pipes and system equipment. provide nuclei for discharge, causing incipient gas gener-
ation in the oil (ANSI 1978). Therefore, the sulfur con-
Coefficient of Thermal Expansion, Specific Heat, and
Thermal Conductivity tent of the oil should be maintained at a low level, and
The coefficient of thermal expansion, the specific heat, this is accomplished in the refining process. Electrical
and the thermal conductivity of mineral oils vary with insulating oils require good electrical properties, oxida-
molecular weight and specific gravity. Table 5-5 shows tion stability, and gas-absorbing properties. This
typical values for various applications of mineral oils requires low sulfur and nitrogen content. The sulfur
used in the United States. Although the refining process content of typical mineral oils is shown in Table 5-5.
alters these properties, they are not significantly When tested for corrosive sulfur in accordance with
changed, and more attention is given to obtaining an oil ASTM D 1275-03 (ASTM 2003c), the results should be
with good electrical properties, oxidation stability, and “noncorrosive.” The purpose of this test is to detect
gas-absorbing properties. The specific gravity is gener- objectionable quantities of free sulfur and corrosive sul-
ally lowered during the refining process, indirectly fur compounds.
affecting these properties.
Because polybutene and alkylbenzene liquids are syn-
The above parameters of polybutene and alkylbenzene thesized, their chemical compositions and purity can be
liquids vary with molecular weight and are quite similar controlled more tightly than mineral oils. Consequently,
to mineral oils. Tables 5-6 and 5-7, respectively, show the presence of sulfur in the liquids is not likely. Alkyl-
typical values for the different types used in cables in the benzene, because of its high solvency, may have a ten-
United States. The thermal conductivity is slightly lower dency to contain sulfur indirectly through other
than that of mineral oil; however, the smaller change of contaminants.
viscosity with temperature ensures fluidity over a wider Acidity
temperature range. The specific heat increases with tem- The acidity of mineral oil is due to the presence of trace
perature and is slightly higher than for mineral oils of acid contaminants. The acidity is determined by mea-
the same viscosity. The coefficient of expansion of poly- suring the quantity of a base material (potassium
butene and alkylbenzene liquids is slightly higher than
5-28
hydroxide) required to titrate all acidic constituents tion, generate hydrogen, ethylene, ethane, methane,
present in 1 gram of sample, generally expressed as the carbon monoxide, and other substances of low
neutralization number. Typical maximum allowable molecular weight. These changes have a tendency to
neutralization numbers for mineral oils are given in increase normal ionization by collision in the liquid and
Table 5-5. When oil is aged in the presence of oxygen, thereby increase the dissipation factor and electrical con-
the acidity increases. Acid contaminants, in conjunction ductivity of the liquid, which could lead to ultimate cable
with moisture and solid contaminants, affect the electri- thermal electric failure. The resistance of an insulating
cal properties of fluids and are corrosive to metals. liquid to ionization is increased when it contains
sufficient aromatic compounds. Gaseous ionization in
The acidity of polybutenes must be controlled, because capillary channels and interstices of the fibrous structure
this type of contamination can be introduced in the of the paper also can occur at high-voltage stresses. This
manufacturing process. The presence of hydrogen and is because it is virtually impossible to remove all of the
aluminum chloride introduces chloride compounds and air from the paper during the drying process in the cable
dielectric instability, which cannot be tolerated in manufacturing plant. The gassing properties of typical
impregnants and filling liquids used in transmission mineral oils used in transmission cables are given in
cables. Inorganic chlorides and sulfates should not be Table 5-5. Although numerous test methods have been
present in the liquid. Organically bound chlorides must devised, one commonly used is described in ASTM
be controlled, and the neutralization number must be D 2300 (the “Modified Pirelli Method”) (Pirelli is now
low. The oxidation stability of polybutene is low, and the Prysmian Cables & Systems) (ASTM 2000b). This
liquid oxidizes readily under normal conditions to form method is used to measure the rate at which gas is
acids. The addition of antioxidant and passivating addi- evolved or absorbed by dielectric liquids when subjected
tives reduces the rate of oxidation to an acceptable level to electrical stress of sufficient magnitude to cause
for transmission cables. ionization in the liquid. The test is used as a measure of
stability and as a means to determine if the liquid
The acidity of alkylbenzene liquids must be controlled absorbs gas. Mineral oils generally tend to evolve gas in
similarly to polybutene liquids in the manufacturing this test and are refined to contain aromatics or blended
process. Also, because alkylbenzenes have high solvent with liquids that absorb gas.
power, greater care must be taken when handling this
liquid to avoid contaminants that may introduce acids. Typical gassing properties of polybutene liquids under
electrical stress are shown in Table 5-6. Tests show that
Water Content
polybutenes are good gas absorbers when used at volt-
The contamination of mineral oil by water is a common
age stresses characteristic of transmission cables. The
and natural occurrence. Proper refining and careful han-
gas absorption of polybutene liquids tends to decrease,
dling of new oil generally keep the moisture content
however, with an increase in viscosity.
below 50-100 ppm. Moisture is easier to keep low in
polybutene and alkylbenzene liquids. However, for all
Alkylbenzene liquids absorb gas at a much higher rate
three liquids, it is imperative to protect the liquid from
than mineral oils and at a higher rate than polybutene
moisture ingress and to remove the moisture from the
liquids. This high absorption characteristic exists even at
liquids prior to their use. Moisture increases electrical
high operating and overload cable temperatures. The
conductivity and dissipation factor and decreases break-
use of alkylbenzene in place of mineral oil significantly
down voltage; it also accelerates aging due to oxidation
reduces degradation during cable insulation aging at
and aging of paper, which readily absorbs moisture. The
high temperature and high-voltage stresses.
amount of water in insulating liquids is usually deter-
mined by the Karl Fischer Method (ASTM D 1533-00,
2000a), and typical maximum values allowable for the 5.6 QUALITY CONTROL
new insulating liquids are given in Tables 5-5, 5-6, and
5-7. Impregnating liquids are carefully dehydrated prior 5.6.1 Cable Manufacturing Quality Control
to cable impregnation. In general, quality-control procedures closely follow the
requirements established for producing paper-insulated
Gassing
transmission cables, as described in Section 5.4.2. How-
Gas evolution may occur in the impregnant or filling
ever, since LPP insulation is a more complex dielectric
liquid under the influence of the high-voltage field, such
than cellulose paper, with more sensitive electrical prop-
as occurs in a termination. This can lead to partial
erties, controlling LPP quality assumes a greater impor-
discharges and ionization. Also, chemical changes, such
tance in the overall cable production program.
as polymerization of the dielectric liquid and condensa-
5-29
With regard to LPP laminate quality control, the fol- periods of time. Synthetic liquids are man-made and
lowing basic properties are closely checked on every composed of “tailored molecules”; therefore their char-
production lot: acteristics are easier to control and maintain. Table 5-8
• thickness summarizes the relevant ASTM standards.
• polymer ratio The liquid may be purchased untreated or treated and

• cavities ready for installation in the cable system. When the liq-
• apparent density uid is supplied, treated, and ready for installation (with-
out field treatment), it is pretreated by the supplier and
• surface friction angle shipped in clean and dry tank trucks or rail cars under
• tensile strength slight nitrogen pressure (about 3 psig [21 kPa]) to pro-
• elongation tect it from contamination during shipment.
• dissipation factor When the liquid is not treated by the supplier and
• dielectric breakdown shipped untreated without a protective pressurized
• adhesion (delamination) nitrogen blanket, it should be treated at the job site
prior to being pumped into the system. The treating
consists first of heating the liquid, then filtering it to
Dissipation factor, dielectric breakdown, and adhesion
remove foreign matter that may have accumulated dur-
tests are conducted after the LPP is impregnated with
ing shipment. The liquid is sprayed into a vacuum dehy-
dielectric fluid. In view of the very low dissipation
dration tank to remove moisture and to free dissolved
factor (tan δ), LPP is highly sensitive to contamination
gases. It is then cooled and transferred to the cable sys-
or other deleterious reactions affecting the quality of
tem. Polybutenes, especially, require careful selection
the laminate.
and control of temperature, and vacuum levels to avoid
loss of the more volatile components (the “light ends”).
The laminar materials, which are processed into LPP
Acceptance tests performed in the field are generally
laminate by the tape supplier, meet rigid and compre-
dielectric strength (ASTM Method D 877-02, 2002b),
hensive specifications, both for high-purity kraft paper
dissipation factor (ASTM Method D 924-03, 2003b),
and polypropylene film. Quality control at the LPP fab-
moisture content (ASTM D 1533-00, 2000a), color
ricating plant is initially focused on the physical proper-
(ASTM Method D 1500-02, 2002c), and visual inspec-
ties of the product, so that mechanical features are
tion. The procedures for obtaining liquids for testing are
immediately certified to comply with the cable manufac-
very critical because the liquid may become contami-
turer’s specifications covering LPP laminate production.
nated or, if exposed to an atmosphere of high humidity,
In addition, representative sample sheets from each new
absorb moisture. As a result, the electrical properties
production run are promptly treated, impregnated, and
can be drastically changed. Devices for sampling electri-
checked for dissipation factor, dielectric breakdown,
cal insulating liquids and procedures for their use are
and adhesion, to provide complete approval for accep-
described in ASTM D 923 (ASTM 1997).
tance to cable manufacturing stock. The dissipation fac-
tor of each cable reel and sample length is closely
When the liquid is supplied to the cable system by the
monitored during LPP cable manufacturing as well.
cable manufacturer, it is generally the practice to pro-
vide a liquid free of added oxidation inhibitors. The
5.6.2 Quality Control of Dielectric Liquids
cable manufacturer generally checks properties of the
The desired quality of liquids used in cables and acces- liquid to an extent depending on the past history of the
sories is maintained over time with the aid of test speci- liquid supplier and internal policy. The tests most often
fications that provide acceptable values for electrical, conducted in the field prior to using the liquid are
chemical, and physical properties. These specifications dielectric strength, dissipation factor, moisture content,
and related testing procedures have been prepared by color, and visual inspection. More extensive testing is
cable manufacturers and ASTM Committee D 27 on performed by the cable manufacturer when a change in
Electrical Insulating Liquids and Gases, in conjunction processing is made by the liquid supplier or if a change
with electrical insulating liquid suppliers. Continuity is made in the handling of the liquid.
tests to ensure quality of a given liquid are important
for both mineral oils and synthetic liquids. Since min- When make-up liquid is supplied to the cable system by
eral oils are complex mixtures of different hydrocarbons the manufacturer, it is expected that laboratory tests
and are of natural origin, it is extremely difficult to specified by the purchaser will be conducted on the
obtain consistently uniform characteristics over long
5-30
liquid. These tests may be as specified in Tables 5-5, 5-6, viscosities of any make-up electrical insulating liquids as
or 5-7, as applicable. In addition, specifications to closely as possible to the original insulating liquids in
ensure the quality of make-up electrical insulating the cable system.
liquids for SCFF (ASTM D 1818-90, 1990a) and HPFF
(ASTM D 1819-90, 1990b) systems prepared by ASTM The AEIC has issue d sp ecifi cation CS 31-95,
Committee D 27 are given in Table 5-8. These specifica- “Specifications for Electrically-Insulating Pipe Filling
tions are useful as a guide for new cable installation as Liquids for High Pressure Pipe-Type Cable,” which
well as for liquid make-up purposes. They are designed addresses polybutenes and alkylbenzenes for both static
to ensure continuity of the quality of the liquids used for (no fluid circulation) systems and systems designed for
cable systems. It is recommended to match the fluid circulation.
Table 5-8 Specification Properties to Ensure Quality of Liquids Used in Transmission Cables
ASTM D HPFF
Test Method Pipe-Filling Liquid
100°C, 60 Hz, max. 924 0.60
Dielectric strength, kV, 60 Hz, 0.1 in.
(2.5 mm) spacing
25°C, min. 877 30
Dc volume resistivity, x 1012 ohms-cm
100°C, min. 1169 2.5
Chemical Properties
Inorganic chlorides and sulfates 878 Nil
Neutralization number, initial, max., mg KOH/g 974 0.05
Sulfur, corrosive 1275 Noncorrosive
Total sulfur, max., weight % 129 0.35
Physical Properties
Viscosity, (a)cSt at 88, 45, 2161
40°C (104°F) 140-150
100°C (212°C) 12 max.
Pour point, °C, max. 97 -21
Flash point, °C min. 92 193
Water content, ppm max. (as shipped in drums) 1533 45
Water content, ppm max, (as shipped in tank trucks or rail
1533 30
cars under nitrogen pressure)
Specific gravity, 15.6°C 1298 0.895-0.940
Color, Initial, max. 1500 2
Specific optical dispersion, 1807
25°C, min. 110 x 10-4(b)
113 x 10-4(c)
Aging Tests (after 96 h at 115°C) 1934A
Color, max. 4
Neutralization number, max., mg KOH/g 0.1
Dissipation factor, % (100 tan δ) 2.8
Resistivity, min., x1012 ohm-cm, 100°C 0.5
a. To convert viscosity to SUS, see ASTM D 2161.

b. Acceptable.
c. Desirable.
5-31
5.7 SPECIFICATIONS AND STANDARDS ASTM D 1533-00 (2000). Test Methods for Water in
RELEVANT TO PIPE-TYPE CABLES Insulating Liquids.
Chapter 2 describes the industry organizations that
develop specifications and standards that apply to ASTM D 2300-00 (2000). Test Method for Gassing of
underground transmission cable systems. This section Electrical Insulating Liquids Under Electrical Stress
lists the specifications and standards that are relevant to and Ionization (Modified Pirelli Method).
pipe-type cables.
ASTM A 523-96 (2001). Specification for Plain End
5.7.1 AEIC Specifications Seamless and Electric Resistance-Welded Steel Pipe for
High-Pressure Pipe-Type Cable Circuits.
AEIC CS2-97 (6th Edition—March 1997). Specifica-
tions for Impregnated Paper and Laminated Paper
ASTM D 202-97 (2002). Test Method for Sampling and
Polypropylene Insulated Cable High-Pressure Pipe-
Testing Untreated Paper Used for Electrical Insulation.
Type.
ASTM D 877-02 (2002). Test Method for Dielectric
AEIC CS31-95 (2nd Edition—December 1995). Specifi-
Breakdown Voltage of Insulating Liquids Using Disk
cations for Electrically Insulating Pipe Filling Liquids
Electrodes.
for High-Pressure Pipe-Type Cable.
ASTM D 1500-02 (2002). Test Method for ASTM
AEIC CG1-96 (3 rd Edition—April 1996). Guide for
Color of Petroleum Products (ASTM Color Scale).
Application of AEIC Maximum Insulation Tempera-
tures at the Conductor for Impregnated-Paper-Insulated
ASTM D 92-02 (2003). Test Method for Flash and Fire
Cables.
Points by Cleveland Open Cup Tester.
AEIC CG3-2005 (2nd Edition—January 2005). Guide
ASTM D 924-03 (2003). Test Method for Dissipation
for Installation of Pipe-Type Cable Systems.
Factor (or Power Factor) and Relative Permittivity
(Dielectric Constant) of Electrical Insulating Liquids.
5.7.2 ASTM Standards
ASTM D 1818-90 (1990). Specification for Continuity ASTM D 1275-03 (2003). Test Method for Corrosive
of Quality of Electrical Insulating Oil for Low-Pressure Sulfur in Electrical Insulating Oils.
Cable Systems.
5.7.3 IEEE Standards
ASTM D 1819-90 (1990). Specification for Continuity
IEEE 1406 (1998). Guide to the Use of Gas-In-Fluid
of Quality of Electrical Insulating Oil for High-Pressure
Analysis for Electric Power Cable Systems.
Cable Systems.
IEEE 1425 (2001). Guide for the Evaluation of the
ASTM D 923 (1997). Standard Practice for Sampling
Remaining Life of Impregnated Paper- Insulated Trans-
Electrical Insulating Liquids.
mission Cable Systems.
ASTM D 971-99 (1999). Test Method for Interfacial
Tension of Oil Against Water by the Ring Method.
5-32
REFERENCES ASTM. 2001. ASTM A 523-96. Standard Specification

AEIC. 1966. “Field Research on 345-kV Underground for Plain End Seamless and Electric-Resistance-Welded
Cable Systems.” Joint AEIC-EEI-Manufacturers Steer- Steel Pipe for High-Pressure Pipe-Type Cable Circuits.
ing Committee. IEEE Transactions on Power Apparatus ASTM International. West Conshohocken, PA.
and Systems. Vol. 85. No. 4. April.
ASTM. 2002a. ASTM D 202-97. Test Method for Sam-
Allam, E. M. 1982. Development of Low-Loss 765 kV pling and Testing Untreated Paper Used for Electrical
Pipe-type Cable. January. EPRI Report No. EL-2196. Insulation. ASTM International. West Conshohocken,
PA.
Allam, E. M., J. H. Cooper, and J. F. Shimshock. 1986.
“Development and Long Term Testing of a Low-Loss ASTM. 2002b. ASTM D 877-02. Test Method for
765 kV High Pressure Oil Filled Pipe Cable.” Paper No. Dielectric Breakdown Voltage of Insulating Liquids
21-06. CIGRE. Paris. Using Disk Electrodes. ASTM International. West Con-
shohocken, PA.
Allam, E. M., A. L. McKean, and F. A. Teti. 1988.
“Optimized PPP-Insulated Pipe-type Cable System for ASTM. 2002c. ASTM D 1500-02. Test Method for
the Commercial Voltage Range.” IEEE/PES Transmis- ASTM Color of Petroleum Products (ASTM Color
sion & Distribution Conference. Anaheim, California. Scale). ASTM International. West Conshohocken, PA.
September.
ASTM. 2003a. ASTM D 92-02. Test Method for Flash
ANSI/ASTM. 1978. ANSI/ASTM D 1807. Standard and Fire Points by Cleveland Open Cup Tester. ASTM
Test Methods for Refractive Index and Specific Optical International. West Conshohocken, PA.
Dispersion of Electrical Insulating Liquids. ASTM
International. West Conshohocken, PA. ASTM. 2003b. ASTM D 924-03. Test Method for Dissi-
pation Factor (or Power Factor) and Relative Permittiv-
ASTM. 1990a. ASTM D 1818-90. Specification for ity (Dielectric Constant) of Electrical Insulating
Continuity of Quality of Electrical Insulating Oil for Liquids. ASTM International. West Conshohocken, PA.
Low-Pressure Cable Systems. ASTM International.
West Conshohocken, PA. ASTM. 2003c. ASTM D 1275-03. Test Method for Cor-
rosive Sulfur in Electrical Insulating Oils. ASTM Inter-
ASTM. 1990b. ASTM D 1819-90. Specification for national. West Conshohocken, PA.
Continuity of Quality of Electrical Insulating Oil for
High-Pressure Cable Systems. ASTM International. Bankoske, J. W., et al. 1978. “Mechanical Effects of
West Conshohocken, PA. Load Cycling on Pipe-Type Cable.” IEEE Transactions
on Power Apparatus and Systems. Vol. 97. No. 3.
ASTM. 1997. ASTM D 923. Standard Practice for Sam- May/June.
pling Electrical Insulating Liquids. ASTM Interna-
tional. West Conshohocken, PA. Bartnikas, R., and K. D. Srivastava, eds. 1987. Power
Cable Engineering. Waterloo, Ontario. Sanford Educa-
ASTM. 1999. ASTM D 971-99. Test Method for Inter- tional Press.
facial Tension of Oil Against Water by the Ring
Method. ASTM International. West Conshohocken, Casey, J. P. 1961. Pulp and Paper. Vol. 3. New York.
PA. Interscience.
ASTM. 2000a. ASTM D 1533-00. Test Methods for Chamberlin, D. M. 1980. “Thermomechanical Bending
Water in Insulating Liquids. ASTM International. West Problems on the Northfield Mountain 345 kV Pipe
Conshohocken, PA. Cable.” IEEE Power Engineering Society, Insulated Con-
ductors Committee. April.
ASTM. 2000b. ASTM D 2300-00. Test Method for Gas-
sing of Electrical Insulating Liquids Under Electrical Constantjnov, T. E. 1964. “Dielectric Materials Mea-
Stress and Ionization (Modified Pirelli Method). ASTM surement, and Applications.” lEE Conference. London.
International. West Conshohocken, PA.
Dieterle, W. 1960. “Contribution to the Knowledge of
the Impulse Strength of Insulating Paper.” S.E.I. Bulle-
tin. No. 13. p. 637.
5-33
Eager, G. S., and D. A. Silver. 1972. “Reduction of Transactions on Power Apparatus and Systems. Vol. 83.
Losses in Conductors and Pipe of Pipe-Type Cables.” Pt. 3. February.
IEEE Underground Transmission Conference. Pittsburgh
Grant, J. 1957. A Laboratory Handbook of Pulp and
Eich, E., E. M. Allam, and W. H. Cortelyou. 1976. Paper Manufacture. London. Edward Arnold & Co. p.
Some Theoretical Considerations Affecting Design of 221.
Lapped Plastic Insulation for Superconducting Power
Transmission Cables. Appendixes to December 1976. Grosser, E. H., and E. J. Steeve. 1973. “Isolated Phase
EPRI Report No. EL-269. Pipe-Type Cable in Aluminum-Fiberglass Pipe.” IEEE
Transactions on Power Apparatus and Systems. Vol. 92.
Emanueli, L. 1930. High Voltage Cables. New York. March/April.
John Wiley & Sons, Inc.
Hall, H. C., and E. Kelk. 1956. “Physical Properties and
EPRI. 1982. Determination of AC Conductor and Pipe Impulse Strength of Paper.” Proceedings of the lEE. Vol.
Loss in Pipe-Type Cable Systems. February. EPRI 103A. p. 564.
Report EL-2256.
IEEE (Japan). 1968. “Improved Aging Characteristics
EPRI. 1988. Thermomechanical Bending Effects on of Oil-Filled Cable Dielectrics Due to Alkyl Benzene as
Extra-High Voltage Pipe-Type Cables. July. EPRI Low-Viscosity Impregnant.” IEEE (Japan) Conference
Report EL-5880. Publication. No. 44. September.
EPRI. 1990. Evaluation of Pipe-Type Cable Joint Kang, B. P. 1963. U.S. Patent No. 3,077,514. February
Restraint Systems. March. EPRI Report EL-6760. 12.
Garton, C. G. 1941. “Dielectric Loss in Thin Films of Kelk, E. and I. O. Wilson. 1965. “Constitution and
Insulating Liquids.” Journal of lEE. Vol. 88. Pt. 2. pp. Properties of Paper for High Voltage Dielectrics.” Pro-
103-120. ceedings of the IEEE. Vol. 112. No. 3. pp. 602-612.
Gazzana-Priaroggia, P., and G. Palandri. 1955. Kojima, K. et al. 1965. “The Influence of Carbon Paper
“Research on the Electric Breakdown of Fully Impreg- Screen on the Power Factor vs. Voltage Characteristics
nated Paper Insulation for High Voltage Cables.” Trans- of Oil-Filled Cable.” Sumitomo Electric Technical
actions of the AIEE. Vol. 74. Pt. III. p. 1343. Review. No. 5. January.
Gazzana-Priaroggia, P., et al. 1960a. “A Brief Review of McAvoy, F. M., and R. C. Waldron. 1971. “EEl-Manu-
the Theory of Paper Lapping of a Single-Core High- facturers 500/550 kV Cable Research Project. Cable A-B
Voltage Cable.” Proceedings of the IEE. Monograph no. High Pressure Oil Paper Pipe Type.” IEEE Transactions
390S. p. 25. on Power Apparatus and Systems. Vol. 90. No. 1. p. 205.
Gazzana-Priaroggia, P., et al. 1960b. “The Influence of Prime, J. B. et al. 1981. “Development of a Reduced
Aging on the Characteristics of Oil-Filled Cable Dielec- Wall 138 kV HPOF Pipe-Type Cable and Joints for Re-
tric.” Institution of Electrical Engineers (lEE). Paper Conductoring of Existing 69 kV Lines on the Florida
No. 3348S. November. Power & Light Company System.” IEEE Transactions
on Power Apparatus and Systems. Vol. 100. No. 7. July.
Gazzana-Priaroggia, P., E. Occhini, and N. Palmieri.
1964. “Dielectric Phenomena Related to the Use of Car- Prime, J. B. et al. 1986. “Update on the Development of
bon Black Paper Screening in High-Voltage Fully Reduced Wall 138 kV HPOF Cables and Joints for Re-
Impregnated Paper Cables.” IEEE Transactions on Conductoring of Existing 69 kV Lines on the Florida
Power Apparatus and Systems. Vol. 83. Pt. 3. February. Power & Light Company System.” IEEE Transactions
on Power Delivery. Vol. 1. No. 4. October.
Gemant, A. 1933. Liquid Dielectrics. New York. John
Wiley & Sons, Inc. Roughley, T. H. et al. 1974. “Design and Installation of
a 138 kV, High-Pressure, Gas-Filled Pipe Cable Utiliz-
Gooding, F. H., and R. B. Blodgett. 1964. “Parameters ing Segmental Aluminum Conductors.” IEEE Transac-
Affecting Increase in Dielectric Loss, Caused by Carbon tions on Power Apparatus and Systems. Vol. 93.
Black Paper Screen, for Oil Paper Dielectrics.” IEEE March/April.
5-34
Salvage, B., and J. A. M. Gibbons. 1960. “The Impulse McKean, A. L., and E. I. Merrell. 1960. “Correlation
Strength of Impregnated Paper Dielectrics as Used in Between Impregnant Stability Under Vacuum Dis-
High Voltage Cables.” Proceedings of the lEE. Vol. charge and Life of Solid-Type Oil-Impregnated Paper
107A. p. 405. Cable on Accelerated High Voltage Load Cycle Test.”
AIEE Conference Paper. No. CP 60-394.
Sticher, J., et al. 1948. “High-Pressure Gas-Filled Cable
Impregnated with Extra-High-Viscosity Oil.” AIEE Melchiore, J. J., and I. W. Mills. 1967. “Factors Affecting
Paper no. 49-84. Winter Power Meeting. New York. Stability of Electrical Insulating Oils.” IEEE Transac-
tions on Electrical Insulation. Vol. EI-2. No. 3.
Tanaka, T., and A. Greenwood. 1983. Advanced Power December.
Cable Technology. Boca Raton, Florida. CRC Press.
Meyerhoff, L., and G. S. Eager. 1949. “AC Resistance of
Whitehead, J. B. 1935. Impregnated Paper Insulation. Segmental Cables in Steel Pipe.” AIEE Transactions.
New York. John Wiley & Sons, Inc. Vol. 68.
ADDITIONAL READING Race, H. H. 1931. “Some Electrical Characteristics of

Ball, E. H., and G. Maschio. 1968. “The AC Resistance Cable Oils.” AIEE Transactions. No. 5.
of Segmental Conductors as Used in Power Cables.”
IEEE Transactions on Power Apparatus and Systems. Reynolds, E. H., and S. C. Clarke. 1959. “Influence of
Vol. 87. Constitution of Oils on Their Gassing Under Electrical
Stress.” Electrochemical Society Symposium on Electri-
Bartlett, S. C. et al. 1957. “Improved Cable Insulating cal Insulation. Philadelphia. May.
Oils.” AIEE Conference Paper no. 57-858. Montreal.
June. Riley, T. N., and R. T. Scott. 1928. “Insulating Oils for
High Voltage Cables.” Journal of Institution of Electrical
Basseckes, H., and M. W. Barnes. 1957. “The Gassing of Engineers. Vol. 66.
Liquid Dielectrics Under Electric Stress—The Influence
of Voltage and Pressure.” AIEE Conference Paper no. Sommerman, G. M. L. 1937. “Properties of Saturants
57-239. for Paper-Insulated Cables.” AIEE Transactions. Vol.
56. May.
Berberich, L. T. 1938. “Influence of Gaseous Discharge
on Hydrocarbon Oils.” Industrial & Engineering Chemis- Tanaka, T., and A. Greenwood. 1983. Advanced Power
try. No. 30. Cable Technology. Vols. I & II. Boca Raton, FL. CRC
Press.
Clark, F. M. 1935. “The Dielectric Strength of Mineral
Oils.” AIEE Transactions. Vol. 54. Thompson, A. W. 1951. “The Application of Petroleum
Oils in the Manufacture of Power Cables.” Proceedings
Clark, F. M. 1940. “Effect of Dissolved Moisture on of the Third World Petroleum Congress. Section VII.
Dielectric Strength of Insulating Liquids.” AIEE Trans-
actions. Vol. 59. No. 8. Watson, P. K. 1956. “The Effect of Hydrostatic Pressure
and Applied Voltage Duration on the Breakdown
Eager, G. S., et al. 1978. “Progress in the Determination Strength of Insulating Oil.” IEEE Winter General
of AC/DC Resistance Ratios of Pipe-Type Cable Sys- Meeting Conference Paper. No. 56-424.
tems.” IEEE Transactions on Power Apparatus and Sys-
tems. Vol. 97. No. 6. November/December. Whitehead, J. B., and F. E. Mauritz. 1937. “Oxidation in
Insulating Oils.” AIEE Transactions. Vol. 56.
Graneau, P. 1979. Underground Power Transmission.
New York. John Wiley & Sons. Wiseman, R. J. 1948. “AC Resistance of Large Size Con-
ductors in Steel Pipe & Conduit.” AIEE Transactions.
Kang, B. P. 1970. “Stability of Electrical Insulating Vol. 67.
Oils.” IEEE Transactions on Electrical Insulation. Vol.
EI-5. No. 2. June.
5-35
CHAPTER 6 Cable Construction: Self-Contained

Fluid-filled
Author: Stephen J. Galloway, Cable Consulting International Ltd.
Reviewer: W. Graham Lawson, Energy Cable Consultants, Inc.
Since their invention in the 1920s, self-contained fluid-filled cables have become the pre-
eminent design for HV and EHV applications worldwide. Today they are used on ac and
dc systems and have become the most highly evolved type of cable with low-loss dielec-
trics and robust constructions. Self-contained fluid-filled cables have been used in instal-
lations with the highest service voltage of 525 kV and in the highest rated circuit at 3800
A. They have been proven to be extremely reliable and versatile, and are in service in a
wide variety of installations such as in long tunnels, in deep shafts, and for use as subma-
rine cables. Today there are several major new projects under construction worldwide.
This chapter gives details of the self-contained fluid-filled cable construction and an
appreciation of the manufacturing path followed by single-core and three-core cables.
The special impregnant is described as well as tests that apply only to fluid-filled cables,
where these are not covered in Chapter 14. The application of this type of cable to dc sys-
tems is outlined, as are the special features applicable to installation in a submarine envi-
ronment (see Chapter 7). Throughout the text, references to supporting information are
made, to provide the reader with a greater depth of knowledge than can be imparted in
this chapter alone.
Stephen J. Galloway is a Senior Consulting Engineer with Cable Con-

sulting International Ltd. He was awarded a Bachelor of Science
honours degree from the University of Southampton, U.K. in 1976. Mr.
Galloway joined BICC Cables Ltd. in 1976 and moved to BICC's
Supertension Cable factory in 1977. There he held a series of senior
posts in the development of cables and accessories in the voltage range
of 33 to 525 kV, and was at one time responsible for installation engi-
neering design and technical specifications. Later he was part of a small
team dedicated to expanding the business into supertension long-length submarine cable
manufacture. In 2001, Mr. Galloway joined Cable Consulting International, where he
carries out advanced finite element analysis and cable engineering design and installation
studies. Mr. Galloway is a Chartered Engineer and a member of the Institution of Engi-
neering and Technology.
6-1
Chapter 6: Cable Construction: Self-Contained Fluid-filled EPRI Underground Transmission Systems Reference Book
6.1 INTRODUCTION 6.2 SYSTEM DEVELOPMENT

The self-contained fluid-filled (SCFF) cable is the most
widely used cable type for HV and EHV cable systems in 6.2.1 SCFF Cable Landmarks
the world and has held this preeminent position since The invention of the first SCFF cable is credited to
the first cables were installed. SCFF cables are in service Luigi Emanueli (1883-1959), who in 1920 designed and
in every developed country in the world. In the U.K. manufactured an SCFF cable to operate at 80 kV. The
system alone there were over 625 miles (1000 km) of field trials on this cable were so successful at 132 kV
SCFF transmission cable in 2004 (ENA 2006). Today that, in 1926, 36 miles of single-core 132-kV cable were
major new installations involving hundreds of miles of ordered by the New York Edison Company (Hunter and
SCFF cable are in progress in the Far East and in the Hazel 1956).
Middle East (IPG 2006).
After this excellent start the SCFF cable saw rapid
Some of the main characteristics of an SCFF cable sys- development, especially in the post-war years when the
tem include the following: requirement for cables for transmission voltages rose
• SCFF cable systems are extremely reliable—many sharply. Development kept pace with demand, and the
first 275-kV fully developed SCFF cable entered opera-
systems have been in service for over 40 years.
tional use in 1959. This was followed by the first 400-kV
• Very large conductor sizes are available, and the rating SCFF cable entering service in 1969. Today SCFF
can be enhanced using special bonding techniques. cables are commonplace at 500 kV, and they are in ser-
• The system may be used in conjunction with a variety vice at 525 kV in North America (see Figure 6-1) (Ray et
of forced-cooling methods. al. 1973; Foxall et al. 1984); these installations have the
• The system operates at the equal highest continuous highest service voltage of any cable type worldwide.
conductor temperature of any cable system.
As well as keeping up with transmission voltages, the
• Higher voltage SCFF cable systems use a low- SCFF cable current-carrying capacity has also risen
dielectric-loss insulation material. with demand. One submarine installation has a
• The cable is of small diameter; hence long lengths 6000-kcmil (3000 mm2) conductor, operates at 500 kV
may be transported and installed. dc, and is rated 2800 MW (Sugata et al. 1996).
• The standard low-pressure system needs no rotating
parts such as pumps in order to operate.
• The system may be installed directly buried, in air, in
ducts, in troughs, in tunnels, in deep shafts, and on
the seabed.
• A full range of accessories is available.
Section 6.2 provides a few details of some of the land-

marks in the development of self-contained fluid-filled
cables and a general description of how the cable works.
Sections 6.3 and 6.4 describe the conductor and conduc-
tor shield. Sections 6.5 and 6.6 review the insulation and
insulation shield. Sections 6.7 through 6.10 cover the
construction processes following insulation—including
laying up, drying and impregnation, metal sheathing,
reinforcement, and jacketing or oversheathing. Sections
6.12 through 6.15 discuss activities related to quality
control—routine tests, shipping, electrical tests, and
hydraulic and fluid quality tests. Section 6.16 describes
cable fluid. Sections 6.17 and 6.18 discuss special appli-
cations, specifically cables for submarine use and cables
for operation on dc. Figure 6-1 SCFF cable for 525-kV operation (courtesy
author).
6-2
EPRI Underground Transmission Systems Reference Book Chapter 6: Cable Construction: Self-Contained Fluid-filled
Self-contained fluid-filled cables have also been installed cable with this type of additional reinforcement. Two
as part of the highest rated transmission line in the other cable systems at 330 kV of similar construction
world at 2640 MVA (3800 A) (Arkell et al. 1977). A are also installed in vertical shafts. These installations
cross-section of the cable is shown in Figure 6-2. For have given over 40 years of trouble-free service (Arkell
this installation, the cables were installed inside rein- and Blake 1968).
forced glass-fiber water-cooling pipes. Extra cooling for
the joints was provided by longitudinal movement of I n 1 9 8 8 , 5 0 0 - k V S C F F c abl e s wi t h 5 0 0 0 - k c m i l
fluid through the cable duct. (2500 mm 2 ) conductors were installed in a route that
crossed ten bridges. The bridges also carried road traffic
The SCFF cable has been installed in conjunction with and railway tracks. The recurrent longitudinal change of
integral pipe-cooling systems, external pipe-cooling sys- the bridge due to thermal movement alone was over
tems, and trough- and weir-cooling systems. Tests have 20 in. (500 mm), and the total longitudinal change due
been successfully conducted on an internally forced- to all factors including the traffic was almost 4 ft 11 in.
cooled design with a duct internal diameter of 2 in. (50 (1500 mm). Accommodation of longitudinal movement
mm). This design is capable of matching a 2200 MVA in a SCFF cable requires careful consideration so that
overhead line rating (Brotherton et al. 1977). Even so, the metal components do not fail through over-bending
most SCFF cable is installed buried direct where the or fatigue, and the insulation remains in good condition
cable is naturally cooled, or in air. without displacement of the insulation tapes. In this
installation, a special link mechanism was developed,
Additionally, SCFF cable has been installed in many tested, and installed, which controls the bending of the
onerous installation conditions. There are several major cables during bridge movement (Minemura and
submarine links in service around the world, many of Maekawa 1989).
which are listed in Chapter 7.
Major developments to every aspect of the cable design
SCFF cable has also been installed and given satisfac- have taken place. Some major developments so far have
tory service in the vertical and inclined, deep shafts been:
often found in hydroelectric generating facilities. In one • The development of the Milliken conductor, which
such installation, 275-kV SCFF cables with reinforced
significantly reduces the ac resistance of large con-
lead sheaths are installed in a vertical shaft 1060 ft
ductors.
(323 m) deep and are designed to withstand a pressure
of 425 psi (3000 kPa). In addition to circumferential • The increase of the continuous operating temperature
reinforcement, longitudinal reinforcement is provided of the conductor to 90ºC.
by aluminum wire armor. Figure 6-3 shows a 400-kV • The dielectric losses in the insulation were signifi-
cantly reduced by the adoption in the 1980s of lami-
nated paper polypropylene (LPP) insulation. This
Figure 6-2 400-kV SCFF cable for 3800 A forced-cooled Figure 6-3 400-kV SCFF cable with LPP insulation and
rating (courtesy author). aluminum wire armor for high-pressure operation
(courtesy author).
6-3
material also enables a higher design electrical stress Design studies and experimental tests on cables
to be used, hence reducing insulation thickness. This intended for use at 600 kVdc have shown that the system
results in a smaller cable diameter and facilitates is suitable for dc systems at 1000 kV (Arkell and Gre-
longer reel lengths. gory 1984).
• The development of extruded aluminum sheaths in
the 1950s led to a more robust design, and the number 6.2.2 Description of Design
of instances of fluid leakage dropped dramatically. The self-contained fluid-filled cable is designed on the
• The replacement of taped servings with extruded basis that the insulation is at all times completely
impregnated. Previous designs of taped-paper-insulated
jackets, initially of polyvinyl chloride (PVC), but more
cable were impregnated with a thick viscous material,
recently of polyethylene. This has reduced corrosion,
which gave rise to contraction voids in the body of the
and has been of sufficient electrical quality such that
insulation when the cable was subjected to thermal
advantage may be taken of the increase in transmis-
cycling in service. At the higher voltages, these voids
sion capacity associated with special bonding.
would allow partial discharge to occur and eventually
• The replacement of mineral oil impregnants with lead to cable failure. This early type of cable design was
alkylbenzene cable fluid led to a more environmen- thus limited in terms of operating stress, and this, in
tally friendly cable system. turn, limited the size of the highest voltage practical for
cable manufacture and installation. The SCFF cable
Developments have also taken place on SCFF cables eliminated this limitation.
outside the present voltage range requirements. Cables
have been tested for use at 765 kV (Couderc et al. 1996), In the SCFF cable, the central duct is filled with fluid,
and full-scale tests have been completed on cables for which is allowed to flow through the wires of the con-
use at 1100 kV (Farneti et al. 1984). Figure 6-4 shows ductor into the insulation. The volumetric change in the
the reels of cable at the test laboratory, and a cross sec- fluid due to the heating effects of load currents is com-
tion of the cable tested. The cable operated at a pressure pensated by connecting the duct in the center of the
of 200 psi (1.40 MPa), which enabled the kraft paper conductor to fluid reservoirs or tanks at intervals along
insulation to operate at the high electric stress of the route using special “stop” or feed joints or at the
760 V/mil (30 kV/mm). The program was positively con- ends of the route. A general description of joints and
cluded, and it was demonstrated that a three-phase terminations for SCFF cables is given in Chapter 8. For
forced-cooled 1100-kV underground SCFF cable system a more detailed description of all accessories for SCFF
can be capable of transmitting a continuous load of cables and for the method of determining the size and
3000 MVA during normal operation and a theoretically location of fluid reservoirs, refer to the Electric Cables
continuous emergency overload of 8000 MVA. The Handbook (Moore 1997).
overload capacity cannot be considered as a continuous
load for practical reasons as the external and internal When the cable cools, the fluid is driven back into the
force cooling systems cannot be run continuously with- cable by maintaining the reservoirs at a positive pressure
out maintenance. at all temperatures. The pressure is kept as low as possi-
Figure 6-4 Installation of the test circuit and cross section of 1100-kV SCFF cable (courtesy
Prysmian Cables and Systems).
6-4
ble, consistent with the profile of the route, with the series of dies to reduce the overall diameter of the con-
minimum pressure at 3 psig (20.7 kPa). A low-viscosity ductor. This type of stranded conductor is also used on
impregnant is chosen to ensure that the fluid will flow other types of cable such as XLPE and HPFF.
readily through the duct, especially during heating and
cooling transients and thus maintain the internal Conductors for single-core cables have a central fluid
pressure within the design range. As the fluid flows duct, whose function is to provide a low-impedance
freely throughout the insulation, the cable has some self- path for the fluid impregnant throughout the length of
healing properties, because there are no voids in the the cable.
insulation or joints. If any air is inadvertently included,
this may dissolve into the fluid or may be removed by Small conductors consist of a number of layers of round
flushing fluid through the system. Additionally, as the wires, each layer stranded in the opposite direction to
cable fluid is common with the accessories, it insulates the one beneath, and stranded onto a steel spiral duct.
the interface between them, thus making the SCFF
system very tolerant of variations in accessory design as Larger conductors consist of a number of keystone-
well as variations in the proficiency of the installation shaped copper segments, which form a self-supporting
jointing personnel. The use of a fluid in common with central duct, over which further layers of flat wires or
the accessories also makes the attainment of higher segments are stranded until the required resistance is
insulation stresses feasible. reached. The wires have a transverse indentation at reg-
ular intervals to promote the flow of fluid between the
Figure 6-5 details the component parts of an SCFF central duct and the insulation. In some designs, the
cable with a corrugated aluminum sheath. segments are provided with an additional longitudinal
tongue-and-groove interlocking system to give greater
6.3 CONDUCTORS stability to the conductor. This type of conductor is
sometimes called a “Conci” conductor, and the inter-
Self-contained fluid-filled cables have conductors
locked version is shown in Figure 6-6A.
stranded from copper or aluminum wires.
Large conductors are of the Milliken type. Conductors
The wire is drawn down from rod delivered to the fac-
above about 2000 kcmil (1000 mm 2 ) are of sufficient
tory in coils of typically 0.3 to 0.4 in. (8 to 10 mm) diam-
diameter for the skin and proximity effects to be signifi-
eter. The rod is then pulled through a series of tungsten
cant factors in the calculation of the conductor ac
carbide dies of progressively decreasing diameter until
losses. These effects are reduced by the Milliken conduc-
the required size is reached. The wire is then annealed in
tor, which consists of a number of smaller, individually
line with the wire-drawing machine by passing an elec-
stranded and shaped conductors laid up together to
tric current through it between two pulleys in the
form the overall conductor. For SCFF cables, the num-
annealer. Aluminum wires do not need to be annealed.
ber of individual conductors or segments is usually six.
Alternate segments are lightly insulated using two insu-
Conductors for three-core cables are stranded from cir-
lating paper tapes. The segments between are lapped
cular wires. Very small (up to 190 kcmil [95 mm2]) con-
with two semiconducting carbon papers, which provide
ductors for three-core aluminum cables may be of solid
electrical continuity between the conductor wires and
rod. During stranding, the wires are passed through a
the shield, and also serve to confine any minute particles
that may have been created in the stranding process to
within the individual segment. A six-segment Milliken
conductor is shown in Figure 6-6B.
Figure 6-5 400-kV self-contained fluid-filled cable

(courtesy author). Figure 6-6 Hollow-core Conci and Milliken conductors.
6-5
The Milliken segments are laid up around a steel spiral high temperature and pressure to remove the impurities
duct and bound together with nonmagnetic tapes such such as lignin and resin. The beating of the pulp and the
as tin-bronze or austenitic (300 series) stainless steel. amount of subsequent calendaring control the impor-
The steel duct is present to give stability to the structure tant properties of the insulation paper, such as the thick-
of the conductor under bending. Where it is important ness, density, and impermeability. Paper for fluid-filled
that the duct presents as low an impedance to the flow cables is washed with deionized water to achieve the
of fluid as possible, the duct may be omitted, as in the highest purity and thus the lowest dielectric loss angle.
cable shown in Figure 6-2. The paper is usually of the two-ply type, and thicknesses
vary from 3 mils (75 µm) to about 8 mils (200 µm); den-
Milliken conductors for cables insulated with LPP often sity varies from 3.6 lbs/in3 to 2.5 lbs/in 3 (100 g/cm3 to
have a thicker metal binder applied to withstand the 70g/cm3). The impermeability is a measurement of the
crushing force that may be imparted by any swelling of porosity and can be varied independently of density.
the polypropylene material. Other parameters such as tensile strength and elonga-
tion are important, and have to be controlled, bearing in
6.4 CONDUCTOR SHIELD mind the width of tape used and the required bending
performance of the final product.
The conductor shield consists of a number of carbon or
metallized carbon tapes applied around the conductor
For SCFF cable insulation, the higher the density and
to eliminate concentrations of electrical stress caused by
impermeability of the paper, the higher the electric
the individual wires of the strand. The conductors
strength. However, once the insulation is impregnated,
intended for use on LPP insulated cables may have extra
the gaps between adjacent turns of tape take on a major
carbon tapes to absorb the higher interfacial pressures
significance, because these gaps are filled with the
associated with this type of insulation.
impregnant alone, and the electric strength of the
impregnant reduces with increasing gap. It is thus com-
In some designs of extra high voltage cable, a duplex
mon practice to apply the higher-density, thinner papers
carbon-black paper tape is applied on the outside of the
near to the conductor, where the electrostatic stress is
carbon tapes. The duplex tape consists of an insulating
highest, and the lower–density, thicker papers towards
tape and carbon-black paper tape, which are bonded
the outside of the insulation. Additionally, for satisfac-
together. The duplex tape is applied with the carbon-
tory bending performance, the surface finish of the
black side against the other shielding tapes. It is claimed
tapes is highly important, and the tapes on the inside are
that duplex tapes assist with meeting the dissipation fac-
narrower than those on the outside.
tor (dielectric loss angle) requirements of modern cable
specifications. However, it is recognized in the specifica-
To construct an insulation with satisfactory bending
tions that cables shielded with carbon-black paper will
performance, it is necessary that the interfacial pres-
have a slightly higher rise in dissipation factor with volt-
sures between the layers of tapes are controlled. It is
age than those shielded with metallized tapes (AEIC
possible to calculate the interfacial pressures from the
1993; IEC 1993a).
tension in the tapes. However, the paper material
shrinks when it is dried, which makes the interfacial
6.5 INSULATION pressures change dramatically. To overcome this effect,
The insulation of SCFF cables consists of either a high- the paper is predried before application, which ensures
quality kraft paper or laminated paper-polypropylene, that most of the shrinkage has taken place before appli-
which is applied over the conductor shield in the form of cation to the cable. The final drying then takes place in
helically applied tapes. Successive layers are applied until the drying vessel. Based on the known and stable char-
the specified insulation thickness is reached. The thick- acteristics of the material, the tension in the tapes can be
ness of the insulation is based on the electric strength of optimized so as to give a pressure distribution through-
the insulation under impulse voltage conditions. out the insulation with the maximum resistance to buck-
ling during bending (Heyda 1967; Gazzana–Priaroggia
The kraft paper is manufactured from all-wood pulp, et al. 1961).
mostly from North America and Scandinavia, using the
kraft or sulphate process (from the German word For cables operating above 220 kV, laminated paper
“kraft,” meaning strength). The pulp is digested with polypropylene (LPP) is an alternative to kraft paper.
sodium sulphide and sodium hydroxide (caustic soda) at
6-6
LPP consists of a film of polypropylene with a layer of where increasing dielectric losses of the paper insulation
paper bonded to both sides, as illustrated in Figure 6-7. restrict power rating. Figure 6-8 shows a machine for
the manufacture of LPP laminate. The figure shows the
The lapping process for LPP differs from that for paper two sheets of paper entering the machine from either
because the material swells when in contact with hot side and going through guides towards the base. The
hydrocarbon fluids. A value of between 3% and 7% head of the polypropylene extruder is seen in the center
swelling under constant pressure has been reported of the picture, withdrawn from its position when the
(Endersby et al. 1992). Different manufacturers have machine is running.
different methods of counteracting this effect. One
method is to lap the cable in a manner similar to kraft
paper cable with the objective of achieving the same Table 6-1 Typical Properties of 4 mils (100 µm) Paper and
bending performance during installation. When the LPP Insulation (Moore 1997)
cable is heated in service, the effect of the impregnant is Property LPP Paper
to produce prospective swelling. In this construction, Tensile strength, MN/m2 50 110
the two layers of paper withstand a hoop stress, which Elongation at break, % 2.0 2.5
prevents swelling, and thus the diameter of the insula- Air impermeability G s Infinity 15,000
tion remains unchanged. Many circuits in successful Density, g/cm3 0.9 0.9
operation have been manufactured with this technique. Relative permittivity at 90°C 2.7 3.4
Dielectric loss angle at 90°C 0.0008 0.0023
Another method is to provide space within the cable for Dielectric loss factor at 90°C 0.0021 0.0078
the polypropylene layer to expand in service. This is Impulse strength, kV/mm 160 135
accomplished by either humidifying the LPP laminate Short time ac strength, kV/mm 55 50
or by spraying droplets of water onto the material such
that a raised pattern is produced on the surface. A dis- Note: The reference from which the values in Table 6-1 are
advantage of this method is that interfacial pressures are taken gives slightly different values to those quoted in Tables
reduced when the cable is lapped and dried. 3-9 and 3-11.
Other techniques of reducing prospective swelling have

been investigated and include the use of bidirectionally
orientated polypropylene in the film, and by heating the
material to achieve a crystalline phase change.
Compared with paper, LPP can be operated at a higher

electrical stress, and has a lower dissipation factor and a
lower permittivity (see Table 6-1). These characteristics
make it attractive at the higher operating voltages,
Figure 6-8 Machine for the manufacture of laminated

paper polypropylene (courtesy author).
Figure 6-7 Laminated paper polypropylene (courtesy
author).
6-7
Compared to paper, the benefit of the lower dielectric head is important to avoid the successive alignment of
loss angle and lower permittivity on the current rating is butt spaces in the radial direction. Figure 6-11 shows a
shown in Figure 6-9. At a voltage of 1100 kV, there is high-definition X-ray of fluid-filled paper insulation.
still a useful current capability, and there are significant T h i s s h ow s t h e bu t t s p a c e s, t h e ove rl ap o f t he
improvements down to 200 kV. metallized shielding tape, the copper continuity wires
between shield and metal sheath, and the corrugated
The higher impulse strength of LPP has allowed 500-kV aluminum sheath.
cables operating at a working voltage stress of 463 V/mil
(18.2 kV/mm) to be installed (Endersby et al. 1993a). Precision in the tensioning of each tape is provided
either by a friction braking system, or sometimes for
In the factory, the insulation process begins with predry- EHV cables, by electroservo control, which maintains
ing the rolls of paper. The paper is passed over heated uniformity of tension whether the machine is stopped,
rollers and stored in a low humidity enclosure. The accelerating, or running. Figure 6-12 shows a head of an
moisture content of the rolls of paper when received is insulating machine with electromechanical control.
about 7% by weight, and the predrying process reduces
this to about 1.5%. The tapes are then slit onto narrow 6.6 INSULATION SHIELD
reels or “pads.”
The insulation shield provides a conducting grounded
electrode around the insulation. It is important that the
The insulating operation takes place in a controlled
shield is continuous and smooth to ensure the electric
humidity atmosphere on a multiheaded lapping
field is constrained to the insulation. Gaps in the shield
machine in which successive heads of tapes are applied
will cause localized increase in the electric field, which
in alternate directions, until the specified thickness is
may eventually lead to failure of the cable insulation.
achieved, as shown in Figure 6-10. Each head applies a
number of tapes, usually either 12 or 16, in the opposite
For single-core cables, the insulation shield or screen
direction to its predecessor. A standard 400-kV cable
often mirrors the conductor screen and consists of car-
designed for a working stress of 381 V/mil (15 kV/mm)
bon and metallized carbon paper tapes used in combi-
and with a 1300-kcmil (630-mm2) conductor will need in
nation. Metallized carbon papers are usually either
excess of 150 tapes.
embossed or perforated to facilitate the extraction of
water vapor during the drying process, and to allow flow
Each tape is applied with a small gap, or butt space,
of impregnant during impregnation.
between the turns so that when the cable is bent, the
tapes will slide over each other without wrinkling or
creasing. The relative registration of the tape in each
Figure 6-9 Variation of current rating with operating

voltage for paper and LPP insulation; conductor size 5000
kcmil (2500 mm2) (Moore 1997) (courtesy Blackwell Figure 6-10 Lapping machine (courtesy Prysmian Cables
Science). and Systems).
6-8
Three-core cables go through a laying-up process (see also included to ensure the finished construction has a
Section 6.7) before impregnation and sheathing, and in circular shape. It is possible to include metal sheathed
some cases a more robust shield is applied for this optical fibers into the construction at this point if
design of cable to ensure that the shield goes forward required. Cables with a corrugated aluminum sheath do
undamaged. Each core is provided with a carbon paper not need spiral ducts incorporated into the assembly,
followed by a metal tape, either copper or aluminum, because the aluminum sheath is sufficiently robust to
applied overlapped and intercalated with a paper or car- retain its circular shape during the bending operations
bon paper tape. The metal tape is overlapped to ensure associated with subsequent manufacture and installa-
complete coverage of the insulation beneath, and the tion on site.
intercalated tape allows for outward flow of moisture
vapor during drying and for flow of impregnant. Two techniques for laying up are in use today. The first
is called the “sun and planet” technique. In this method,
6.7 LAYING UP individual bobbins of insulated and screened core,
together with the ducts and fillers, are loaded onto the
Following insulation, the cores for three-core cables
carriage of the machine. The carriage is rotated about its
must be laid up to form the three-core cable, similar to
longitudinal axis as the components are drawn off,
that in Figure 6-16. At this time, ducts and fillers are
Figure 6-11 X-ray of cable insulation and corrugated aluminum sheath (courtesy author).
Figure 6-12 Lapping head with electromechanical control (courtesy author).
6-9
taped together, and wound onto a drum mounted on a ated and heated externally in an attempt to remove any
conventional fixed take-up. moisture that it might have “picked up” awaiting or dur-
ing sheathing. Clean, degasified fluid was then admitted
The second technique is called the “drum-twister” tech- to the cable via one of the end fittings, while vacuum
nique. In this method, the bobbins of core and fillers are was maintained at the other. Once the cable came under
set on fixed pay-off stands on the factory floor, and the pressure, impregnation was complete.
components drawn off, taped together, and taken up
onto a drum, which rotates about its transverse axis as The drying and impregnating process was greatly
well as its longitudinal axis. improved in the 1960s by the adoption of mass impreg-
nation. In this process, the drum of cable is placed into
Both techniques lay the cores together without the vessel and, after drying, the entire vessel is filled with
imparting a twist. This is fundamental, because it the impregnating fluid. The cable is then fed into the
prevents distortion of the conductor and creasing in the lead or aluminum extruder under fluid via a siphon
tapes of the insulation. The cores and fillers are bound tube. The cable is thus sheathed under low pressure of
together with a fabric tape binder. This binder has one fluid, which prevents the insulation from becoming con-
or more copper wires woven into it to provide continuity taminated with moisture. This technique produces insu-
for the charging current between the dielectric shields lation with a slightly better dielectric loss angle than the
and the metal sheath. earlier method. However, cleanliness of the equipment is
paramount if low-dielectric-loss angles are to be main-
6.8 DRYING AND IMPREGNATION tained and so, to facilitate cleaning, the process drums
and impregnation vessel are manufactured from stain-
Following insulating in the case of a single-core cable, or
less steel.
the laying up of a three-core cable, the manufacturing
process continues with loading of the cable into the dry-
For paper insulation, the temperature of impregnation is
ing and impregnating vessel. The cable is then dried by
not critical, and the insulation may be allowed to cool
being heated under vacuum to 120 to 130°C. The heat-
before admitting the impregnant if required. For cables
ing is provided by radiant heat from the vessel walls, and
insulated with LPP, it is common practice to allow the
it may be supplemented by passing a dc current at low
insulation to cool to near-ambient temperature before
voltage through the conductor. This latter technique
admitting the fluid, so that the possibility of the
speeds up the drying process as the cable core is heated
polypropylene swelling in the fluid is kept to a minimum.
from the inside as well as the outside. The dryness of the
cable may be measured directly by checking the capaci-
Where the cable is to be sheathed with extruded alumi-
tance of the insulation. Alternatively, the amount of
num, impregnation may, alternatively, take place after
moisture in the entire vessel may be checked by isolating
sheathing. In this process, the cable is dried in a mass
the vessel from the vacuum pumps and measuring the
impregnation vessel and, before impregnation, is then
pressure rise over a fixed time. Where it is believed that a
removed to a transfer vessel behind the aluminum
cable is properly dried, but is failing the pressure rise
extruder. The transfer vessel is connected via a stainless
test, it may be that the impregnation vessel itself is leak-
steel tube to the extruder, and the vessel and tube are
ing, and allowing small quantities of air to enter. In
evacuated. The cable is then aluminum sheathed under
these cases, an Equilibrium Vapor Pressure (EVP) test
vacuum. Once sheathing is complete, one end of the
may be conducted. This test samples the vacuum in the
cable is opened to vacuum, and clean, degasified fluid
vessel and compares the vapor pressure of the sample to
impregnant is admitted to the other end until the cable
the same vacuum that has had all moisture removed. A
comes under pressure, at which time impregnation is
desiccant such as phosphorus pentoxide may be used for
complete. This process is called “vacuum sheathing,”
this purpose.
and it produces an insulation with a loss angle at least
equal to one that has gone through the mass impregna-
Early processes then removed the reel of dried cable
tion route.
from the vessel and placed it on a turntable or stand on
the factory floor. It was then immediately sheathed with
For details of the impregnant, please refer to Section
lead to prevent moisture ingress. Once sheathed, the
6.16.
cable was placed in another vessel, where it was evacu-
6-10
6.9 METAL SHEATHING length of loose aluminum tube slipped over the exposed
The metallic sheath of an SCFF cable has two main pur- ends of the cable cores themselves.
poses. The first is to retain the fluid impregnant, and the
second is to conduct short-circuit current. The most Additionally, metallic sheaths manufactured from cop-
common types of sheaths for SCFF cables are lead alloy per are available, and stainless steel has been used on an
or corrugated aluminum. Early lead-sheathed cables experimental basis. Both of these are applied corrugated
were sheathed using a ram press. This had the disadvan- and welded longitudinally. The use of stainless steel gen-
tage that “stop marks” were produced in the sheath erally requires an additional copper conductor to carry
when the ram of the press was raised to allow the barrel fault current.
to be recharged with metal. Modern fluid-filled cables
are sheathed with a continuous lead extruder such as 6.10 REINFORCEMENT
that shown in Figure 6-13. Stop marks do not occur Cables sheathed with lead alloy are invariably provided
with these machines. The lead alloy is chosen to mini- with a taped reinforcement to contain the internal fluid
mize creep and fatigue deterioration arising from instal- pressure. This reinforcement is almost always applied
lation configurations along the cable route. under the anticorrosion polymeric jacket, but installa-
tions do exist where the tapes are outside the jacket (Ice-
Where three-core cables are terminated into potheads, it ton 1985).
is necessary to separate the cores of the cable using a tri-
furcating joint. As the conductors of the individual The reinforcement tapes are most commonly either a
cores do not have a central fluid duct, the single-core tin-bronze alloy or austenitic (300 series) stainless steel.
cables (or “tail” cables) between the joint and the O t h e r m at e ri a l s i n c l u d e c a d m i u m c o p p e r a n d
potheads have a specially extruded lead sheath with galvanized steel (three-core cables only). For normal
flutes on the inside to provide for a flow of fluid beneath systems, up to a maximum internal static pressure of 76
the sheath. psi (525 kPa), one layer of tapes is sufficient. For
higher pressures, or where the installation conditions
Continuously extruded aluminum provides a much are considered to be onerous such as for submarine
stiffer sheath than that manufactured from lead alloy, installations, it is common practice to apply a double
and aluminum sheaths are normally corrugated to pro- layer of tapes.
mote flexibility. The corrugations are rolled into the
sheath in a helical manner to provide for fluid flow A fabric bedding tape is provided between the lead alloy
under the sheath. Where special installations dictate, it sheath and the tapes to prevent the edges of the tapes
is possible to apply a smooth sheath with increased from damaging the sheath during subsequent bending.
thickness—for example, for use at the bottom of deep- The reinforcement tapes are applied in pairs and with a
shaft installations. double start. This technique balances the sideways
forces on the cable due to the tension in the tapes and
The metallic sheath for the single-core tail cables of prevents excessive bending during application. The
three-core, aluminum-sheathed cables consists of a short tapes are covered with a coating of bitumen compound
and with a fabric binder tape. The bitumen compound
enhances anticorrosion performance and longitudinal
water blocking. The fabric tape helps keep the tapes cov-
ered with compound during their service life and pro-
vides a smooth surface for the application of the
anticorrosion jacket.
Joints in tin bronze tapes may be soldered or brazed;

joints in stainless steel tapes should be spot welded.
For cables with very high internal pressure (150 psi

[1034 kPa]), it is possible to apply a layer of armor
wires to give longitudinal reinforcement to the lead
sheath. For single-core cables, aluminum armor wires
Figure 6-13 Continuous lead extrusion machine are used in preference to galvanized steel to avoid
(courtesy author). losses due to magnetic hysteresis. Figure 6-3 shows a
6-11
400-kV cable with aluminum wire armoring for longi- mechanical strength and absorb moisture. These charac-
tudinal reinforcement. teristics put these materials at a disadvantage for heavy
transmission-class cables and those intended for use in
Once the reinforcement process is complete, it is good specially bonded systems where the long-term electrical
practice, especially for large three-core cables, to apply a integrity of the jacket is important.
high internal pressure to the cable. This pressure forces
the lead sheath quickly to expand until restrained by the Antitermite additives can be incorporated into polyeth-
reinforcement. The sheath expansion gives a good circu- ylene or PVC.
lar shape to the cable for the application of the poly-
m e ri c j a c ke t o r ove r s h e at h , a n d h e l p s p reve n t Once extruded, one or more lines of text can be
intercrystalline cracks forming in the lead material due embossed or printed onto the jacket. Additionally, a dif-
to slow creep. ferent colored stripe may be coextruded into the jacket
using a “piggy-back” extruder.
Sheaths manufactured from aluminum, copper, or stain-
less steel do not require additional reinforcement. The Following an in-line spark test, a conductive coating is
manufacturing process produces a strong and flexible applied to the outside of the oversheath to facilitate
sheath, which is inherently strong enough to retain the high-voltage quality control testing. These tests are car-
internal pressure. One installation has a cable with an ried out during manufacture and on site.
extruded smooth aluminum sheath without additional
reinforcement and operates at 360 psi (2500 kPa) 6.12 ROUTINE TESTING
(Moore 1997).
Routine tests for fluid-filled cables are specified in
AEIC CS4 (AEIC 1993) and in IEC 60141-1 (IEC
6.11 JACKETING OR OVERSHEATHING 1993a). Other national specifications exist, and these
SCFF cables are provided with a high-grade extruded and the specifications quoted specify the same generic
polymeric jacket or oversheath. In addition to providing electrical tests:
corrosion protection, it forms the insulation required for • Conductor resistance
insulated sheath systems. It, therefore, has to be of the
highest quality and integrity. • Insulation capacitance
• Insulation dielectric loss angle (ionization factor test)
Before the jacket is applied, the metallic sheath is given • High-voltage withstand
a thin, complete coating of bitumen compound. This
coating acts as a corrosion inhibitor and helps prevent • Dc withstand on anticorrosion polymeric jacket
longitudinal water penetration in the event of jacket
damage. Additionally, checks on the dimensions of the cable
components and on the cable fluid are carried out.
The jacketing material in most common use is medium- Where the tests are specific to fluid-filled cables, these
or high-density polyethylene. PVC has seen common use are described in Section 6.14 for electrical tests and in
in the past, and rubber materials such as neoprene have Section 6.15 for hydraulic and fluid tests. An overview
also been used. Polyethylene provides a very robust of Routine Testing may be found in Chapter 14.
oversheath with a relatively low thermal resistivity, and
is the most common material where the cables are direct 6.13 SHIPPING
buried or installed in ducts. PVC is used where cables After tests are complete, a pulling eye is normally fitted
are installed in buildings such as substations or in tun- to one or both ends of the cable, and these are fixed to
nels because it can be formulated with a reduced flame the sides of the reel. Cables for land routes are shipped
propagation performance and with low smoke and fume on robust reels made from steel or timber. Each length is
emission characteristics. Heavily filled jacketing materi- fitted with a pressure-compensating tank to keep the
als that exhibit low-smoke performance under fire con- cable under positive pressure for all expected ambient
ditions and do not emit acidic halogens, although in temperatures. It is good practice to fit a small pressure
common use at lower voltages, have not found favor at gauge to the pipe work so that this aspect can be regu-
the higher voltages, because they tend to have lower larly checked until the cable is installed.
6-12
6.14 ELECTRICAL TESTS APPLICABLE TO Note that, in AEIC CS4, slightly smaller conductor
FLUID-FILLED CABLES sizes are given for three core cables at 69 and 138 kV.
The test voltages are the same whether the cable is of a
6.14.1 High-Voltage Withstand Test single-core or of a three-core design.
This is a routine test and is carried out on every reel
before shipment. For this test, the cable has to with- 6.14.2 Measurement of Dielectric Loss Angle
stand the test voltage for 15 minutes without break- This test, also known as the dissipation factor test or
down. Table 6-2 shows IEC and AEIC values of the tan δ measurement, is a very important test, because it
test voltage. measures the inherent electrical properties of the insula-
tion. Requirements for paper and LPP cables are shown
in Table 6-3. The specifications give values for the
Table 6-2 Routine Test Withstand Voltages
“base” or initial measurement at working voltage and
Cable Sys- IEC 60141-1 AEIC CS4
tem Voltage
also the amount by which the loss angle may rise with
V kV Test Volt- Conductor Size Test voltage. The base value is a function of the characteris-
age kV
kcmil
Volt- tics of the insulation material and the way it has been
mm2 (a) age kV processed. For example, a cable that has not been
60 to 69 (b) 82 4/0 to 4000 110 to 2030 85.5 properly dried will have a relatively high value of base
350 to 750 180 to 380 130.5 loss angle.
110 to 115 (b) 138
800 to 4000 400 to 2030 117.0
350 to 750 180 to 380 135.0
120 -
800 to 4000 400 to 2030 121.5 Once the base value has been measured, the voltage is
500 to 900 250 to 460 149.0 increased, and the rise in loss angle is measured. This
130 -
1000 to 4000 510 to 2030 129.0 test, called the “tan delta difference” or “ionization
132 to 138 (b) 162
550 to 900 280 to 460 151.5 factor” test, checks for ionic impurities in the paper or
1000 to 4000 510 to 2030 136.5 impregnant. Additionally, if air is trapped in the
750 to 900 280 to 460 177.0 insulation, this will cause partial discharge and thus
150 to 161 (b) 184
1000 to 4000 510 to 2030 159.0
high ionization.
1000 to 2000 510 to 1010 228.0
220 to 230 (b) 220
2250 to 4000 1140 to 2030 186.0
275 to 287 275 - - - 6.15 HYDRAULIC AND FLUID QUALITY TESTS
1000 to 2000 510 to 630 310.5
330 to 345 (b) 325
1500 to 4000 760 to 2030 276.0 6.15.1 Expulsion Test
380 to 400 375 - - -
This test is designed to check that the cable is com-
2000 to 2500 1010 to 1270 402.0
500 495
2500 to 4000 1270 to 2030 375.0 pletely impregnated. AEIC CS4 specifies that an Expul-
sion Test shall be carried out on reels of complete cable.
a. Approximate conversion kcmil to square millimeters
Fluid is withdrawn from the pressurized cable until all
b. Where a range of system voltage is shown, the AEIC value
corresponds to the highest value in that range. flow ceases. The volume of fluid and the drop in pres-
Table 6-3 Dielectric Loss Angle Requirements for Paper and LPP Cables at Routine Test (IEC 60141-1)
System Rated Cables Not Screened with Carbon Black Paper Cables Screened with Carbon Black Paper
Voltage, Voltage,
V (kV) Vg (kV) Max Dielectric Loss Angle Max Difference Max Dielectric Loss Angle Max Difference
x 10-4 x 10-4 x 10-4 x 10-4
At Vg At 1.67 Vg Between Vg and At Vg At 1.67 Vg Between Vg and
1.67 Vg 1.67 Vg
Paper LPP Paper LPP Paper LPP Paper LPP Paper LPP Paper LPP
220-230 127 30 14 34 16 5 4 30 14 36 19 7 6
275-287 160 30 14 34 16 5 4 30 14 36 19 7 6
330-345 190 28 14 31 16 4 4 28 14 34 19 7 6
380-400 220 28 14 31 16 4 4 28 14 34 19 7 6
500 290 28 14 31 16 4 4 28 14 34 19 7 6
6-13
sure are measured, and the Expulsion Constant (Ke) is pass value can be more readily achieved for aluminum-
calculated using Equation 6-1. sheathed cables than for lead-sheathed cables due to
the greater inherent stiffness of the aluminum sheath
dV material.
Ke = 6-1
dP × A × L × 51.7 However, this is a searching test and is very sensitive to
Where: pressure variations associated with ambient temperature
dV = Volume of expelled fluid (cm3). changes. It is much less easily measured on completed
dP = difference in pressure before and after installations, where the temperature rise associated with
expelling fluid (psi). the sun coming from behind some clouds during the test
A = cross sectional area of the cable inside the has been known to cause the expected pressure drop
metallic sheath (in2). actually to become a pressure rise. It is, therefore, pru-
L = length of the cable under test (x103 ft). dent to conduct this test when the ambient temperature
is at its most stable, and thus the test is often conducted
The value of Ke may not be greater than 0.065. in the early hours of the morning.
6.15.2 Impregnation Coefficient Test 6.15.3 Flow Test

The Impregnation Coefficient Test is a similar test to the The purpose of this test is to ensure that no abnormal
Expulsion Test and may also be conducted on every reel restriction to fluid flow is present in the cable duct, joints,
of cable before shipment from the factory. In addition, and associated pipework. This is achieved by measuring
the British Standards Institution specification BS 7922 the drop in pressure for a measured flow rate of fluid.
(British Standards Institution 2004) requires that the This value is then compared to a calculated value. The
test is also repeated on completed installations at the flow rate may be calculated from Equation 6-3.
time of commissioning, and Energy Networks Associa-
tion ENA C84/2 (ENA 2000) recommends it whenever Pr 4
the cable is taken out of service or repair work is carried Q= 6-3
out. Like the Expulsion Test, this test is intended to 25.5ηL
check that there is no free air trapped in the insulation Where
or other space contained within the metallic sheath. It Q = Flow rate (l/sec).
differs from the Expulsion Test in that only a relatively P = Total pressure difference on the hydraulic
small volume of fluid is withdrawn, and thus there is no section (including allowance for route pro-
risk of losing all pressure from the cable. Also a more file) (bar).
precise value for the volume of fluid in the system is r = internal radius of the cable duct or feed
used, which allows a tighter pass level to be set. pipe (mm).
η = Viscosity of the fluid at the test tempera-
To conduct the test, the feed tanks are first isolated from ture (centipoise).
the cable. A measured quantity of fluid is then with- L = length of the cable section (m).
drawn from the cable over a fixed time, and the conse-
quent drop in pressure noted. The flow rate varies with the fourth power of the radius
of the duct, and thus the calculation is very sensitive to
The impregnation coefficient (Ki) is then defined as: this parameter. It is, therefore, necessary to split the cal-
culation into as many parts as there are pipe radii (with
dV 1 their associated lengths). Thus one part would be for the
Ki = × 6-2 flow in the cable duct, another for the feed pipework,
V dP
Where: and perhaps another for a length of cable with a differ-
dV = the volume of fluid withdrawn from the ent sized duct.
system (l).
V = the total volume of oil in the installation For three-core cables, the value obtained from Equation
(l). 6-3 is multiplied by three.
dP = the measured drop in pressure (bar).
The measured value rarely correlates precisely with the
The specifications give a pass value for Ki of 4.5 x 10-4 calculated value, and the specifications give no pass
bar -1 for all sizes and constructions of cable. In the level. The figures obtained serve as a guide to ensure
factory where individual reels of cable are tested, the that no undue obstructions are present in the system.
6-14
6.15.4 Residual Gas Pressure Test The fluid to be sampled is sprayed in a fine jet into a
This test takes the Impregnation Coefficient Test one vacuum enclosed by the domed end of a glass tube. The
step further by measuring the amount of air dissolved in fluid falls under gravity to the base of the tube. The tube
the fluid. It is also useful as a tool to check the quality stands in a larger glass outer tube to which it is com-
of the fluid in feed tanks, and the output of fluid treat- mon at the base. The outer tube fills with the fluid from
ment units, which by their nature, cannot be checked the inner tube to a predetermined level set by an over-
using the Impregnation Coefficient Test. flow or drain tube. The space above the fluid in the
outer tube is a vacuum common with the reservoir at
The test finds most use in systems where the electric the base of the apparatus and is thus an adjustable and
stress in the cable and joint insulation are at their measurable vacuum. As the fluid is sprayed into the
highest and, therefore, more prone to deterioration due inner tube, the gas comes out of solution such that the
to contamination of the fluid with air. The SCFF cable pressure in the upper part of the tube is the same as that
system will age naturally mainly due to deterioration of in the fluid at the base. The vacuum in the outer tube is
the insulation material. This is not critical until the varied until the level of fluid in the outer tube is the
pressure of dissolved gas approaches that of the system same as the level in the inner tube. At this equilibrium
itself. When diversionary or remedial work is carried point, the pressure in the two tubes is the same, and this
out, a balance should be struck between the value when can be measured in the outer tube. This pressure is
the system was new and that practicably achievable in known as the RGP of the fluid.
the time available. For guidance, a value of 10 Torr
(mm Hg) should be achievable for new installations at 6.15.5 Fluid Cleanliness Test
system voltages above 200 kV, and a value expected for For cables intended for use at 200 kV and above, it is
such installations that have been in service is 50 Torr good practice to check the quality of the fluid for partic-
(mm Hg). After flushing the installation, values of ulate contamination. The distribution of particle size in
twice that achieved for new installations are acceptable. the fluid is such that it is only necessary to specify the
Values for lower system voltages are less stringent number of particles in the smallest size range to ensure
(ENA 2000). that the overall cleanliness of the fluid is satisfactory.
This is set at 50,000 particles per 100 ml of fluid in the
Different cable installation companies use different 2-5 µm range (ENA C84/2 [ENA 2000]).
designs of specially constructed measuring equipment,
one of which is described here (see Figure 6-14). Other
equipment works on similar principles.
Figure 6-14 Schematic of residual gas pressure (RGP) test apparatus (courtesy author).
6-15
6.15.6 Fluid Electrical Tests system has been installed in vertical shafts as well as
Common electrical tests on the cable fluid are the cable tunnels. Figure 6-15 shows fire-retardant 400-kV
volume resistivity test, the dielectric loss angle test, and SCFF cables installed in a 1968-ft (600-m) deep vertical
the measurement of breakdown strength. AEIC CS4 shaft in a power station in Italy.
requires the latter two tests to be carried out as a routine
test on samples of fluid extracted from the cable on its Alkylbenzene liquids are synthetic hydrocarbons. They
reel before dispatch from the factory. The resistivity and are produced by first making the appropriate aromatic
loss angle tests are carried out at the operating and n-paraffin, or olefin, and then reacting the
temperature of the conductor in a laboratory. The polymer with benzene in the presence of a catalyst such
electric breakdown strength is also used as an as aluminum chloride or hydrofluoric acid. The
acceptance test for drums of fluid delivered to site and is product is then purified by the removal of the catalyst
carried out at ambient temperature in a portable test cell. and acid by-products. This is followed by distillation,
filtration, and drying to produce an acceptable
It is possible to dramatically improve the electrical prop- electrical insulating liquid.
erties of the fluid by treating it with Fuller’s earth or
molecular sieve. An impregnant consisting mainly of dodecylbenzene is
now primarily used for SCFF cables, and either
branched chain or linear dodecylbenzene may be used.
6.16 CABLE FLUID
In general, the linear type is preferred, because it is
Highly refined low-viscosity mineral oils were used as more readily available and is able to be classified as
cable impregnants for many decades; however, alkylben- “readily biodegradable” (Endersby et al. 1993b).
zene materials were introduced into fluid-filled cables in
the late 1960s and 1970s as an alternative. Alkylbenzene
fluids are a synthetic material and are, therefore, per-
ceived as a more consistent product. They have signifi-
cantly superior gassing properties relative to mineral
oils, because their ability to absorb gas under electrical
stress increases with temperature, and they have a high
resistance to oxidation. At present, synthetic alkylates of
the dodecylbenzene type are the preferred impregnant
for SCFF cables, and today dodecylbenzene has effec-
tively completely replaced mineral oil (Moore 1997).
In the 1980s, development work was carried out to find

an alternative impregnant that had a superior perfor-
mance under conditions of fire. The resulting cable sys-
t e m wa s p ri n ci p a lly i nt en de d fo r u s e i n t un n el
installations. In Europe, type approval testing was suc-
cessfully conducted on such an SCFF cable system with
a low-flammability impregnant. The system comprised
standard construction cable and accessories, impreg-
nated with a silicone oil, which had an additive to Figure 6-15 400-kV SCFF fire-retardant cables installed
improve the gassing characteristics (Lanfranconi 1986 in a vertical shaft (courtesy Prysmian Cables and
and US Patent). The type tests were successful, and the Systems).
6-16
Typical electrical, chemical, and physical properties of seabed. It is, however, recommended to bury the cable in
dodecylbenzene fluid commercially used in transmission sections of the route that are prone to damage, espe-
cables are shown in Tables 6-4 and 6-5. cially where the cable crosses busy shipping channels
and at the shore ends.
6.17 SPECIAL CONSTRUCTIONS FOR
SUBMARINE USE The length of a SCFF submarine cable is often limited
by the need to feed the full length of the route with fluid
SCFF cables have seen extensive use on ac and dc
to maintain it under positive pressure at all times. The
submarine systems, and dominate the market at the
most onerous condition is during the cooling transient.
present time. Very large new projects are being installed
This is easier to achieve with three-core cables, because
(IPG 2006).
these designs have three ducts, and thus their hydraulic
impedance is much less than their single-core counter-
SCFF cables intended for submarine use are invariably
parts. The length of single-core cable that can be
armored to provide tensile strength and protection dur-
installed may be considerably extended by increasing
ing laying and to prevent damage when installed on the
the diameter of the central fluid duct. It is also possible
to use a special ultra-low-viscosity impregnant to extend
Table 6-4 Typical Properties of Dodecylbenzene Cable the length of cable that can be installed. Some long-
Impregnating Fluid length submarine installations are equipped with
Property Test Method Value pumping plants to maintain the system under pressure
Density, gm/cm3 at 20°C ASTM D 1298 0.86 at all times.
Kinematic viscosity – centis-
ASTM D 445
tokes
The remainder of the construction of a submarine fluid-
At 40°C 4.0 – 4.5
filled cable is more robust than that intended for use on
At 60°C 2.5 – 3.0
land. The operating stress of the insulation is slightly
At 90°C 2.0 – 2.2
lower. The lead sheath is slightly thicker, and this is rein-
Flash point, °C ASTM D 93 130
forced with a double layer of tapes (Foxall et al. 1984).
Pour point, °C ASTM D 97 -45
The anticorrosion polymeric jacket may be protected
Corrosive sulfur ASTM D 1275 Noncorrosive
with thin metal “anti-teredo” tapes. These tapes are usu-
Neutralization value, mg KOH/g ASTM D 974 0.01
ally made of copper or brass. However, it is the armor
Expansion coefficient, per °C ASTM D 1903 0.00080
layer, or layers, which constitute the major difference
Specific heat capacity, kJ/kgK ASTM D 2766 1.67
between submarine and land cables.
Gassing under electrical stress, ASTM D
-0.02 (min)
ml/min 2300B
Biodegradability (linear dode- Readily biode-
Most submarine cables are armored with a single layer
OECD 301D of round, galvanized steel wires. A single layer of wires
cylbenzene) gradable
allows the cable to be coiled, thus avoiding the need for
specialized cable-laying vessels with turntables (Arkell et
Table 6-5 Quality Control Properties of Dodecylbenzene Cable Fluid

Property Test Method Value
Material Supplied in Material Supplied in Output from Fluid
Bulka Barrelsa Treatment Unit
Dielectric loss angle 90°C and
ASTM D 924 0.001 0.002 0.003
40 – 60 Hz
Volume resistivity at 90°C, 1010
ASTM D1169 100 50 50
Ω-m
Breakdown voltage, kV ac ASTM D 1816 40 30 50
Moisture content, ppm ASTM D 1533 75 75 20
Less than 50,000 parti-
Particle count IEC 970 No requirement No requirement cles in 2-5µm range
per 100 ml
Residual gas pressure, mm Hg Special apparatus Not applicable Not applicable 1.0
a. Values from IEC 60867 (IEC 1993b).
6-17
al.1989). However, the act of coiling imparts a twist to 6.18 SELF-CONTAINED FLUID-FILLED CABLES
the cable, and it is important to ensure that the lay direc- FOR OPERATION ON DC
tion and length of all components are correct to ensure The most significant differences between cables that
that damage does not occur. operate on dc compared with those that operate on ac
are the insulation and armoring. The significance of the
Where submarine cables have a double layer of armor, armoring is discussed in Section 6.17.
then these are usually applied with opposite lay
directions; however, it is not possible to coil this design. In an ac cable, the electrostatic stress in the insulation
It is, however, much easier to produce a design that is does not vary with temperature. For a single-permittiv-
torque balanced and, therefore, less prone to twist and ity insulation, it is at all times governed by the dimen-
to throw a loop in the catenary between the cable vessel sions of the cable. The magnitude of the stress in zones
and the seabed. of the insulation may be varied slightly by altering the
permittivity of the insulation. The equations are given in
Cables with either a single- or a double-layer of wires Chapter 3.
may be wound onto a turntable, and this is generally
accepted as being the safest option. The thickness of the insulation in an ac cable is designed
such that the cable will withstand the lightning impulse
Cables intended for operation in a dc system may have test carried out as part of type approval testing. The
steel wire armor applied without loss of rating. In these normal operating voltage is a small fraction of the
systems, there is no alternating magnetic field external impulse test voltage, and the associated electrical stress
to the cable to create hysteresis losses in the armor. under normal operating conditions is well below the
withstand capability of the insulation. Figure 6-17
Cables for ac systems with steel wire armor must either shows a 270-kV dc SCFF cable.
be of a three-core construction (see Figure 6-16)
because the magnetic field external to the three cores in
trefoil formation is minimal (Arkell et al. 1989), or, for
single-core designs, a copper return conductor must be
provided beneath the steel wires to reduce the magnetic
field to an acceptable level (Grzan et al. 1992) and thus
minimize hysteresis losses in the armor. In some cases, a
copper armor is applied, and the wires thus have a dual
role (Foxall et al.1984; Dominguez Miguel et al. 1994).
Armor wires are bound down with multiple polypropy-

lene strings. Strings are preferred on submarine cables,
because extruded finishes are likely to slip through the
laying machinery, and may split if they sustain local
damage. Strings provide a good gripping surface and
contain any damage to a small area that may be more
easily repaired. The use of multicolored strings allows
easy observation of the cable through cameras
attached to remotely operated cable-burying machinery
on the seabed.
For long lengths it is necessary to bond the armor wires

to the metal sheath to prevent overvoltages appearing
across the polymeric jacket. These are installed about
every 3 miles (5 km). Alternatively, a semiconducting
anticorrosion jacket may be provided (Foxall et al. 1984). Figure 6-16 Three-core armored submarine SCFF cable
(courtesy J-Power Systems).
6-18
For cables for use on dc, there is no lightning impulse tapes are similar to those used on ac. The insulating
test requirement; however, cables are designed to with- paper used for dc cables has a density in the range 62
stand a polarity reversal that can arise with some types lbs/ft3 to 44 lbs/ft3 (1000 kg/m3 to 700kg/m3), and thick-
of dc converters. For dc cables under steady-state condi- nesses from 4 mils (100 µm) to 8 mils (200 µm).
tions, the stress in the insulation is dependent on,
among other factors, the electrical resistivity of the insu- It is also possible to use LPP as an insulating material
lation. The insulation resistivity itself is highly depen- for SCFF dc cables. The first commercial use of LPP in
dent on the temperature and to a lesser extent on the dc SCFF cable is in the 500-kV dc submarine cable sys-
applied electrical stress. The electrical stress distribution tem across the Kii channel in Japan. Many development
across the insulation is, therefore, influenced by the tests were carried out, and it was concluded that, under
thermal characteristics of the insulation and by the heat dc conditions, LPP had a superior performance com-
generated by the load current (see Figure 6-18). Analy- pared to kraft paper, as had been found with ac cables.
sis of these effects leads to the well-known phenomenon Figure 6-19 shows the improved dc breakdown strength
of “stress inversion,” whereby the stress on the outside of LPP compared to paper. The tests were carried out
of the insulation is greater than the stress adjacent to the on models consisting of three thicknesses of 5-mil
conductor when the cable is under load. It is, therefore, (125-µm) laminate with a hole in one tape to simulate a
good practice to grade the insulation so as to take into butt space. The results show the dc breakdown strength
account this effect rather than using a conventional ac
cable paper-grading design. The impulse and long-term
dc strength of these papers increase with increasing den-
sity and with decreasing thickness.
As well as electrical performance, it is necessary for the

cable to have a satisfactory mechanical performance.
For large insulation thicknesses, it is usually necessary
to have thicker paper tapes on the outside of the cable to
achieve this. Thus the density of the papers may be var-
ied in the wall of the insulation. In one design (Arkell
and Gregory 1984), four different densities of paper, but
only three different paper thicknesses, are used.
It is customary to use the same insulating machine for

dc as for ac cables, and thus the dimensions of the paper
Figure 6-18 Stress distribution in the insulation of the

cable in Figure 6-17 (Arkell et al. 1984) (courtesy IEEE).
Figure 6-19 Dc breakdown strength of LPP and kraft paper

measured on models of 5-mil (125-µm) tapes (Sugata et al
Figure 6-17 270-kV dc 1600-kcmil (800-mm2) SCFF cable 1996) (courtesy author).
(courtesy author).
6-19
of LPP to be 1.6 times that of kraft paper at the cable is believed to be the highest rated, naturally cooled cable
conductor operating temperature (Sugata et al. 1996). in the world and demonstrates how the SCFF cable has
evolved with LPP insulation, not only on low-loss ac
This particular dc installation operating at 500 kV, with systems, but also on dc systems at the highest level.
a 6000-kcmil (3000-mm2) conductor and LPP insulation
6-20
REFERENCES Endersby, T. M., B. Gregory, and S. G. Swingler. 1992.

AEIC. 1993. AEIC Standard CS4-93. Specifications for “The Application of Polypropylene Paper Laminate
Impregnated-Paper-Insulated Low and Medium Pressure Insulated Oil Filled Cable to EHV and UHV Transmis-
Self-Contained Liquid Filled Cable. 8th Edition. Associ- sion.” Paper No. 21-307 CIGRE. Paris.
ation of Edison Illuminating Companies. New York.
Endersby, T. M., B. Gregory, and S. G. Swingler. 1993a.
Arkell, C. A., and W. E. Blake. 1968. “Installation of E. “Polypropylene Paper Laminate Oil Filled Cable and
H.V. Oil Filled Cables in Deep Shafts.” IEE Conference Accessories for EHV Application.” IEE Third Interna-
Proceedings: Progress in Overhead Lines and Cables tional Conference on Power Cables and Accessories
200 kV and Above. London. 10 kV-500 kV. Conference Publication No. 382.
Arkell, C. A., W. E. Blake, A. D. R. Brealey, K. J. H. Endersby, T. M., S. J. Galloway, B. Gregory, and N. C.

Hacke, and G. E. A.Hance. 1977. “Design and Con- Mohan. 1993b. “Environmental Compatibility of
struction of the 400 kV Cable System for the Severn Supertension Cables.” IEE Third International Confer-
Tunnel.” Proceedings IEE. Vol. 124. No. 3. ence on Power Cables and Accessories 10 kV-500 kV.
Conference Publication No.382.
Arkell, C. A. and B. Gregory. 1984. “Design of Self
Contained Oil-Filled Cables for UHV DC Transmis- ENA. 2000. Energy Networks Association. “Engineering
sion.” Paper No. 21-07. CIGRE. Paris. Recommendation C84/2. Code of Practice for the Main-
tenance of Self-contained Fluid-filled Pressure Assisted
Arkell, C. A., B. Gregory, and J. E. Hawkes. 1984. Cable Systems up to 400 kV Operating Voltage.”
“Design and Testing of 270 kV D.C. Self-Contained Oil-
Filled Cable for the Land Section of the UK-France ENA. 2006. Energy Networks Association Website.
Connection.” IEEE/PES Transmission and Distribution 2006. www.energynetworks.org.
Conference. Kansas City, Missouri.
Farneti, F., P. Menesatti, E. Dotti, and G. Luoni. 1984.
Arkell, C. A., S. J. Galloway, E. B. Parsons, B. G. Wood- “Testing of a 1100 kV, 3 to 9 GVA Underground Trans-
cock, and D. E. Woolmer. 1989. “Design, Manufacture mission System.” IEEE/PES Winter Meeting. Dallas,
and Installation of 150 kV Submarine Cable System for Texas. Paper 84 WM 190-5.
the Java-Madura Interconnection.” IEE Proceedings.
Vol. 136. Pt. C. No. 3. Foxall, R. G., K. Barlow-Larsen, and G. Bazzi. 1984.
“Design, Manufacture and Installation of a 525 kV
British Standards Institution. 2004. “BS 7922 Electric Alternating Current Submarine Cable Link from Main-
Cables: Fluid–filled Paper- and PPL-insulated, Metal land Canada to Vancouver Island.” CIGRE. Paris.
Sheathed Cables and Accessories for Alternating Volt- Paper No. 21-04.
ages up to and Including 400 kV (= 420 kV). Require-
ments and Test Methods.” Gazzana-Priaroggia, P., E. Occhini, and N. Palmieri.
1961. “A Brief Review of the Theory of Paper Lapping
Brotherton, W., H. N. Cox, R. F. Frost, and J. Selves. of a Single Core High-Voltage Cable.” IEE Proceedings.
1977. “Field Trials on Internally Oil-Cooled Cables.” Vol. 108C.
IEE Proceedings. Vol. 124. No. 3.
Grzan, J., E. I. Hahn, R. V. Casalaina, and J. O. C. Kan-
Couderc, D., Q. Bui Van, A. Vallee, R. Hata, K. sog. 1992. “The 345kV Underground/Underwater Long
Murakami, and M. Mitani. 1996. “Development and Island Sound Cable Project.” IEEE.
Testing of a 800 kV PPLP-Insulated Oil-Filled Cable
and its Accessories.” Paper No. 21/22-04. CIGRE. Paris. Heyda, P. G. 1967. “Theoretical Aspects of Mechanical
Stresses in a Paper-Lapped Cable Insulation.” IEE Pro-
Dominguez Miguel, M., M. Ruiz Urbieta, E. Gallango ceedings. Vol. 114. No. 10.
Faraco, A. Benchekroun, and O. El-Kindi. 1994.
“Technical Requirements for the Submarine Electrical Hunter, P. V., and J. T. Hazell.1956. Development of
Interconnection Spain-Morocco.” Paper 21-205. Power Cables. George Newnes Limited. London.
CIGRE. Paris.
6-21
Iceton, J. W. 1985. “The 50 MW Jersey – France 90 kV Minemura, S. and Y. Maekawa. 1989. “500 kV Oil-
A.C. Line.” IEE Fourth International Conference on Filled Cable Installed on Bridges.” IEEE/PES Transmis-
AC and DC Power Transmission. Conference Publica- sion and Distribution Conference. New Orleans, Louisi-
tion. No. 255. ana. Paper 89 TD 352-2 PWRD.
IEC. 1993a. International Electrotechnical Commis- Moore, G. F. 1997. The Electric Cables Handbook.
sion. Standard 60141-1. “Tests on Oil-Filled and Gas- Blackwell Science. Oxford, UK.
Pressure Cables and their Accessories Part 1: Oil filled,
Paper–Insulated, Metal Sheathed Cables and Accesso- Ray, J. J., C. A. Arkell, and H. W. Flack. 1973. “525 kV
ries for Alternating Voltages up to and Including Self-Contained Oil-Filled Cable Systems for Grand
400kV.” Coulee Third Powerplant – Design and Development.”
IEE London Paper No T73.
IEC. 1993b. International Electrotechnical Commis-
sion. Standard 60867. “Insulating Liquids. Specifica- Sugata, T., Y. Koga, H. Takashima, T. Imajo, R. Hata,
tions for Unused Liquids Based on Synthetic Aromatic O. Fujii, S. Yoshida, and M. Shimada. 1996. “Develop-
Hydrocarbons.” ment of 500 kV DC PPLP-Insulated Oil-Filled Subma-
rine Cable.” CIGRE. Paris. Paper No. 21-302.
IPG. 2006. International Power Generation Website.
www.ipg.antfx.com. United States Patent 4967039. 1990. “Insulating Liquids
and Electric Cables.”
Lanfranconi, G. M., and B. Vecellio. 1986. “Fire Retar-
dant Oil-Filled Cables.” CIGRE Report No. 21-11. Ple-
nary Session. Paris.
6-22

Special Applications
Authors: Phil Bolin, Mitsubishi Electric Power Products, Inc. (Section 7.2)
W. Graham Lawson, Energy Cable Consultants, Inc. (Section 7.3)
Allen MacPhail, BC Hydro (Section 7.3)
Ernesto Zaccone, Prysmian Cables and Systems (Section 7.3)
David Lindsay, Southwire (Section 7.4)
Reviewer: Pierre Argaut, Silec Cables
A number of cable types are used for special applications. These installations typically
involve high costs and require a great deal of engineering analysis and design.
One of these cable types is gas-insulated lines (GIL). First installed in 1972, GIL is pres-
surized with sulfur hexafluoride (SF6) gas and is used in high-power applications.
Special-application cables also include dc and submarine cables. The demand for HVAC
and HVDC submarine power transmission links has continued steadily since the 1950s.
A review of worldwide experience presented in Section 7.3 shows that, while the maxi-
mum route length of HVAC links approaches approximately 65 miles (100 km) at the
60–90 kV level or about 25 miles (40 km) at the 500-kV level, longer lengths are possible
if intermediate islands allow the use of shunt reactor compensation. Submarine HVDC
links have continued to increase in length, reaching approximately 350 miles (560 km) in
2007, and a voltage level of 500 kV in 2006.
The basic cable types employed in submarine HVAC and HVDC cable systems are gener-
ally the same as those employed in HVAC underground systems. Therefore, Section 7.3
focuses on differences in design and testing requirements for HVAC and HVDC subma-
rine cable applications.
A third area of special application cables includes superconducting cables. Since the dis-
covery of high-temperature superconducting (HTS) materials in 1986, there has been
continuing interest in developing underground superconducting cables that would have
the capability of carrying three to five times more power than conventional cables. Today
there are a number of projects worldwide demonstrating HTS cables.
Phil Bolin is General Manager of the Substation Division of Mitsubishi

Electric Power Products, Inc., responsible for the supply Gas Insulated
Substations (GIS), large power transformers, generator circuit break-
ers, and Static Var Compensators (SVC) on both an equipment and
turnkey basis. Previously Mr. Bolin worked for 20 years in the develop-
ment and production of GIL for High Voltage Power Corporation and
Westinghouse Electric. He has 16 patents and published many papers
related to GIL. He is a Fellow of the IEEE and is active in the IEEE
Substations Committee, CIGRE, and IEC. Mr. Bolin has a B.S. and M.S. from MIT in
Electrical Engineering.
7-1
Chapter 7: Cable Construction: Special Applications EPRI Underground Transmission Systems Reference Book
Graham Lawson graduated with a a CIGRE member and past Chair of AEIC’s Task
B.Sc. in physics from Edinburgh Uni- Group for developing CS9 “Specifications for Extruded
versity and a Ph.D. in electrical engi- Insulation Power Cables and Their Accessories Rated
n e e r i n g f ro m t h e U n iv e r s i t y o f Above 46 kV Through 345 kVac.” Since retiring from
Southampton prior to joining Pirelli BC Hydro in 2006, Mr. MacPhail has formed a small
Cables (now Prysmian Cables & Sys- consulting company, Cabletricity Connections, special-
tems) in 1970. His career in Research, izing in power cable application engineering.
Development and Engineering
(RD&E) with Pirelli involved extended periods in Italy, Ernesto Zaccone is currently with the
Brazil, the United Kingdom, and North America, where R & I ( Re s e a rc h a n d I n n ovat i o n )
he held the position of V.P. RD&E. In 1992, he joined department of Prysmian Cables and
Power Technologies, Inc. (now Siemens PTI, Inc.) as System (formerly Pirelli Cables) based
Manager of the Underground Cables Group. In 1997, in Milan, Italy. He obtained a degree
he founded Energy Cable Consultants, Inc., a small in Electrotechnical Engineering in
business corporation registered in New York State. 1969. During the same year, he joined
Since 1992, Dr. Lawson has provided consulting services the power cable industry, where he was
in the field of underground and submarine cables to a engaged in a number of different activities such as
worldwide client base. Among his most recent U.S. projects in high-voltage testing laboratories and high-
activities, Dr. Lawson has provided consulting services voltage cable system design. In particular, he was
to the NeptuneRTS Project, the 500-kVdc, 660-MW sub- involved in the study, design, and realization of some
marine cable link between New Jersey and Long Island, major EHV (Extra High Voltage) extruded cable
and the 138-kV, 300-MVA Connecticut to Long Island projects. He was also the author of several studies on the
submarine cable link that is to replace the 1969 SCFF application and impact of EHV cables when integrated
cable circuits. Dr. Lawson is a Fellow of the IET (UK), a in transmission systems. Currently he is involved in a
Senior Member of the IEEE, and an individual Member number of innovative underground cable projects and in
of CIGRE. international standardization activities. Mr. Zaccone is
the Chairman of the CEI CT 20 Italian standardization
G. Allen MacPhail graduated from the body for power cables. He is the Co-convenor of the IEC
University of British Columbia, Van- TC20 WG 16, which is responsible for the international
couver, with a BApSc (EE) in 1972. standardization of high-voltage cables and their accesso-
He then joined BC Hydro, specializing ries, and is an active member of CIGRE and IEEE.
in underground and submarine cable
engineering for most of his career. Mr. David T. Lindsay is the Business Man-
MacPhail’s experience includes most ager for HTS Cable Systems at South-
phases of a typical project cycle from wire Compa ny. I n this role he is
feasibility studies through to commissioning tests and responsible for all business and techni-
maintenance, for cable systems ranging from 15 to cal activities related to development
525 kVac and up to +/-450 kVdc. It also includes HV a n d c o m m e rc i a l i z at i o n o f h i g h -
and EHV cable condition assessment, life expectancy temperature superconducting (HTS)
and extension; up-rating feasibility studies; real-time cable technology. David joined South-
dynamic rating and fiber optic distributed temperature wire Company in 1998 as a Development Engineer to
sensing systems. In addition to in-house responsibilities, work on new product designs and manufacturing pro-
Allen has been active with BC Hydro International sub- cesses. Since that time he as been involved with all HTS
marine cable projects in Central America, Alaska, Ire- activities at Southwire. David is also Director of the
land, Egypt, and Jordan, at voltage levels from 35 to joint venture Ultera through which Southwire partners
400 kV. Activities also include acting as an advisor to with nkt cables of Germany to develop HTS cables. He
EPRI’s Underground Transmission Task Force. He is a holds a Bachelors Degree in Materials Engineering from
contributor to the IEEE ICC as past Vice Chair of the the Georgia Institute of Technology and Masters of
Accessories Subcommittee, past Chair of the Cable Business Administration from the University of West
Construction and Design Subcommittee, and partici- Georgia. He is a member of the IEEE Power Engineer-
pant in many discussion and working groups. He is also ing Society.
7-2
EPRI Underground Transmission Systems Reference Book Chapter 7: Cable Construction: Special Applications
7.1 INTRODUCTION 7.2 GAS-INSULATED LINES (GIL)

Extruded-dielectric, pipe-type, and self-contained fluid- GIL is described in general in Section 2.8.1. This section
filled ac cables account for the majority of installed describes the construction of two presently available
lengths of transmission cable systems throughout the types of GIL—factory assembled and tested shipping
world. However, a number of other novel cable types units, and field assembled and tested lines. Both use sul-
also exist—including gas-insulated lines, dc cables, long- fur hexafluoride (SF6) gas (or SF6-N2 mixtures) at a
length submarine cables, and superconducting cables. moderate pressure (about 60 psig [500 kPa]) to insulate
These cables are being used, or developed for use, in a high-voltage aluminum conductor supported by epoxy
special applications, and have unique aspects to their spacers inside a grounded aluminum enclosure.
design, installation, and specifications.
7.2.1 SF6 Gas and SF6-N2 Mixtures for High-
Gas-insulated lines (GIL) are used in applications where Voltage Insulation
high-power capabilities are required, such as substation
SF6 gas is widely used for high-voltage insulation in
ties and through tunnels. GIL is similar to the SF 6 -
circuit breakers and Gas-Insulated-Substations (GIS)
insulated metal-clad bus bar connected to switchgear.
for its excellent electrical insulation and arc interruption
capabilities. For a GIL, the conductor/enclosure diame-
Submarine cables may be either HVAC or HVDC, with
ter ratio is approximately 3 for a near-optimum electric
most HVDC cables being installed in longer length
field distribution. The overall size is chosen to pass the
(≥ 25 miles [40 km]) submarine crossings. Submarine
lightning impulse withstand voltage type test while
cables have many design aspects in common with under-
avoiding condensation of SF6 at low temperatures.
ground cables. Their mechanical design, however, must
See Table 7-1 for typical GIL sizes and electric field
permit very long splice-free lengths and allow installa-
stress levels.
tion at significant water depths.
SF6 Physical and Environmental Characteristics
Research has been underway for several decades on SF6, a sulfur atom surrounded by six tightly bonded flu-
underground transmission cables utilizing supercon- orine atoms, is a dense, colorless, odorless, nontoxic gas
ducting materials. Superconducting cables offer the that is stable up to more than 300ºC. It is an excellent
potential for lower losses, high efficiencies, and greater electrical insulator due to having heavy molecules with a
utilization of rights-of-way due to the higher power den- strong electronegative characteristic. When exposed to
sity at a given voltage level. an electric discharge or arc, it decomposes into reactive
sulfur (S) and fluorine (F) based components that will
This chapter describes these cable types. Section 7.2 cov- react with contaminants (H2O, O2) to form small
ers GIL. Section 7.3 discusses dc and long-length sub- amounts of HF, SOF2, SO2, and so on, but mostly the
marine cables. Section 7.4 describes superconducting decomposed gas quickly recombines into SF6.
cables.
SF 6 does not affect the earth’s ozone layer, but is a
For each of these cable types, Chapter 7 provides infor- strong greenhouse gas, having an anthropogenic green-
mation on application experience, design basics, con- house gas potential of about 23,000 times that of CO2
s t r u c t i o n , i n s t a l l at i o n p r a c t i c e s, a n d i n d u s t r y on a weight basis. SF6 is included in the 1996 Kyoto Pro-
specifications and standards. The sections also discuss tocol as one of the six greenhouse gases slated for emis-
special issues that are unique to each cable type. sion reduction. Although prior standards allowed as
Table 7-1 Size and Electric Stresses for GIL (courtesy Mitsubishi Electric)
Voltage Operating Factory Test Ratio of Factory/ Max Stress
Class kV Std. Factory Conductor OD Enclosure ID Stress kV/mm Stress Operating at BIL kV/mm
(L-L) BIL kV Test kV mm mm RMS kV/mm RMS Stress Peak
72 350 160 51 152 1.49 5.75 3.85 12.57
145 650 310 89 229 1.99 7.37 3.70 15.46
242 900 425 102 292 2.60 7.92 3.04 16.78
362 1300 500 127 362 3.14 7.52 2.39 19.54
550 1550 740 178 495 3.49 8.13 2.33 17.03
800 1800 860 178 610 4.21 7.85 1.86 16.42
1200 2175 1100 279 749 5.03 7.98 1.59 15.79
7-3
much a 3%/year leak rate, the present IEC 62271-100 tle effect on the insulation level. This figure also shows
(IEC 2006a) leak rate limit is 0.5%. The objective for that most of the benefit of using a mixture of SF6 with
manufacturers of SF6 high-voltage equipment is a limit air (or N2) is obtained with only 20% SF6.
of 0.1%/year. Mixtures of SF 6 and N 2 further reduce
SF6 emission potential. Governments presently address The cleanliness of the interior of the GIL is extremely
concerns with SF6 for high-voltage equipment in rela- important, because small conducting particles can ini-
tion to global warming through voluntary SF6 emission tiate an internal electric arc. Figure 7-3 shows the effect
reduction programs (EPA n.d.). of free-conducting particles on a coaxial gas gap. Small
particles below 3.2 mm (0.125 in.) in length cause break-
Contamination Control—Purity, Moisture, and down at test-voltage levels, while larger particles cause
Conducting Particles breakdown at operating-voltage levels. There are strong
SF6 is easily handled in a liquefied state at pressures of pressure and time dependencies for particle effects. A
600 psig (4137 kPa) at 20ºC. It is produced by reacting free-conducting particle exposed to the electric field
fluorine with sulfur, filtered and furnished in cylinders gains a charge from the electrode that it is contacting,
at 99.99% purity and a very low moisture level (less than proportional to the electric field at that location and
10 ppmv). The GIL interior must be dry enough so that time. The electric force on this induced charge is also
after evacuation of air to below 200 microns and filling proportional to the product of the charge and the elec-
with SF6, the moisture level will stay below 1000 ppmv tric field, so the force becomes proportional to the
for the life of the installation. This will ensure that the square of the electric field.
dew point of the gas inside the GIL will only occur at
temperatures lower than the freezing temperature for
water. If the moisture content is high enough for liquid
water to condense on the surface of an epoxy spacer, a
conductor-to-enclosure arc (flashover) will occur, as
shown in Figure 7-1. However, the formation of ice is
not a problem, as shown in Figure 7-1.
SF 6 can be reused using commercially available gas-

handling systems that liquefy the SF6 for compact stor-
age and include filters to remove reactive contaminants
and moisture. Reuse criteria are established by IEC
60480 (IEC 2004a). The required purity level is 98%.
Figure 7-2 shows that small amounts of air have very lit-
Figure 7-2 Effect of mixing air with SF6 gas
(courtesy Mitsubishi Electric).
Figure 7-1 Effect of moisture in GIL (courtesy Figure 7-3 Effect of free-conducting particles on
Mitsubishi Electric). breakdown voltages (courtesy Mitsubishi Electric).
7-4
When the force becomes greater than the force of gravity tendency for particles to move until they abruptly stop
on the particle, the particle lifts and moves in the direc- in low-field regions. Deliberate particle traps are simple
tion of the electric field. For a coaxial system under dc metal pieces placed in the lowest region of the enclosure,
voltage, with the particle initially resting at the bottom creating areas partially shielded from the electric field.
of the system (enclosure side), the particle travels to the They may be continuous along the length of the cable,
highly stressed conductor and, if conditions are suitable, but experiments and experience have shown that placing
initiates a breakdown by acting as a sharp projection on particle traps at each spacer is sufficient. A typical
the conductor surface. Under ac voltage, the electric particle-trap system associated with a spacer is shown in
field reverses direction each half cycle, but a particle Figure 7-4.
already moving freely in the gas gap cannot change the
polarity of its charge, so it is decelerated and perhaps 7.2.2 Epoxy Spacers
even driven back to the electrode that it just left. Spacers are used at intervals of about 20 ft (6 m) to sup-
Impulse and switching-surge voltages are over before port the conductor in the center of the enclosure. Epoxy
the particle has time to move, so they are less affected by spacers include Bisphenol A and Cycloaliphatic. Pow-
particles unless the particle already happens to be in a dered fillers—such as aluminum oxide, fused silica, and
high-field region (for example, deposited by vibration or aluminum trihydrate—are used for rigidity and resis-
an ac voltage on a spacer surface near the conductor). It tance to damage by electric discharges. The temperature
is obvious that the effects of particles are both impor- limit of the epoxy formulation is above 110ºC. Vacuum
tant and complex. Techniques to deal with particles casting is used to ensure that there are no voids inside
include cleanliness, conditioning, and particle traps. the spacer that might have electrical discharges at the ac
system voltage. Electric field stress inside the spacer is
Cleanliness is fundamental, and it is not difficult to kept below 127 V/mil (5 kV/mm) rms for a service life
achieve. The objective is to keep the gas gap between the on the order of 50 years. Surface electric field stresses
conductor and enclosure (especially the spacer surface) are generally lower. Each insulator is tested for freedom
free of particles above a critical size. Since particles can from partial discharges at an ac voltage of from 1.9 to
collect together and act as a larger single particle, the 3.8 times the operating voltage stress.
need for cleanliness extends down from the theoretical
critical size, from about 40 mils (1 mm) to dust size. The The critical “triple point,” where the metal of the con-
initial assembly, the shipped units, the field-assembled ductor or enclosure meets the epoxy spacer and insulat-
system, and the operating cable must remain clean. ing gas, is shielded by shaping the adjacent metal or
Conditioning in gas insulation is the application of elec-

tric stress in such a way as to raise the subsequent break-
down voltage of the system. The physical basis is that, in
many gases, a low-energy-level breakdown will destroy
the initiating disturbance (free-conducting particle or
projection on the electrode surface), while the gas itself
will recombine without leaving behind any by-products
that would reduce the breakdown voltage. SF6 is excel-
lent in this respect. Conditioning without breakdown is
achieved by raising the voltage slowly enough or in steps
so that free-conducting particles have time to migrate
into low-field regions without causing breakdown.
Particle traps are deliberate or natural low-field regions

in the system where the motion of the particle is inhib-
ited or prevented. Because particle motion is propor-
tional to the square of the electric field, there is a strong Figure 7-4 Particle-trap system (courtesy CGIT Westboro).
7-5
epoxy material to reduce the electric stress. Electric field used—various polishing and washing techniques
plots for typical designs are shown in Figure 7-5. are used.
Conical spacers are used when a gas barrier is required Enclosure section length is limited for convenience of
to limit the amount of gas to a practically handled transport to about 60 ft (18 m). Flanges or butt welds
amount of less than 2205 lbs (1000 kg). The conical join the sections. Flanges have “O” ring seals. Butt
spacer can be a gas-pass-type by making it with holes. welds use a backup ring with special high-temperature
Post-type spacers have cost and particle control advan- seals to prevent contamination of the interior of the
tages, and are used only where support of the conductor GIL during welding. Welding of flanges or butt welds
is needed—in most designs the post space assembly is can be automated, but manual welding by a suitably
fixed to the conductor and slides or rolls on the inside of trained welder is adequate. Both metal-inert-gas and
the enclosure. tungsten-inert-gas welding are used.
GIL spacers are entirely encapsulated inside the enclo- Selection of enclosure wall thickness is a result of manu-
sure, immune to degradation from water, sunlight, or facturing limits, pressure vessel requirements, and electri-
other environmental factors. cal resistance. The resulting relatively thick wall tolerates
shipping and handling stresses without difficulty.
7.2.3 Enclosures, Enclosure Joints, Mechanical
Design, and Corrosion Protection The enclosure is effectively mechanically decoupled by
Enclosures are extruded, longitudinally rolled, and the spacer design from the much more flexible conduc-
welded or spiral welded from strong, high- tor, so thermal expansion and flexibility of the enclosure
electrical-conductivity aluminum alloys (6061-T6, can be considered separately. For direct-buried GIL, the
6063-T6, AlMg3). Extrusions are limited to about 14 in. enclosure is locked in position by the sand backfill, so
(350 mm) diameter tubes with relatively thick walls of the enclosure and its joints must be able to withstand
around 0.3 in. (8 mm). Spiral welded enclosures are the resulting thermal expansion stress. GIL enclosures
made on special machines using continuous automated above ground or in a tunnel accommodate thermal
welding, and can be as large in diameter with as thin a expansion (and contraction) using flexibility from direc-
wall as desired. Longitudinally welded enclosures have tion changes or periodic stainless steel bellows. The
only been used in countries where spiral welding is not amount of thermal expansion movement is about
accepted by local pressure vessel codes. The interior of +/-4 in. per 330 ft (100 mm per 100 m) of GIL length.
the enclosure must be smooth and clean before being
Direction changes with a radius of more than 650 to
1300 ft (200 to 400 m) (depending on enclosure diame-
ter) are accommodated by flexing of the enclosure. For
sharper direction changes, prefabricated “elbows” are
added to the straight bus sections, as shown in Fig-
ure 7-12.
Enclosure aluminum alloys do not need any added cor-

rosion protection for atmospheric exposure. For buried
applications, a thick plastic corrosion protection coating
is used in conjunction with cathodic protection. For
Figure 7-5 Typical electric-field designs for GIL (courtesy Figure 7-6 Plug and socket contact for GIL
Mitsubishi Electric). conductor joint (courtesy CGIT Westboro).
7-6
high current applications in air, special paint that resistance, leakage, and high-voltage ac withstand with
reflects sunlight but emits infrared radiation is used. partial discharge monitoring. After testing, the section
is filled with dry air or nitrogen, provided with shipping
Enclosures are bonded to each other at each end and caps, and packed in containers or truckload-sized fix-
periodically as required by the installation situation. tures for shipping for field–joining, as shown in
The bonding connection allows circulating currents Figure 7-7. A post-type spacer mounted on the conduc-
approximately equal to the conductor current to flow tor that can roll on the inside of the enclosure to accom-
through the enclosures. These circulating currents cause modate differential thermal expansion of the conductor
losses and heating of the enclosures, but reduce external relative to the enclosure is shown in Figure 7-8. A gas-
magnetic fields and allow thorough grounding for per- barrier-type insulator is shown in Figure 7-9.
sonnel safety.
7.2.4 High-Voltage Conductor and Conductor

Joints
High-electrical-conductivity extruded aluminum alloys
(6101-T64, for example) are used for the high-voltage
conductor. The production length is up to 60 ft (18 m)
for shipping and handling convenience. Wall thickness is
chosen to minimize electrical resistance. Walls thicker
than 0.6 in. (15 mm) are not used due to current distribu-
tion skin effects at 50 or 60 Hz system frequency. The
interior of the conductor should be reasonably clean and
dry. The exterior that will be exposed to the highest elec-
tric field stress must be smooth, without significant elec-
Figure 7-7 Factory-assembled GIL sections being
tric field-enhancing protrusions. It is usually polished, joined by bolted flanges for the enclosure and plug-
cleaned, and then handled with care to avoid scratches. socket conductor contact (courtesy CGIT Westboro).
Conductor sections are joined by welding, bolting, or

sliding plug-in contacts. The overall system design dic-
tates the techniques used. Sliding plug-in contacts are
used as needed for the convenience of field-joining sec-
tions and to accommodate the differential thermal
expansion of the warmer conductor relative to the
cooler enclosure. The plug-in contact has multiple low-
force silver-plated copper fingers transferring current
from the socket to plug. With high-quality silver plat-
ing, the socket and plug material can be either copper or
aluminum. Specially chosen lubricants are used to
ensure a long life. The contact assembly is covered by a Figure 7-8 Factory-assembled and -tested spacer
with particle trap (courtesy CGIT Westboro).
shield to prevent enhancement of the electric field and
capture any particles that might be created by contact
motion, as shown in Figure 7-6.
7.2.5 Factory-assembled and -Tested GIL

Factory-assembled and -tested GIL requires the follow-
ing steps: receiving and storing conductor and enclosure
material, casting and proof-testing epoxy spacers,
assembling conductor contacts, cutting enclosure and
conductor tubing to length, preparing the ends for con-
tact assemblies or flanges, welding the contacts or
flanges in place, polishing and cleaning the conductor
and enclosure, mounting spacers on the conductor, and
inserting and fixing the conductor inside the enclosure.
The assembled section is then tested to check contact Figure 7-9 Factory-assembled and -tested
gas barrier spacer (courtesy CGIT Westboro).
7-7
Constructing a GIL from factory-assembled and tested vals, special conductor contacts, enclosure bellows,
shipping units is simple in terms of site work. The ship- elbow, gas barrier, and test sections are added, as shown
ping units are unloaded, moved into position, shipping in Figures 7-11 through 7-14.
caps removed, and the shipping unit joined to the next
section. The conductor contact plugs in while the bolted 7.2.7 Field Test
or welded enclosure joint is made. A factory technician Measurement of conductor resistance (including the
ensures that the work is done with appropriate attention plug-in contacts) is done after installation to confirm
to avoiding contamination of the interior and achieving proper assembly and verify the condition of the plug-in
a leak-free seal. conductor contacts. Gas moisture level is also measured.
Practical shipping lengths are limited to 60 ft (18 m), so A high-voltage ac withstand test is done to “condition”
that a 0.6-mile (1-km) circuit length requires about 300 the GIL in terms of particle control (see “Contamina-
factory units, including some allowance for elbow and tion Control” in Section 7.2.1) and to prove voltage
termination sections. About 18 truckloads will bring the withstand. Series resonant test sets with either variable
GIL sections to the site. A site crew of five can install inductance or variable frequency are commercially
about five sections per day, so 60 work days (12 weeks) available. Based on observations of particle movement
are needed. Another week will be used for evacuation, during testing, a gradual increase in test voltage to the
gas filling, and high-voltage test. final level in a two-hour sequence of steps is effective
and efficient.
7.2.6 Field-assembled and -Tested GIL
Field-assembled and -tested GIL saves most of the fac- High-voltage impulse voltage tests have not proven to
tory work described above by having 40- to 60-ft (12- to add significant value and would be very difficult for
18-m) long enclosure and conductor sections shipped long GILs.
directly to the site from the tubing manufacturer. The
time required to assemble the GIL in the field depends 7.2.8 Operation and Maintenance
on site conditions, the skill of the workers, and the effi- Other than periodic checking of a gas density monitor
ciency of automated joining equipment. Components that will alarm if the GIL leaks about 5% of the initial
such as spacers are separately shipped to the site. Enclo- charge of gas, there are no operation or maintenance
sure sections are joined into long lengths by automated tasks. GIL is a sealed system without any need for inte-
welding. Conductor sections are similarly joined by rior inspection. The exterior should be visually
welding, provided with simply mounted spacers and inspected once a year to check for damage or corrosion.
inserted into the enclosure. A typical construction
sequence is shown in Figure 7-10. At appropriate inter-
Figure 7-10 Field-assembled installation and test of GIL (courtesy Siemens).
7-8
7.3 DC CABLES AND LONG-LENGTH

SUBMARINE
7.3.1 Introduction
There are considerably more HVAC underground and
submarine cable installations in the world compared to
HVDC, simply because the latter are relatively expen-
Figure 7-11 Field-assembled GIL typical straight section sive in comparison. This is mainly because of the high
(courtesy Siemens). (Note long stroke of the conductor
plug-in contact.) cost of ac/dc converters to transform ac to dc and back
again, usually exceeding $100M (~€75M), depending on
voltage and power. However, when cable lengths exceed
the critical length for effective transmission of ac real
power, dc systems become necessary.
The choice of a dc power transmission cable system in

preference to the more conventional ac cable option
would generally be made in cases where the power trans-
fer requirement is greater than 150–300 MW and one or
more of the following characteristics apply:
• Long length (typically ≥ 25 miles [40 km]) submarine
Figure 7-12 Field-assembled GIL elbow section for abrupt cable link or interconnection, with length limits
direction changes (courtesy Siemens). mainly dependent on system voltage and ampacity.
• Intermediate length (typically 5 to 25 miles [8 to
40 km]) submarine cable interconnections between
two large ac transmission networks where power
transfer control is a potentially serious problem. A dc
cable system provides an asynchronous or flexible
transmission interconnection.
• Reinforcement of a long-length ac transmission
system in areas of high load density (cities) without
increasing the interrupting duty of ac circuit
breakers.
Figure 7-13 Field-assembled GIL bellows section to
absorb thermal expansion and contraction of the enclosure
(courtesy Siemens). (Note the long conductor contact The relationship of cable length to the choice of an ac or
overlap and movement capability.) dc transmission voltage lies in the capacitance of the
cable. As the ac cable length and voltage increase, the
capacitance and hence the ac charging current, increases
in proportion (charging current is equal to the voltage
divided by the cable’s capacitive reactance). At the so-
called critical length, the capacitance charging current
equals that of the thermal current rating of the cable,
and no real power can be transmitted. For short- to
medium-length ac cable systems, the charging current
can be compensated for by the use of shunt reactors at
the cable terminations, or in the case of submarine
Figure 7-14 Test port section for GIL.The conductor section
cables, at intermediate islands. For long cable lengths,
in the center can be unbolted and a test bushing installed
on the top port to test in either direction (courtesy Siemens). however, this becomes impractical, and dc power trans-
mission is necessary.
7-9
Figure 7-15 summarizes transmission distances and a lower investment level than conventional HVDC
capacities for various ac and dc cable types. power transmission systems (Ruter 2002). Later subma-
rine applications are in service for the ± 150 kV, 330
As regards the use of a dc cable system as a reinforce- MW Long Island Sound and the ± 150 kV, 350 MW
ment of an ac system in an area of high load density, the Finland-Estonia crossings.
only known application is the 36.5-mile (59-km) long
±266 kV, 640 MW link between the Kingsnorth power The following subsections are devoted primarily to a
plant and the Willesden substation in the city of Lon- review of the state-of-the-art in ac and dc submarine
don. This trial installation commenced service in 1971, cable technology and, to a lesser extent, because of their
but was retired a few years later (Last et al. 1970). limited use, to the state-of the-art in dc underground
cables. Section 7.3.2 reviews worldwide submarine cable
Prior to 1997, with one exception, all dc cables at volt- experience, while Section 7.3.3 reviews HVDC under-
ages ≥ 100 kV were submarine cables with, in some ground cable experience. Sections 7.3.4 and 7.3.5 discuss
cases, associated land cable installations. The exception the design and testing aspects of submarine and dc
was the 1993 St. Lawrence River ± 450 kV, 2000 MW cables that are unique to these cables. Because of the
project (Bell et al. 1992), which was installed in a tunnel high cost and long outage times associated with subma-
and in short direct-buried sections. In 1997, a new volt- rine cable system failures, reliability in service is of
age source ac/dc converter (VSC) and XLPE dc cable major importance. As a result, Section 7.3.6 concludes
technology (Apland et al. 1998) was introduced, which with a discussion of submarine cable hazards and miti-
has resulted in the implementation of several under- gation methods.
ground cable projects. The most notable of these
projects is the ± 150 kV, 220 MW Murraylink Project Note: The units quoted throughout Section 7.3 are
(Australia), which was installed in 2002 (Normark those taken from the original references. Approximate
2002). The route length is 112 miles (180 km), which is conversions to U.S. units can be obtained as follows:
to date a world record for long-length underground • km x 0.62 = miles
cable installations. Among the stated advantages for the
new technology are a significant reduction in project • mm x 40 = mils
lead time due to modularized manufacturing and instal- • mm2 x 2 = kcmil
lation of VSC converters, and a capability to transmit • kV/mm x 25 = V/mil
electricity over long distances with efficient operation at
Figure 7-15 Transmission cable system selection criteria for various

cable types and capacities (courtesy Prysmian Cables and Systems).
7-10
7.3.2 A Review of Worldwide Submarine Cable SCFF cables were introduced in the early 1920s and
Experience today are used extensively, as indicated above, at all HV
and EHV levels up to 525 kV in both underground and
Ac Submarine Cables submarine transmission line applications. Various prob-
In the HV and EHV ranges—that is, from 110 kV and lems, such as fatigue failure of lead sheaths and thermo-
upwards—ac submarine cable links on the order of 10 mechanical failure of splices, have had to be overcome
km or more are dominated by the self-contained, fluid- over the years, as the service voltage and power trans-
filled cable type (SCFF) with XLPE cables that are nor- mission capacities were increased. In consequence of a
mally used up to 150 kV and are beginning to be applied rather long development and service history, SCFF
at the 230-kV to 420-kV levels. Up to a voltage of cable technology is now mature and highly reliable.
150 kV, submarine cables have, in principle, a three-core
configuration to facilitate the installation in one opera-
tion, thereby reducing the cost for cable protection. For
higher voltages, the cable design is typically single core.
Figure 7-16 shows a sample of the submarine cable con-
necting the islands of Sardinia (Italy) and Corsica
(France). The cable connection is 15 km long, laid at a
depth of 75 m, and buried in the seabed. At the shore
ends, additional protective cast iron shells have been
applied. The cable weight is 75 kg/m in air and 48 kg/m
in water, and the overall diameter is 207 mm. The cable
is rated 150 kV for the transmission of 150 MVA. Each
copper conductor’s cross-sectional area is 400 mm2. A
fiber optic element has been placed in the cable inter-
stices for data and signal transmission. Similar cables
have been used for the connection of large off-shore
wind farms in the European North Sea, as indicated in
Table 7-2. Figure 7-16 HVAC 150-kV XLPE submarine cable
(courtesy Prysmian Cables and Systems).
Table 7-2 Listing of Many of the World’s Major Ac Submarine Cable Links
Voltage Conductor Cable Length Max. Water Depth
Name of Link Date (kV) (mm2) (km) (m) Cable Type
Vancouver Is. 1956 138 225 24 + 4 190 SCGF
Mallorca-Menorca 1973 132 500 4 x 42 90 SCFF
Long Island Sound 1977 345 2500 MCM 3 x 2.1 30 HPFF
Prince Edward Is. 1977 138 3 x 240 3/C x 14 30 SCFF
Norway (Oppland Fylkes Elverk) 1981 145 630 4 x 1.2 N/A XLPE
Vancouver Is. 1984 525 1600 4 x 39 400 SCFF
Karlskrona 1986 145 500 3x7 N/A XLPE
Labuan-Beaufort (Malaysia) 1989 132 3 x 190 3/C x 15 N/A SCFF
South Padre Is. 1991 138 380 3 x 14 1.5 XLPE
Long Island Sound 1991 345 2000 4 x 13 35 SCFF
Negros-Cebu (Philippines) 1993 138 300 4 x 18 60 SCFF
Leyte-Cebu (Philippines) 1995 230 630 4 x 32.5 280 SCFF
Penang Island (Malaysia) 1996 275 800 6 x 14 20 SCFF
Spain-Morocco 1997 400 800 4 x 26 630 SCFF
Gulf of Aqaba (Egypt-Jordan) 1997 420 1000 4 x 13 840 SCFF
Isle of Man (UK) 2000 90 3 x 300 3/C x 105 40 XLPE
Galveston Island (USA) 2001 138 3 x 630 3/C x 4.6 <15 XLPE
Horns Rev Wind Farm (Denmark) 2002 150 3 x 630 1 x 21 <20 XLPE
Seas Roedsand Wind Farm (Denmark) 2003 132 3 x 760 1 x 22 <20 XLPE
Sardinia-Corse Islands 2005 150 3 x 400 1 x 15 75 XLPE
Norway-Gossen Is. 2006 420 1200 4 x 2.2 210 XLPE
7-11
Cable and accessory designs and manufacturing pro- this project was successfully carried out at a 900-m
cesses are now more or less standardized at an interna- water depth.
tional level. • Isle of Man, United Kingdom, 2000: The world’s long-
est XLPE HVAC submarine cable is the Isle of
A listing of some of the major ac submarine cable links Man—U.K. Mainland link, which is 105 km and 90
is presented in Table 7-2. The following details for some kV. Single-core flexible factory-installed splices were
of the more significant links are provided to supplement spaced at intervals of 5–7 km.
the information listed in the table:
• Norway-Gossen Island, 2007: The world’s highest
• Mallorca-Menorca, 1973: The longest (42 km) SCFF voltage ac XLPE submarine cable system is 2.2 km
ac submarine cable link ever installed. The 90-m-deep long, 210 m deep, and 420 kV.
link was designed for both 132 kVac and ± 200 kVdc
operation. Dc Submarine Cables
Four types of submarine power cable can be considered
• Vancouver Island, 1984: This double-circuit, 325-m-
for high-voltage, direct-current (HVDC) operation,
deep link continues to hold the world record for the
namely:
highest voltage (525 kV) and power transmission
capacity (1200 MW/circuit). The 30 km and 8 km • Self-contained, fluid-filled cables (SCFF cables)
long routes, including an intermediate island with • Self-contained, mass-impregnated cables (MI cables)
shunt reactive compensation, place this link among
the longest ac EHV transmission cable links ever
• Self-contained, pre-impregnated, gas-filled cables
installed (see Figure 7-17). (SCGF cables)
• Long Island Sound, 1991: The 20-km submarine por- • Self-contained, solid dielectric cables (XLPE and
EPR cables)
tion of this 345-kV, 750-MVA interconnection con-
sists of four cables, with one being kept as a spare for
The first three cable types in the above list all have insu-
use in case of failure on one of the energized phases.
lation consisting of paper tapes impregnated with a
The cable is the largest and most heavily armored sin-
dielectric fluid, but are distinguished either by the type
gle-core ac submarine cable ever manufactured with a
of impregnant or by the use of a fluid-pressurizing sys-
diameter of 170 mm and a mass of 83.5 kg/m.
tem. In the case of the SCFF cables, a low-viscosity syn-
• Gulf of Aqaba, 1997: The 13-km link consists of three thetic fluid—usually an alkylbenzene—is maintained
420-kV cables plus a spare. The transmission capacity under pressure by oil-pumping stations placed at one or
is 550 MW. Cables were laid in a maximum water both cable ends. The MI and SCGF cables both employ
depth of 840 m, which is considered to be close to a high-viscosity compound. In the latter case (as the
the limit for SCFF cables, although a sea trial for name implies), the paper tapes are impregnated with the
compound prior to being applied to the cable conductor,
and the resulting spaces between tapes are filled with
nitrogen gas at a pressure of around 1300–1400 kPa
(200 psi). The impregnation process for the MI cable is,
on the other hand, carried out on the entire cable core
after application of the paper tapes—hence the name
“mass-impregnated.”
The use of HVDC solid dielectric cables has been

delayed due to complications with insulation design,
which are caused by the buildup of space charges in the
insulation and their subsequent distortion of the electri-
cal stress distribution. This phenomenon manifests itself
in a significant reduction in the dc and impulse electric
strength of the insulation when the temperature
increases and a 30–40% decrease in the dc electric
strength when the polarity is reversed (Ruter et al. 2000).
Recent progress in the development of modified XLPE
Figure 7-17 HVAC 525-kV SCFF copper flat wires
armored submarine cable connecting Vancouver using special “functional groups” (Satoru et al. 1999)
Island with the Canadian (BC) mainland (courtesy has apparently been successful in solving the space-
Prysmian Cables and Systems). charge problem. As a result, the first modified-XLPE
7-12
HVDC submarine cable link (the 150-kVdc, 330-MW • The state of the art in depth of laying is the 1000 m
Cross Sound Cable Link) was installed in Long Island deep Italy-Greece link, but a 1650 m depth project at
Sound in 2002. The second modified-XLPE submarine the voltage of 500 kV is in progress, between Sardinia
cable link between Finland and Estonia, a distance of and Italy.
some 75 km, was commissioned in 2006, operating at • The 150-kV, 330-MW Cross Sound cable system is
150 kVdc and transporting 350 MW of power.
the first dc submarine cable system to operate with
the relatively new voltage source control (VSC) con-
Table 7-3 lists the majority of important HVDC subma-
verter technology (Apland et al. 1998).
rine cable links in service worldwide at the present time.
As can be seen by referring to this table, the total
7.3.3 A Review of HVDC Underground Cable
installed cable length is on the order of 3,000 km, and
Experience
the total service experience can be summarized as
approximately 52,000 km-years. Other significant con- HVDC underground cables can be discussed under two
clusions include the following: major categories, namely:
• MI cables account for approximately 80% of the • Cables on land connecting converter stations to the
entire installed cable length. submarine cable segments of the circuits
• SCFF cables are only used for relatively short routes • Cables not associated with submarine cable links
(a few 10s of km).
In the former case, the distance from the shoreline often
• The state of the art in long-length MI cable technol- determines the type of cable to be employed. If the
ogy is the 450-kV, 700-MW, 580-km Norway to distance is not too great, the submarine cable can be
Netherlands link in the North Sea.
Table 7-3 Listing of Many of the World’s Major Dc Submarine Cable Links
Name of Link Date Voltage (kV) Power (MW) Length (km) Max. Water Depth (m) Type
Gotland 1 1956 100 20 100 160 MI
Italy - Sardinia 1965 200 100 2 x 118 450 MI
Cook Strait 1 1965 250 300 3 x 39 300 PIGF
Konti-Skan 1 1965 285 300 64 80 MI
Vancouver Is. 1 1969 300 156 3 x 27 200 MI
Skaggerak 1,2 1976 263 250 2 x 125 600 MI
Vancouver Is. 2 1976 300 185 2 x 35 200 MI
Hokkaido/Honshu 1980 250 150 2 x 42 290 SCFF
Gotland 2,3 1983 150 160 2 x 100 160 MI
Cross-Channel 2 1986 270 250 8 x 50 55 MI
Konti-Skan 2,3 1988 285 300 2 x 64 80 MI
Fenno-Skan 1989 400 500 200 117 MI
Cook Strait 2 1991 350 500 3 x 40 300 MI
Skagerrak 3 1993 350 500 125 500 MI
Cheju (Korea) 1993 180 150 2 x 96 160 MI
Baltic Cable 1994 450 600 250 60 MI
Sweden - Poland 1999 450 600 253 90 MI
KII Channel Japan 2000 500 2800 4 x 49 70 SCFF
Italy - Greece 2001 400 500 1 x 160 1000 MI
Moyle (UK) 2001 250 500 2 x 55 100 MI
Cross Sound (USA)a 2002 150 330 2 x 42 40 XLPE
Bass Link (Aus) 2005 400 500 1 x 290 75 MI
Norway-Netherlands 2007 450 700 1 x 580 410 MI
New Jersey - Long Island (Neptune) 2007 500 660 1 x 84 25 MI
2008 500 1 x 420
Sardinia-Italy 500 1650 MI
2010 500 1 x 420
a. First HVDC Submarine Cable System with VSCs and XLPE Insulation.
7-13
pulled ashore from the cable-laying vessel, as shown in work control in a region with an extensive use of
Figure 7-18. windmill generation.
• Direct Link in New South Wales, Australia: Installed
An example of the use of an HVDC submarine cable for in 2000, this 84-kVdc, 160-MW project is described
the underground connection to the converter station is as an entrepreneurial project designed to take advan-
found in the 450-kV Baltic Cable Project (Eriksson et al. tage of different tariffs.
1993). In this case, the link from the Swedish converter
station is routed southwards, first via a 12 km overhead • Murraylink, Victoria, Australia: With a route length
line then underground for a distance of approximately of 180 km, the Murraylink Project is the world’s long-
5 km. The submarine MI cable was installed in a precut est underground cable project. The operating voltage
trench using a combination of 1600 regular and 100 syn- is 150 kVdc, and the power transfer is 200 MW. The
chronized motor-driven cable rollers. installation date was 2000.
An example of the use of a different cable type for the The stated advantages of the VSC/HVDC XLPE tech-
underground section of an HVDC interconnection is nologies are as follows:
provided by the 270-kVdc Cross Channel Project • Rapid control of both active and reactive power
(Arkell et al. 1986). An SCFF cable was selected for the • Minimal environmental impact
18-km underground section from Folkestone to the con-
verter station at Sellindge in the United Kingdom. The • Connection to weak networks
transition joint to the MI submarine cable consisted of a • Preassembled converter enclosures reduce civil work,
disconnectable back-to-back SF6 termination/splice. installation, and commissioning
• No ac magnetic field and low dc magnetic field,
As regards the use of HVDC underground cables not because by necessity, the VSC technology requires
associated with submarine cable links, there are only a two cables operated in a bipolar mode with current
few examples. The only SCFF cable project ever to be flow in opposite directions, and the cables are usually
implemented was the trial 266-KVdc Kingsnorth, Lon- installed close together.
don Project commissioned in 1971 and retired a few
years later (Last et al. 1970). More recently, the intro- The fact that cables, rather than overhead lines, must be
duction of voltage source converters (VSC) and modi- used the entire distance between converters, to avoid
fied XLPE cable technologies by two competitors possible damage to converters caused by overhead line
(Apland et al. 1998) has resulted in a few HVDC under- pole-to-pole faults, may in some scenarios prove to be
ground projects as follows: a disadvantage.
• The Gotland Island Project: This is a 80-kVdc,
50-MW project linking Nas and Visby, a distance of 7.3.4 HV Submarine and Dc Cable System
70 km, commissioned in 1999, with the scope of net- Design
Introduction
Submarine and dc cables have many of their design
aspects in common with underground cables. For exam-
ple, the insulation design of SCFF submarine cables is
virtually identical to that of SCFF underground cables.
Since these common design aspects are dealt with else-
where, this subsection focuses on design features that
are unique either to submarine or to dc cables. These
design aspects are treated under the following four
headings, namely:
• mechanical design of submarine cables
• insulation design of HVDC cables
• special requirements for the accessories—i.e., splices
and terminations
Figure 7-18 Dc cable destined for installation on land being
pulled ashore from the cable-laying vessel (courtesy • design issues relating to dc ground return current.
Prysmian Cables and Systems).
7-14
Mechanical Design of Submarine Cables S = axial spacing between nearest phases

Historically, submarine cables were generally armored (mm).
with one or two layers of helically applied armor wires DSM = root mean square diameter of metallic
in order to sheath and armor (mm).
• withstand tensile stresses during laying and recovery
From Equation 7-1, it is possible to see that when the
• provide protection against damage from fishing spacing of the phase cables S is of some tens of meters,
trawls and other third-party marine activities as for the water depth, the current in the metallic screen
• prevent cable abrasion by friction against rock, coral and armor becomes equal to the current in the conduc-
reefs, etc. tor. For this reason, in order not to increase dramati-
cally the losses in the metallic screen and armor, it is
Nowadays, the most common method of providing pro- necessary to reduce the armor resistance to a value simi-
tection against mechanical damage in service is by lar or lower than that of the conductor.
burial in the seabed, so that in practice, the role of the
armor is primarily that of accommodating the tensile For this reason, large copper flat wire armors are often
stresses encountered during laying (and recovery should used in order to reduce the heating caused by the circu-
repairs become necessary). lating currents (see Figure 7-17). Due to this kind of
construction, the cable phase inductance is that of a
In the case of HVDC cables, the armor generally con- concentric coaxial cable calculated as per Equation 7-3.
sists of galvanized steel wires, while for HVAC cables,
DSM
both galvanized steel and copper wires have been used, L = 0.2 × ln
the copper wire (round or flat) armor being used where 2 × RcGMR 7-3
necessary to improve the cable circuit power transfer Where:
capability. HVAC cables for voltages higher that 150 kV L = phase inductance for greatly spaced sub-
are generally single core and are laid in the seabed with marine cables in mH/km.
the phases spaced at distances equal or greater than the DSM = root mean square diameter of metallic
water depth to provide a sufficient space for a U-shaped sheath and armor.
laying configuration that may be necessary to repair the RcGMR= geometric mean radius of the conductor
mid cable without overlapping the other cables. In these (commonly = 0.779 Rconductor).
conditions, the induced circulating currents in the
metallic sheath and armor are practically the same as A variation on either copper wire or steel wire armor
the current circulating in the conductor. HVAC cables is a “hybrid” design using a large copper
return conductor over the lead sheath. Its relatively low
The calculation of the current circulating in the metallic impedance attracts much of the conductor return cur-
screen and armor of a submarine cable can be carried rent, resulting in less circulating current and eddy cur-
out by using the Equation 7-1: rent heating in the external steel wire armor. This design
was used for the 345-kVac Long Island Sound crossing
1
Is = I × cables (Arnaud 1992) (see Figure 7-19).
R
1 + ( s )2
Xm 7-1
Where:
Is = Current circulating in the metallic screen
and armor (A).
I = Current circulating in the conductor (A).
Rs = Resistance of the metallic screen (i.e., the
lead sheath) and armor in parallel (Ω/m).
Xm = Reactance of the screen and armor calcu-
lated as in Equation 7-2:
2× S
X m = 2 × ω × 10 −7 × ln( )
DSM [Ω/m] 7-2
Where:
ω = 2·π·f where f is the frequency. Figure 7-19 HVAC 345-kV SCFF Long Island Sound 1991
cable having copper flat wires and steel round wire armor
(courtesy Prysmian Cables and Systems).
7-15
Independently of the type of armor employed, corrosion The factor 1.3 in Equation 7-4 accounts for additional
protection is provided by flooding the armor wires with tensile load caused by laying and recovery forces and
bitumen and applying an overall serving layer of dynamic forces during laying and recovery operations.
polypropylene string.
For water depth higher than 500 m, a more accurate for-
In general, a single layer of round wires is used when the mula should be applied, because the factor 1.3 may not
water depth is less than ~200 m. For deeper waters, a be suitable in all cases, especially for very heavy cables.
double layer of round or flat wire armor is used, with The evaluation of the tension should be carried out
the layers applied with opposite helical lays to provide according to the actual laying conditions, the wave
an antitorsion property and to reduce the mechanical motion, and the cable-laying vessel characteristics.
load on the conductor. In this case, the cable cannot be
coiled on a fixed platform, so rotating turntables must To prove that the mechanical design of a submarine
be used for storage and transport as well as cable laying. cable is suitable for installation in a specified water
Figure 7-20 shows examples of the use of single-wire depth, a prototype submarine cable at least 30 m in
and double-wire armor. length, including one factory joint and one repair joint,
should be subjected to a tensile bend test. During this
According to experience, the maximum pulling tension test, the cable sample should be wound and unwound
encountered during cable laying and retrieval can be three times on a reel with a diameter not greater than the
calculated from the equations given in the CIGRE Elec- diameter of the pay-off sheave of the cable-laying vessel,
tra 171 (CIGRE 1997). For water depths between 0 and while under a tensile load calculated using the above for-
500 m, the following two equations apply, namely: mulae. The test should not give rise to breaks or perma-
T = 1.3 x W x D + H 7-4
nent deformations of the cable conductor or armoring.
H = 0.2 x W x d 7-5
Another very important mechanical aspect that should
Where: be taken into consideration during the mechanical
T = Pulling tension having the same unit of W design of submarine cable is to avoid the hydrostatic
(i.e., Newton). pressure collapse during or after installation at high
W = Weight of 1 m of cable in water. depths. For SCFF cables, the internal oil pressure
D = Maximum laying depth in m. should compensate the external pressure, but the differ-
H = Maximum bottom tension. ences in oil- and water-specific gravities should be evalu-
d = Laying depth with a minimum of 200 m ated and may become critical at elevated water depth.
(if D ≤ 200 m). Heavy oils have been developed for this application. For
mass-impregnated cables, there is no internal pressure,
and the risk of collapse may be very high. Because all
the cable layers are not perfectly tight, the effect of the
external pressure may cause the deformation of the lead
sheath. Although the circular cross section can generally
withstand the higher pressure, once distortion begins
(i.e., lead deformation during thermal loading), the
shape becomes mechanically unstable, and the surface
may pinch or crease. In the past, this risk was avoided
by adopting a slightly oval-shaped design with a poly-
ethylene sheath applied directly over the lead sheath, fol-
lowed by one or two layers of reinforcing tapes and
appropriate servings. Under current loading conditions,
the thermal expansion increases the minor axis of the
lead sheath, the cable tending towards a circular shape.
As the cable cools, the external water pressure causes it
to return to its original oval shape, thus maintaining the
dielectric under pressure at all times. Developments and
experience have demonstrated that special lead alloys
Figure 7-20 Examples of double- and single-wire armored (e.g., 0.4% Sn, 0.2% Sb) have adequate fatigue life to
cables. 450-kVdc Baltic Cable (left) and 100-kVdc Gotland withstand the resultant flexing of the sheath during the
Island Cable (right) (courtesy ABB HV Cables). cable service life.
7-16
Modern paper-lapping technology combined with the Figure 7-22 shows the results of the calculation of the
adoption of the compact keystone (conci) conductor electrical stress distribution in the insulation wall of a
provides a round cable design that is able to withstand 500-kVdc MI cable under the following conditions:
external pressures higher than 175 bar. One installation 1. In the steady-state, full-load condition (curve A)
at 1000-m water depth has been implemented (Giorgi et
2. In the steady-state, no-load condition (curve B)
al. 2002), and two more installations at 1650-m water
depth will be completed during the period 2008 to 2010 3. At the instant of voltage application (curve C)
(see Table 7-3). Figure 7-21 shows a high-depth subma-
rine cable, which depending on the particular case, may As shown in Figure 7-22, the radial field distribution is
have a copper or aluminum conductor. dependent on load. Under a no-load condition, the
maximum electrical stress appears at the conductor
Insulation Design of HVDC Cables shield, as is the case in ac cables, though generally the dc
The insulation design is considerably more complex in stress is lower than the capacitance stress, as is the case
the case of dc cables than for ac cables due to significant here. On the other hand, when the cable is operating
differences in the radial electric stress distribution in the under a steady full-load condition, the maximum elec-
two cases. In the ac case, the distribution is capacitive, trical stress is transferred to the insulation shield. This
and it does not vary significantly with temperature. In phenomenon is called “stress inversion.”
the dc case, the radial field is determined by the electri-
cal resistivity of the insulation ρ. The electrical resistiv- A correct insulation design is established when the max-
ity is not a constant, but varies with temperature and imum dc electrical stress under no-load and full-load
electrical stress according to the following relationship: conditions is less than or equal to the preset design
ρ = ρo exp − (αT + βE) 7-6 limit, and when in addition, the transient stress due to
the superposition of the dc stress maxima and an oppo-
Where: site polarity BIL is less than or equal to a preset max-
ρo = the electrical resistivity at ambient temper- ima. In the case of HVDC MI cables, the following
ature and low electrical stress. typical design parameters apply according to current
T = the temperature at radius r. industry practice:
E = the electrical stress at radius r.
• Maximum dc electrical stress: 1181 V/mil (30 kV/mm)
α = temperature coefficient.
β = stress coefficient. • Maximum transient stress: 3346 V/mil (85 kV/mm)
• Maximum temperature: 50–55°C
Using Equation 7-6, the electrical stress distribution in
the insulation of dc cables can be calculated under In the 500-kVdc cable design used for the calculations
steady-state conditions (steady dc voltage and tempera- presented in Figure 7-22, the dc stress maxima deter-
ture gradient across the insulation), according to the mine the insulation thickness. The maximum transient
method described by Arkell and Gregory (Arkell and stresses based on a BIL of 1000 kV are lower than the
Gregory 1984). 85 kV/mm design limit.
Figure 7-21 High-depth submarine HVDC mass- Figure 7-22 Electrical stress distribution in a 500-kVdc
impregnated round cable design (courtesy Prysmian MI cable.
Cables and Systems).
7-17
Identical design procedures apply for SCFF cables insu- Different arrangements of the HVDC cable transmis-
lated with either paper or laminated paper polypropy- sion are possible. Figure 7-23 shows some typical con-
lene (LPP) tapes (Nakanishi et al. 2000) and for figurations.
modified XLPE (Apland et al.1998) insulated cables.
For SCFF paper-tape insulated cables, temperatures up Earth/sea return systems have been used extensively
to a maximum of 85°C are possible, though designs are since the 1950s and continue to be used. A recent exam-
generally limited to a lower temperature by a 40 kV/mm ple of the use of such a system is the 2001 Italy-Greece
dc stress limitation. For SCFF LPP and for modified- 400-kVdc, 500-MW Project. Two sea electrodes are
XLPE insulated cables, design data are proprietary, and used. The anode consists of 39 bars of titanium and is
there are no de facto industry standards. located near Greece in the Corfu Strait, while the bare
copper ring cathode is located in Italy near Cape
The ampacity I of an HVDC cable is obtained from the Otranto (Giorgi et al. 2002).
following equation:
In the last few years, concern about possible electrolytic
Δθ corrosion of subsea steel structures and adverse environ-
I=
R ⋅ (T1 + T2 + T3 + T4 ) 7-7
mental effects on fish and fauna—although as yet
unproven—have meant that, in some cases, earth/sea
Where:
return paths are no longer accepted. For example, in the
Δθ = conductor temperature rise above ambient.
recently completed 400-kVdc, 600-MW Basslink Project
R = conductor dc resistance per unit length.
(Australia) and in the 500-kVdc, 660-MW New Jersey-
T1 = insulation thermal resistance per unit
Long Island Neptune Project, separate metallic return
length.
cables provide the return current path. In both cases, the
T2 = jacket thermal resistance per unit length.
medium voltage return (MVR) cable is XLPE insulated,
T3 = serving thermal resistance per unit length.
with an insulation thickness corresponding to that of a
T4 = surrounding medium thermal resistance
12/20-kVac cable design. The maximum dc voltage
per unit length. This term may also include
under full-load conditions amounts only to a few kV.
mutual heating of adjacent cables or
The HVDC and the MVR cables are laid in a bundle
heat sources.
together with a fiber optic cable. The close proximity of
the HVDC and MVR cables substantially eliminates
Note: the units are the same as the equations presented
any magnetic field external to the cable system.
in Chapter 11.
Monopolar transmission is more cost effective than
The voltage and current ratings of an HVDC cable are
bipolar at power transmission levels up to approxi-
interrelated, since both are functions of temperature, in
mately 600–800 MW (Giorgi et al. 2002), although a
contrast to the HVAC case, where the voltage and cur-
single monopole has no redundancy in case of cable fail-
rent ratings can be considered independently.
ure. Northern Ireland Electricity decided to adopt an
Catering for the dc Ground Return Current
HVDC transmission can either be bipolar or monopo-
lar. In the bipolar case, the two poles operate at equal
and opposite polarities, and the current circulates
around the bipolar circuit. Therefore, there is no need
for provision of a return current path. In monopolar
transmission systems, on the other hand, a return cur-
rent path is required. This can be provided in at least
three different ways:
• Earth/sea return via an anode cathode electrode
system
• Separate metallic return cable
• Metallic return conductor integrated into the HVDC
cable
Figure 7-23 Typical configurations of HVDC submarine

cable connections.
7-18
integrated metallic conductor solution for their 250-kV, The creepage distance in the case of HVDC termina-
500-MW Moyle submarine cable link between Northern tions for coastal areas is usually calculated on the basis
Ireland and Scotland (Balog and Evenset 2002). This of 40–45 mm/kV, though 50 and even 58 mm/kV have
link consists of two separate 250-kVdc, 250-MW mono- been used. Taking the lowest of the above figures, the
poles, each of which consists of a 250-kVdc IRC (inte- creepage path for a 400-kVdc termination would be
grated return conductor) cable. approximately 16 m, and the overall length in the region
of 3.2 m.
Figure 7-24 shows the construction of an HVDC-IRC
submarine cable. The IRC consists of two layers of flat Other important parameters are the impulse electric
copper strip applied with counter helical lays, which strength of the insulation internal to the porcelain. For
together with the steel wire armor provide a torsion-free example, the stress cone or condenser insulation around
design. The Moyle HVDC-IRC Submarine Cable the cable core and the wall of the porcelain have both to
Project is the only project to date to make use of an IRC. be dimensioned to meet the specified BIL level. To com-
plete the design, the dc stress field of the accessory at the
Cable Accessories – Splices and Terminations service voltage has to be computed to check that preset
Terminations limits are complied with.
HVDC terminations are similar to those used for con-
ventional underground HVAC cable applications. In No particular problems with HVDC terminations have
fact, a 275-kVac cable termination was used in the been reported. Successful experience with 400 and 450
266-kV, 644-MW HVDC cable installed in London in kVdc has been achieved with the Fennoskan (1989) and
1970 (Last et al. 1970). The porcelain insulator was of Baltic (1994) cable projects, respectively, and 500-kVdc
the “anti-fog” type used in normal industrial environ- SCFF cable terminations have been in operation in the
ments. Such terminations have a relatively long “creep- St. Lawrence River project since 1992.
age” distance, achieved by using special sheds with
Splices
ribbed undersides. Cable terminations installed in
Underground cable splices for HVAC and HVDC SCFF
extreme conditions of heavy industrial pollution or near
cables are practically interchangeable in the case of
the coast, where salt fog is a problem, require even longer
equivalent system voltages. For example, 275-kVac stop
creepage paths. Such environments are more hazardous
joints have been used successfully on a 266-kVdc SCFF
in the case of HVDC terminations, because of electro-
underground cable system (Last et al. 1970). These
static attraction effects. HVDC terminations tend, there-
splices are rigid splices having diameters several times
fore, to be much longer than their HVAC counterparts.
that of the cable being spliced.
In general terms, the following three types of splice need

to be considered during the planning phase of a typical
submarine cable project, namely:
• Factory splices
• Transition splices
• Repair and field splices
Factory splices—so called for obvious reasons—are

used to join one or more manufacturing lengths to pro-
duce a continuous shipping or delivery length. In the
case of SCFF or MI submarine cables, typical manufac-
turing lengths are 30–40 km, depending on the cable size
and weight. Shipping lengths of 100–120 km are possi-
ble, depending on the cable plant’s handling and storage
facilities, as well as the loading capacity of the cable-lay-
ing vessel. XLPE cable manufacturing lengths are much
shorter, because the extrusion process has to be inter-
rupted after a specified volume of extrudate has been
processed, for cleaning of the extrusion heads and
replacement of the filter packs (if used). Manufacturing
Figure 7-24 Construction of an HVDC-IRC submarine
cable (courtesy Nexans Norway AS).
7-19
lengths on the order of 4–5 km of 150-kVac XLPE cable Repair splices for water depths greater than approxi-
are feasible. mately 50 m are generally of the flexible type. In the case
of shallow water, it is possible to use a rigid splice. Such
Flexible splices having an outer diameter approximately a splice can be constructed more quickly, but it is diffi-
equal to that of the cable are the standard factory splice cult to lay. Flexible splices are preferred, although con-
used in all types of HVAC and HVDC submarine cables. siderably more expensive because of longer repair times
These splices have equivalent electrical characteristics to and more complex repair vessels. There are many exam-
those of the cable and are capable of withstanding the ples of the use of flexible repair splices; one example is
pure tensile and tensile bending stresses encountered the repair of the 345-kV SCFF cable in Long Island
during cable laying or recovery for repair should the Sound, which was carried out in 2004. Flexible repair
cable be damaged in service. splices differ from flexible factory splices only in that the
jacket has to be remade (in some cases using a heat-
The various phases in the construction of a flexible fac- shrinkable tube), and the individual armor wires have to
tory splice can be summarized as follows: be carefully spliced to withstand laying tensions without
• The cable conductor is spliced using a flush ferrule cable elongation.
(Miranda and Gazzana-Priaroggia 1976) having the
7.3.5 HVDC and Submarine Cable System Tests
same diameter as the conductor. Alternative splicing
techniques include brazing and MIG (metal inert Introduction
gas) welding. From a mechanical strength point of The subject of cable system testing is dealt with fully in
view, the splice is weaker than the conductor itself. Chapter 14, and reference to that chapter should be
Nevertheless, very high tensile strengths can be made for general information on the various test catego-
achieved and a very high safety margin exists. ries, test procedures, and techniques that are common to
• The semiconducting shields and insulating tapes are all types of power cable. In the following, consideration
applied using taping machines. In the case of XLPE is given to the specific test requirements applicable to
splices, the taped splice is subjected to a vulcanizing HVDC and long-length submarine cable systems.
process to fuse and crosslink the materials. Test Standards and Recommendations
• An extruded lead tube, which is slightly larger in At present, there are no U.S. test standards or recom-
diameter than the lead sheath of the cable and which mendations for either HVDC or submarine cables. The
has previously been placed over one of the cable ends, only relevant U.S. document is an IEEE guide to sub-
is placed in position over the splice and swaged down marine power cables (IEEE Std. 1120 [IEEE 2005]).
over the insulation shield. The ends are butt-soldered This guide provides the prospective user with a useful
to the cable sheath. list of factors to be considered when planning, design-
• The polymeric jacket is reconstructed by extrusion ing, permitting, installing, commissioning, and repair-
molding. ing submarine power cables. A short bibliography is
also provided.
Because of the reduced splice diameter, the length of a
flexible splice is much longer than that of a rigid splice. Applicable test specifications in this field consist of the
The length of a flexible splice for a 500-kVdc MI cable is following four documents, all of which have been pub-
~5 m between the lead wipes. lished by CIGRE:
• Recommendations for Mechanical Tests on Subma-
Transition splices are required to join either different rine Cables (CIGRE 1997).
cable types or cables of the same type but with different
• Recommendations for Tests of Power Transmission
sizes. In the former case, back-to-back GIS terminations
DC Cables for a Rated Voltage up to 800 kV
have been used to transition between HVDC MI sub-
(CIGRE 2000b).
marine cables and SCFF underground cables (Last et al.
1970; Carcano et al. 1996). A transition splice between a • Recommendations for Testing Long AC Submarine
1250 mm2 MI submarine cable and a 2000 mm2 MI land Cables with Extruded Insulation for System Voltage
cable was developed for the 400-kVdc Italy-Greece Above 30 (36) to 150 (170) kV (CIGRE 2000a).
Project. This splice was similar to the flexible factory • Recommendations for Testing DC Extruded Cable
splice described above but employed an asymmetric Systems for Power Transmission at a Rated Voltage
compression ferrule (Carcano et al. 1996). up to 250 kV (CIGRE 2003).
7-20
A brief review of mechanical testing of submarine cables Routine Tests:

( C I G R E 1 9 9 7 ) h a s a l r e a dy b e e n p r e s e n t e d i n • High Voltage Test at Works: 1.8 U0 for 15 minutes
Section 7.3.4; the more significant features of the remain-
ing three test recommendations are discussed below. • Conductor Resistance Test
Laminar Cable Tests (CIGRE 2000b)

• Capacitance Test
In contrast to the qualification procedures applied to • Power Factor Test
pressurized SCFF and HPFF HVAC laminar (paper or • Factory Acceptance Test: 1.8 U0 for 15 minutes
LPP tape) dielectric cables (IEC 141-1 1993; AEIC CS2
1990), HVDC qualification or type tests have always Type Tests:
required verification of their life expectancy by carrying
out a series of daily loading cycles at elevated direct • Load Cycle Test: 20 daily load cycles at 1.8 U0
voltage. MI cables are not pressure assisted, and voids in • Polarity Reversal Test: 10 daily load cycles at 1.4 U0
the insulation may be created during the cooling cycle, with reversals every 4 hours
thus leading to the dielectric breakdown if the cable is • Superimposed Impulse Test: 10 negative and 10 posi-
not well designed or manufactured. A type test program tive impulses superimposed opposite polarity U1.8 U0
(see Chapter 14 for the definition) is then always neces-
sary for the approval of the specific cable design and
After Laying Test:
manufacturing process. These tests, which are applied to
all types of laminar dielectric cables, were originally • 1.4 U0 negative polarity for 15 minutes
developed for MI cables, since their life was known to be Testing of Long-Length Ac Submarine Extruded Insulation
related to the voltage V by an equation of the form: Cables (CIGRE 2000a)
Cons tan t These CIGRE recommendations refer to IEC 60840
Life = (IEC 2004b) for all of the Routine, Special (or Sample),
VN 7-8
Type, and After Installation Tests. U.S. users may prefer
Where N, an integer, is the so-called life exponent. to refer to national specifications such as AEIC CS9
(AEIC 2006), although it only applies to underground
The applied voltage was always 2.0 U0, but the number applications.
of daily cycles varied until the issue of the first CIGRE
Recommendations in 1968 (Bjurstrom, 1968). Since Special provisions in view of testing long-length cables
then, the number of daily load cycles has remained at are necessary in the following cases:
20. The test corresponds to a life of 40 years using a
value of N equal to 10. Another feature of the load- • Routine and Special Tests: Normally the cable lengths
cycling tests was the inclusion of a further 10 loading are much longer than those on which it is possible to
cycles at 1.5 U0 with polarity reversals every 4 hours. perform partial discharge (PD) measurements due to
This was to ensure that the insulation design was able to attenuation of the cable. It is, therefore, normal to
withstand the relatively high transient electrical stress perform PD measurements on samples taken from
that occurs, particularly under full-load conditions, at the beginning and end of extrusion lengths. The (Spe-
the instant of polarity reversal. Polarity reversal tests are cial) Impulse Test is also to be performed on these
only relevant to cable systems where reversal of the cable samples.
direction of power flow is foreseen. • Factory Acceptance Tests: If the delivery length is too
long for the High Voltage Test specified in IEC
In the latest revision of the CIGRE Recommendations 60840, other test methods must be agreed upon by
(CIGRE 2000b), the test voltages during the 20-day the customer and the manufacturer.
load-cycling test and the 10-day load-cycle test with
• Type Tests: The water penetration test specified in
polarity reversals have been reduced to 1.8 U 0 and
IEC 60840 is unsuitable for submarine cables. This
1.4 U 0 , respectively. This is in recognition of the fact
test is, therefore, replaced by separate longitudinal
that no electrical aging has been noted in cables after
tests for the conductor and for the jacket. In the first
many years in service.
of these, a sample is submerged in water for 10 days
at ambient temperature and under a water pressure
An outline of the tests specified in the latest revision of
corresponding to the maximum laying depth. After
(CIGRE 2000b) is presented as follows:
completion of the test, the conductor is examined,
and water penetration must comply with the manu-
facturer’s stated penetration distance. In case of the
water penetration under the jacket, a similar test is
7-21
specified with the exception that, prior to the test, the at 1.15 x the maximum impulse voltage that the cable
sample shall be preconditioned by three daily heat experiences in the dc voltage, plus reverse polarity
cycles between ambient temperature and the maxi- impulse test, and a dc voltage test at 1.85 U0 for 15 min-
mum conductor temperature plus 5°C. utes at ambient temperature.
• After Laying Test: Two alternative ac voltage tests are
specified—namely, 1.7 U0 for one hour or 1.0 U0 for If separate or return conductor cables are foreseen,
24 hours. (IEC 60840 specifies 1.7 U0 for 5 minutes or these should also be subjected to Type Tests consisting
of the following elements in sequence:
1.0 U0 for 24 hours.) However, the recommendations
allow other test methods to be agreed between the • Mechanical preconditioning as for the HVDC cable.
manufacturer and customer. • A thermomechanical preconditioning consisting of
Testing of Dc Extruded Cable Systems (CIGRE 2003) 24 daily load cycles from ambient to the maximum
During the 1990s, new crosslinkable insulation systems in-service temperature.
were developed. They did not suffer from the problem • A no-load ac test at 1.15 x the maximum expected ac
of space-charge distortion of the electric field, which transient voltage (due for example to converter com-
occurs in conventional XLPE cables. This led to the mutation failure) for 30 minutes.
introduction of land and submarine cable applications
• Impulse test at 1.15 x the maximum impulse voltage
of the new materials technology. At present, the maxi-
that the cable experiences in the dc voltage, plus
mum voltage in service is 150 kVdc, but within a few
reverse polarity impulse test.
years, it is expected that 250-kVdc systems will be avail-
able. With this in mind, CIGRE developed recommen-
If an integral return conductor is to be used, the return
dations for the testing of dc cables (CIGRE 2003) to
path function is to be tested together with the HV cable
provide a common approach to cable qualification.
in an integrated test program.
The philosophy adopted was, wherever possible, to
Three after-installation tests are specified—namely:
base the tests on existing recommendations, standards,
and practices and to introduce new tests specific to the • High Voltage Test: 1.45 U0 for 15 minutes for the HV
new technology. cable, and a 15-minute test at a dc voltage level to be
agreed upon by the manufacturer and customer in
In line with IEC 62067 for HVAC extruded cables (Sha- case of the return conductor.
piro et al. 1997), the CIGRE recommendations intro- • Jacket Test according to IEC 60840.
duce a Pre-qualification Test, which is to be carried out
• A TDR (Time Domain Reflectometry) Test for engi-
before supplying cable on a general commercial basis.
neering information (if desired).
The duration of the test is 360 days and consists of daily
load cycles and periods of continuous loading and of
7.3.6 Reliability in Service
zero load at 1.45 U0 . In the case of conventional con-
verter systems, there are periods of daily load cycling at Because of the high cost and long outage times associ-
1.25 U0. For the new VSC technology, which does not ated with submarine cable failures, reliability in service
change polarity, for the 360-day test, the loading cycles is a major issue. Experience has clearly demonstrated
with polarity reversals are replaced by an equal number that both HV SCFF and MI cables are mature and
of daily load cycles at 1.45 U0. highly reliable products, and that almost all of the cable
faults experienced in service have been due to external
The Type Test regime, in general terms, adheres to the damage caused by third parties, with most faults being
principles underlying the CIGRE Recommendations for attributed to fishing trawls and ship anchors. Mechani-
laminar dielectric cables. These consist of mechanical cal protection is now considered to be mandatory for a
preconditioning, a 30-day load cycle test at 1.85 U0 /1.45 submarine cable system installed in a high-risk marine
U 0 , without/with polarity reversals, and replace load environment. The preferred protection is cable burial in
cycles with polarity reversals with regular load cycles at the seabed, which is discussed in the following section.
1.85 U 0 for VSC systems, which do not have polarity Cable Burial
reversals. Following completion of load cycling, the This high incidence of cable failures due to third-party
cable system is subjected to dc voltage, plus reverse marine activities has led to widespread use of cable
polarity impulse tests, as is done for laminar dielectric burial in the seabed as a means of providing the neces-
cables. Two additional tests not specified for laminar sary protection against such activities. The effectiveness
dielectric cables are included—namely, an impulse test of cable burial in significantly reducing failures due to
7-22
third-party marine activities is well proven by experi- relatively soft seabed sediment—e.g., sand and soft to
ence, though the absence of up-to-date global failure medium hard clay.
statistics makes it difficult to quantify the benefit as far
as power cables are concerned. However, in this context, The working principle for the jet plow (Figure 7-25) is to
published failure data for submarine telecommunication fluidize the seabed materials in a narrow path and to a
cables is relevant. Shapiro et al. (Shapiro et al. 1997) predetermined depth without displacing the majority of
state that failure rates of telecommunication cables fell the material or causing turbidity in the surrounding
from 0.37 failures / year / 100 km (1959–1979) to
0.04 failures / year / 100 km after 1985 due to burial of
existing and new cable systems.
Indicative burial depths for protection against anchor-

ing given by Shapiro et al. are as follows:
• Anchor Mass 0.2 tons: 0.50 m (1.6 ft)
Typical power cable burial depths are 1.0–2.0 m in

sandy soils and deeper in soft soils. Burial is typically
carried out at the shore ends and down to water depths
of ~100–150 m. Table 7-4 provides some examples of
submarine cable projects where cable protection has
been provided by burial in the seabed.
Notwithstanding the foregoing, cable burial require-

ments are very site-specific. Decisions should be based
on risk analysis studies considering local natural and
man-made hazard investigations.
Cable Burial Equipment
The jet plow and the water-jetting machine are now the
preferred types of equipment for power cable burial in Figure 7-25 Photograph and sketches of a jet
plow (courtesy Prysmian Cables and Systems).
Table 7-4 Burial Depths and Burial Equipment Used in Some Important Submarine Cable Projects
Project Details Date Max. Water Depth (m) Burial Depth (m) and Burial Equipment Used
Langkawi Island, Malaysia 1.5
1986 20
132 kVac SCFF Cable Water-jetting ROV
South Padre Island, USA 1.0
1991 1.5
138 kVac XLPE Cable Water- jetting ROV
NYPA Long Island USA 1.5 – 3.5
1991 40
345 kVac SCFF Cable Water-jetting ROV
Skagerrak 3 (N to DK) 0.5 – 1.5
1992 500
350 kVdc MI Cable Water-jetting ROV
Penang Island, Malaysia 1.0 – 2.5
1996 15
275 kVac SCFF Cable Hydroplow
Honshu-Shikoku Islands (J) 2.5
1999 75
500 kVdc SCFF Cable Hydroplow
Moyle - Scotland to N. Ireland 1.0
2001 N/A
Italy - Greece 1.0 (to 150 m water depth)
2002 1000
7-23
waters beyond 5 m. The method has been positively cult such as inadequate soft sediment, depth, or at pipe-
proven to place fiber optic cables and power cables at a line or cable-crossing locations.
consistent required burial depth in sand and clay bot-
tom conditions. The jet plow is towed by a support ves- 7.4 SUPERCONDUCTING CABLES
sel with dynamic positioning or on a barge using
mooring anchors. The fluidizing effect provides rela- 7.4.1 Superconducting Materials and Cable
tively low and controlled towing forces. The burial depth Systems
achievable is typically around 8 ft (2.4 m).
Underground distribution and transmission cables uti-
lizing superconducting materials have been a research
The water-jetting machine (Figure 7-26) combines the
and development interest for several decades. Supercon-
effect of fluidizing the seabed and hydrodynamic trans-
ductors are those materials that exhibit no electric dc
port of the fluidized material. In the front, low-pressure
resistance when cooled below a material-specific critical
water jets are used to fluidize the seabed. The fluidized
temperature, and have power densities as high as 100x
material is transported backwards utilizing other jetting
that of copper. Benefits to power cables can immediately
nozzles. The cable or pipeline sinks by its own weight
be recognized as lower losses, high efficiencies, and
into the trench before the fluidized material is allowed
greater utilization of rights-of-way due to higher power
to settle and start the back-filling process.
density at a given voltage level.
Water-jetting machines have been in operation since
In general, superconductors are divided into two classes:
1978 and have been used to install many hundreds of
low-temperature and high-temperature superconductors.
km of telecommunication and power cables. Different
Low-temperature superconductors (LTS) have critical
versions are available for fiber optic cables, power
temperatures below approximately -387°F (-233°C,
cables, and pipeline burial. There are various ejector
40 K). This class of superconductors includes metals
modules to handle the different seabed sediment types
such as mercury (Hg) and metal alloys such as NbTi and
encountered and high-pressure or rock-cutting systems.
NbSn. High-temperature superconductors (HTS) have
Machines have been developed for deep-water applica-
critical temperatures above approximately -387°F
tions, where it is necessary to have water pumps on
(-233°C, 40 K) and are cuprate ceramics. Examples
board, with power delivered by an electrical umbilical.
include bismuth strontium calcium copper oxide
Alternative Method of Protection (BSCCO) and yitrium barium copper oxide (YBCO).
Alternative protection methods such as covering the The first HTS materials were discovered in 1986. Some
submarine cables with concrete mattresses or rock HTS materials have the distinct advantage of critical tem-
dumping can also be considered, but these are generally peratures above -320°F (-196°C, 77 K), putting them well
only used in special circumstances where burial is diffi- within the temperature range of liquid nitrogen (LN),
which is a low-cost, environmentally friendly coolant
material. In addition to temperature, superconductors
must be operated within critical boundaries of magnetic
field and current density to exhibit zero resistance.
At a given temperature and magnetic field, the maxi-

mum amount of current that a superconducting wire
will carry is called the critical current, Ic. For example,
BSCCO-2223 wires commonly used in cables may have
a critical current of approximately 130 Amps (60 Hz,
alternating current) at -320°F (-196°C, 77 K), with no
external magnetic field applied and only its self-gener-
ated magnetic field. If additional current above the criti-
cal current value is applied to the wire, it will begin to
behave like a normal material, and electrical resistance
will rise.
Since the discovery of HTS materials in 1986, much

Figure 7-26 Water-jetting machine suitable for deep- development has occurred around the world to manu-
water applications, ready for operation (courtesy Nexans facture long lengths of HTS wires. Availability of these
Norway AS). wires has led to research and development activities for
7-24
power applications that include cables, transformers, duct banks and may reduce substation footprint
generators, motors, fault current limiters, and VAR sup- requirements in dense urban areas where space is a pre-
port devices. HTS cables have been under investigation mium. They will offer retrofit alternatives for aging
in the United States, Europe, and Asia as cable compa- infrastructure. Thermal and electromagnetic emissions
nies pursue this new technology and create cable designs to the surrounding soil and adjacent equipment are also
that can leverage the technical advantages of conductors virtually eliminated with HTS cables. Environmental
that may carry up to 5.0 kA in continuous operation. impacts are further reduced versus other cable types due
to the environmentally friendly nature of LN as the
As of the printing of this text, multiple HTS cables are cooling medium.
in service in live utility grid demonstration projects
MV Cables
around the world. These projects use different geometri-
HTS cables at traditional distribution voltage levels
cal designs of conductors operating at various voltage
(10–46 kV) deliver power equivalent to conventional
and current levels to demonstrate the technical and
subtransmission (69 kV) and transmission level
commercial viability of HTS cables in the utility envi-
(≥ 115 kV) cables. This change in power delivery capa-
ronment. Projects include cables pulled into under-
bility at a given voltage level may offer system flexibility
ground ducts with cable-to-cable joints built in
not previously possible. In dense urban areas, these
underground vaults.
applications include relocating transformers to outlying
History of LTS and HTS Cable Projects areas, deploying station-to-station bus tie on the low-
Prior to the discovery of HTS materials in the mid- voltage side of the transformer, uprating underground
1980s, several LTS cable projects were undertaken substation get-away cables that are thermally saturated,
around the world to demonstrate the technical feasibil- and retrofitting existing pipe-type cable ducts.
ity of underground cables made with LTS wires as
replacement for copper and aluminum. These cables Real estate availability and cost in dense urban areas
were cooled with liquid helium down to temperatures of present challenges for siting new substations and/or
-452°F (-269°C, 4 K). One notable example of such expanding existing stations to serve growing load
projects is the Brookhaven National Lab project in the demand. HTS cable circuits at MV levels operating at
1970s (Forsyth 1986). Although these projects were 3–5 kA continuous, enable step-down transformers to
technical successes, they highlighted the challenges of be relocated outside the downtown areas. Power that
reliable cryogenic cooling to such low temperatures and was traditionally fed into these areas at transmission-
the economic cost to do so. Cooling for LTS materials level voltage and then transformed may be fed at the dis-
requires approximately 200 W of input power for each tribution voltage to a much smaller station equipped
watt of cooling at the operating temperature. solely with bus, breakers, and switchgear.
Interest in HTS cables quickly followed the material dis- The high current capability of a distribution HTS cable
coveries. The mid-1990s and early 2000s saw a flurry of also allows the utility company to use the cable as a sta-
cable projects around the world to design, test, and in tion-to-station bus tie on the low voltage side of the
some cases, install in utility grids, cables made with HTS transformer. The HTS cable will be capable of carrying
wires and cooled with LN. These projects ranged in volt- full station load at the MV level. This increases redun-
age level from 10 to 115 kV and included both warm-
and cold-dielectric cable designs (see Table 7-5). As wire
properties improved and cable designs evolved to higher Table 7-5 HTS Cable Projects
voltages and more economical and efficient designs, Country Characteristicsa
additional projects were undertaken primarily in Japan 66 kV, 100 meters, CD, lab test
the United States in the early 2000s to demonstrate Italy 115 kV, 50 meters, 1 phase, WD, lab test
HTS cables in utility grid applications. All projects have Japan 77 kV, 500 meters, 1 phase, CD, lab test
been the beneficiary of government-sponsored develop- United States 12 kV, 30 meters, CD, grid test
ment programs. Denmark 30 kV, 30 meters, WD, grid test
Applications for HTS Cables Korea 23 kV, 100 meters, CD, lab test
Commercial applications for HTS cables will utilize the China 35 kV, 33.5 meters, WD, grid test
unique features of superconducting materials to deliver United States 35 kV, 350 meters, CD, grid test
more power in a smaller space with reduced impact to United States 13 kV, 200 meters, CD, grid test
the surrounding environment. The higher power density a. CD = Cold Dielectric
of HTS cables allows high power throughput in smaller WD = Warm Dielectric
7-25
dancy and allows transformer capacity to be shared Generation

between stations—increasing asset utilization. In generating stations, HTS cable links may be utilized
at the generating voltage to take power directly off the
HTS cables operate at a constant temperature with a generators and transport it to step-up transformers
high-performance thermal insulation system surround- located further distances away. These links may enable
ing the cable. This insulation system isolates the cable the generation plant to relocate transformers to safer
and its cold-space from adjacent cables and other equip- locations with lower electrical losses.
ment (i.e., steam pipe, other power cables). In situations
where substation get-aways are thermally saturated, 7.4.2 HTS Cables—Design Basics
replacement of one or more circuits with HTS cable will
Cable Design Options
reduce the thermal loading and may relieve the thermal
Multiple options exist for geometrical, structural, and
stress to all adjacent circuits.
electrical design of HTS cables. The electrical insulation
HV/EHV Cables materials may be inside or outside the cryostat, and sin-
HTS cables at transmission-level voltages will offer high gle or multiple electrical phases may be placed inside a
power for bulk power delivery and transport through single cryostat. All HTS cables start at the innermost
congested areas. Operating HV cables at currents of layer with a former. This former may be a stranded cop-
4–5 kA facilitates the reduction in voltage class. Con- per member or a hollow tube. The function of the
ventional projects considered at 345, 400, and 500 kV former includes serving as a structural member on top
may be accomplished at 138 or 230 kV. This reduction of which to strand HTS wires, and in some cases it
in voltage level affects rights-of-way, substation space serves as a flow path for LN. In the case of stranded
requirements, transformers, breakers, and all other asso- copper cores, the copper serves as fault current stabili-
ciated auxiliary equipment. zation to protect the HTS wires in the event of a fault
current on the cable. See the discussion of Thermal
High-capacity, low-resistance HTS transmission circuits Management Issues and Fault Current Protection below
strategically inserted into congested regions may also for further detail regarding the need for copper to pro-
relieve bottleneck situations. The inserted HTS cable tect the HTS wires.
will absorb a large part of the system duty, unloading
other highly stressed assets. Reduction in load on aging On top of the former, one or more layers of HTS wires
cable circuits may have a significant positive impact on are helically wound to make up the phase conductor of
their expected lifetime. the cable. It is these HTS wires that replace the copper
or aluminum conductor in a conventional cable.
Very Low Impedance—VLI
Options for the placement of the dielectric materials dic-
One aspect of superconducting cables is their very low
tate if this cable core becomes a warm-dielectric or cold-
impedance (VLI) as compared to their conventional
dielectric HTS cable.
copper and aluminum cable alternatives. The low
impedance results from thinner dielectric layers and per- Warm Dielectric
mits longer circuit lengths. Some suggest that an HTS The HTS core may be placed directly inside the cryostat.
cable with VLI characteristics, in conjunction with a Conventional dielectric layers are then extruded on the
phase angle regulator, may allow for the control of outside of the cryostat to operate at ambient (warm)
power flow on the ac network much like that which is temperatures. Dielectric materials for this application in
possible in dc networks. previous projects have included EPR and polyethylene.
A conventional concentric wire screen may be applied to
Dc Cables
the outside of the dielectrics. Warm-dielectric cables
Dc superconducting cables are envisioned as a possible
present certain disadvantages, including EMF emissions
future long-distance transmission application.
due to conventional screening methods, and lack of
Extremely high power links for transcontinental trans-
access to the surface of the cryostat for maintenance
mission of electric power may be possible, using dc
and/or repair. Should problems develop with the vac-
cables operating at 50–100 kV and tens of thousands of
uum space of the cryostat to reduce the efficiency of the
amperes. See Chapter 17 for additional discussion.
thermal insulation, the dielectric may be frozen by the
cold operating temperature of the LN, leading to crack-
ing and electrical breakdown.
7-26
This cable design has been successfully demonstrated in Single Phase per Cryostat
two separate projects, each with multiyear operations in These cables place a single conductor core with dielectric
a utility substation (see Figure 7-27) (Tonnesen et al. and superconducting screen layer inside a cryostat (see
2004; Xin et al. 2005). Figure 7-28). Three such cable and cryostats must be
placed in separate ducts to comprise a three-phase circuit.
Cold Dielectric
In cold-dielectric cable designs, the stranded HTS core This cable design has been successfully demonstrated
is wrapped with layers of dielectric materials, which may (Stovall et al. 2001), with the cables providing power to
be made of paper or polymer tapes. These tapes operate an industrial complex for more than 6 years.
with the LN coolant saturating the material, filling butt-
gaps, and serving as part of the dielectric system. The Triad—Three Phases, Three Cores in One Cryostat
LN in these cables serves much the same purpose as The triad arrangement twists three superconducting
impregnating fluids in conventional paper cables. The cable cores to create a triplexed, three-phase structure
fully insulated conductor may have a second layer of (see Figure 7-29). A common cryostat covers the three
HTS wires helically wrapped to serve as the concentric phases. This common cryostat is then pulled inside a
screen. The superconducting screen in these cables may
operate at 100% of the load current, resulting in a fully
shielded cable design with no EMF outside the cable
structure. This lack of EMF eliminates eddy current
heating in surrounding equipment or steel duct. The
shielding protects adjacent HTS phase conductors from
induced magnetic fields that may saturate the HTS wire
to limit their current-carrying capabilities. This allows
shielded cold-dielectric HTS cables to operate at higher
load densities than warm-dielectric cables or non-HTS-
shielded cold-dielectric cables.
Figure 7-28 Single phase per cryostat, cold dielectric
Cold-dielectric cable designs afford the opportunity to cable. Stranded copper core. LIPA, Long Island, NY
(courtesy AMSC).
optimize the three-phase structure to reduce material
usage and physical spacing. Options for cold-dielectric
cables include single-phase per cryostat, three phases
twisted together inside a single cryostat, and three concen-
tric phases on a single core all inside a common cryostat.
Figure 7-27 Warm-dielectric cable installed at Figure 7-29 Triad—Three phase per cryostat.
AMK substation in Copenhagen, Denmark, 2000 Cold-dielectric cable with stranded copper cores.
(courtesy nkt cables). Albany, NY National Grid. In service July 2006
(courtesy Sumitomo Electric).
7-27
single duct. Advantages of the triad arrangement Thermal Management Issues

include reduction in overall size and compaction to a Superconducting cables must be maintained at an oper-
single underground duct. The single cryostat reduces ating temperature below the critical temperature of the
costs and lowers the total cold surface area and, there- given material. Cold-dielectric cable designs have the
fore, decreases cryogenic cooling requirements and added restriction of maintaining an operating condition
ongoing operational costs of the cooling system. whereby the temperature of the LN impregnated into
the laminar dielectric remains below the saturation
This cable design has been successfully demonstrated in curve at any given pressure. Bubble formation in the LN
Japan in a laboratory environment (Honjo et al. 2003), creates localized weaknesses in the dielectric properties
in the United States in a utility grid demonstration of the impregnated laminate-liquid system and may lead
(Weber et al. 2006), and in Korea in a laboratory envi- to breakdown (Sauers et al. 2004).
ronment (Cho et al. 2006).
The requirement to maintain the LN below the satura-
Triax—Three Phases, One Core in One Cryostat
tion curve holds true for normal operations as well as
Placing the three phase conductors and screen concen-
during and immediately following a short-circuit fault
tric on a common core reduces the HTS wire require-
event carried through the cable. If proper thermal man-
ments by a factor of 2. The concentric arrangement of
agement designs are not implemented, the energy dissi-
the phases self-cancels electromagnetic fields generated
pated into the nitrogen during a fault may vaporize
by the phase conductors. This eliminates the need for
some liquid, resulting in bubble formation. These bub-
superconducting screens for each phase. This leads to a
bles could lead to electrical failure if the cable experi-
single copper screen common for all three phases con-
ences a breaker reclosure prior to temperature recovery.
centric on the outside of the cable structure. This
A secondary risk created by vaporization of the LN is
arrangement, like individual cores for each phase with
the resulting pressure increase in the liquid system. If
HTS screens, has no EMF emissions. The single cable
not properly contained, the increasing pressure could
core simplifies the manufacturing, cable installation,
open emergency venting valves, burst disks, or cata-
and accessory installation steps. As with the triad
strophically rupture components.
arrangement, the Triax uses a single cryostat and
achieves similar reductions in cryostat costs and cryo-
The simplified energy balance of the cable system
genic cooling requirements and operating costs (see Fig-
(Demko et al. 2003) is represented by Equation 7-9,
ure 7-30).
where ρ is the cable density, CHTS is the heat capacity, z
is the coordinate direction along the cable axis, k is the
This cable design has been successfully demonstrated in
thermal conductivity, AHTS is the cable cross-sectional
the United States in a utility grid demonstration project
area, THTS is the average temperature in the cable, and
(Demko et al. 2006).
Q′conv,i is the convection heat-transfer to liquid nitrogen.
∂THTS
ρC HTS
∂τ
∂ ⎛ ∂T ⎞
= ⎜ kAHTS HTS ⎟ + P' AC −∑ Q'conv ,i
∂z ⎝ ∂z ⎠ i 7-9
Total thermal input due to heat leak through the cry-

ostat, terminations, splices, and ac losses of the system
must be balanced against the refrigeration capacity of
the cooling system.
Fault Current Protection
If the available fault current is high enough to exceed
the superconducting capabilities of the conductor
design, copper stabilization methods must be employed
to protect the HTS wires from damage as well as main-
tain an acceptable LN temperature below the saturation
Figure 7-30 Triax—three phases on single core. curve. The specific superconducting capability of a given
Cold-dielectric, hollow former. AEP Bixby station, cable is a result of the detailed design and rating of the
Ohio. In service August 2006 (courtesy Southwire). cable. Each HTS wire carries a finite amount of current
7-28
in the superconducting state (see discussion of critical Designs that use three separate cryostats for the three
current in Section 7.4.1). The ultimate capability of the p h a s e s m ay f l ow i n o n e di re c t i o n t h ro u g h t wo
cable is the summation of the capability of the individ- phases with a combined return flow through the third
ual wires making up the phase, minus defects and dam- cryostat. This layout results in nonsymmetrical temper-
age due to manufacturing and/or installation. atures and flow rates through the three phases. This
must be taken into consideration when designing ther-
Multiple methods exist to incorporate copper stabiliza- mal management.
tion into the conductor. One common approach is to
use a stranded copper core as the former of the cable. 7.4.3 Construction
The copper core provides a path for excessive fault
Superconducting Materials
current without damaging the HTS wires in the cable.
The copper cross section required will depend on the HTS Wires
exact fault-current tolerance requirements of the spe- HTS wires are available in several formats and chemical
cific case. Additional copper wires or tapes may need to makeup. In general, the types of wire with application
be added to the outer neutral layers of HTS wires in for power cables are referred to as first- and second-
cold-dielectric cables. generation wire: 1G and 2G.
1G Wires
Energy deposited into the system during a given fault-
current, I, event may be calculated using Equation 7-10 These wires are manufactured with a powder-in-tube
based on the cable resistance, R. method by which the superconducting powder is packed
into silver-alloy tubes. These tubes are drawn and
t =τ
t =τ mechanically worked to reach a final flat tape shape.
E = R∫ I 2 dt = ∫ VIdt (MW) 7-10 The ceramic superconducting material is bismuth stron-
t =0
t =0 tium calcium copper oxide (BSCCO-2223). These tapes
Where are commercially available from multiple vendors in
τ = Maximum duration of fault current (s). lengths in excess of 3280 ft (1000 m). Standard wire
I = Maximum magnitude of fault current dimensions are approximately 157 mils (4 mm) wide by
(kA). 7.8 mils (0.2 mm) thick.
V = System voltage (kV).
2G Wires
R = Resistance (Ohms).
E = Energy (mega-Watts). These wires are made by placing multiple thin-film lay-
ers of buffer materials and superconducting materials
System Thermal Impact on metal alloy films or foils. 2G wires are also called
coated conductors. The ceramic superconducting mate-
Hydraulics of Pumping LN
rial is yitrium barium copper oxide (YBCO). The final
The fluid mechanics of pumping liquid nitrogen long
tape thickness is much thinner than the 1G counterpart
distances must be considered and may affect cable and
and therefore promises to yield a much higher engineer-
cryostat design. Adequate cross-sectional area must be
ing current density (Je). At the time of publication,
allocated as a flow path for the liquid. Pressure drops
2G wires are being fabricated in lengths up to 328–656 ft
must be controlled to ensure proper thermal capacity
(100–200 m), with critical current values approximately
and heat removal at the end of a cable run, away from
50–75% of that available from 1G wires. 2G wires have
the cooling skid. These hydraulic issues affect the ther-
the promise of higher critical current density and better
mal design of the cable and its ability to ride through
in-field performance at a lower price than 1G. 2G wires
faults with minimal recovery times.
are made in wide strips and slit to the desired width.
Separate Return Line vs. Bidirectional flow vs. 2x1 Flow
Three Boundary Conditions (Temperature, Magnetic Field,
One variable to consider when designing the fluid flows
Current Density)
through the system is the layout for “go” and “return”
Three boundary conditions limit the ability of these
paths for the liquid to flow out through the system and
HTS materials to actually become superconducting.
then return to the cooling skid. Cable designs with one
HTS wires have critical temperatures, magnetic field sat-
cryostat such as the Triad or Triax may require a sepa-
uration, and current density levels. Each of these prop-
rate return pipe or cryostat. If the cable core is hollow,
erties must be maintained below its critical level for the
one option to eliminate the extra return line is bidirec-
HTS wires to transport current through a power cable
tional flow. LN may travel one direction through the
with near-zero resistance.
annulus around the conductor core and back in the
reverse direction through the hollow former.
7-29
Wire Properties for Cable Applications and fluid-filled cables may be used. Laminated paper
Typical minimum values for critical tensile stress and polypropylene (LPP) is also an option. Some manufac-
strain are 14,500 psi (100 MPa) and 0.3%, respectively. turers have developed synthetic lapped dielectric materi-
These requirements are set at liquid nitrogen tempera- als specifically tailored for use in LN.
ture (-321°F, -196°C, 77 K) at one atmosphere pressure.
Minimum bending diameter varies depending on the Cryostats
requirements of a specific manufacturer or cable design, The cryostat provides the thermal insulation. It consists
but may be as low as 2 in. (50.8 mm) for 1G and 1 in. of two metal corrugated tubes inside one another. Stain-
(25.4 mm) for 2G. less steel has been the most common material for this
application, but aluminum tubes have also been utilized
HTS cables typically operate such that the HTS wire is in Korea. Between the two tubes is wrapped metallized
immersed in pressurized LN. To prevent penetration of mylar, commonly known as “superinsulation.” Spacers
LN into the wires, the wire must be liquid impervious are placed between the tubes to maintain the proper
and hermetically sealed. Penetration of subcooled and spacing. A high-level vacuum is then pulled in the gap-
pressurized nitrogen has the potential to balloon the space to create a high-quality thermal barrier to prevent
wire and cause catastrophic damage when the nitrogen heat penetration into the cold space inside the cryostat.
expands during a rapid warmup. HTS wires must be cer- Performance of 0.0254–0.0381 W/in. (1.0–1.5 W/m) is
tified as hermetically sealed through pressure tests for common in cryostats of the size range of interest to
extended time durations at a pressure above the maxi- allow placement of the HTS cable inside.
mum operating pressure of the system.
These vacuum-insulated, flexible pipes are common for
Mechanical tolerances and hermeticity may be achieved transport of cryogens in the industrial gas, aerospace,
by various methods. Some wire manufacturers achieve and other industries. For cable applications, the stainless
these limits through the design of the HTS wire and pro- steel or aluminum tubes may be formed around the con-
cessing steps. Other manufacturers laminate the bare ductor if desired.
HTS wires with metal strips such as stainless steel, cop- Former Options
per, or brass to achieve these limits. The central member of the HTS cable is called a former.
The former may have multiple functions depending on
Critical current of a given wire is measured as the dc the exact design of a given manufacturer. In its most
current at which the wire develops 2.54 μV/in. (1 μV/cm) basic roll, the former serves as a structural member on
at -321°F (-196°C, 77 K) and 1 atm pressure. This repre- which the HTS wires are wrapped and provides strength
sents the common definition of the transition point for pulling and mechanical stresses.
from the superconducting state to the normal state for a
wire. It is at this point that the critical current density of Some manufacturers use stranded copper as the former.
a given wire is exceeded. Additional current applied to In this case, the former also is used as the fault current
the wire results in normal behavior and resistive heating path to protect the HTS wires from damage. In other
of the wire. If splices are permitted in delivered continu- cases a hollow, corrugated metal tube has been utilized.
ous lengths of wire, the splice resistance at -321°F Here the former may also serve as a flow-path for the
(-196°C, 77 K) and 1 atm is typically on the order of LN coolant.
100–200 nano-ohms.
Cryogenic Systems
Dielectric Options The cryogenic cooling plant serves the main function of
Warm-dielectric cables utilize conventional dielectric cooling the cable to its nominal operating temperature
and semiconducting materials extruded on the outside and pressure and removing heat from the system. The
of the cryostat. EPR and polyethylene are suitable for system must include a circulation pump and heat
this application. exchanger. Typical HTS cable operating conditions
range from -337 to -324°F (68–75 K) and 5–15 atm. The
Cold-dielectric cables must employ a lapped dielectric method of cooling, or heat removal, may be an open
that is suitable in pressurized LN and capable of many cycle, where boiling nitrogen is vented to atmosphere, or
thermal cycles from ambient conditions to -321°F a closed cycle, where mechanical refrigerators are used
(-196°C, 77 K) and back to ambient. Common kraft to provide cooling. Each method has pros and cons that
paper of familiar application in conventional pipe-type should be considered.
7-30
Cryogenic cooling plants must be located in a substa- expansion. Each method or cycle has unique advantages
tion at one or both terminations of the cable. Single- and disadvantages. Although the details of each are out-
point cooling (liquid flows out and back to the same side the scope of this document, each cycle uses a
cooling plant) is practical for circuit lengths up to method of expansion and/or compression of gases to
approximately 1 mile (1.6 km). Longer distances may produce cooling. Detailed discussions of cryogenic cool-
require additional cooling or pumping stations. These ing and engineering may be found in many references
additional stations may be distributed along the circuit, including (Flynn 1997).
or concentrated at the far termination. Limits on dis-
Accessories
tance between cooling/pumping plants are driven pri-
marily by the hydraulics of pumping liquids over long Splices, Cable Joints
distances and allowable system changes in temperature Superconducting cable joints are used to join multiple
for a given design. A larger cross-sectional area available lengths of cable together such that circuit lengths in
inside the cryostat for liquid flows will allow for longer excess of what is possible to put on one reel may be
distances between cooling/pumping stations. implemented (see Figure 7-31). The cables are joined
together to allow all layers of the cable structure to be
Open-cycle Cooling
connected at the normal operating temperature and
This method uses the boiling of LN under a partial
pressure for the superconducting materials. Supercon-
vacuum to reduce the temperature of the liquid. The
ducting tapes may be soldered together with a minimum
LN flow stream that goes to the cable is run through a
1 in. (25.4 mm) overlap joint. The resulting HTS-tape-
heat exchanger in the subcooled LN bath, thereby
to-HTS-tape joint has resistance on the order of several
removing heat.
nano-ohms. Dielectric tapes in cold-dielectric cables are
wrapped to control electrical stresses.
Open-cycle cooling has the advantage of not using
mechanical compression refrigerators operating at cryo-
Across the HTS cable joint, the cryostats must also be
genic temperatures. The cooling is provided by simple
joined. Multiple sections of vacuum-jacketed stainless
boiling of LN at a partial vacuum, which is achieved by
steel pipe can be connected, and a new vacuum insu-
common industrial vacuum pumps. This feature makes
lated section created to provide thermal insulation.
open-cycle cooling relatively simple and highly reliable.
The disadvantage of open-cycle cooling is the high Terminations
usage rate of LN. The liquid boiled under vacuum is HTS cable terminations must incorporate a material
vented to atmosphere and must be replenished regularly. transition from HTS to copper to allow the power to be
brought out to ambient conditions, where connections
Closed-cycle Cooling
to conventional station equipment are made (see
Closed-cycle cooling techniques employ a mechanical
refrigerator to recool, or recondense nitrogen. The act
of expansion during a compression-cycle absorbs heat
and provides cooling. The cooling may be used to recon-
dense evaporated nitrogen to a colder temperature, or
applied to a heat exchanger to directly cool the LN flow
to the cable.
Closed-cycle cooling has the advantage of not venting

large quantities of nitrogen to the atmosphere. However,
the electric power consumption of the mechanical
refrigerators may be higher than that of open-cycle
methods. Maintenance and reliability are also concerns
for machines with moving parts operating at cryogenic
temperatures.
Several methods exist for achieving closed cooling such Figure 7-31 HTS cable joint built in underground vault
as Stirling, Gifford-McMahon, Pulse Tube, and turbo (courtesy Sumitomo).
7-31
Figure 7-32). In addition to material transitions, the ter- HTS cables should be designed such that adequate
mination must provide for the thermal transition from structural support is included in the cable components
the normal operating temperature of approximately to withstand the mechanical stresses during installation.
-321°F (-196 °C, 77 K) to ambient. Cables with stranded copper cores as formers may
depend on this copper member as the means to pull the
Dielectric stresses must be controlled along these mate- cable. Conductors with hollow formers must utilize the
rial and thermal transitions. This presents challenges of hollow tube or other factors. Cables that have the cry-
finding suitable dielectric materials that can operate in a ostat formed over the conductor and delivered to the
liquid nitrogen environment and undergo extreme ther- job site as a single unit have the option to transfer more
mal transitions and in some designs thermal gradients of the pulling stresses to the cryostat and reduce pulling
along the given part. Thermal contraction of materials tensions on the HTS cable itself. Conventional wire or
must be closely matched with other components such nylon mesh cable-pulling grips may be used to secure
that differential expansion and contraction of compo- the pulling rope to the cable or cryostat.
n e n t s d o n o t c a u s e m e c h a n i c a l d a m ag e d u ri n g
cooldown or warmup of the system. Materials com- In addition to pulling tensions, the HTS conductor must
monly available as catalog items to be used for cryogenic be designed to withstand the sidewall bearing pressures
dielectric applications include glass-fiber-reinforced created through the wall of the cable. Cold-dielectric
polymer composites such as G-10 and Ultem® (James et cables must be designed and manufactured such that
al. 2006). these compressive stresses on the radial thickness of the
lapped dielectric tapes do not adversely affect their per-
7.4.4 Installation formance. Maximum compressive stresses must also be
considered on the HTS wires, and care should be taken
Cable Pulling and Mechanical Issues to ensure these forces do not exceed the tolerance of the
Installation of superconducting cables into under- superconducting materials. The absolute value of com-
ground duct banks is achieved using much the same pressive stress limits on the HTS wires vary depending
techniques employed for conventional cables. Pulling on the material (BSCCO vs. YBCO) and manufacturer
equipment, such as winches and tuggers, are used to pull of the wire.
the cable and/or cryostat into ducts (see Figure 7-33).
Care must be taken not to exceed the maximum pulling Thermal Independence from Surrounding
tensions for these cables because superconducting wires Environment
and cryostat thermal envelopes cannot sustain the same Installation of superconducting cables has the advan-
level of longitudinal stresses as conventional cables. tage of independence from the thermal characteristics of
However, since the weight of HTS cables is substantially the surrounding environment. Unlike conventional
reduced, pulling tensions should be lower for compara- cables that rely on the thermal resistivity of the soil to
ble pulls. dissipate heat generated due to resistive electric losses in
the cable, HTS cables have a self-contained thermal
environment created by the cryostat. HTS cable installa-
tions do not require special thermal backfill or other
Figure 7-32 Cable termination, AEP Bixby station, Ohio Figure 7-33 Pulling of HTS cable into cryostat during
(courtesy Southwire). installation (courtesy Southwire).
7-32
measures to ensure proper heat dissipation. Because Some parts of existing XLPE or paper cable standards
HTS cables do not require heat to be dissipated through apply to testing and requirements of HTS cables such as
the soil, depth of burial or directional boring is also impulse tests for specific voltage levels. Discrepancies
independent of the cable rating and adds a degree of between XLPE and paper cable standards with regards
freedom for installation parameters. HTS cables may be to ac withstand tests, partial discharge, and other
placed deep in the ground if required to go under such requirements make determining the proper test require-
obstacles as freeways, roads, train tracks, or buildings. ments vague and unclear for the HTS cable manufac-
turer. Parts of these standards simply do not apply, such
HTS cables are also not adversely affected by paralleled as thermal cycling of cables. In addition, new parame-
or intersected steam lines, water lines, or other electrical ters unique to superconducting cables need to be
circuits. Ampacity ratings of HTS cable do not need to addressed, such as the critical current, thermal stability
be derated due to these factors. Extreme exterior heat at operating temperature and pressure, and pressure
loads such as steam lines may add an incremental ther- containment and relief of the cryostat and termination
mal load for the cooling system that must be accounted enclosures.
for in thermal budgets, but these factors do not detract
from the cable performance. All U.S.-based in-grid demonstration projects ongoing
at the time of publication have made every effort to test
7.4.5 Standards cable and accessories to applicable portions of IEEE,
IEC, AEIC, and ICEA standards for ac withstand, BIL,
No HTS Cable-specific Standards partial discharge, and other requirements.
At the present time, no standards exist that specifically
address superconducting cables. The IEEE, IEC, AEIC, At this time, close communication and cooperation are
and other organizations have not written such stan- required between the manufacturer and end user to
dards, although working groups are investigating these ensure a level of comfort with qualification, plant, and
issues. As HTS cable technology evolves, and more installation tests.
projects are installed in utility grids around the world,
common testing parameters and requirements will need
to be set up to ensure performance and safety.
7-33
REFERENCES CIGRE. 1977 Electra No. 171. “Recommendations For

AEIC. 1990. CS2-90. Specification for Impregnated Mechanical Tests on Submarine Cables.”
Paper and Laminated Polypropylene Insulated Cable
High-Pressure Pipe-Type. CIGRE. 2000a. Electra No. 189. “Recommendations
for Testing Long AC Submarine Cables with Extruded
AEIC. 1993. CS-7. Specifications for Crosslinked Poly- Insulation for System Voltage Above 30 (36) to
ethylene Insulated Power Cables Rated 69–138 kV. 150 (170) kV.”
Third Edition.
CIGRE. 2000b. Electra No. 189. “Recommendations
Apland, G., et al. 1998. “DC Transmission based on for Tests of Power Transmission DC Cables for a Rated
Voltage Source Converters.” Report No. 14-302. Voltage up to 800 kV.”
CIGRE 1998 Plenary Session. Paris.
CIGRE. 2003. Technical Brochure No. 219. “Recom-
Arkell, C. A., and B. Gregory. 1984. “Design of SCFF mendations for Testing DC Extruded Cable Systems for
Cables for UHVDC Transmission.” CIGRE Paper Power Transmission at a Rated Voltage up to 250 kV.”
21-07. Paris.
Demko, J. A., et al. 2003. “Cryostat Vacuum Thermal
Arkell C. A, E. H. Ball, K. J. H. Hacke, N. W. Water- Considerations for HTS Power Transmission Cable Sys-
house, and J. B. Yates. 1986. “Design and Installation of tems.” IEEE Transactions on Applied Superconductivity.
the UK Part of the 270 kVDC Cable Connection Vol. 13. pp. 1930–1933.
Between England and France Including Reliability
Aspects.” CIGRE Paper 21-02. Paris. and Demko, J. A., et al. 2006. “Triaxial HTS Cable for the
AEP Bixby Project.” IEEE Applied Superconductivity
Arnaud, U., A. Bolza, F. Magnani, and E. Occhini. Conference. Seattle. 2LA06.
1992. “Long Island Sound Submarine Cable Crossing.”
CIGRE Paper 21-306. Paris. EPA. n.d. U. S. Environmental Protection Agency. SF6
Emission Reduction Partnership for Electric Power Sys-
Balog, G. E., and G. Evenset. 2002. “Integrated Return tems. Memorandum of Understanding.
Conductor Cable.” CIGRE Paper 21-301. Paris. www.epa.gov/electricpower-sf6/.
Bell, N., et al. 1992. “± 450 kV DC Underwater Cross- Eriksson, K., G. Hjalmarsson, and K. Lampe. 1993.
ing of the St-Lawrence River of a 1500 km Overhead “Baltic Submarine HVDC Link is the World’s Long-
Line with Five Terminals.” Report No. 21-301. CIGRE est.” HVDC Transmission Magazine. pp. 47–55.
1992 Plenary Session. Paris.
Flynn, T. M. 1997. Cryogenic Engineering. New York,
Bex, S. et Al. 2006. “Basslink HVdc Design Provisions New York: Marcel Dekker.
Supporting ac System Performance.” CIGRE paper
B4-301. Paris. Forsyth, E. B. and R. A. Thomas. 1986. “Performance
Summary of the Brookhaven Superconducting Power
Bjurstrom, B. 1968. “Progress Report of Study Commit- Transmission System.” Cryogenics. Vol. 26.
tee No. 2 (HV Cables) – Appendix II – Recommenda- pp. 599–614.
tions for Tests on DcC cables for a Rated Voltage up to
300 kV.” CIGRE Paper 21-01. Paris. Giorgi A. et. al. 2002 “The Italy-Greece HVDC Link”
CIGRE Paper 14-116. Paris.
Carcano, C. et al. 1996. “Qualification Program for the
400 kVdc Deep Water Interconnection Between Italy Hjalmarsson, G. et al. 1992. “After-service Analysis of
and Greece.” CIGRE Paper 21-304. Paris. the 32 year old HVDC Cable Gotland 1.” CIGRE
Paper 21-302. Paris.
Cho, J., et al. 2006. “Design and Experimental Results
of a 3 Phase 30 m HTS Power Cable.” IEEE Transac- Honjo, S., et al. 2003. “Electric Properties of a 66 kV 3-
tions on Applied Superconductivity. Vol. 16. No. 2. Core Superconducting Power Cable System.” IEEE
pp. 1602–1605. Transactions of Applied Superconductivity. Vol. 13.
pp. 1952–1955.
7-34
IEC. 1993. Publication 141-1. “Tests on Oil-filled Nakanishi, Y. et al. 2000. “Installation of 500 kV DC
Cables and their Accessories.” Submarine Cable in Japan.” CIGRE Paper 21-304.
Paris.
IEC. 2004a. Publication 60480. “Guidelines for the
Checking and Treatment of Sulfur Hexafluoride (SF6) Normark, B. 2002. “Cross Sound Goes Underground.”
Taken from Electrical Equipment and Specification for PEI Magazine. pp. 22–25. December.
its Re-use.”
Ruter, F., J-L Parpal, and S. G. Swingler. 2000. “A
IEC. 2004b. Publication 60840. “Power Cables with Review of HVDC Extruded Cable Systems.” CIGRE
Extruded Insulation and Their Accessories for Rated Panel Session. Paper 2-03. Paris. August.
Voltages Above 30 kV up to 150 kV.”
Satoru, M. et al. 1999. “Development of XLPE Cable
IEC. 2006a. International Electrotechnical Commis- Under DC Voltage.” Paper B7.3. JiCable Conference.
sion. IEC Standard 62271-100. High-voltage Switch- France.
gear and Controlgear.
Sauers, I., et al. 2004. “High Voltage Breakdown, Par-
IEC. 2006b. Publication 62067. “Power Cables with tial Discharge and Aging in Lapped Tape Insulated
Extruded Insulation and Their Accessories for Rated Cold Dielectric Model Cables.” Advances in Cryogenic
Voltages above 150 kV (Um = 170 kV) up to 500 kV Engineering. Vol. 50. pp. 889–895.
(Um = 550 kV)—Test Methods and Requirements”.
Shapiro, S. et al. 1997. “Threats to Submarine Cables.”
IEEE. 2005. Std. 1120™. “IEEE Guide for the Planning, pp. 742–749. Sub Optic Conference. San Francisco.
Design, Installation, and Repair of Submarine Power
Cable Systems.” IEEE PES – ICC Committee. March. Stovall, J. P, et al. 2001. “Installation and Operation of
the Southwire 30-meter High-Temperature Supercon-
James, D. R., et al. 2006. “Breakdown and Partial Dis- ducting Power Cable.” IEEE Transactions on Applied
charge Measurements of Some Commonly Used Dielec- Superconductivity. Vol. 11. pp. 2467–2472.
tric Materials in Liquid Nitrogen for HTS
Applications.” IEEE Applied Superconductivity Con- Strandem, A. and J. E. Larsen. 1992. “Installation of
ference. Seattle. 3LF06. Skagerrak 3—A 350 kVdc Link Between Norway and
Denmark.” CIGRE Paper 21-308. Paris.
Jue, J. 1992. “Report on a Test of a 260 kV HVDC Sub-
marine Cable.” Paper Presented at the IEEE-ICC. Tonnesen, O., et al. 2004. “Operation Experiences with
a 30 kV/100 MVA High Temperature Superconducting
Last, F. H. et al. 1970. “The Underground HVDC Link Cable System.” Superconducting Science Technology.
for the Transmission of Bulk Power from the Thames Vol. 17. pp. S104–S105.
Estuary to the Center of London.” Paper 71 TP
181-PWR. IEEE Winter Power Meeting. New York. Weber, C., et al. 2006. “Albany Cable Project Progress
Update.” U. S. Department of Energy Superconductiv-
Minemura, S. et al. 1980. “250 kV Direct Current Sub- ity Program Annual Peer Review. Washington, DC.
marine Cable for Hokkaido–Honshu Link.” CIGRE
Paper 21-03. Paris. Xin, Y., et al. 2005. “Introduction of China’s First
Live Grid Installed HTS Power Cable System.” IEEE
Miranda, F. J., and P. Gazzana-Priaroggia. 1976. Transactions on Applied Superconductivity. Vol. 15.
“SCFF Cables–A Review of Progress.” Proc. IEE. pp. 1814–1817.
Vol. 123. pp. 229–238.
7-35
CHAPTER 8 Cable Systems

Accessories
Authors: Pierre Argaut, Silec Cable
Milan Uzelac, G&W Electric Co.
Ernesto Zaccone, Prysmian Cables and Systems
Reviewer: Steven P. Walldorf, USi (Underground Systems Inc.)
High-voltage and extra-high-voltage power cable systems form part of the backbone of
transmission systems, and therefore, reliability considerations are of the highest priority.
Accessories (joints and terminations) are very important components of the cable system.
Their behavior affects the full reliability of the whole system.
High-voltage transmission cable systems operate at elevated electric stresses. Further,

the thermal performance of joints and terminations and installation factors can limit
the ampacity rating of the overall cable system. As a result, the appropriate design,
qualification, and the right selection of the most suitable accessories are of
fundamental importance.
This chapter provides details on the types of accessories used for the most common cable
types in use today. It provides guidance on the right selection of accessories as a function
of their application and describes the qualification and testing according to the relevant
national and international standards.
Pierre Argaut has been working with Silec Cable since 1988 and is cur-
rently Vice President in charge of R&D. He received a degree in Electri-
cal Engineering in 1971 from the University of Grenoble (France).
Most of Mr. Argaut's career has been in the HV-EHV area, first in the
EHV switchgear and GIS Industry from 1971 to 1980, and then in the
cable industry from 1988 to the present. From 1980 to 1988, he worked
as Operation Manager of SEPL (South European PipeLine), the largest
crude oil pipeline in western Europe. Mr. Argaut is currently a Member
of the Strategic Advisory Group and Convener of the Tutorial Advisory Group of
CIGRE Study Committee B1 (Insulated Cables), and is involved in several Discussion
Groups within ICC.
Milan Uzelac graduated from the Electro-technical University of Bel-

grade, Serbia, in 1968 with a major in Electrical Power Engineering. He
was employed with Minel-Elektrooprema, Belgrade, as a design engi-
neer, senior design engineer, and the head of R&D Department. His
achievements include development of a line of vacuum circuit breakers
from 12 to 36 kV and air-break disconnecting switches from 72.5 up to
420 kV. Milan relocated to the United States in 1989 and continued his
career with the G&W Electric Company of Chicago as an R&D engi-
neer, product engineer, and chief R&D engineer. His responsibility has been the develop-
ment and design of accessories for high-voltage and extra-high-voltage cables, both with
extruded and impregnated paper insulation. His expertise includes analysis of electrical
field and thermal studies with evaluation of ampacity ratings of cable accessories. Mr.
8-1
Chapter 8: Cable Systems Accessories EPRI Underground Transmission Systems Reference Book
Uzelac is involved in IEEE and CIGRE working groups projects in high-voltage testing laboratories and high-
for developing industry standards for high-voltage cable voltage cable system design. In particular, he was
accessories. He currently chairs the IEEE Working involved in the study, design, and realization of some
Group on revising the IEEE 1300 Standard for cable major EHV (Extra High Voltage) extruded cable
connections in gas-insulated switchgear. projects. He was also the author of several studies on
the application and impact of EHV cables when
Ernesto Zaccone is currently with the integrated in transmission systems. Currently he is
R & I ( Re s e a rc h a n d I n n ovat i o n ) involved in a number of innovative underground cable
department of Prysmian Cables and projects and in international standardization activities.
System (formerly Pirelli Cables) based Mr. Zaccone is the Chairman of the CEI CT 20 Italian
in Milan, Italy. He obtained a degree standardization body for power cables. He is the Co-
in Electrotechnical Engineering in convenor of the IEC TC20 WG 16, which is responsible
1969. During the same year, he joined for the international standardization of high-voltage
the power cable industry, where he cables and their accessories, and is an active member of
was engaged in a number of different activities such as CIGRE and IEEE.
8-2
EPRI Underground Transmission Systems Reference Book Chapter 8: Cable Systems Accessories
8.1 INTRODUCTION IEC 60840, and IEC 62067 for extruded-dielectric

cable systems.
This chapter provides theoretical design and practical
information on cable system accessories—namely joints • Depending on the application, the user may purchase
(also called splices) and terminations (also called pot- cable, accessories, installation design, and field instal-
heads or sealing ends). Also included is information on lation separately or as a complete cable system. In
other accessories: sheath-bonding systems, metallic either case, compatibility of all the components must
screen/sheath (also called shield) sectionalizing, link be ensured—whether by the “system provider,” the
boxes, sheath voltage limiters (also called surge divert- user, or a third-party integrator.
ers), and bonding leads. • For extruded-dielectric cable systems above 170 kV,
IEC 62067 details the testing that has to be per-
Section 8.2 covers joints; Section 8.3 reviews termina- formed to demonstrate the compatibility of cable and
tions; and Section 8.4 discusses other accessories. accessories, and stipulates the range of approval of
type testing and prequalification testing. CIGRE
Chapter 8 discusses differences in joints and termina- (CIGRE 2006) prepared the recommendations for
t i o n s fo r t h e va ri o u s c abl e t y p e s, a n d p rovid e s additional testing to take into account the evolutions
information to help the design engineers choose the of an already prequalified cable system, and espe-
optimal joint or termination design. It also includes test cially the changes in the accessories.
requirements for prequalification of the accessories
according to the most relevant national and interna-
8.2 JOINTS
tional standards.
8.2.1 Theory and General Description
Related information is covered in other chapters. Chap-
ter 7 discusses submarine cable accessories in more Function
detail. Chapter 12 provides additional information on The basic function of a joint is to join cable sections.
metallic screen/sheath bonding on cable systems. Joints are required when the length of the circuit
exceeds the allowable continuous cable lengths, which
The following general points can be made at the outset: may be limited by manufacturing, shipping, or installa-
• The performance of a cable circuit is dependent on tion considerations. To fulfill its purpose, a joint must
several factors, including: provide a robust conductor connection surrounded by a
dielectric system that meets the same performance
- Performance and reliability of the cable requirements as the cables and within a suitable ground-
- Performance and reliability of any single acces- ing enclosure or covering. Joints are installed in man-
sory holes (vaults) or tunnels, or they are directly buried in
- Skill and experience of the persons assembling trenches. They often perform one or more of the follow-
the accessories onto the cable ing additional tasks:
- Suitability of the installation design to take into • Isolate and insulate metallic screen (sheath/shield)
account mechanical, thermomechanical, ther- sections (Miranda 1976). See discussion on screen
mal, and electrical constraints of the particular interruption joints in Sections 8.2.2 and 8.4.1.
cable system. • Provide a hydraulic barrier to sectionalize hydraulic
• Cables and accessory materials are often manufac- systems (see pipe-type cable stop joints in Section
tured under controlled process conditions. Cables are 8.2.3; self-contained cable stop/feed joints in Section
factory-tested before shipping, while only prefabri- 8.2.4; and transition joints between oil- or gas-filled
cated accessories can be factory-tested in the same cables and extruded cables in Section 8.2.2).
way. However, factory testing of accessories does not • Mechanically support or restrict the cable as required
in itself guarantee that an installation will be
to support the weight of vertical lengths of cable or to
successful. There are many installations of both
limit the expansion and control the bending of cable
extruded and taped cable systems without factory-
when appropriate.
tested accessories that have been successfully in
service for many years. Design Approach
• Compatibility of cable and accessory design in the The ideal joint for any cable type would result in no
particular service application should be validated by mechanical, thermal, or electrical discontinuity. Such a
appropriate electrical tests as specified in U.S. and joint would represent a reconstituted cable, which would
international standards such as IEEE 48, IEEE 404, have the same dimensions, thermal properties, and elec-
8-3
trical properties as the cable. Such joints are technically cally use hand-taped joints or prefabricated (epoxy)
possible, and in fact are sometimes constructed to fac- joints for special applications.
tory-join submarine cable lengths or to make repairs of
Joints for Extruded-dielectric Cables
submarine cables when no increase in diameter can be
Taped joints adopt elastomeric semiconducting and
tolerated for mechanical reasons. However, they require
insulating self-amalgamating tapes to rebuild the con-
a long cable length for pencilling and replacing the insu-
ductor screen, insulation, and insulation screen. The
lation; they take longer than conventional splices to
construction of these joints is conceptually similar to
construct; and they require the highest level of skill to
the hand-taped joints used for laminar-insulated cables
install (Miranda et al. 1976; Takaoka et al. 1987).
and is limited to the HV range due to the poorer dielec-
tric strength than cable insulation—and hence large
For practical reasons, limitations on length and cost
diameter—and the risk of leaving voids between tapes.
usually dictate joints of larger diameter than the cables
that they join. Increasing the radial dimensions simpli-
Field-molded joints are typically used for high-voltage
fies the design and construction of suitable connectors
extruded cables. In the most common type (see Figure
and lowers the operating stresses in the dielectric, when
8-1), the joint is built by taping over the connector and
properly done. However, the increase in the radial
cable chamfer with unvulcanized tapes. The tape may be
dimensions of the joint and its individual parts intro-
applied by hand or by machine. Molding is accom-
duces longitudinal components of the dielectric stress.
plished after taping by placing a metal mold over the
Extruded or molded dielectrics can well tolerate such
taped buildup and heating the entire mass for several
longitudinal stress, but laminar or multilayer dielectrics
hours. The pressure generated by the thermal expansion
are weak when subjected to a stress acting along their
consolidates the taped material into a void-free mass,
interfaces. As a result, these stresses must be taken into
which is similar to cable dielectric (Hahm et al. 1988).
consideration during the design.
Other external means (hydraulic or pneumatic) can be
applied outside the joint to prevent the formation of
A larger-diameter joint can also cause a localized ther-
voids during the vulcanization process.
mal restriction on the cable system, so the thermal per-
formance of the joints must also be carefully considered
One other method places the insulation compound over
during the design of the cable system.
the connector and cable chamfer by extruding the com-
pound into a suitable mold. This technique is used in
For a given cable type, the joint design and choice of
Japan to joint 500-kV XLPE cables (Kubota et al. 1994;
materials may be influenced by the installation condi-
Takeda et al. 1995; Okada et al. 1999). After extrusion,
tions (buried, manhole, or tunnel) and the thermal char-
the joint is heated, and the insulation is pressurized with
acteristics of that installation environment.
nitrogen gas to crosslink the joint material.
Technologies
Extruded-dielectric cables may use taped, field-molded, Other types of joints, called prefabricated joints, adopt
heat-shrink sleeve, prefabricated/premolded or prefabri- premolded and/or thermoset resin components that are
cated/composite joints. Laminar-insulated cables typi- preformed and tested in the factory.
Figure 8-1 Field-molded joint for extruded-dielectric cables.
8-4
Prefabricated/premolded joints include one sleeve (see The joints are generally protected by a metallic shell
Figure 8-2) (Parmigiani et al. 1987; Leufkens et al. 1992; sealed to the cable sheaths by different methods (see
Chatterjee et al. 1995; Peschke et al. 1996) or more elas- following section entitled “Sheath Closures”). The joint
tomeric sleeves (Nelson et al. 1978). The pressure shell is isolated from the environment by different
between the sleeve and the cable, and between different systems described in the following section entitled
sleeve interfaces, is maintained by interference fit and by “Anticorrosion Protection”. However, if the extruded
the elastic memory of the sleeve. cable does not have a metallic sheath, but only a wire
shield, heat shrink tube protection is generally used for
In another design, called a prefabricated composite the joint.
joint, two factory-premolded elastomeric stress cones
Joints for Laminar-insulated Cables
are inserted into a central insulator of cast thermoset
Hand-taped joints were the first to be utilized and are
resin (see Figure 8-3). The pressure of the stress cone
the predominant type for laminar-dielectric cables
against the central insulator and the cable core is
(Short 1949). Their quality is highly dependent on the
ensured by a compression device, which is usually
skill of the splicer. Impregnated kraft paper or lami-
made of metallic springs (Nakanishi et al. 1995; Okada
nated paper polypropylene (LPP) insulation is applied
et al. 1999).
over the shielded connector and stepped (pencilled)
cable insulation. Insulation buildup at the edge of the
Conductors are joined by various designs of connectors
cable shielding is applied in the form of the stress cone
including compression, soldering, welding, and combi-
for proper shaping of the electrical field. The outer sur-
nations of these techniques (see following section enti-
face of the joint insulation is covered with semiconduc-
tled “Connectors”).
tive and metallic screen (shielding) that is eventually
connected to the ground. Some applications require
Figure 8-2 One-piece premolded joint for extruded-dielectric cables, 69-500 kV.
Figure 8-3 Section view of a prefabricated composite joint (courtesy of Prysmian Cables and Systems).
8-5
interruption of the screen (shield break), as shown in be defined in order to maintain the radial and longitudi-
Figure 8-4. nal stresses within safe values.
For single-conductor self-contained fluid-filled (SCFF) Part of the joint designer’s responsibility is the control
cables, each single-conductor joint is contained in a sin- of cable movement and support of the cable adjacent to
gle metallic housing, which may incorporate a sheath the joint. This involves controlling thermally-induced
insulator (Watanabe et al. 1986). For pipe-type cable, mechanical forces, which tend to concentrate stress
the three single-conductor insulated joints are contained within or adjacent to the joints.
in a single, welded steel casing (see Section 8.2.3, “Pipe-
type Joints”). Minimizing hotspots in the joint is a function of joint
materials and construction as well as the installation
Quartz-filled epoxy castings are utilized in prefabricated method. The flow rate also plays an important role in
or partly prefabricated joint designs (Gazzana- fluid-circulating cable systems. Thermal performance is
Priaroggia et al. 1956). The prefabricated component generally improved by reducing dimensions to their
may incorporate an embedded electrode with a diameter minimum consistent with dielectric requirements, or by
greater than the cable insulation to shield the end of the selecting thermally superior materials.
insulation and the connector area. In addition, precast
epoxy components can serve as a pressure barrier, which It is evident that these challenges are interrelated. The
can hydraulically isolate adjacent cable sections in pipe- use of higher-strength insulating material results in
type and SCFF cable systems and in transition joints. smaller joints, which operate at lower temperatures.
They are also used to provide a channel for feeding
Electrical Design
dielectric fluid to the conductor of SCFF cables. Precast
The joint design procedure normally starts after estab-
epoxy joints are also used to obtain a positive grip on
lishing the allowable stresses in the joint insulation
the connector and offer the means to anchor a cable by
materials and at interfaces based on breakdown tests on
its conductor, which is used to control axial motion and
models or full-size accessories.
to hang cables when large vertical drops are encoun-
tered. Although these prefabricated joints may be larger
The first approach utilizes the aid of a computer pro-
than taped joints, the epoxy components have low
gram to calculate the equipotential line distribution (see
thermal resistivity, and the temperature rise of the
Figure 8-5) and the actual stresses in the critical region
conductor and connector in the joint relative to that in
of the prefabricated joint—i.e., over the embedded elec-
the cable is minimized.
trodes and the interfaces between different materials. If
Design Challenges these stresses are higher than the allowable stresses, the
The use of joint materials with characteristics that are joint design is revised by changing some dimensions and
different than the factory insulation in terms of dielec- materials, and the process is repeated until the correct
tric constant and dielectric strength results in design stresses are obtained.
challenges. In fact, longitudinal components of the elec-
trical stresses acting along the interfaces of the lami- Analytical methods are applied for simple geometries
nated insulation (for taped joints) or of joint like for taped joints where the critical electrical stresses
components (for prefabricated joints) must be taken occur at the connector (radially, and axially along its
into account. The overall dimensions of the joint should taper), at the base of the stress cone (radially), or along
Figure 8-4 Hand-taped joint for laminated cables (courtesy of G&W Electric Company).
8-6
Figure 8-5 Equipotential line distribution in a premolded joint.

the interface between cable and joint insulation (axi- the ground or with close clamping). In fact, the unbal-
ally). Analytical formulas for computing these stresses anced thrust transmitted to the joint by the rigidly
can be found in References (Short 1949; Wakabayaski et installed cable can damage the joint unless special pre-
al. 1970). In addition, numerical techniques for solving cautions, like cable offsets, are taken to reduce the thrust
for the stresses in nonuniform electrical fields can be to the joints (CIGRE 2001b).
found in Reference (Vitkovitch 1966).
When cables are installed into pipes, the tendency for
The connector diameter and the dielectric strength of cable movement to concentrate at or near joints has
the joint material determine the diameter of taped
joints. A stress cone is then designed, taking the axial
dielectric strength of the joint insulation into account. A
shape described by a “log-log” taper (Short 1949) is
normally adopted in order to maintain constant the
axial stress along the outer stress cone profile. This
shape is applied to the design of flat paper or paper-
polypropylene laminate-insulated joints. If crepe paper
or self-amalgamating tapes are to be used, their
overlapping construction should not create axial
surfaces, except at the interface with the cable core.
The design of the insulation taper or pencil depends on Figure 8-6 Joint showing “log-log” outer taper and “log”
the relative dielectric constants of cable and joint insula- insulation taper.
tion. For taped cables, a low-dielectric strength tangen-
tial to the interface between cable and joint along the
surface of the pencil is assumed (Wakabayaski et al.
1970). By using a “log” taper design, it is possible to
make this stress constant over the length of the pencil.
Although not directly related to the axial stress in the
cable or joint material along the pencil, the use of a
“log” pencil shape has become standard practice in
North America (see Figure 8-6).
For extruded-dielectric cable molded joints, the dielec-

tric strength of the insulation is theoretically isotropic,
and shorter stress cones and pencils are possible.
Mechanical/Hydraulic Design
During their operation, joints are subjected to thermo-
mechanical forces caused by the current circulating in
the conductor. These forces also depend on the installa-
tion conditions of the cable on both sides of the joint.
The worst condition occurs when the joint is at the tran- Figure 8-7 Installation of 220-kV XLPE cable joints in a
sition between a flexible cable system (cable installed manhole (courtesy of Silec Cable).
with snaking) and a rigid cable system (cable buried in
8-7
been a major problem for all types of cable systems, but because the circulating fluid tends to maintain an axially
particularly for extra-high-voltage, pipe-type cable sys- uniform temperature at the surface of the joint shield that
tems installed in the 1960s and 1970s (Mcllveen et al. is the same as that in the adjacent upstream cable. Since
1978). This problem has been solved by using a rigidized the joint thermal resistance is greater than that of the
design or supplementary mechanical reinforcement cable, an increase occurs in the conductor temperature
and/or providing offsets for accommodating expansion within the joint relative to the adjacent cable. Designs to
in the adjacent cable (Minemura et al.1990). minimize over-temperature for hand-taped cable joints
are discussed in References (EPRI 1988; Allam et al.
Where pipe-type cable joints are installed in cable sec- 1986). Various joint designs for pipe-type cable are
tions with appreciable differences in elevation, special compared thermally in Reference (Engelhardt 1972).
“anchor” and “skid” joints have been utilized. Section Two-dimensional resistance network analogs have been
8.2.3 provides further discussion of these joints. employed to study splice conductor and connector tem-
peratures (Engelhardt 1972; Weedy et al. 1967).
If differences in elevation result in hydraulic pressures
exceeding the limitations of sheaths, pipes, or terminals, Auxiliary cooling of joints has been employed for self-
a full-stop joint, which provides a pressure barrier, can contained cable (Weedy et al. 1967), and such tech-
be used. Full-stop joints are also used for routine sec- niques could be applied to pipe-type or extruded cable
tionalization of SCFF cables when multiple pressurizing systems when the situation warrants the added complex-
and reservoir sections are required. In this case, precast ity (ELECTRA 1983, 1990b).
epoxy splices are used to both anchor the conductor
and provide hydraulic sectionalization for pipe-type Prefabricated joints, which employ cast epoxy compo-
cable systems. nents, or premolded joints, which incorporate a heat
sink device at the connector, tend to have a thermal
Thermal Design resistance comparable to that of the cable and provide
Joints generally have diameters greater than the cables the closest match to a cable’s thermal performance. This
that they join and therefore have a thermal resistance characteristic has been used to advantage in the layout
greater than the cable. This is normally offset by a lower of some heavily forced-cooled, pipe-type systems.
thermal resistance from the joint to ambient tempera-
ture, particularly for joints installed in manholes. When Although the joints of a heavily loaded transmission
cable and joints are directly buried, the overheating of cable can limit the normal loading capability of the cir-
the joints is avoided by increasing the spacing with cuit, it should be noted that the short-term emergency
respect to the cable. overload capacity of joints may be closer to that of the
overall cable system. This is due to the larger thermal
An important exception is for pipe-type cables when mass of the joint, which gives them a longer time con-
circulation or oscillation of the dielectric fluid is stant (see Figure 8-8) (ELECTRA 2004). Hence, their
employed. The movement of fluid, although designed to temperature rise when exposed to a step increase in
average out hotspots along the cable, has the effect of loading is slower than other parts of the system. How-
producing a hotspot at the joint (EPRI 1984). This is ever, hand-taped joints on laminated cables exhibit
Figure 8-8 Typical heating curve of a 132-kV extruded cable and premolded joint.
8-8
higher dielectric losses due to the inevitable moisture component of the electrical stress introduced by its
picked up during the taping operations. This causes a protrusion with respect to the conductor.
more sharply rising dielectric loss with increasing tem-
perature, and can lead to thermal runaway in the joint at The compression of the connector tube can be carried
a lower temperature than the cable itself can tolerate out by indentation with mono or multi ram presses or
(McIlveen et al. 1978). Operation of joints on extra- with hexagonal dies used with a hydraulic press. A duct
high-voltage cable systems above emergency tempera- pin is used with the hollow-core conductors used in
tures must be avoided. SCFF cables to ensure the continuity of the oil flow. For
compact copper conductors, a round compression can
Connectors be used with a copper tube presenting suitable humps.
The connector joins the conductors of the two cables As a variation of this process, a flush compression con-
electrically and mechanically. Thus the connector must nector can be obtained by stepping the conductor and
exhibit good electrical conductivity to avoid a higher the bore of the connecting tube. This design is normally
temperature than the cable's conductor in any operating adopted in submarine cable joints.
condition and present sufficiently high mechanical pull-
out strength to withstand mechanical stresses during Aluminum compression connectors are used for alumi-
operation. The following types of connectors are used num conductors (up to approximately 1250 kcmil
both for extruded and laminated cables. [~630 mm2]), while welding is preferable for larger cross
Compression Connectors sections. In fact, with aluminum connectors, a creep
A compression connector includes a tube (ferrule) of the phenomenon can take place as an effect of loading
same material as the cable conductor into which the cycles, leading to a reduction of the contact pressure
conductors to be joined are inserted (see Figure 8-9). between connector and conductor, with corresponding
The tube is then compressed by a hydraulic press tool. increase of the electrical resistance, and ultimately
resulting in thermal failure.
The compression (crimp) connector is the type most
commonly used (up to approximately 5000 kcmil When a copper conductor has to be joined to an alumi-
[~2500 mm2] cross sections), because it is easy to install num conductor, a special bimetallic connector is used.
and does not require heat. The cross section of the These connectors are half copper and half aluminum.
connector is equal to or slightly less than the cross The two connector halves are joined in the factory by
section of the conductor that it is joining. When the friction welding or high-speed inertial welding.
connector is exposed to the electric field, like in taped
Mechanical Bolted Connectors
joints, it is necessary to provide suitable chamfers at
In mechanical bolted connectors, the compression of
both ends in order to minimize the longitudinal
the conductors inside the ferrule is made by tightening
threaded bolts. The bolts are then finished flush with the
surface of the connector (see Figure 8-10).
Mechanical connectors, extensively used in MV accesso-

ries, can also be used in HV joints, but only for relatively
small conductor cross sections (up to about 1250 kcmil
[~630 mm2]), due to their limited short-circuit current
Figure 8-9 Hydraulic press tool on a compression

connector.
Figure 8-10 Mechanical bolted connector.
8-9
capability. These connectors have a larger diameter than ductors, a mixture of combustible material and pow-
compressed connectors. dered metal that melts and fuses around the two cable
conductors in an exothermic reaction. This weld is not
Soldered or Brazed Connectors
flush and requires additional shaping in the field. Care
To make a soldered or brazed connector, a metal tube
must be exercised in cleaning the conductors, especially
with a top cut is placed around the two conductors, and
conductors with filled and/or insulated strands; other-
it is filled with a melted lead alloy (melting temperature
wise, there is a risk of obtaining a porous weld.
about 190°C) or a high-temperature melting alloy.
Plug-in Connectors
Soldered connectors are used with small conductor Plug-in connectors consist of two metal connectors that
cross sections (below approximately 1250 kcmil terminate the conductor and that are mated through the
[~630 mm2 ]) and with paper-insulated cables having a interposition of elastic or spring-loaded contacts able to
short-circuit current temperature below 160°C), because carry the current. Locking pins can be used to anchor
the solder can become soft during the cable system the two parts together.
operation. Brazed connectors do not present this prob-
lem but are more difficult to install. Plug-in connectors are used in SCFF cable stop joints
and for extruded cable plug-in type premolded joints
Welded Connectors
and back-to-back joints.
For welded connectors, the two conductors are fused
together by the application of molten metal. The Metal Metallic Sheath Closures
Inert Gas (MIG) or Tungsten Inert Gas (TIG) welding The connection between the cable sheath and the joint
process is adopted in this case. shell must fulfill the following functions:
• contain the fluid pressure for HPFF, SCFF, HPGF,
The welding process is adopted for large aluminum con-
and fluid-filled polymeric accessories. For dry
ductors and for enameled wire copper conductors. In
extruded cable accessories, the seal must be tight so
fact, the burning of the enamel ensures a good contact
that water is not sucked into the accessory during a
among strands. This technology requires a highly skilled
cable-cooling cycle.
operator and is time consuming. Due to the high tem-
perature developed during the process, special cooling • conduct the capacitive and short-circuit current of
clamps are required on both sides of the weld in order the cable system
not to damage the cable insulation. In the case of fluid- • withstand the sheath thermomechanical loads in ser-
filled cables, vacuum is applied to the clamps in order to vice.
remove the oil from the strands and simultaneously cool
the area. In case of an extruded-dielectric cable, the Three main methods, or combinations of them, are
clamps may be cooled with circulation of air or water. used—plumbing, welding, and mechanical clamping.
The conductors to be joined are cut in a wedge shape, Plumbing

and then buttered with weld metal, and the two “V” Plumbing is the most common method used for con-
grooves are filled until the conductor diameter is necting the joint shell to the lead, copper, or aluminum
reached. As a result, this connection is flush with the sheath of cables. The cable sheath is tinned to remove
conductor. The weld provides an electrical connection the oxide layer. In case of an aluminum sheath, this pro-
that is equivalent to the conductor and that is not sub- cess involves a vigorous brushing, and heating and rub-
ject to instability due to decrease of contact pressure bing of a thin stick of lead metal alloy on the surface.
from load cycling. However, the tensile strength of the Then a lead-based alloy in stick form is softened to a
welded connector is significantly (50 to 60%) lower than plastic state using a gas torch and is put into position on
the ultimate tensile strength of the conductor due to the the sheath and joint shell end until a “wipe” is obtained.
annealing of the conductor near the weld. If necessary, If the plumb is made outside the joint shell, it is called
mainly for submarine cables, the tensile strength can be “external plumb,” and it is submitted to a hoop stress
improved by round-compressing the conductor and the due to the internal pressure. If the joint is operated at a
weld (hardening process). pressure above 40 psi (275 kPa), the wipe must be rein-
forced using epoxy-resin-impregnated glass-fiber tape,
Exothermic Welding to avoid creep phenomena at elevated temperatures. The
Exothermic welding (also known as Thermite, Ther- same reinforcement is also adopted with thick alumi-
moweld, or Cadweld®) is mostly applicable to extruded num sheaths to support the high longitudinal forces that
cable accessories. It is formed by igniting into a tailor- this type of sheaths can transfer to the plumb.
made crucible, placed around the butted squared con-
8-10
If the plumb is located inside the joint, it is called an • Protecting the joint metal work from corrosion
“internal plumb.” It works under compression forces • Preventing water from entering in the joint
and can successfully support pressures up to 400 psi
(2760 kPa) and more. • Electrically insulating from earth the insulation
screen, shield wires, and metallic joint shell depend-
For extruded cables, the “external plumb” is generally ing on the joint type.
adopted. In the case of XLPE aluminum-sheathed
cables, when the sheath is very tight on the cable core, it The most commonly used design of joint shell protec-
is very important to control the heat applied during tin- tion, especially for sectionalized and stop joints having
ning and plumbing to prevent damage to the cable core. bonding leads and feeding pipes, is a glass fiber box
filled with hydrophobic compound (bitumen, polyure-
When the extruded cable has only a copper wire shield or thane resin, etc.). In this design, the box is only a con-
a copper wire shield associated with a thin metallic tainer for pouring the resin, and the watertightness is
sheath, which is not contributing to the short-circuit cur- achieved by the compound. In some cases a molded box
rent, the copper wires are generally plumbed to a copper is used, which, in combination with heat-shrink tubes, is
tube fitted on the cable core, and then the joint shell is also watertight. So a double barrier is ensured.
plumbed to this tube to simulate a conventional sheath.
When the joint shell has a regular profile, as in straight
Welding joints, heat-shrink tubes or adhesive tapes can
Welding is used with thick-wall aluminum-sheathed be adopted.
SCFF and HPGF cables to avoid the creep phenomena
that can take place with nonreinforced plumbed clo- Another possibility is to cover the joint shell in the fac-
sures. TIG or MIG welding techniques are generally tory with a coating of insulating material (PVC dipping)
adopted. The weld is generally performed on the crest of and to isolate the junctions on-site with the cable jacket
the corrugation, for corrugated sheaths, or by belling and between the two halves using heat-shrink tubes
out the end of the sheath, for smooth cable sheaths. This or tapes.
technique is also applicable to aluminum-sheathed
extruded cables. According to IEC and IEEE specifications, a water-
Mechanical Clamping immersion test, including loading cycles in water and a
With the new type of thin laminated aluminum or cop- final impulse test, are required to check the correct
per sheaths used today with extruded cables, especially if design of the anticorrosion protection of the joint.
associated with a bonded plastic jacket, when the thick-
ness is on the order of 0.039 in. (1 mm) or less, plumbing 8.2.2 Joints for Extruded-Dielectric Cables
is no longer applicable. In fact, it is difficult to remove General
the jacket and to do the plumb without damaging the An important difference between pipe-type or SCFF
sheath. In addition, the annealed sheath could hardly cables and extruded-dielectric cables is that the
support the thermomechanical loads transferred by the extruded-dielectric cables do not require insulating
jacket during loading cycles. For these reasons, a fluid. This class of cables and accessories must be essen-
mechanical clamping device is generally used. It consists tially void-free as manufactured, and must be operated
of a supporting ring, which is fitted underneath the thin without the benefits of ionization suppression or dielec-
sheath and jacket of the cable, and external sectors that tric strength enhancement by means of pressurized
are pressing the jacket in order to ensure a good electri- dielectric liquid or gas. Accordingly, few accessories are
cal contact of the sheath with the supporting ring. The required for pressurizing, accommodating the thermal
supporting ring is normally part of the joint shell to expansion and contraction of insulating fluid, or moni-
ensure the electrical continuity of the cable sheath. toring and alarming the pressure and volumes. Thus the
accessory system is essentially reduced to terminations
Because three longitudinal cuts are made in the cable and joints, and in the case of specially bonded systems,
sheath to be able to fit the supporting ring, special pro- the associated bonding and sheath voltage limiters
visions must be taken to seal the clamping area. (surge diverters).
Anticorrosion Protection
The anticorrosive protection is the outer covering of the It is essential to select the design of accessory to be
joint, which may have multiple functions depending upon exactly compatible with the particular extruded cable
the particular application in the cable systems, including: type and the particular service application. Compatibil-
ity should be validated by electrical-type approval tests
8-11
and be supported by either prequalification tests or ser- • insulation screen

vice experience of the specific systems (cable and acces- • screen wires, if any
sories). The performance of the accessory is dependent
on the quality, skill, and training of the jointing person- • longitudinal water blocking, if any
nel and on the use of the specialized tools required for a • metallic sheath, if any—for example, whether an
particular accessory. extruded sheath, a welded sheath, or a laminated foil
barrier. Also whether cylindrical or corrugated
For satisfactory service performance, it is most impor- • over-sheath (i.e., jacket)
tant that the correct size of accessory is selected to suit
the particular cable. The outer diameter of the cable • armor, if any
insulation and its tolerance and shape are particularly • special features (e.g., presence of optical fiber or pilot
important in the selection of an accessory employing a wires).
premolded component, such as an elastomeric stress
cone or an elastomeric joint molding. Such components Environmental Control during Splicing and
are designed to fit a specific range of diameters of pre- Terminating
pared cable insulation (that is, with the insulation screen In general, for voltage ratings through 138 kV, only min-
removed and the insulation smoothed and shaped). The imal environmental control of the splicing and terminat-
components must not be used outside this range. The ing areas is necessary. The exposed cable in the work
minimum cable diameter is determined by the need to area must be protected from free water, especially water
achieve sufficient pressure to eliminate voids at the that might be entrained in the cable conductor, at inter-
interface with the cable insulation. The maximum cable faces between the cable and splice insulation, or
diameter is determined by such considerations as: (a) between the cable and the termination stress control
preventing damage by over-stretching during assembly, device. Dehumidification of the area is generally not
and (b) limiting the maximum pressure at the interface required, except to provide a reasonably comfortable
such that the compression set of the cable insulation and environment for personnel to limit contamination by
molded insulation is minimized. heavy perspiration.
The diameter and tolerance of the conductor and of its It is necessary to control the cleanliness of the working
compaction (the ratio of the effective cross-sectional area during certain critical operations—for example,
area of the metal to the total area occupied) are needed when the cable insulation shield has been removed, and
in selecting a connector that will exhibit stable conduc- the insulation has been finally smoothed and cleaned for
tivity and high mechanical strength. splicing or terminating, and before the restoration of the
shielding is completed. Particulates contaminating the
The diameters and tolerances of the cable metallic insulation or interfaces can result in unacceptable con-
sheath and over-sheath are needed to ensure that acces- centrations of electrical stresses. Particulate control
sory metallic flanges and other components can be practices range from using dry (sometimes deionized)
passed back over the cable during assembly. air or nitrogen-pressurized glove boxes, to less elaborate
“good-housekeeping” practices, such as plastic
As a minimum requirement, the following dimensional coverups, cleanup procedures using vacuum sweepers,
and constructional details should be obtained by the and occasionally moistening floors to keep dust levels
user from the manufacturer to ensure compatibility of down. Whatever the practice, scrupulous inspection for
the accessory with the cable. Diameters, maximum and contamination of any sort must be enforced.
minimum tolerances, eccentricity dimensions, construc-
tion, and material need to be obtained for each of the Many of the steps in the splicing process, irrespective of
following cable components: the method employed, are common to all types, so that
apparent differences in time between the various proce-
• conductor and special features (e.g., water blocking), dures for restoring shields and insulation may not yield
if any significant savings in overall splicing time. These com-
• conductor screen mon steps include the usual need for heat-annealing and
• insulation (ovality and eccentricity dimensions are straightening the cable ends for splicing, stripping the
tightly bonded extruded insulation shield, painstakingly
required)
cleaning and smoothing the stripped ends, training and
• maximum allowable interface pressure racking the completed splice, etc.
8-12
Types of Joints can also be produced to combine different functions,

The following five basic classes of joints are in use today including stress control, insulation, etc.
for extruded cables, each with a number of design and
material permutations: A heat-shrink sleeve joint consists of several sleeves
• Heat-shrink sleeve joints that are positioned over the prepared cables to be
joined and that are “heat-shrunk” once the conductors
• Field-taped joints have been connected.
• Field-molded joints
• Prefabricated joints (premolded and composite) A typical construction of a heat-shrink sleeve is shown
in Figure 8-11.
• Back-to-back joints
Field-taped Joints
Heat-Shrink Sleeve Joints These joints (schematically shown in Figure 8-12) use
Heat-shrinkable elements are polyolefin components hand-applied tapes. Tapes can be adhesive, but better
heated in a factory process above their crystalline melt- performance has been obtained with self-amalgamating
ing point, then expanded and cooled. These compo- ethylene-propylene rubber (EPR) or polysobutene rub-
nents are supplied expanded. When heated again above ber tapes. They have been generally used to 69 kV, but
the crystalline melting point, typically 120°C, the com- they were proven satisfactory in the Waltz Mill test pro-
ponent contracts (“shrinks”). The shrink ratio is typi- gram at 138 kV on a number of the tested systems. They
cally 3 to 1. have been in generally successful commercial operation
on several U.S. installations at 115 kV and 138 kV, but
Heat-shrink sleeves can be used for external protection, with relatively low insulation screen stress. In fact, the
but also, when appropriate fillers are incorporated, as radial stress at the beginning of the stress cone, which
stress control elements. Multilayer heat-shrink sleeves corresponds to the cable screen stress, cannot be miti-
Figure 8-11 Heat-shrink sleeve joint (courtesy CIGRE).
Figure 8-12 Field-taped joint (courtesy CIGRE).
8-13
gated. Field-taped joints have the advantage of being • Mold-Extrusion. The insulation is heated and melted
made from readily available and economical materials in an extruder and then injected into a heated mold.
and represent the simplest approach with only minimal In “crosslinked extrusion-molded joints,” the insula-
special tooling and training requirements. The principal tion is crosslinked at an elevated temperature, which
disadvantage is that they have inherently larger diame- activates a chemical agent present in the extruded
ters, and are longer and more laborious to tape, owing material.
to their basically laminar design and the need to employ • Block-Molding. The insulation is applied in the form
modest radial and longitudinal design stresses.
of two half moldings, which are fitted to the pencilled
Field-molded Joints cable ends. Then a mold is installed and heated to
In field-molded joints, the field-applied insulation is melt and bond to the cable. In “crosslinked block-
melted and bonded to the pencilled extruded cable ends molded joints,” the insulation is crosslinked at an
in various forms of molds (see Figure 8-13). This elevated temperature, which activates a chemical
technique is available for both PE and XLPE cables as agent present in the premolded material (Tsuchiya et
well as for EPR cables. In the case of crosslinked al. 1999).
extruded-dielectric cables (i.e., XLPE and EPR cables),
the insulation must be heat-and-pressure cured. Using proper technique, mold design, and process tem-
Connector and insulation semiconducting screens are perature and pressure control, it is possible to achieve
generally applied separately. nearly perfect fusion of field-applied material, and good
bonding to the extruded conductor shield and insula-
Insulation can be applied on the prepared cable in dif- tion of the cable. The result is a void-free joint, having
ferent ways: high and essentially isotropic (longitudinal: radial)
dielectric strength. This permits the use of high radial
• Taping. The insulation is applied in the form of layers
and longitudinal design electrical stresses with signifi-
of tapes, then softened and melted by heat. In
cant reduction in splice length and diameter. When a
“crosslinked tape-molded joints,” the insulation is
“flush-type” connector is used, the splice can be built to
crosslinked at an elevated temperature, which acti-
essentially the same dimensions as the cable and can be
vates a chemical crosslinking agent present in the
used as a duct or repair joint, or a flexible submarine
tape material.
cable joint. However, the manufacture of this joint is
• Mold-Injection. The insulation is heated and melted complicated and long. In addition its electrical perfor-
in a cylinder or a pot in a factory or field process and mance is very sensitive to assembly mistakes and to the
then injected into a heated mold by application of influence of the field environment. Highly skilled joint-
fluid pressure. In “crosslinked injection-molded ers are required to assemble, operate, and disassemble
joints,” the insulation is crosslinked at an elevated the complex equipment required at each jointing loca-
temperature, which activates a chemical agent present tion, usually in a manhole.
in the factory-prepared material.
Field-molded joints, which have been used in the past
from 115 kV up to 500 kV, are today seldom used to join
Figure 8-13 Field-molded joint (courtesy CIGRE).
8-14
land cables, because prefabricated joints that are factory • Application of a mechanical force using springs, as
molded and routine tested are preferred. The extrusion- shown in Figure 8-15.
molded (also known as field vulcanized) design is still • Use of the elasticity of the premolded body expanded
used in Japan for joining 500-kV extruded cables; how-
on the cable insulation (see Figures 8-16, 8-17, and
ever, the jointing time is very long—on the order of
8-18).
more than one week per joint (Kubota et al. 1994).
Prefabricated Joints The same principle is applicable and used for the pre-
Prefabricated joints are usually made with premolded- molded stress cones of cable terminations.
rubber-insulated components, which, for some designs,
are used in conjunction with cast epoxy components. The three main families of this type of joint are compos-
(In CIGRE terminology, both premolded joints and ite prefabricated joints, one-piece premolded joints, and
composite prefabricated joints are referred to as prefab- three-piece premolded joints.
ricated joints.) When used in conjunction with cast • Composite prefabricated joint.
epoxy components, they are referred to as composite
joints. Premolded joints could be of one-piece, two- Two premolded stress-cones made of elastomeric
piece, or three-piece type (CIGRE 1994, No. 89). This material are positioned on the prepared cable ends
class of joint is available through the 500-kV rating. and then inserted into a cast epoxy insulator. The
Both one-piece and multipart designs of premolded pressure at the cable core interface, as well as at the
joints have been developed and tested. Prefabricated stress-cone to central epoxy insulator interface, are
joints combine the two advantages that they are manu- obtained by means of compression devices such as
factured in a factory in strictly controlled conditions springs (Tanaka et al. 1992).
and that they are routine-tested before delivery like the • One-piece premolded joint.
associated cables. Another advantage is that the In this joint, the stress control is ensured by semicon-
assembly of the joint over the cable core is facilitated by ductive elements: connector screen, stress control
using special assembly tools. The principal disadvan-
tages compared to field-molded joints are their high
first cost and their larger size.
In prefabricated joints, the allowable stress at the inter-

faces is correlated to their smoothness and relative pres-
sure (CIGRE 2002). According to the Paschen curve
(Kreuger 1989), the smaller the cavity (i.e., the rough-
ness of surfaces), the higher the breakdown stress. The
sensitivity of the interfaces to irregularities depends on
the interface pressure and on the physical characteristic
of the materials that are interfacing. A high interfacial
pressure minimizes the size of microcavities. Figure 8-14
shows the effect of the pressure on the electrical strength
of the interface between a silicone rubber (SIR) sleeve
and an XLPE cable (Kunze et al. 2000; Tanaka et
al. 2000).
Figure 8-14 Electrical interface strength vs. interface
Two different methods are commonly used to achieve pressure and smoothness.
the required interface pressure:
8-15
profile screens, and insulation screens, which are cable using different methods (“slip-on joint”) or pre-
factory-molded and form, together with the insula- expanded in the field or in the factory over a carrier
tion, a single premolded sleeve (Parmigiani et tube and parked on one of the two cable ends (“cold
al.1987; Chatterjee et al. 1995; Peshke et al. 1996; shrinkable joint”). After the connector and the
Nakamura et al. 2003). In order to follow the cable connector screen have been applied, the sleeve is
insulation diameter variation with temperature and positioned at the center of the joint. The elastic
maintain a positive pressure, the sleeve is based on an memory of the sleeve ensures the required pressure at
interference fit over the cable insulation. For this the interface with the prepared cable. The physical
purpose, the sleeve can be slipped over the prepared characteristics of the rubber, especially a low-stress
Figure 8-15 Prefabricated composite joint (courtesy CIGRE).
Figure 8-16 One-piece premolded joint (courtesy CIGRE).
Figure 8-17 Plug-in-type one-piece premolded joint (courtesy CIGRE).
8-16
relaxation, are essential in order to maintain the tralizes and anchors the conductor. This insulator can
pressure for the whole cable system life. In order to be fluid-filled, with the same or different insulating fluid
prevent the shrink-back of the cable insulation during as the enclosure. In some designs, the terminations are
load cycling, some designs incorporate a longitudinal directly immersed in the enclosure. In some construc-
locking arrangement that fits into a groove machined tions, the terminations are without insulating fluid.
in the cable insulation. See Figure 8-16.
A variation on this design is the one-piece plug-in Different kinds of back-to-back joints are available. The
joint, shown in Figure 8-17 (Leufkens et al. 1992). In main ones are:
this case, the two cable ends are prepared separately, • Back-to-back joint with two insulators (Figure 8-19).
eventually in the factory, and are provided with a
In this family, the insulators can be used to segregate
plug-in connector. Then, with a special pulling tool,
the SF6 gas or the insulating liquid from each one of
the cable ends are pushed inside the rubber sleeve,
the cables that are joined. The enclosure fluid can be
expanding the rubber sleeve at the same time. A
any of the insulating fluids from each of the cable ter-
metallic insert embedded in the central semiconduct-
minations or even another one.
ing electrode is used to lock the cable ends through
special spring-loaded pins. The advantage of this • Back-to-back joint with one insulator (Figure 8-20).
design, compared to the conventional one-piece pre- In this family, one termination is directly immersed,
molded joint, is that it can have a different diameter and the single insulator anchors both conductors and
bore on each side, so it is able to join two cables with centralizes corona shields within the main insulation
different dimensions. fluid. The directly immersed cable must be protected
• Three-piece premolded joint. against water ingress or fluid leak.
In the three-piece premolded joint, each cable has its • Back-to-back joint without insulator.
own adaptor molding, making the transition between In some cases, both terminations are directly im-
different cross-section cables easier (Nelson et mersed. In these cases, both immersed cables must be
al. 1978). protected against water ingress or fluid leak. Provi-
sions need to be taken to handle any thermomechan-
Back-to-back Joints
ical cable movement.
It is common to put in series two lengths of LV or MV
cable by just connecting outdoor terminations. The Screen Interruption Joints
same principle is sometimes used at high voltages using Most designs are available with shield interruptions for
metal-enclosed GIS terminations or oil-immersed ter- use as insulating joints in specially bonded systems.
minations to constitute what is called a “back-to-back Details vary widely, but the design must provide for
joint.” Just as metal-enclosed terminations can be single interrupting the metal shield/sheath continuity, as well
phase or three phase, the “back-to-back” joints may as interrupting the bonded (polymeric) insulation shield
also be single or three phase. while maintaining its essential electrostatic shielding
integrity. The former requirement is typically met by
Metal-enclosed terminations generally consist of an inserting an insulating flange or molded insert of suit-
insulator that houses the prepared cable ends, and cen- able polymer/epoxy in the joint casing. Fulfilling the lat-
Figure 8-18 Three-piece premolded joint (courtesy CIGRE).
8-17
ter requirement usually requires creating a gap in the central epoxy insulator, which comprises an HV elec-
restored insulation shield of the splice, and shielding the trode in which the three conductors are plugged.
gap with overlapped layers of semiconducting tapes
Transition Joints
and/or paint separated by a thin layer of insulating
There has been considerable interest in joint designs for
material. Some designs incorporate a layer of metal-
connecting extruded-dielectric cables to pressurized,
oxide-loaded polymeric material having a steep current
impregnated-paper cables (Gregory 1991 et al.; Chu
vs. voltage characteristic as a surge suppressor at this
1992 et al.; Gahungu et al. 1993; De Ridder et al. 2002).
critical discontinuity. Likewise, the insulator in the
Generally this interest stems from providing for possible
splice housing must be protected against transient over-
future replacement of sections of older paper cables with
voltages by means of sheath voltage limiters (also
extruded-dielectric cables as leakage problems, reliabil-
known as surge diverters) usually incorporated in the
ity concerns, and maintenance costs due to corrosion
bonding or link boxes, between sheaths and ground (see
become factors.
Section 8.4.1).
Wye (Y) Branch Joints The two different cables must be completely segregated.
There are applications where three cables have to be The easiest way to make a transition joint is to use a
connected. All the techniques employed for straight fluid-immersed termination on each cable connected
joints may be considered: heat-shrink sleeves, taped, back to back in the same enclosure, as illustrated in Fig-
field-molded, premolded/prefabricated, or back-to- ures 8-19 and 8-20. However, this solution is bulky and
back. Nevertheless, the most common construction type may be expensive.
of Y-branch joints is the composite prefabricated type
(see Figure 8-21). Three stress cones are inserted into a A more compact design is achieved by adopting a solid
barrier, generally precast in epoxy resin, which is
Figure 8-19 Back-to-back joint with two insulators (courtesy CIGRE).
Figure 8-20 Back-to-back joint with one insulator (courtesy CIGRE).
8-18
designed to withstand the pressure in the fluid-filled (Figures 8-22 and 8-23). When necessary, annular chan-
cable. Connection of conductors from both sides can be nel can be provided in the laminated cable termination
made by means of insulated metallic parts embedded in to permit oil or gas to be fed to the central conductor
the epoxy insulator. On each side of the transition joint, duct (Figure 8-24).
all available techniques of terminations can be used
Figure 8-21 Prefabricated composite Y-branch joint (courtesy CIGRE).
Figure 8-22 Polymeric extruded cable to oil or gas-filled paper cable transition joint three-core
type (courtesy CIGRE).
Figure 8-23 Polymeric extruded cable to oil or gas-filled paper cable transition joint single
core–non fed type (courtesy CIGRE).
8-19
The IEEE 404 standard (IEEE 2000) recommends the depicts the most common components. Three single-
test methods and requirements for transition joints (see phase joints are bound together with Nylon tape and
Section 8.2.5). A new CIGRE working group has also supported with “spider” assemblies. Reducers at each
been launched to prepare international recommenda- end of the joint are used as interface between the pipes
tions for these tests. of the cable line and the joint housing.
8.2.3 Pipe-Type Joints Each single-phase joint (Figure 8-26) includes a connec-
tor, hand-taped insulation, and semiconductive and
General Description metallic shielding.
Pipe-type joints are used on laminar-dielectric cables
both with kraft paper insulation and laminated paper Connectors can be pressed, welded, or soldered to cable
polypropylene (LPP) insulation. High-pressure fluid- conductors. Two cable conductors can also be welded
filled (HPFF) pipe-type joints are in commercial service to each other, avoiding the use of a connector. Since
up through 345 kV, although they have been success- connectors represent high-voltage electrodes, they must
fully tested to 550 kV and 765 kV with LPP cables. be of proper shape and with a smooth surface, which is
obtained by different methods, including application of
Ty p i c a l l y a p i p e - t y p e j o i n t c o n s i s t s o f t h r e e lead tape, semiconductive tape, etc.
single-conductor cable joints in a common carbon steel
housing. Cables are individually spliced, forming three The insulation of each cable needs to be stripped back
single-phase joints that are mechanically reinforced to by “penciling” the insulation in steps at the connector
prevent damage due to bending or cable movement. A before taping. The hand-taped materials are applied in
more detailed description and design of each three stages. The first fills the areas to each side of the
component are provided later in the text. Figure 8-25 connector until the connector outside diameter (OD) is
Figure 8-24 Polymeric extruded cable to oil or gas-filled paper cable transition joint single core-fed type.
Figure 8-25 Typical normal HPFF joint in housing (courtesy of USi).
8-20
reached. The second continues the buildup to the origi- shown in Figure 8-27. The diameter of the joint varies
nal cable OD, known as the “factory diameter.” Finally with the cable size, connector type, and the dielectric
a “jacket” is applied over the factory diameter to the properties of hand-applied insulating tape.
final OD of the joint insulation. The jacket is extended
longitudinally to the vicinity of the semiconductive High-voltage pipe-type joints are typically purchased in
screens of both cables. The ends of the jacket must be kit for m, where the kit includes all components
carefully graded down to the cable insulation in the including connectors, housings, and spiders etc. The kit
form of stress cones. Since all tapes are applied by hand, also includes all required consumables such as
the quality of the joint is highly dependent on the skill impregnated insulating and semiconductive tapes,
of the splicer. shielding braids, cotton tapes and installation instruc-
tions. Cable suppliers and experienced joint suppliers
The last step in the construction of the single-phase can provide knowledgeable guidance on design and con-
joint is to apply a semiconductive and metallic screen struction of the joints.
over the jacket OD and stress cones to mate with the
cable shields. The construction of a typical normal pipe-type joint
involves the following basic steps:
As described in the section on the electric design of • Installing pipe reducers
joints (Section 8.2.1.), the diameter of a joint depends
on the outer diameter of the connector screen (high- • Sliding housings into their temporary position over
voltage electrode) and the dielectric strength of the insu- cable line pipes
lation between the conductor screen and outer jacket • Determining cable cut locations, and making final
screen (ground electrode). The total length of the joint cable cuts
depends on the length of the stress cones and pencilling. • Installing the connector
Typical longitudinal dimensions of 345-kV joints are
Figure 8-26 Components of the taped joint (courtesy of G&W Electric Company).
Figure 8-27 Typical longitudinal dimensions of single-phase 345-kV joint.
8-21
• Applying the connector screen • between 400 and 600 psi** (2760 and 4140 kPa) on
• Stepping the cable insulation typical cable systems with high-speed circulation.
• Applying impregnated insulating tape to build the • up to 950 psi** (6550 kPa) on circuits with high-speed
jacket circulation, where extremely long circulation loops or
severe elevation differences are involved.
• Applying semiconductive tapes and copper mesh
over each leg **These higher pressures are normally on the smaller-
• Binding joints and installing spiders (cable supports) diameter return pipe on circulated systems. The
• Positioning cable housings potheads and the cable pipes do not typically see
extreme pressures.
• Welding the housings
• Pressure and leak testing of the housings System electrical data and cable information are relevant
for the electrical design of the joint (connector, insula-
Design Requirements for Pipe-type Joints tion stepping at connector, stress cone at cable shields.
Joints for pipe-type cables are not “off-the-shelf ” items,
but rather custom-designed specifically for the cables to Pipe and manhole data is needed to design the fittings
be joined. The designs of both the electrical insulation (reducers) that are used to connect the pipe-type joint
of the joint and the mechanical components must be housings to the incoming line pipes (Figure 8-25) and to
coordinated with the cables and pipes to be joined and select the outer diameter of the housings.
the design of the manholes.
Table 8-1 summarizes the information required for the
Manhole size is needed to select a design of the housing design of pipe-type joint.
to accommodate space restrictions. Figure 8-28 shows
an example of a pipe-type joint in a manhole. Development of a pipe-type joint design involves evalu-
ation of dielectric materials, dimensions of the joint,
The mechanical, rather than the electrical, aspects dif- time and cost to install, expertise of available jointers,
ferentiate pipe-type design requirements from the other etc. Although various design trade-offs can be consid-
high-voltage cable systems, because pipe-type high- ered, a pipe-type joint should first be designed based on
pressure systems must consider the mechanical rein- overall system requirements.
forcement of the three individual cables and the design
of reducers and housings for the fluid pressures in the There are a number of design trade-offs for the joints.
cable system. The housings of pipe-type joints are For example, conservatively designed electrical stress
designed to withstand the high pressures of the levels provide some protection against problems with
insulating fluid. Typically fluid pressures are: workmanship and poor environmental conditions dur-
• between 200 and 400 psi (1380 and 2760 kPa) on nor- ing splicing. However, they result in longer inner and
mal self-cooled circuits and circuits with slow or outer joint tapers and a larger joint outer diameter.
moderate circulation rates.
Figure 8-28 Typical 345-kV pipe-type LPP cable joint assembly in manhole (courtesy of USi).
8-22
This, in turn, means a larger joint, longer time to splice, thermal performance and allows use of smaller-diame-
and a higher thermal resistance at the joint. ter pipe.
Components of the Pipe-Type Joint—Materials and
Another design trade-off occurs in selecting joint hous-
Design Practices
ings. Two-part housing is the simplest, lowest-cost
A number of different materials and design issues are
option and involves the least amount of field welding,
important in designing, manufacturing, and installing
but it requires the most manhole space on either side of
pipe-type joints. This section reviews the materials and
the reducers. Use of a multipart joint housing, on the
designs of components that are generally used and
other hand, such as a three-part or a five-part, saves
discusses some of the considerations that should be
manhole space, but involves more costly housings and
evaluated.
more field welding.
Connectors
For example, crepe paper, the most widely used joint The use of compression-type connectors is most com-
insulation in the past for pipe-type cables, does not have mon for pipe-type cables. Other types of connectors that
the best dielectric characteristics available, but it is the are also utilized, depending on application, are
easiest and most foolproof to install. Selecting the insu- described in Section 8.2.1. Copper connectors for instal-
lation material requires considering dielectric properties, lation with a 4-ram press are typically 6.5-7.5 in. (16.5-
type of cable to be spliced, thermal properties of the 19 cm) long, providing space for two or three sets of
dielectric, ultimate joint dimensions, and installation indents per cable. Once one set of indents is pressed, the
requirements. LPP is presently the most popular joint tool is rotated 45°, and the next set of indents pressed
insulation and is used from 138-kV and above. LPP approximately 0.75 in. (2 cm) from the first set and so
allows smaller-diameter joints, which provides better on. Increasing the connector length, so that additional
sets of indents can be made, increases the pullout
Table 8-1 Information Relevant for Design of Pipe-Type strength of the connector. Connectors installed with hex
Joints and round dies show somewhat higher pullout strengths
System Electrical Data Pipe Data than those installed with a 4-ram press, but generally
Cable system nominal voltage Pipe material these differences are not significant.
Cable system maximum voltage Size (OD and wall thickness)
Basic impulse level (system BIL)
Load requirements (Normal and
Four types of copper crimp connectors are in common
Emergency, Long Term, and use. They differ in design of taper angles at their ends
Short Term) and in the time and skill required for installation:
Cable Data Manhole Data
• Conventional indent-type 4-ram press connector
Cable design type Manhole inside dimensions
Conductor material (particularly length) • 4-ram connector with preformed tapered ends
Conductor construction and size Joint Type
Conductor shield material and • Compression connector for installation with hex,
thickness Normal; Stop; Semi-Stop;
Anchor; Skid; Trifurcating; rounding, and finishing dies
Insulation type and thickness
Insulation screen material and Wye, H, Back-to-Back; etc. • Field-machined (FM) connector.
thickness Termination Data
Type and material of skid wires Pressure rating, nominal,
Pressurizing medium (liquid or gas) The conventional 4-ram press-type connector is gener-
maximum continuous, and
Maximum operating and transient test ally manufactured with tapered ends with angles
pressures between 8 and 12°. Lead tape is hand-applied to the
Number of cables to be joined (two Environmental
or more) Route profile ends of the connector during installation to reduce the
Other subsurface structures effective taper angle to 4-7°.
nearby, particularly heat-
generating sources
Water crossing or other envi- The 4-ram press connector with factory-formed tapered
ronmentally sensitive cross- ends is similar to the standard 4-ram press connector,
ings except that instead of using hand-applied lead tape to
Earth thermal characteristics
form the final tapers, the connector is made with ends
fabricated to the final taper angle. This minimizes the
amount of lead foil-tape that is hand-applied and conse-
quently minimizes the variation in taper angle due to the
splicer’s installation technique.
8-23
The connector installation with hex, rounding, and fin- The current-cycling tests with compression aluminum
ishing dies provides somewhat higher pullout strength connectors with large aluminum conductors have shown
compared to the 4-ram press-installed connectors. an increase of electrical resistance between cable con-
Obtaining a round, smooth, straight connector after ductor and connector after several cycles. This causes
pressing depends to a high degree on the splicer’s skill increased losses and hence unstable thermal behavior.
and experience. Lead tape at the ends must also be As a result, the preferred method of joining aluminum
hand-applied to obtain proper shape. conductors is welding. Additional crimping after weld-
ing enhances pullout strength. That is one of the reasons
The field-machined (FM) connector is also applied with why commercial installations of underground cable with
compression dies, but it is then machined in the field large aluminum conductors only began after methods
with a router on a fixed guide to provide a finished con- for field-welding the conductors were developed
nector with an effective angle of only 1 or 2°. This con- and proven.
nector requires time-consuming field machining and a
high degree of installer experience. It is popular at Large aluminum conductors are joined by an elaborate
345 kV, however, because it provides one of the lowest process, with a number of critical steps involving careful
profile connectors. removal of oil from the strands, attachment of a small
vacuum pump for removal of creeping oil from the
Conventional preformed and field-machined connectors strands, a pressing connector, setting-up cooling fans,
are shown in Figure 8-29. cutting a wedge in the strands at the center of the sleeve,
buttering the exposed strand ends with weld metal, then
completing the weld and smoothing the surface (Gold-
lust et al. 1967). The weld provides an electrical connec-
tion that is not affected by load cycling. Compression of
the connector improves the tensile strength of the joint,
since the strength of the welded joint alone is limited
due to annealing of the conductors. MIG welding is
used to limit the heat input to the conductors and insu-
lating tapes of the cables. The tensile strength of the
welded connectors is approximately 50-60% of the ulti-
mate tensile strength of the conductor, while the electri-
cal conductance of the joint equals that of the
conductor. The welded connectors are well proven, with
successful installations dating from the late 1960s.
Finally, connectors must be installed so that sharp

changes in the contours of the connector are avoided
and that no sharp or jagged edges exist, because high
electrical stresses at the asperities could cause the joint
to fail. The indents formed during pressing of the con-
nector are filled with either lead or epoxy compound. If
lead is used, it is peened into the indent. In the final
steps, the connector must be polished to remove any
asperities. Fine, nonconducting abrasive cloth, such as
aloxite, is used for this procedure (see Figure 8-30).
Insulation
Three types of oil-impregnated dielectric tape are used
to insulate high-voltage joints: crepe paper, flat paper,
and laminated paper polypropylene (LPP).
In the past, the most widely used insulation for pipe-

type joints was low-density, deionized-water-washed cel-
lulose paper, which is slit into tapes, dried, and vacuum-
impregnated with oil. Crepe tape was generally used at
Figure 8-29 Types of copper connectors for pipe-type
230 kV and below, while flat-paper (kraft paper that is
cables.
8-24
not creped) tape was used above 230 kV. Crepe paper LPP is the most widely used dielectric material today for
was popular because it is easier to apply than the flat pipe-type cables and joints. This is a flat tape made from
paper. Although crepe-paper joints have been success- a sandwich of paper-polypropylene-paper. It is attrac-
fully tested at 345 kV and are installed on commercially tive because it has lower losses and capacitance and
operating cables, flat-paper tapes have been more com- higher breakdown strength than conventional cellulose-
monly used at this voltage. This is because crepe paper paper insulation. LPP insulation allows the use of
results in a larger joint diameter, a uniform outer jacket smaller joint diameters, which reduces splicing time,
buildup is more difficult to obtain, and its dissipation decreases joint lengths, and often allows use of smaller
factor is slightly higher than flat paper. Typical specifi- pipe diameters. All of these benefits allow a more cost
cations for both the dry and impregnated paper tapes attractive design.
are shown in Table 8-2.
Figure 8-30 Conventional 4-ram press connector (courtesy Dominion-Virginia Power).
Table 8-2 Typical Characteristics of Dry and Impregnated Paper Tapes

Tape Property Crepe Paper Flat Paper
0.003 in. 0.005 in.
(0.076 mm) (0.127 mm) 0.005 in.
Before Creping Before Creping (0.127 mm) Flat
Tape Dry
Finished thickness, nominal range, mils 10-13 15-19 4.5-5.5
Elongation, % min. 50 50 2.5
Tensile strength, machine direction
lb/in., min. 12 20 60
lb/in., nominal - 20 25
Apparent density, g/cc, nominal 0.28 0.34 0.72-0.80
Water extract conductivity,
30 30 5.0 Max.
micro-mhos, max.
Moisture, % max. 7 7 6
Ash, % max. 1 1 0.4
Tape Impregnated
Dissipation Factor at 80°C,
% max. 0.25 0.30 0.20
% nom. 0.18 0.22 0.15
Dielectric breakdown, kV min. 6.0 8.0 8.0
Dielectric breakdown under 30%
elongation, kV, min. overall 4.0 4.5 N/A
Tensile strength, lb/in, width, min. 6 10 60
Elongation, %, min. 40 40 2.5
8-25
Typical characteristics of LPP are shown in Table 8-3. to ease installation. Narrower tapes should be provided
to tape around the connector and to fill in the inner
Unlike the cable itself, which is machine-taped, the joint jacket between the stepped cable insulation, because
is hand-taped, and installation requires the use of there is limited space to place the tapes. Wider tapes can
trained, qualified splicers. be used on the outer jacket, because there is a long,
unobstructed length to be taped, and the wide tape
Insulating tapes are factory-impregnated with dielectric speeds splice completion.
fluid similar to that used to fill the cable pipe. Unlike the
Semiconductive Shield Tapes
fluid used to impregnate the cable, which must be rela-
Similar to the conductor and insulation screen of the
tively viscous at room temperature so that it does not
cable, semiconductive tapes are applied over exposed
drain or migrate out, the fluid used to impregnate the
cable conductor and connector and over the taped insu-
splice tapes (the oil in the can with the tapes) has a rela-
lation of the joint to form electrostatic screens (shields).
tively low viscosity. Splice tape impregnants include
The carbon-black (semiconductive) tape is usually butt-
100% polybutene, a mixture of polybutene and alkyl-
lapped over the connector and terminated to make a
b e n z e n e, o r n ap h t h e n i c m i n e ra l o i l . T h e p o ly -
smooth transition between the connector shield and the
butene/alkylbenzene mixture is by far the most popular
cable conductor shield. The carbon-black tape of the
tape impregnant. It is more readily available than naph-
joint is half-lapped over the joint insulation and
thenic mineral oil and has better oxidation stability and
smoothly tied to the cable semiconductive insulation
gassing characteristics than both the mineral oil and the
screen. It is essential that the tapes are smoothly
straight polybutene.
applied, with no asperities or jagged edges, which would
create high stress points within the joint.
The tape is dried in heated retorts under vacuum, and
then it is impregnated with a suitable dielectric fluid.
Semiconductive tapes are filled with carbon-black and
The impregnated tapes, in fluid-filled cans, are sealed
have volume resistivity between 105 and 108 Wcm. For
under vacuum. Frequent testing of the tape and impreg-
easier application, a crepe tape is used. Semiconductive
nant during processing is essential to ensure that the
tapes are impregnated with dielectric fluid similar to
final product is free of moisture and contaminants, and
that used with the insulating paper tapes.
meets the specifications. Tapes are provided in different
thicknesses and widths based on the location in the joint Metal Shield Tapes
where the tape is to be used. Good design practice calls A metallic shield tape is applied over the carbon-black
for thinner tapes in areas of the joint that are subject to insulation shield of the taped joint. This metallic shield
higher stresses, such as near the conductor or at the tape conforms well to the varying profile of the joint.
start of the outer taper. Thicker tapes are used on the Metallic shields provide a conductive path for capacitive
lower stress regions, because they allow the joint to be charging current during normal operation, preventing
built up faster, reducing splicing time and exposure of overheating of the semiconductive tape. It also provides
the insulation to the atmosphere. Tape width is selected a path for short-circuit current in case of a line-to-
Table 8-3 Typical Characteristics of LPP Tape Insulation

Tape Property 0.004 in. 0.006 in.
(0.101 mm) (0.152 mm)
Thick Thick
Tape Dry
Finished thickness, nominal range, mils 3.92-4.08 5.93-6.07
Tensile modulus, kg/cm2x103
E1—MD 77 64
E2—CMD 43 36
Compressive modulus, kg/cm2 55 56
Coefficient of friction—CMD (Dry) 0.42 0.37
Polypropylene ratio, %
49 52
(Film/Laminate thickness)
Tape Impregnated
Dissipation factor at 80°C, % nom. 0.073 0.069
SIC (relative dielectric constant) 2.8 2.8
Dielectric breakdown, volts/mil 3200 2880
8-26
ground fault in the system. The metallic shield is solidly shown to be prone to damage by thermomechanical
grounded to the pipe and steel sleeves through jumper bending (TMB). The cable expands, contracts, flexes,
cables that are tied to the reducer. The metallic shield is and moves due to changing thermal conditions during
a tinned, all-copper mesh tape. load cycling. The cable is constrained in the line pipe,
limiting the degree to which it can bend, but this is not
Spiders (Cable Supports)
necessarily so within the joint casing. By tying the cable
After the three cable phases are joined, the individual
phases together and tying them to spiders, TMB dam-
joints are tied together using cable supports (“spiders”)
age is avoided. Second, spiders hold the joint off the
and cotton or nylon tape. Figure 8-31 shows a cross sec-
bottom of the steel sleeve, eliminating the possibility of
tion of a spider with the cable phases tied to it.
damaging the joint insulation from the high tempera-
ture of the steel housing during welding operations. Fig-
Using spiders and tying the cables together serves two
ure 8-32 shows typical locations of spiders for a
purposes. First, thick-walled, pipe-type cables have been
345-kV joint.
Spiders are made of cast aluminum. One leg of the

three-legged spider is removable (on steel pins), so that
the spiders can be easily inserted among the three
phases after the splicing has been completed. Spiders
come in assorted sizes and widths. They are sized to fit
the sleeve in which they are to be installed, with a close
enough fit to allow them to slide back and forth within
the joint casing but not to twist out of their straight
installation position (see Figure 8-33). The design of
Figure 8-31 Three 1/C joints secured to aluminum spider. spiders gets more elaborate in case of Wye or “H” joints.
Figure 8-32 Typical spider locations for 345-kV LPP insulated joint (courtesy of USi).
Figure 8-33 Field installation of 230-kV normal joint with five spiders (courtesy of Dominion-Virginia Power).
8-27
Housings and Reducers vidual casing sections are lighter and easier to handle.
HPFF joints are installed in a steel housing that is Typical multipart nested housings and related manhole
designed to withstand operating and transient pressures length requirements are shown in Figure 8-34.
of the system medium (liquid or gas) with certain design
margin. Figure 8-25 shows a normal joint within its The design of the joint sleeves, with one sleeve swaged to
housing. There are two reducers in a typical joint, one mate with the next, permits nesting of the sleeves during
on each end, where the line pipes meet the joint. The splicing, and accommodates welding requirements.
joint itself is enclosed in steel sleeves between the With one sleeve end overlapping the next section or the
two reducers. reducer, the casings are joined with a filet weld, circum-
ferentially around the casing. The filet weld does not
The reducer is made from hot-rolled (HR) steel such as require highly skilled welders and also ensures that
ANSI 1020 or 1015 or ASTM A-36. welding slag does not enter and damage the joint.
Reducers are multifunctional: First, the reducer is The sleeves must be clean when they are installed.
welded to the end of the line pipe and capped with a test Surface rust, for example, would contaminate the pipe-
plate (a gas-tight plate with a small valve in the center) filling fluid. Thus, sleeves should be supplied with
used to cap the line during construction of the system. suitable coatings inside and out. Typically, the inside of
When capped, the line pipe can be pressure- and leak- the casings is coated with an epoxy paint, which
tested, and it will allow the pipe to be evacuated and prevents rusting of the interior before installation and
filled with dry air or nitrogen to keep the pipe dry and does not react with or dissolve in the dielectric pipe-
clean for the cable. filling fluid. The exterior is often field-coated after the
sleeves are welded in place. To prevent rusting prior to
Second, the reducer is used for securing cable skid-wires installation, Phenolated Alkyd Primer is generally
to ensure that the joint and cable shields are all solidly applied to the exterior of the casings. The ends of the
grounded to the pipe enclosure. Proper grounding is casings to be welded should not be coated with paints,
essential for preventing damaging voltage drop to
develop between the pipe and cable shielding during
normal and transient voltage conditions.
Finally, reducers are used as an interface between the

cable pipe line and housing sleeves of the joint. The ends
of steel sleeves are welded to the reducers after the splic-
ing has been completed. The slot on the outer edge of
the reducer acts as a heat sink, so the cables in the line
pipe are not damaged by heat due to welding.
The sleeves that make up the body of the joint casing are
made of carbon-steel pipe, which is generally specified
to be ASTM A-53 or A-106, Grade B. The ends of the
sleeves are sized to allow welding to the next sleeve
section or the reducer. The wall thickness of the casings
is based on the maximum expected pressure in the cable
system. As the casing diameter increases, thicker-walled
steel sleeves must be used due to the increasing hoop
stress. Casings between 12 and 18 in. (30.5 and 45.7 cm)
OD are typically 0.375 in. (9.5 mm) thick. Larger-
diameter casings, such as 24-28 in. (61-71 cm) OD
sleeves used on wye joints, are 0.5 in. (12.7 mm) thick.
The sleeves are designed to slide to either side of the

joint during splicing. Since manhole size is often lim-
ited, multiple nesting sleeves are often provided. Multi-
part casings, such as three- or five-part sleeves, require
more field-welding than a single- or two-part casing, but
they can be installed in a smaller manhole, and the indi- Figure 8-34 Typical multipart nested housings.
8-28
because they often will give off noxious odors or resistivity of only 20-30% of that of impregnated paper.
hazardous gases when subjected to welding The cast-epoxy joints should, therefore, operate with a
temperatures. Thus, it is common practice to apply only lower conductor temperature. Epoxy castings have also
deoxaluminate primer on any areas of the sleeve to been used in splices requiring a full fluid stop. Installa-
be welded. tion of epoxy castings into a steel plate in the joint cas-
ing permits complete segregation of the fluid from one
Joint casings often have valves attached via weldolet fit- section of the cable system to the next. Unlike the taped
tings. A 2 in. (5.1 cm) valve is placed on top of the cas- semi-stop joints discussed below, which permit some
ing for evacuation, fluid filling, or gas-in-oil testing. A 1 fluid to weep through the taped structure of the joint, an
or 2 in. (2.54 or 5.1 cm) valve is placed on the bottom epoxy casting permits a complete fluid seal.
for draining fluid.
While the concept of using the Voltalit in normal joints
Cast-Epoxy Insulation Components
appeared attractive, experience has shown that the eco-
Using cast-epoxy components for electrical insulation
nomics have not been attractive. The only significant use
of impregnated-paper cable has been considered as an
of Voltalit® joints is in full stop joints, where there is no
alternative to a totally hand-taped joint. This concept
comparable alternative when high differential pressures
was first presented in a patent disclosure in the mid-
are involved. The use of the casting does not eliminate
1950s for self-contained fluid-filled joints and has been
the need for experienced splicers. Great care must be
extended to pipe-type applications. These joints utilize
used when installing the casting. If the cable is not per-
an epoxy-resin casting to replace a major portion of
fectly straight, the outer insulation of the cable can be
what, on a conventional splice, would be hand-applied
damaged when the casting is installed.
paper tape. The design most widely known is called the
“Voltalit®,” which is shown in Figure 8-35.
Finally, since LPP insulation provides a smaller OD and
a cost similar to paper-tape insulation, the thermal ben-
A Voltalit® casting is used on each of the three cables
efits of cast epoxy have become less important. The pri-
within the joint. The Voltalit® has a hollow center bore,
mary use of cast-epoxy components in joints has only
through which the cables can be inserted. It includes a
been for application in full stop joints.
stress control electrode (corona shield) imbedded in the
center of the casting. There is a groove in the corona Types of Pipe-Type Joints
shield to receive spring-loaded stoppers that are built in Many different joint types are utilized in pipe-type cable
a specially designed connector. Stoppers are used to systems based on their role in the system apart from the
provide positive contact between the connector and the basic function to join cables.
corona shield and to properly position the insulator
onto the cable. The Voltalit®-style joint eliminates the If the joint has to stop the flow of insulating fluid in the
need for polishing the connector, applying semiconduct- cable system at the point of joint installation this joint is
ing shielding tapes to the connector, and stepping down called a stop joint.
the cable insulation. The casting does not totally elimi-
nate the need for hand taping, since the outer stress con- Likewise, semistop joints partially limit the flow of the
trol cone build-up from the cable insulation to the top of oil.
the casting is still necessary.
Based on the number of cables jointed, the joints are
The epoxy resin generally used for these castings is classified as normal (for two cables), wye (for three
loaded with quartz or silica flour and has a thermal cables), and “H” (for four cables).
Figure 8-35 “Voltalit®” cast-epoxy joint insulation.
8-29
A joint is called trifurcating when the pipe line with LPP-taped insulation has higher radial and longitudinal
three single-phase cables is at one end of the joint and breakdown strength than conventional kraft paper,
three pipe lines each for the single-phase cable are at the allowing LPP-insulated joints to be smaller than compa-
other end of the joint. rable paper-insulated joints of the same voltage. An
LPP-insulated joint on an LPP-insulated cable offers the
In the case of a very steep incline of the pipe, an anchor maximum benefit in terms of reduced joint size, because
joint is used to carry most of the weight of the armored both joint length and outer diameter are reduced in
cable, while a skid joint is employed to limit thermal comparison with a paper-insulated system. An LPP-
movement of the cable. insulated joint on a paper-insulated cable offers a bene-
fit in reducing the joint outer diameter. Although the
If the joint is used to connect two cables with different outer diameter can be reduced, the joint length cannot
size conductors, the joint is called reducing. be reduced, because the inner stepping involves the
paper insulation of the cable and is limited by the longi-
Some joints can be used to perform several functions. In tudinal breakdown strength of the paper, not the LPP.
that case the name of the joint type is compounded—
Wye and "H” Joints
e.g., wye, trifurcating and reducing joint.
“Wye” joints are used to connect sets of three single-
conductor cables to permit a tap of a cable circuit. The
Some of these joints are described below.
splice gets its name because it looks like the letter “Y,”
Normal Joint with two cables on one side and one cable on the other.
An example of a normal joint is shown in Figure 8-25. Likewise, an “H” joint connects four cables, two at each
Normal joints are designed for different insulation end of the joint. The design of a 138-kV wye joint is
tapes. For example, a normal joint can be designed for shown in Figure 8-36 and an “H” joint in Figure 8-37.
paper cable with kraft-paper joint insulation, paper (Note that only one phase of the three is shown on both
cable with LPP joint insulation, or LPP-insulated cable of these drawings for clarity.)
with LPP insulation of the joint.
Figure 8-36 Design of 138-kV wye joint (courtesy of USi).
Figure 8-37 Design of a 138-kV H-type joint (courtesy of G&W Electric Company).
8-30
Wye and “H” joints are large and time-consuming to phases in the line pipe to be separated and brought up,
splice, involve unwieldy and heavy steel housings, and through the riser pipes, to the three cable terminations.
require good splicing skills. The joints require a greater
buildup of insulation than a normal joint. For example, The trifurcating joint is the same as a normal joint with
each phase in a 138 kV, 2000 kcmil (1014 mm2) normal regard to the design and method of splicing the cables.
joint has an outer insulating jacket diameter of approxi- Unlike a normal joint, which has a standard reducer on
mately 4 in. (10.2 cm), whereas the outer diameter of the each end attached to the line pipe, a trifurcating joint
connector in a wye joint is 5.75 in. (14.6 cm) or 40% has a standard reducer on the side to the main feeder
greater. Insulating the wye and “H” joints are also more run and a trifurcating manifold on the other side. The
difficult than a straight joint, particularly around the trifurcating manifold allows the individual stainless-
crotch of the connector. To allow a tight insulation steel riser pipes to be attached. A trifurcating manifold
buildup in this area, in addition to normal tape wrap- design is shown in Figure 8-38.
ping, paper triangles must be applied.
A spreaderhead assembly or trifurcating assembly is
Despite the disadvantages noted above, the wye and used in place of the trifurcating joint at the end of the
“H” joints can provide substantial cost savings and pro- circuit when the cables must be separated and brought
vide a wide variety of installation alternatives when a up to the terminations when a joint is not required. The
tap of a circuit is required. For example, if the wye joint spreaderhead assembly accomplishes the same function
were not available, a tap of a pipe-type circuit would as the trifurcating manifold, except that it does not have
require installing another circuit from the closest sub- a splice casing attached to it. The riser pipes are welded
station. The wye joint eliminates the need for termina- to it on one end, and the line pipe is welded to it on the
tions, station bus-work, substation monitoring and other end. The spreaderhead has a large window cut
alarms, substation real estate, and the additional pipe- into the body so that the three cables from the individual
type cable that a tap from the station would involve. riser tubes can be gathered together. This service open-
Depending on the distance of the tap from the closest ing is covered after installing the cable with either a
substation, the potential cost savings of a wye joint can welded window (steel plate) or a dresser coupling.
be significant.
Spreaderhead and trifurcator assemblies are basically
Wye and "H" splices have been successfully used on the same, only that a spreaderhead is installed above
pipe-type cables at 69 and 138 kV for many years. The ground, typically directly under the pothead support
connector of the wye and “H” joint is manufactured structure, while the trifurcator is installed below ground
from cast copper and has compression-type connector in a manhole. Figure 8-39 shows the spreaderhead
ends. The connector is typically pressed on with a 4-ram installed in a substation.
press. The housings are bulky, with diameters of 24-
30 in. (61-76 cm) and weights of approximately 475 lbs The body of the spreaderhead or trifurcator assembly is
(216 kg) for each sleeve section. made of carbon steel. Wherever a single phase passes
through metal, such as where the cables enter the riser
Trifurcating Joints and Spreaderheads
pipe, that plate must be nonmagnetic stainless steel to
Trifurcating joints or spreaderhead assemblies are used
avoid overheating by magnetic losses. The stainless-steel
at the ends of each cable circuit to allow the three cable
Figure 8-38 Example of trifurcating manifold joint design.
8-31
components, including the reducer and couplings, are tubes are installed in each of the three holes. The cable
typically specified as Type 304, nonmagnetic, since outer diameter is built up with hand-applied tapes in
stainless steels are not all nonmagnetic. the area of the tubes, and then each cable is inserted
into one of the tubes. By compressing the tubes with
Semistop and Stop Joints
stainless-steel binding clamps, a relatively good fluid
Pipe-type cables are generally not limited by hydrostatic
seal is provided across the plate area. Semistops
head pressure; they can be installed over long distances
generally allow some amount of fluid to pass through
involving elevation changes. This creates a potential
the tape structure of the cable and between the barrier
problem. If the cable pipe is accidentally broken, with
tube and the cable—hence the name “semistop.”
no internal obstructions to segregate sections, fluid loss
could be catastrophic. For example, a 5-mile-long
The semistop function is usually only desirable during
(8 km) 345-kV feeder in a 10-in. (25.4 cm) diameter pipe
an emergency situation, such as a pipe break or during
would hold approximately 72,000 gallons (272,520 l ) of
maintenance, when the feeder pipe is opened and fluid
free fluid. If the cable was installed in a location with a
loss is being prevented by application of a fluid freeze.
gradual, consistently increasing elevation change from
Under these circumstances, limiting hydraulic head
one end to the other and there was a breach of the pipe
pressure is essential. The situation is made more
at the low end of the circuit, it would be possible to lose
complex because, during normal conditions, a semistop
all of the fluid in the system. Semistop joints can signifi-
is not desirable and, in fact, could cause operational
cantly limit the amount of fluid flow from one side of
problems. During normal operation, for example, the
the semistop to the other.
fluid in the system expands and contracts in response
to temperature changes due to the electrical load
The splicing of the three phases is no different from
variation on the circuit. The fluid must be able to flow
that of a normal joint. The difference lies in the
from the pressurization plant into and out of the
semistop plate within the casing. In a typical design,
circuit with no constraints. If fluid feed is constrained
there is a nonmagnetic stainless-steel plate installed on
due to semistop fixtures, starvation could occur, and an
one side of the joint casing, usually bolted to an inter-
electrical breakdown of the cable is possible. Another
mediate reducer. Fluid flow between the reducer face
example of a problem is forced circulation or cooling.
and the semistop plate is prevented by means of an O-
Many circuits depend on forced circulation or cooling
ring or flat gasket seal. There are three holes in the
to obtain required ampacity. Obviously, forced
plate to allow each cable phase to be installed, and
Figure 8-39 Spreaderhead installed in a substation (courtesy of USi).
8-32
circulation cannot be provided if semistops are Anchor and Skid Joints

installed. These problems are solved by installing Anchor and skid joints are used with armored cable
bypass piping around the semistop. when cable is installed on a very steep incline or in long
vertical risers. The anchor joint is installed on the top
Bypass piping is provided to allow pressurization of all and supports most of the weight of the cable. The skid
sections of the circuit and to allow slow- or high-speed joint is installed at the bottom of the incline or riser and
fluid circulation during normal conditions. During an permits the cable to expand and contract with load
emergency, the bypass piping is valved closed. The valv- changes. Cable installations at hydroelectric sites and
ing of the bypasses can incorporate manual valves, tunnels are typical application areas for these joints.
motor-operated valves, or hydraulic fuses. Bypass pip-
ing installed around a 345-kV semistop joint is shown in As in the case of the semistop joint, the anchor and skid
Figure 8-40. joints use conventional splicing techniques. The casing
design is modified to accommodate the mechanical
Semistops can be designed to withstand differential components. An anchor joint is shown in Figure 8-42.
pressures of 30-200 psi (260-1380 kPa), depending on As in the case of the semistop joint, a nonmagnetic
the design of the semistop tube and the materials used. stainless-steel plate is installed in an intermediate cas-
Higher differential pressure ratings would require the ing. The plate has three holes drilled through it, and
use of a full-stop joint. This is possible by using joints three stainless steel tubes extend from the holes on the
that incorporate epoxy insulating bushings that are cast side toward the joint. The cables are each passed
into a stop plate, such as a Voltalit®-type joint. A com- through one of the tubes on the plate before splicing
bination Full Stop Anchor joint with diffusion cham- begins. The armor wire is brought back to the stainless
bers and bypass piping installed is shown in Figure 8-41. steel tube, wrapped over it, and bound down to the tube.
It is this armor wire that supports the cable, not the con-
ductor/connector. The anchor plate is solidly bolted to
Figure 8-40 Semi-stop joint with diffusion chambers and bypass piping installed (courtesy of USi).
Figure 8-41 Full Stop Anchor joint for 345-kV LPP insulated cable shown with bypass piping installed (courtesy of USi).
8-33
Figure 8-42 Anchor joint detail (courtesy of USi).
Figure 8-43 Skid joint detail (courtesy of USi)
an intermediate reducer. The skid joint at the bottom of • Single-conductor, hollow-core cable with a metallic
the riser is similar to the anchor, but the plate holding sheath
the cables is allowed to move somewhat. This accommo- • Compressed or welded connector with central duct
dates expansion and contraction of the cable, but pro-
pin, which is a means of attaching the conductors
tects the lower cable sections in case of failure of the
and controlling the fluid flow in the hollow core
anchor at the top. A skid joint design is shown in Figure
8-43. Additional holes are provided in the anchor and • Electrical insulation and electrical shielding
skid plates to permit free fluid flow past the assembly, • Metallic casing, which operates with fluid pressurized
eliminating the need for external by-pass piping. between 15 and 100 psi
(105-690 kPa)
The fluid velocity in rapid circulation systems requires • Corrosion protective covering.
that either a baffle plate or diffusion chamber be utilized
where fluid-inlet piping attaches to cable or riser pipes.
The decision to use a simple baffle plate or more com-
plex diffusion chamber depends on the flow rate of the
fluid. Diffusion chambers prevent the rapidly circulat-
ing fluid from impinging directly on the cable and even-
tually damaging the top papers or shaking it apart. A
diffusion chamber for installation on a stainless-steel
riser is shown in Figure 8-44.
The high flow rates for forced-cooled circuits can cause

high differential pressures around the cable system.
Casing designs should be reviewed to ensure that sleeves
are adequate to withstand the higher operating
pressures. See Chapter 9 for additional discussion of
forced cooling.
8.2.4 Self-Contained Fluid-Filled Joints

Normal Joint General Description Figure 8-44 Diffusion chamber for installation in a
A Self-Contained Fluid-Filled joint consists of: stainless-steel riser.
8-34
The normal joint is used at all locations where no fluid pressed on. The two male pieces are then joined together
supply or signal alarm connections are required. The over a female center piece. This design permits easier
metallic screen interruption joint is used to limit sheath replacing of one cable section without disturbing the
losses, and to permit sheath voltage equalization by spe- other. Single-piece compressed connectors are generally
cial bonding methods. A self-contained oil-filled joint used outside North America. Solder connectors are rec-
containing the screen interruption is shown in Figure ommended as alternatives only on conductor sizes
8-45. The insulating gap is omitted or shorted for a non- below 750 kcmil (380 mm2).
insulating joint.
The 4-ram indents are filled with lead plugs or epoxy
Self-contained normal joints are available up to 1100 kV filler to smooth the connector surface. A thin piece of
and can be installed on either paper (Occhini et al. 1978) copper is then applied over the indents to further
or LPP insulation (Couderc et al. 1996). Self-contained smooth the surface.
LPP cables, with LPP joints, have been installed com-
mercially outside of North America (Minemura et al. Electrostatic screens using carbon-black crepe tape are
1989). Self-contained, three-conductor normal joints normally applied over the conductor and connector and
have been used in the past at voltages from 15 to 69 kV over the taped insulation of the joint. The carbon-black
and up to 138 kV outside North America. Conductor crepe tape is preimpregnated with the dielectric fluid of
sizes up to 1000 kcmil (507 mm2) have been installed. the cable.
Cable Connectors
The semiconducting tape is normally butt-lapped over
In jointing SCFF cable, it is essential that the cable be
the connector. The carbon-black tape is half-lapped
kept under positive fluid pressure at all times to prevent
over the joint insulation and terminated at the cable
the entrance of air and moisture. The conductor of a
insulation shield.
fluid-filled cable is a hollow core, which can be either
self-supporting or supported by a steel spiral. See Chap- Insulation and Shielding
ter 3, Conductors. SCFF joints follow the design principles described in
Section 8.2.1. Figure 8-45 shows a typical SCFF splice
The copper connectors are sometimes two pieces, each construction. Most SCFF joints are insulated using
half having a conductor socket and a valve to shut off half-lapped preimpregnated crepe paper tape to build
the fluid flow temporarily, with a telescopic fit between the desired contour. Preimpregnated paper tapes and
the two halves. A flanged steel core support is inserted paper rolls are also used for insulating materials. In
into each end of the cable hollow core to support pres- order to correctly apply the paper tapes in the connector
sure from the ram indents. The connectors are generally area, the oil pressure at the joint position must be lower
pressed on to the conductor with a 4-ram press. than about 40 psi (275 kPa). If this limit is exceeded, it is
necessary to freeze the cable near the joint.
The telescopic connection between halves is made by
pressing the outer shell into preindents milled into the A tinned, all-copper mesh tape is applied over the
male portion of the connectors. Another common carbon-black tape on the joint. This metal tape is
design uses three pieces. Each cable end has a male piece normally wrapped over the metal cable sheath at both
Figure 8-45 Self-contained screen interruption joint (courtesy of G&W Electric Co.).
8-35
ends. The tape is soldered to the sheath to provide Corrosion Protection Covering
good grounding. The joint casing must be protected from the external
environment to avoid corrosion of the casing and of the
On insulated normal joints, the outer carbon-black tape exposed cable sheath. In case of specially bonded sys-
and the copper mesh braid applied over the insulation tems, the corrosion protection must also ensure a suit-
must be terminated in an insulated gap. Refer to “Sec- able insulation. Different approaches are used,
tionalizing” in Section 8.4.1 for a more detailed expla- including self-amalgamating tapes, heat-shrinkable
nation. The gap is constructed using half-lapped layers tubes, plastic coating of the sleeves, and fiberglass or
of crepe tape or paper roll between and over the shield- rotomolded boxes filled with hydrophobic materials.
ing materials. These solutions are generally combined.
Stop Joint Description
After the shielding is completed, spacer blocks are tied
A self-contained fluid-filled stop joint consists of:
in place on the shielding using cotton twine. The cotton
twine is factory-impregnated with a dielectric fluid. New • Single-conductor hollow cable with a metallic sheath
designs use nylon string, which does not require impreg- • Fluid valve connector, which attaches the conductors
nation. These spacer blocks center the joint and prevent and controls the fluid flow in the hollow core
the shielding braid from contacting the casing, which
could cause undesired circulating currents.
• Stop insulator, which fully terminates each cable end
• Electrical insulation and electrical shielding
Joint Casing
Joint casings are made up of two or three nonmagnetic • Metallic casing, which is filled with fluid between 15
metallic sleeves. In the case of an insulated joint, one and 100 psi (105-690 kPa)
sleeve is equipped with an insulator. In many cases, util- • Corrosion protective covering.
ities standardize on insulated joint casings, then connect
a jumper across the insulator when it is not needed. Stop joints are used to divide a cable circuit into inde-
pendent, hydraulically fed sections, with each section
The joint sleeve is generally manufactured from a being fed at the stop joint location from gravity (Type
copper alloy with an opening that will allow it to slide CC) or pressurized (Type AC or DC) reservoirs. Refer
over the cable sheath on one end and the completed to Chapter 9 for a description of reservoirs. Stop joints
joint on the other end. The copper sleeve is hot-tinned can be also used to reduce the hydrostatic head pressure
to provide a good soldering surface and corrosion resis- on the cable system. Self-contained stop joints are
tance. The sleeve is attached to the metallic sheath by a available up to 750 kV and can be installed on either
solder wipe, or by a mechanical means in the case of paper or LPP insulation. Self-contained LPP stop joints
some aluminum sheath cables. Reinforced wipes are have been installed outside of North America. Three-
required with double-lead sheaths and with reinforced- conductor stop joints have been used for voltages
lead construction at operating pressures above 40 psi b e t we e n 1 5 a n d 1 3 8 k V. C o n d u c t o r si z es u p t o
(275 kPa). A lead wedge is used to center the sheath in 1000 kcmil (507 mm2) are in use. A typical stop joint for
the wiping sleeve before the wipe is made. A threaded SCFF cable is shown in Figure 8-46.
cylindrical ground connection boss is commonly sup-
Design and Materials
plied, silver brazed to the casing to provide easy ground-
ing or bonding of the cable system. This section reviews the design and materials issues of
the major components that are important in applying
In the case of a three-piece casing, a center sleeve is sol- and evaluating self-contained stop joints.
dered or connected with bolted flanges to each of the Cable Connectors
end sleeves. In a normal joint, the hollow-core connector permits
only axial flow of the dielectric fluid through the joint.
The insulated sleeve generally consists of an epoxy ring In a stop joint, the copper connectors have several dif-
with embedded inserts for fixation to copper alloy ferent openings that allow fluid flow inside the stop
sleeves. Two “O”-ring gaskets placed in a groove pro- insulator end. The valve inside the connector shuts off
vided in the copper sleeve’s flange ensure the oil sealing. the hollow core and permits the evacuation of each end
Another design utilizes a high-alumina porcelain insula- through the vents provided. After filling, the valve is
tor with nickel iron sleeve at each end. One sleeve is sil- opened and fluid can flow from the cable core. A
vered soldered to the housing, and the second to the flanged steel core support is inserted into the hollow
wiping sleeve. All the metal parts are hot-tinned. core to support pressure from a hydraulic compression
8-36
tool, then a thin copper shield is generally applied over must be connected to a pressurized reservoir or to the
the connector in order to obtain a smooth surface. cable with higher fluid pressure.
Insulation and Shielding Corrosion Protection Covering
Insulation and shielding materials and application are The function of this protection is the same as described
identical to those for SCFF normal joints. earlier for normal joints. However, due to the irregular
profile of the joint casing, the solution with a fiberglass
Stop-Insulator
box filled with a hydrophobic resin is generally adopted.
The stop insulator may be either cylindrical or conical
in shape. The more recent insulators are made of cast- Arc-Proof Covering
molded epoxy. The conical stop insulator is assembled Arc proofing is normally applied to protect the cables
to the end casing using a retained gasket, split clamp, and joints against arcs and fires in the manholes and
and bolts. The connector stem is gasketed to the stop vaults. Arc proofing is applied over the corrosion pro-
insulator. The connector gasket is compressed using a tection covering. One half-lapped layer of arc-proofing
large nut threaded on the connector stem. After the stop tape is applied from duct mouth to duct mouth.
insulator is assembled, a solder wipe is made between
the casing and the metallic sheath. Internal Pressure Characteristics
Internal pressure characteristics of self-contained joints
Center-Section Materials are critical to system operation because the joints are
The center section of the stop joint is assembled after usually at the low elevations of the cable system. The
both stop ends have been completed and filled with operating pressures in these cases may limit cable
dielectric liquid. A copper stress-cone shield is slipped system operations.
over each stop insulator and attached to each casing
end. A split-clamp connector is used to connect the The internal pressure characteristics are determined by
two connector stems together, and a copper shield is the thickness of the copper sleeve assembly. Table 8-4
placed over this sleeve. Half-lapped layers of crepe
paper tape are applied over the connector area and the
stop tube ends. Table 8-4 Typical Operating Pressures for Self-Contained
Joints
The clearance between the connector and the casing is Standard Medium High
psi (kPa) psi (kPa) psi (kPa)
filled with a set of concentric, split insulating tubes held
Nominal operating pres-
together with a pair of insulating rings. Insulating spac- sure
15 (105) 40 (275) 100 (690)
ers are used between the insulators. These tubes provide Max. continuous operat-
the radial electrical strength required due to the large ing pressure (over 10 22 (150) 55 (380) 138 (950)
volume of liquid in the casing. hours)
Max. transient pressure
Joint Casing 24 (165) 60 (415) 150 (1035)
(up to 10 hours)
Joint casings are made up of three sleeves. In most cases, Max. peak transient 32 (220) 80 (550) 200 (1380)
the stop joint is also equipped with an insulator. Details Max. field (proof) test 28 (195) 70 (485) 175 (1205)
are the same as described earlier for a normal joint. The Factory test (1 hour) 40 (275) 100 (690) 250 (1725)
center casing is vacuum-filled with dielectric fluid and
Figure 8-46 Self-contained fluid-filled stop joint (courtesy of Prysmian Cable & Systems).
8-37
shows the internal pressure characteristics for standard, The following tests are required by IEEE Standard 404-
medium, and high-pressure SCFF joints. 2000:
• Production tests, performed by the manufacturer on
8.2.5 Industry Specifications and Standards 100% of premolded and cold shrinkable (prefabri-
Transmission cables joints are covered by a number of cated) joints produced. Production (routine) tests
standards both national and international. The most involve partial discharge and ac withstand or impulse
commonly used transmission joint standards and their withstand voltage tests.
major requirements and characteristics are described
• Design tests on the joints rated 69 kV and above are
below.
carried out on a total of four samples (two samples in
National Standards air and two in water) of joints in order to qualify each
The national standard that is used in North America individual type and to demonstrate the compliance of
and some other countries is the IEEE Standard 404- the joints with the standard. The tests in water are
2000 “IEEE Standard for Extruded and Laminated not required if the joint design incorporates a solid-
Dielectric Shielded Cable Joints Rated 2500 V to metal housing that is welded or soldered to a solid
500000 V.” The subsequent revision of this standard, cable sheath or pipe. Tables 8-5 and 8-6 show the
currently in process, will likely have more extensive requested test values.
requirements for these joints. This standard is based on
functional and design test requirements and covers both When transition joints are being tested, all electrical
normal joints and transition joints. test values are based on the cable with the lower test
requirements.
Table 8-5 Voltage Ratings and Test Levels for Extruded-dielectric Cable Joints Rated 69-500 kV
Voltage Voltage BIL and full AC withstand voltage and DC Minimum
rating rating wave time withstand partial
phase-to- phase-to- (kV crest) Column A Column B voltage discharge
phase, ground, 15 min 6h 15 min voltage
U U0 @3 U0 @2.5 U0 (kV) level,
(kV rms) (kV rms) (kV rms) (kV rms) 1.5 U0
(kV rms)
69 39.8 350 120 100 240 60
115 66.4 550 200 166 300 100
138 79.7 650 240 200 315 120
161 93.0 750 280 232 375 140
230 132.8 1050 400 332 525 200
345 199.2 1300 600 500 650 300
500 288.7 1550 870 725 775 435
Table 8-6 Voltage Ratings and Test Levels for Laminated Dielectric Cable Joints Rated 69-500 kV
Voltage rating Voltage rating BIL and full AC withstand test DC withstand
phase-to-phase, phase-to- wave (kV voltage
U ground, crest) 15 min
(kV rms) U0 Time Voltage (kV)
(kV rms) (h) (kV rms)
69 39.8 350 24 100 175

115 66.4 450 24 170 225
120 69.3 550 24 170 275
138 79.7 650 24 200 325
161 93.0 750 24 230 375
230 132.8 1050 24 330 525
345 199.2 1300 24 500 650
500 288.7 1550 24 720 775
8-38
The standard specifies a design test sequence for each - Partial-discharge test
joint type (extruded-dielectric cable joint, laminated- - Voltage test
dielectric cable joint, and transition joint). The 30-day
The test requirements are shown in Table 8-7.
cyclic aging test at double line-to-ground voltage and
conductor temperature of the cable within 5°C of its • Sample tests on accessories. These tests are to be car-
rated emergency operating temperature for a period of ried out by the manufacturer on those accessories
at least six hours in each cycle, is mandatory for that cannot be routine-tested (i.e., taped or field-
extruded-dielectric cable joints and transition joints. It molded). The accessory shall be fully assembled in its
is not required for laminated cable joints. final configuration. The number of accessories to be
tested is one per contract if the number of accessories
More details on the test requirements are given in the is above 50.
Chapter 14. - Partial-discharge test
International Standards - Voltage test
Based on experience and CIGRE studies, IEC interna- The test requirement are shown in Table 8-7.
tional standards have been made available for HV • Type test. These tests are to be carried out in order
transmission cables. The IEC standards are arranged to qualify each individual type of accessory and to
so that the test for cable accessories is included in the demonstrate satisfactory performance characteris-
same standard as the cable. There is a subdivision of tics to meet the intended application and compliance
standards between extruded, oil-filled, internal pres- with the standard. The following tests can be carried
sure gas-filled, external pressure gas-filled, and high- out on the complete cable system or on the
pressure pipe–type. individual accessory:
IEC Standards for Extruded Cable Accessories - Partial-discharge test at ambient temperature.
Two standards cover transmission extruded cable acces- In the case of testing the cable system, the cable
sories for rated voltages above 30 kV up to 500 kV. is to be subjected to a bending test before the
installation of accessories.
The first standard is IEC 60840-2004 – “Power Cables - Tan delta measurement at high temperature
with Extruded Insulation and Their Accessories for (only for test as a cable system)
Rated Voltages Above 30 kV (Um = 36 kV) up to 150 kV
- Heating cycle voltage test
(Um = 170 kV) - Test Methods and Requirements.”
- Partial-discharge test at ambient and at high
This standard covers the test requirements for cables temperature
and accessories, based on the complete cable system - Lightning impulse voltage test, followed by a
approach or the individual approach. The test require- power frequency voltage
ments are the same as those of the cable but with some - Test of outer protection for buried joints
peculiarities concerning the routine and sample test.
- Examination of the accessories (or of the cable
The test procedures and sequences are the following:
system) after the completion of the above test
• Routine test on prefabricated accessories. These tests The tests requirements are shown in Table 8-7.
are to be carried out by the manufacturer on the
main insulation of each accessory.
Table 8-7 IEC 60840 Voltage Ratings and Test Voltages

Rated voltage Highest volt- Routine volt- Partial dis- Tan δ measure- Heating Lightning AC Voltage test after
phase-to- age for equip- age test for charge test ment cycle volt- impulse volt- installation for 1hour
phase ment 30 minutes 1.5 Uo Uo age test age test kV
U Um 2.5 Uo kV kV 2Uo kV
kV kV kV kV
60 to 69 72.5 90 54 36 72 325 72
110 to 115 123 160 96 64 128 550 128
132 to 138 145 190 114 76 152 650 132
150 to 161 170 218 131 87 174 750 150
8-39
Prior to the electrical-type tests of the accessories, the The prequalification test is to be carried once unless
insulation thickness and the cable dimensions are to be there is a substantial change in the cable system with
measured to evaluate the electrical gradient at inter- respect to material, manufacturing process, and design
faces. This is one of the factors necessary for evaluation and design levels. The test is to be considered valid for
of the validity of the range of type test approval. all those cable systems having the same or lower voltage
ratings as long as the calculated electrical stresses at the
The second standard for extruded transmission cables is insulation screen are equal to or lower than that of the
IEC 62067-2006 – “Power Cables with Extruded Insula- tested cable system. More details are available
tion and Their Accessories for Rated Voltages above in Chapter 14.
150 kV (Um = 170 kV) up to 500 kV (Um = 550 kV) -
IEC Standards for Laminated Dielectric Cable Accessories
Test Methods and Requirements.”
A number of IEC standards are applicable to all types
of laminated dielectric cables. These standards have
The requirements of this standard are very similar to
been issued for many years and now are included in the
those of the previous IEC 60840 with some
maintenance list of IEC. The standards are listed below:
major differences:
• IEC 60141-1 – 1998. Tests on oil-filled and gas-
• The type test procedure is carried out on a cable sys-
pressure cables and their accessories - Part 1: Oil-
tem basis only, the procedure and test sequence are
filled, paper or polypropylene paper laminate
the same as that of IEC 60840, but in addition, a
insulated, metal-sheathed cables and accessories for
switching impulse test is required for cable systems
alternating voltages up to and including 500 kV.
having Um 300 kV.
• An additional long-term prequalification test is
pressure cables and their accessories. Part 2: Internal
requested in order to demonstrate satisfactory long-
gas-pressure cables and accessories for alternating
term performance of the complete cable system (cable
voltages up to 275 kV.
and accessories).
Table 8-8 indicates the requested test values. Particular pressure cables and their accessories. Part 3: External
attention is to be paid to the prequalification test proce- gas-pressure (gas compression) cables and accessories
dures where the test circuit arrangement is to be repre- for alternating voltages up to 275 kV.
sentative of the installation design condition in order to • IEC 40141-4 – 1980. Tests on oil-filled and gas-
take into account the thermomechanical forces acting pressure cables and their accessories. Part 4: Oil-
on cable accessories. impregnated paper-insulated high pressure oil-filled
pipe-type cables and accessories for alternating
voltages up to and including 400 kV.
Table 8-8 IEC 62067 Voltage Ratings and Test Voltages

Rated Highest Routine Partial dis- Heating cycle Lightning Voltage test Switching AC Voltage test
voltage voltage for voltage test charge test voltage test impulse after light- impulse after installa-
phase-to- equipment kV/minutes 1.5 Uo 2Uo voltage test ning impulse voltage test tion for 1hour
phase Um kV kV kV kV voltage test kV kV
U kV 2Uo
kV kV
220 to 230 245 318/30 190 254 1050 254 - 180
275 to 287 300 400/30 240 320 1050 320 850 210
330 to 345 362 420/60 285 380 1175 380 950 250
380 to 420 420 440/60 330 440 1425 440 1050 260
500 550 580/60 435 580 1550 580 1175 320
8-40
8.3 TERMINATIONS The termination design must take into account that the
maximum electrical stress occurs at the end of the cable
8.3.1 Theory and General Description screen termination rather than at the cable conductor.
The sharp edge effect of the screen termination intro-
Function duces a considerable field enhancement, especially along
The basic function of a cable termination is to transition the surface of the insulation. A stress-control device
from an underground cable to another electrical compo- must be used in the termination to gradually decrease
nent while withstanding both mechanical and electrical the magnitude of the longitudinal and radial stresses.
stresses. This control device for its intrinsic construction has a
larger diameter than the cable.
There are three types of terminations:
• Outdoor/indoor termination, which connects the Three methods are commonly used to maintain the
cable to an aerial bus bar or to an overhead line stresses below the permitted limits for the dielectric
• Metal-enclosed GIS termination, which connects the concerned:
cable to a bus bar insulated with pressurized gas (usu- • Geometrical-capacitive stress control, accomplished
ally SF6 or N2-SG6 mixture). through a molded elastomeric stress cone or thermo-
• Oil-immersed transformer termination, which set resin stress cone, incorporating a suitably shaped
connects the cable to metal-clad bus bar isolated with electrode;
oil. • Capacitive grading control, using either a cylindrical
capacitor cone incorporating a series of cylindrical
To fulfill its purpose, each type of termination must conductive foils, or individual toroidal capacitor
provide: “doughnuts” connected in series; and
• High-current connection from the cable conductor to • Resistive or high-permittivity stress control accom-
an external electrical component (e.g., bus bar) plished through a layer of elastomeric material pre-
• Insulation between the cable conductor and ground loaded with a nonlinear resistance or high-
with the same performance as the cable permittivity filler. This solution is an extrapolation of
the stress control used on MV extruded-dielectric
• Provision to support the cable and the ability to with- cable terminations and is applicable only up to 60 kV
stand cable thermomechanical loads and external due to the relatively low strength and high losses of
forces such as wind, ice, and weight of the bus if ter- the nonlinear material.
mination is used to support rigid bus connection
Technologies
• Protection to the cable insulation and sheath against
Different stress relief devices as well as different insula-
the ingress of water and the ingress of pressurized
tors that may be porcelain, composite, or epoxy, are
dielectric fluid or gas from adjacent metal-clad bus
available.
bar housing
• High-current connection to permit the flow of short- Outdoor/indoor Terminations
circuit current from the cable metallic sheath or A typical stress relief cone for high-voltage extruded
shield wires to the system earth. cable and an insulator for an outdoor termination is
shown in Figure 8-56. The semiconducting stress-con-
The termination may also perform the additional task trol electrode is incorporated in a high-strength dielec-
of insulating the metallic sheath or shield wires from tric material, and the stress cone is interference-fitted
earth potential—in case of special sheath-bonding con- over the cable, so that the electrode becomes the termi-
nections, or in case it is necessary to check the integrity nation of the cable screen. This design is more com-
of cable oversheath. monly used and has been installed on cable systems up
to 500 kV (Rosevear et al. 1995; Beker et al. 1998; Van
Design Approach Der Wijk et al. 1998; Attwood et al.1998).
The termination has to accomplish the transition
between the high cylindrical electric field in the cable Another design called prefabricated composite and
(up to 381 V/mil [15 kV/mm] at operating voltage) and insulator outdoor termination incorporates spring load-
the relatively low operating stress of the metal-enclosed ing or other forms of mechanical force to compress the
apparatus (up 76.2 V/mil [3 kV/mm] at operating volt- stress cone against the cable and a conically shaped
age) or the much lower operating stress of the bus bar in epoxy casting inside the insulator (see Figure 8-58)
air (about 25.4 V/mil [1 kV/mm]). (Takeda et al. 1995; Choi et al. 1998).
8-41
In the case of a laminated cable, the electrode of the connector is required to transmit current from the con-
stress cone is incorporated into an epoxy resin casting ductor to the high-voltage electrode.
or is obtained by application of semiconductive and
metallic screens over properly shaped paper buildup Another design, which is today preferred for extruded
(paper roll) at the edge of the cable shielding (see Sec- cables up to 500 kV, is the prefabricated composite dry
tion 8.3.3). metal-enclosed GIS termination shown in Figures 8-47
and 8-62. The interface dimensions of GIS terminations
For these types of outdoor/indoor terminations, the with the switchgear have been standardized (IEC 60859
outer housing is a porcelain insulator or a composite for extruded and SCFF cable terminations, and IEEE
insulator filled with an insulating liquid or gas. The 1300 for both extruded cable and laminated cable termi-
composite insulator has a rigid cylindrical or conical nations) in order to facilitate interchangeability and
core of thermoset resin reinforced with glass fiber onto define the area of responsibility in case of problems. A
which elastomeric sheds have been fitted or molded. The shorter length is standardized for dry GIS terminations
sheds are usually formed from silicone rubber with anti- compared to stress cone and insulator wet types.
tracking characteristics. The insulator profile and creep-
age length generally depend on the pollution level Oil-immersed transformer terminations have the same
specified for the installation site (see Section 8.3.5). construction as GIS terminations, except for a larger
top corona shield in order to reduce the stress in the
The capacitor cone and insulator outdoor termination transformer oil.
shown in Figure 8-59 has been adopted for 400-kV
Design Challenges
XLPE cable systems and for laminated cable up to
Although the cable insulation is only subjected to a
500 kV. For extruded cables, due to the large core-
radial stress, both radial and longitudinal stresses are
diameter variation versus temperature, it is necessary to
present in the termination.
adopt crepe-paper bedding between the bore of the pre-
fabricated condenser cone and the cable to minimize the
In order to correctly design the termination, it is
longitudinal oil channel, which is a critical interface.
necessary to calculate these stresses and compare them
with the ones allowed by the different materials and
In the prefabricated elastomeric sheds outdoor termina-
components used in the termination. Most critical are
tion, shown in Figure 8-53, the antitracking elastomeric
the stresses acting along the interfaces. The electrical
housing and the stress cone are generally molded in one
strength of the interfaces is relatively low, and depends
piece, which is interference-fitted over the cable core. In
on different factors including the quality of the
this way, a completely “dry” design is obtained, but the
interfaces, the pressure between the interfacing
termination may not be self-standing, so in this case it
materials, and the lubrication of the interface (CIGRE
must be suspended to the overhead line conductor or to
2002). The overall dimensions of the termination are
a supporting insulator, such as a surge arrester (Le Peu-
established by keeping the radial and longitudinal
rian et al. 1992; De Ridder et al. 1999).
stresses within safe values.
Self-supporting dry terminations are being developed.
One design is an extrapolation of a dry GIS termination
(Amerpohl et al. 2002), and another design replaces the
oil filling with a gel (Dejean et al. 2003).
Metal-enclosed GIS/ Oil-immersed Transformer Terminations
Figure 8-61 shows schematically a stress cone and insu-
lator metal-enclosed GIS termination, which is concep-
tually similar to the stress cone and insulator outdoor
termination described above. In this case, the insulator
is usually formed of cast thermoset resin, and offers a
relatively short creepage due to the high strength of the
pressurized SF6 gas surrounding the termination. In
some designs, especially for laminated dielectric cables,
the insulator is provided with an embedded solid high- Figure 8-47 Dry-type termination for SF6 gas-insulated
voltage electrode in order to prevent any ingress of the system connection (courtesy of Prysmian Cables &
pressurized SF6 gas into the cable. In this case, a plug-in Systems).
8-42
It is necessary to control the thermomechanical forces reduce this conductor thrust and prevent a possible flex-
transferred to the termination by the cable during its ion of the cable inside the insulator, a suitable offset is
operation, and to check that the insulator is able to with- provided for the cable entering the termination.
stand the pressure of the fluid inside or outside the ter-
mination and the mechanical stresses transferred to the Different approaches are adopted concerning the cleat-
termination by the surrounding environment (e.g., wind ing of the cable in air below the termination. Generally
or ice loading on bus bars, earthquake, vibrations). for oil-filled cables, a rigid cleating to a bend-controlling
beam is used in order to avoid the risk of fracture of the
The designer must also check that the termination, due to metallic sheath due to fatigue and the leaking of the
its much larger diameter than the cable, is not a thermal pressurized dielectric fluid. In comparison, for extruded
bottleneck for the current rating of the cable system. cables, the bend is generally suspended and free to move,
so a limited thrust is transferred to the termination.
Electrical Design
As for joints, a computer program is used to calculate Laminated cable terminations must be able to withstand
the equipotential line distribution (see Figure 8-48) and the pressure of the dielectric fluid from the cable insula-
the actual stresses in the critical regions of the termination and allow feeding dielectric fluid back into the
tion—i.e., over embedded electrodes, at the interfaces cable system.
between different components and materials and in the
fluids inside and outside the insulator. If these stresses When the extruded cable termination is filled with a
are greater than the allowable ones, the designer must dielectric liquid, it is necessary to allow for fluid expan-
review the termination drawing, and the electric field sion due to temperature variation. The expansion vol-
calculation must be repeated a number of times until the ume may be accommodated using an air-filled space
stresses are within the allowable range. usually at the top end of the insulator (only for verti-
cally installed terminations), by an external header
Analytical methods are used to define the dimension tank, by an external pressurized feed tank, or by an
and position of the metal foil of the condenser cones internal flexible accumulator (generally placed at the
used as stress control in some types of terminations. bottom of the termination).
A “log log” profile is generally used at the termination To avoid uncontrolled loads being transferred to the ter-
taper of the stress cone in laminated cable termina- mination from the GIS or transformer bus bar, flexible
tions in order to keep constant the axial component of connections must be utilized.
the stress.
Mechanical/Hydraulic Design In areas of high seismic risk, it is necessary to verify the
The cable conductor thrust in terminations is generally design of the outdoor termination insulator and of its
transferred to the supporting structure via the insulator supporting structure.
(conductor-locked designs). Thermal Design
Although terminations have a larger diameter than the
In the case of outdoor/indoor terminations, which cable, their ampacity rating is generally greater than that
present a long and unrestrained cable end, in order to of the cable that they connect. In fact, terminations do
Figure 8-48 Equipotential line distribution in an outdoor termination.
8-43
not have mutual heating effects and the heat is transferred The lower metalwork includes at least an end bell (wip-
mainly longitudinally through the conductor to the air ing sleeve or entrance housing) and a base plate (mount-
(for outdoor terminations) or to the dielectric fluid (for ing plate).
GIS/transformer terminations) outside them and to the
bus bar connection on top of the termination. In addi- Mounting plates are normally nonmagnetic (aluminum
tion, for fluid-filled designs, the dielectric fluid inside the or stainless steel), but also galvanized steel can be used
insulator also contributes to the heat dissipation. if the plate has a radial slot or the current is below 1000
A and the opening in the plate is relatively large.
Special installation conditions—for instance, outdoor
terminations installed in hot climates or transformer Mounting plates are generally insulated from the sup-
terminations in hot oil—may require specific consider- porting structure to allow flexibility in connecting cable
ation during the design of the cable system. metallic sheaths to ground. For this purpose, in outdoor
terminations, four stand-off (pedestal) insulators are
For fluid-filled terminations, forced cooling of the fluid generally used (see Figure 8-49), while an epoxy ring is
or other means, such as cooling pipes, can be adopted to adopted for SF6 and oil-immersed terminations.
improve the thermal rating, if necessary.
As far as the seal between the metallic sheath of the
The basics on ampacity (thermal) analysis of the cable and the termination end bell is concerned, the
pipe-type cable terminations are described below in same solutions described in Section 8.2.1 for joints (see
Section 8.3.3. Sheath Closure section) are adopted.
Main Components of the Termination Installation
Termination designs must be coordinated with the type
Connectors
of cable to be used. Table 8-9 indicates the type of infor-
The same type of connectors described in Section 8.2.1
mation required to order an HV termination.
for joints are used to join the cable conductor to the
stalk of the termination. In outdoor terminations, the
System electrical information and cable data are
part of the stalk exposed to the environment is generally
required to select the correct termination. The same
copper to resist corrosion. If the cable has an aluminum
information is essential for any high-voltage termina-
conductor, a bimetallic rod of aluminum to copper is
tion. Additional information varies with the type of sys-
usually adopted.
tem to be used.
Top Metalwork
The top metalwork of a termination has two functions: The construction of a typical high-voltage termination
• Seal the top of the insulator and the conductor stalk, involves the following steps: (1) install the mounting
plate and any cable supports to the mounting structure;
and
(2) determine the cable cut-off length; (3) install the con-
• Anchor the cable conductor through the conductor nector; (4) remove the cable shielding tapes, jackets etc.;
stalk. (5) install the stress cone assembly; (6) lower the outer
A nonmagnetic material (stainless steel, aluminum, or

brass) is generally used. If the metalwork presents sharp
edges, a corona shield is used to avoid radio interference
effects.
Lower Metalwork
The lower metalwork of a self-standing termination has
the following characteristics:
• Seal the bottom of the insulator to the cable metallic
sheath,
• Anchor the termination to the supporting structure,
and
• Ensure the grounding of the cable sheath, and collect Figure 8-49 Outdoor termination showing the mounting
the insulator leakage current. plate with the supporting insulators (courtesy of Prysmian
Cables & Systems).
8-44
housing over the stress cone assembly; (7) complete As indicated in Section 8.3.1, there are three families of
installation of all termination seals; (8) pressure-test the terminations:
integrity of the termination (pressurized systems only); • Outdoor terminations to connect the cable to the
(9) fill the unit with dielectric fluid if required; and (10)
overhead network,
assemble the aerial lug to the cable connector.
• Metal-enclosed terminations to connect the cable to
8.3.2 Terminations for Extruded-dielectric Cable a gas-insulated substation (GIS), and
An extruded-dielectric termination consists of a single- • Oil-immersed transformer terminations to connect
conductor, extruded-dielectric cable, a stress-relief the cable to a transformer.
device, and an insulated enclosure.
Outdoor Terminations
As the outdoor termination interfaces with air-insulated
Table 8-9 Information Required to Specify a High-Voltage switchgear or equipment, it is subjected to full climatic
Termination conditions.
System Electrical Data
Terminations above 138 kV are generally filled with
Nominal system voltage, kV
unpressurized liquid dielectric.From 69 to 138 kV, both
Maximum system voltage, kV
fluid-filled and dry terminations are available. Outdoor
Lightning impulse (BIL), kV
porcelain insulator assemblies are available from 69 to
Minimum insulator creepage distance. in. (cm) for an installation in
a heavily contaminated area
500 kV (see Figure 8-50).
Switching impulse (BSL), kV for system voltages of 345 kV and
above
Cable Design Data
Type of cable (pipe-type, extruded-dielectric, self-contained, etc.)
Conductor size, material and construction
Shielding material and thickness
Insulation type and thickness
Jacket material and thickness
Additional Information
Factory pressure test requirement for pressurized systems, psi
(kPa)
Maximum allowable continuous working pressure
Maximum transient pressure and allowable time of application
Type of aerial connection
Mounting-plate insulation requirements
Riser pipe size (pipe-type only), in. (cm)
Pressure alarm systems
Type of cable support system (wire armor clamps, kellems grips,
cable clamps, etc.)
Note: Some types of extruded-dielectric terminations require actual

measured diameter over the insulation for correct sizing of the
stress cone. Conductor size and nominal insulation thickness are
all that are required for HPFF and SCFF cables.
Figure 8-50 420-kV XLPE cable outdoor terminations with
porcelain insulator.
8-45
Nonceramic (also called composite) insulators are avail- Terminations are generally prepared at ground level and
able for outdoor terminations up to 500 kV (see then installed with a crane on the pole/tower.
Figure 8-51).
Different kinds of outdoor terminations can be found
Composite insulators are much lighter than porcelain depending on the stress relief device and the antitrack-
insulators. They can be easily installed on poles or towing covering. The main ones are:
ers, requiring no fenced area. Transition towers or poles • Dry-type terminations:
at 230 kV are nowadays very common (see Figure 8-52).
- Prefabricated elastomeric sheds and stress-cone
outdoor termination
- Heat-shrink sleeve outdoor termination
- Elastomeric sleeve outdoor termination
• Fluid-filled-type terminations:
- Stress-cone and insulator outdoor termination
- Deflector and insulator outdoor termination
- Prefabricated composite and insulator outdoor
termination
- Capacitor-cone and insulator outdoor termina-
tion
- Prefabricated composite and capacitor-cone
outdoor termination
Dry-type Terminations
In these terminations, protection against adverse cli-
matic conditions is provided by polymeric sheds. Differ-
ent stress-control devices are adopted depending on the
kind of termination.
Figure 8-51 420-kV XLPE cable outdoor terminations with
composite insulator (courtesy Silec Cable). Dry terminations are available up to 138 kV. Generally
not self-supporting, they require an external support,
which can be a surge arrester or a station insulator.
A l t e r n a t i v e l y, t h e y c a n b e s u s p e n d e d t o t h e
overhead conductor.
• Prefabricated elastomeric sheds and stress-cone out-
door termination.
In prefabricated elastomeric sheds and stress-cone
outdoor terminations, the stress-control device is a
premolded stress cone that is slipped with interfer-
ence on the prepared cable. Individual premolded
sheds are then slipped on the cable before sealing of
the upper part and connection with the conductor
stalk (see Figure 8-53). In some constructions, the
stress cone and the shed form one single premolded
component.
• Heat-shrink sleeve outdoor termination.
In the heat-shrink sleeve outdoor terminations, the
stress control is achieved by means of a resistive or
high-permittivity sleeve, or tape or mastic. Then a
Figure 8-52 Overhead/underground transition of a
230-kV cable line, with porcelain insulators mounted on sleeve is heat-shrunk along the prepared cable, and
the end pole (courtesy of Silec Cable).
8-46
sheds are heat-shrunk onto this sleeve (see Figure stretched onto the sleeve before sealing and installa-
8-54). tion of the conductor stalk (see Figure 8-55).
• Elastomeric sleeve outdoor termination. Fluid-filled-type Terminations
In this termination, which is very similar to the previ- In outdoor fluid-filled terminations, an insulator is
ous one, the longitudinal sleeve is a premolded elas- subjected to adverse climatic conditions and houses the
tomeric sleeve, which is expanded on a support and stress-control device and the prepared cable. This
cold-shrunk onto the cable, and the sheds are insulator can be made from ceramic or thermoset resin,
or may be a composite insulator composed of a rigid
hollow-core insulator covered by premolded polymeric
sheds, which have been extruded, fitted, or molded on
the rigid insulator.
The following fluid-filled outdoor terminations differ by

the stress-control device.
• Stress cone and insulator outdoor termination.
For a stress cone and insulator outdoor termination,
the electrical field is controlled by a stress cone that is
fitted on the cable (see Figure 8-56).
It can be obtained by field molding, using one of the
techniques that have been described for field-molded
joints, by field-taping, or by a factory premolded
part. The insulator is filled using oil or gas. It is
important for the fluid to be compatible with the
stress-cone and cable material.
• Deflector and insulator outdoor termination.
Figure 8-53 Prefabricated elastomeric sheds and stress- For this termination, the electrical field is controlled
cone outdoor termination (courtesy CIGRE). by a metallic or semiconducting elastomeric stress-
control profile deflector fitted on the prepared cable
and immersed in the fluid that is filling the insulator
Figure 8-55 Elastomeric sleeve outdoor termination

Figure 8-54 Heat-shrink terminations (courtesy CIGRE). (courtesy CIGRE)
8-47
(see Figure 8-57). In some cases, some high- ping insulating and conductive tapes on the prepared
permittivity material is applied at the semiconduc- cable.
tive end of the cable as an adaptor. This very simple • Prefabricated composite and capacitor cone outdoor
design is appropriate only for relatively low- termination.
insulation screen stresses.
For a prefabricated composite and capacitor cone
• Prefabricated composite and insulator outdoor termi- outdoor termination, the electrical stress is controlled
nation. by a combination of the two techniques described
For this termination, one premolded stress-cone above (see Figure 8-60). This kind of termination is
made of elastomeric material is positioned on the adopted in Japan for 500-kV rated systems.
prepared cable end and then inserted into a cast
Metal-enclosed GIS Terminations
epoxy insulator (see Figure 8-58). The pressures at
The stress-control techniques are very similar for metal-
the cable core interface, as well as at the stress-cone-
enclosed GIS terminations as to those used in outdoor
to-central-epoxy-insulator interface, are obtained by
terminations, even if the stress distributions in both
means of compression devices such as springs.
types of terminations are not exactly the same. The
• Capacitor cone and insulator outdoor termination. main types are:
In some traditional designs of capacitor cone and • Stress cone and insulator termination
insulator outdoor terminations, the stress control is
achieved by means of capacitors (see Figure 8-59). • Deflector and insulator termination
These capacitors can be either prepared coils of • Prefabricated composite dry-type termination
capacitors dimensioned to achieve a linear voltage • Capacitor cone and insulator termination
distribution, or they can be made on-site by wrap-
• Directly immersed termination.
Figure 8-56 Stress cone and insulator outdoor terminations (courtesy CIGRE).
8-48
Stress Cone and Insulator GIS Termination Deflector and Insulator GIS Termination
In this termination, the electrical field is controlled by a As for outdoor terminations, the electrical field is con-
stress cone that is fitted on the cable (see Figure 8-61). trolled by a metallic or semiconducting elastomeric
stress control profile deflector fitted on the prepared
The stress cone can be obtained by field-molding, using cable and immersed in the fluid that is filling the insula-
one of the techniques described for field-molded joints, tor. In some cases, some high-permittivity material is
by field-taping, or by a factory-premolded part. The applied at the semicon end of the cable adaptor. This
insulator is filled with oil or gas. When SF6 gas is used, it very simple design is appropriate within certain limits of
can be provided by external piping or by the GIS the electrical field over insulation.
through a hole on the insulator. In this case, the
Prefabricated Composite Dry-type GIS Termination
p r e s s u r e m o n i t o r i n g i s p rov i d e d b y t h e G I S
For this design, one premolded stress-cone made of
compartment pressure-monitoring device. The insulator
elastomeric material is positioned on the prepared
anchors the cable and prevents any longitudinal
cable end and then inserted into a cast epoxy insulator
movement. It also keeps the cable in the central axis of
(see Figure 8-62). The pressures at the cable core inter-
the enclosure. Cast epoxy resin or porcelain is usually
face, as well as at the stress-cone-to-central-epoxy-insu-
used. IEC 60859/PART 1 and IEEE 1300 specify the
lator interface, are obtained by means of compression
interface between the GIS and the cable termination
devices such as springs. In some constructions, a plug-
with regards to dimensions, limit of supply (who is
in technique is used to connect the cable to the GIS. In
responsible for the supply of the GIS part and of the
this case, the cable can be installed and connected to
cable system part), and testing requirements.
the GIS without the presence of the GIS manufac-
turer. IEC 60859/PART II specifies the interface
Figure 8-57 Deflector and insulator outdoor termination (courtesy CIGRE).
8-49
between the GIS and the dry-type termination with Directly-immersed GIS Termination
regards to dimensions that are not the same as in Part In this design, the stress-control device and the prepared
I, limit of supply and testing requirements. cable are directly immersed in the SF6 of the GIS with-
out using the axial insulator, as in all the constructions
Capacitor Cone and Insulator GIS Termination
previously described (see Figure 8-64).
The stress cone of the “stress cone and insulator GIS
termination” is replaced in this construction by a capac- In this case, the mechanical function of the insulator has
itor voltage-grading device, either prepared in the fac- to be ensured by one of the insulators of the GIS. The
tory and made of several different coils of polymeric cable must be sealed to prevent any leak of the SF6.
films fitted onto the cable or prepared on-site with tapes
of polymeric films (see Figure 8-63).
This construction offers the possibility to reduce the
diameter of the GIS enclosure.
Figure 8-58 Prefabricated composite and insulator outdoor termination (courtesy CIGRE).
8-50
There is no international standard covering this con- Their description is the same as that of the GIS. In con-
struction. Agreement between the accessory manufac- structions using an insulator, the insulator can be porce-
turer and GIS manufacturer must be reached. lain or thermoset resin. The insulating fluid inside the
termination is generally different from that of the trans-
Oil-immersed Transformer Terminations former, except in the directly-immersed type.
Very similar to GIS terminations, oil-immersed trans-
former terminations connect a cable to a transformer. For safety reasons, an SF6 compartment can be used to
Depending on the stress-control device, the following
separate the cable termination from the transformer. In
types can be found:
these cases, a bushing is installed on the transformer
• Stress cone and insulator oil-immersed termination and connected to the SF6 housing of the intermediate
• Deflector and insulator oil-immersed termination enclosure. The cable terminations are one of the types
described in the GIS terminations section.
• Prefabricated composite dry oil-immersed termina-
tion 8.3.3 Pipe-type Cable Terminations
• Capacitor cone and insulator oil-immersed termina- A pipe-type termination is a single-phase device that
tion normally operates with liquid or gas pressure between
• Directly-immersed oil-immersed termination 200 and 410 psi (1380-2830 kPa) depending upon
Figure 8-59 Capacitor cone and insulator outdoor termination (courtesy CIGRE).
8-51
whether it is a standard or high-strength termination. of the termination components that need to withstand
For HPFF pipe-type cable systems that are to utilize rather high operating and emergency pressures of the
high-speed circulation over extended distances, operat- fluid (or gas) of the pipe-type cable system. Although
ing pressures as high as 650 psi are possible under spe- similar, there are some differences in design of termina-
cific contingencies. tions for high-pressure fluid-filled (HPFF) and high-
pressure gas-filled (HPGF) terminations. These differ-
The three single-phase terminations are spaced apart to ences are highlighted later in the text.
provide the necessary electrical clearances as dictated by
Stress Cone for Pipe-type Cable Terminations
the external dielectric (air, SF6, or liquid dielectric). The
As previously stated in Section 8.2.1, the stress cone for
separation of the steel pipe containing the three cables
laminar cable is made of paper insulation applied over
to the three single steel pipes, each containing one cable,
cable insulation in close proximity to the edge of the
is done below the terminations in a spreaderhead as
cable insulation screen. A typical design of the stress
shown in Figure 8-39, or a trifurcating joint.
cone formed in double logarithmic (L-L) shape is shown
in Figure 8-65.
Similar to terminations for extruded cables, the main
task of the pipe-type cable termination is to shape the
A roll of paper that is perforated and impregnated at the
electrical field at the edge of the cable insulation screen.
factory is slipped over prepared cable with the bottom
The difference is in the use of material for the stress
edge of the paper roll in the vicinity of the cable
cone that needs to be compatible with laminar insula-
insulation screen and tightened onto the cable. Once the
tion of pipe-type cable and in the mechanical structure
Figure 8-60 Prefabricated composite and capacitor stress cone outdoor termination
(courtesy CIGRE).
8-52
paper roll is tight on the cable, excessive paper is torn Paper Roll Design
from the paper roll along perforations exposing the It is important that the direction of the lay of the paper
stress cone slope. The path of the perforation is carefully roll and the lay of the last paper layer of the cable insu-
designed and cut with computer-guided perforators at lation is the same. Otherwise, tightening of the paper
the factory in a such way that the slope is a double- roll onto the cable will cause unraveling of cable insula-
logarithmic shape. A copper band that is positioned tion. This is the reason why perforations must be sym-
near the end of the L-L slope at the outer diameter of metrical to the center of the paper roll, so that it could
the paper roll ensures that the edge of the ground be flipped (if required) in the field to match the lay of
electrode is smooth. Semiconductive tape is applied the cable insulation.
from the bottom edge of the copper band along the L-L
slope and the exposed cable insulation overlapping the While the overall shape of the paper roll depends on the
cable semiconductive tapes. Metallic braid, which is termination in which it is used, the shape of the L-L
applied over the semiconductive tape and soldered to slope depends on the cable construction. A longer L-L
both the copper band and the cable metallic tapes, slope is required for cable with thinner insulation of the
provides a conductive path for capacitive current, same voltage class and conductor size. It is possible to
preventing the semiconductive tapes from overheating. design paper rolls that can be used on several cable sizes.
This solution is particularly useful for customers who
Figure 8-61 Stress cone and insulator metal-enclosed GIS termination (courtesy CIGRE).
8-53
carry spare units in their stock. It is important that full paper rolls. Only the paper roll at the edge of the cable
information is available for all cables on which the spare insulation screen requires the L-L slope.
paper roll could be used. Although an identical paper
Manufacturing and Storage of the Paper Rolls
roll can be used for some number of different cable sizes,
Paper rolls are made of electric-grade kraft paper and
it is unlikely that preparation of each cable will be the
dried at elevated temperature under vacuum in specially
same. A detailed cable preparation drawing needs to be
built impregnation retorts. The drying process is rather
provided for preparation of each cable.
long and can take several weeks, depending on the size
of each paper roll and number of rolls in retort. Once
The design of the paper roll with double-logarithmic
the paper rolls are dry, they are impregnated with
stress cone is also applicable for cables with laminated
impregnating fluid, similar to processing paper cables.
paper polypropylene (LPP) insulation. Paper rolls for
For example, impregnation fluid for paper rolls for
both paper and LPP cables are made of the same paper
HPGF cables is higher in viscosity than for paper rolls
and impregnated in the same impregnating fluid. The
for HPFF cables.
only difference is in the L-L slope due to the thinner
insulation wall of LPP cable for the same voltage class
Each impregnated paper roll is placed in an aluminum
and conductor size and somewhat higher dielectric con-
can that is filled with specified insulating fluid. The can
stant (SIC) of LPP cable insulation.
is capped to provide a hermetical seal. The shelf life of
the paper roll is not limited as long as there is no leak of
Some terminations require more paper rolls in series.
the liquid from the can. This is why the cans are stored
For example, 345-kV outdoor terminations require two
in a vertical position with the lid facing down; any oil
Figure 8-62 Prefabricated composite dry-type metal-enclosed GIS termination (courtesy CIGRE).
8-54
leak would be visible. A paper roll from such a can with relatively small outer diameters for HPFF cable
should not be used. terminations. This is achieved by utilizing an inner
stress-control porcelain that is placed at the edge of the
Outdoor Pipe-type Cable Terminations 69 kV copper band of the paper-roll stress cone. The bottom
through 161 kV section of the insulator features a conductive glaze,
Although there are differences between termination which is connected to the ground, thus further alleviat-
designs based on the voltage class, rated operating pres- ing electrical stress at the copper band.
sure, and type of cable system (HPFF or HPGF), there
are a number of components that perform the same High Pressure Gas-Filled terminations require a paper
function regardless of the termination design. Figure 8- roll with a large outer diameter. This is required to
66 illustrates such common components and shows the reduce the electrical stress at the end of the paper-roll
differences between outdoor terminations for HPFF stress cone to such a level as to prevent ionization of
and HPGF cables for voltages 69 through 161 kV. Fig- the gas.
ure 8-69 shows a typical design of HPFF terminations
for 230 kV and 345 kV voltage class cables for outdoor Another difference between HPFF and HPGF termina-
application. There are no applications of High-Pressure tions is in the requirement for filters and check valves in
Gas-Filled cable systems for voltage classes 230 kV and the semistop housing of the HPFF units. This is
above in the United States. described later in the text.
The main difference between HPFF and HPGF termi- Stand-Off Insulators
nations is in the design of the paper roll. The goal in Stand-off insulators are mounted between the support
designing any paper roll is to make it as small as possi- structure and the termination mounting plate. Insula-
ble, since it is easier to install smaller paper rolls than tors support the weight of the termination and provide
large and bulky ones. It is possible to make paper rolls insulation between the structure that is connected to the
Figure 8-63 Capacitor stress cone and insulator metal-enclosed GIS termination (courtesy CIGRE).
8-55
station ground and the riser pipe that is normally epoxy suitable for outdoor application (e.g., cycloa-
grounded via the cathodic corrosion-protection system. lyphatic epoxy).
Insulators are typically rated at 5 kV. They are also
Mounting Plates
exposed to mechanical stresses due to the cantilever
A mounting plate is used to attach the pipe stub assem-
force at the termination coming from the connection
bly to the support (mounting) structure and to provide
between the termination aerial lug and station bus. Insu-
support for the stand-off insulators. Mounting plate
lators are made either from porcelain or from filled
materials include aluminum, type 304 nonmagnetic
Figure 8-64 Directly-immersed metal-enclosed GIS termination (courtesy CIGRE).
Figure 8-65 Typical double-logarithmic stress cone (courtesy of G&W Electric Company).
8-56
stainless steel, and galvanized steel. If made of magnetic pressure and vacuum testing of the cable pipeline even
steel, the design of the plate must incorporate a gap to before the cable is pulled. Once the cable is pulled, a
prevent circulating currents and heating of the plate due convenient “night-cap” can be slipped over the cable
to the magnetic field generated by load current in the stub and attached to the base plate, providing a
cable conductor. vacuum-tight seal.
It is important that both the structure and mounting The material for the pipe stub subassembly is typically
plate are properly leveled. If this is not the case, exten- specified as ASTM Type 304 nonmagnetic stainless
sive mechanical stress could result in damaging the steel. Other nonmagnetic materials such as copper and
stand-off insulators. aluminum have been used for special applications.
Pipe-Stub Subassembly Semistop Assembly
A pipe-stub subassembly consists of a pipe or tube A semistop assembly centers the cable and, in the case
welded to the base plate of the termination. This subas- of HPFF cable terminations, controls the movement of
sembly is the first termination component that is fluid between the termination and pipe. A packing of
attached to the cable system. While the base plate is rubber gasket in the collar of the semistop housing
attached to the top surface of the mounting plate by provides a relatively good seal between the collar and
bolts, the pipe stub subassembly is welded to the riser varnished Dacron glass (VDG) tape, which is applied
pipe using a coupling over the joint area. over the cable. Although some oil can weep along paper
and VDG tape, the seal stops free movement of fluid
Similar to the reducer plate of pipe-type joints, the base (hence, semistop).
plate of the termination can be capped, allowing
Figure 8-66 HPFF and HPGF outdoor terminations for 69 to 161 kV voltages
(courtesy of G&W Electric Company).
8-57
Thermal expansion and contraction of fluid require oil use on outdoor terminations with aluminum conduc-
to move back and forth between termination and pipe. tors. This method provides for a copper connector out-
Since the oil in the pipe could be contaminated, and the side the termination.
termination requires clean oil for its operation, filter
units with built-in “incoming” check valves are installed The connector stalk (stud) is threaded and secured with
into the flange of the semistop housing, forcing the oil a threaded hood at the termination top plate, ensuring
from the pipe to go through the filter before entering the that the cable will not slip due to its own weight and the
termination. There are also “outgoing” check valves in weight of the paper roll. A combination of longitudinal
the semistop housing flange that allow oil to move from and radial holes allows applying the vacuum on the ter-
the termination into the pipe. To prevent sudden loss of mination from the connector top, which could be bene-
oil from the termination in case of a sudden drop of oil ficial in some procedures for filling the termination with
pressure in the pipe, “outgoing” check valves are set to dielectric fluid. Since the top connector section (stud) is
open at 20 psi differential pressure between termination not sensitive to the cable size, many users select two-
and pipe. piece connectors, as shown in Figure 8-67. The studs
can be stocked as spare items and combined with a
A gas-filled termination uses the semistop for centering cable-size-sensitive bottom section of the connector (fer-
only. Through-holes in the semistop housing flange allow rule) as needed. These two connector sections can be
communication of gas between termination and pipe. either brazed or crimped to each other in the field.
Termination Body Vent Valves
The termination body, also known as the spool, is used Small vent valves are typically installed on the top plate
to provide room for the paper roll stress cone. It is of HPFF terminations or at the top of a hollow connec-
designed to accommodate the longest possible stress tor. The vent valve is used during filling and for venting
cone for a particular voltage class and range of cable gases after a prolonged period at pressures lower than
sizes. The bodies are made of nonmagnetic stainless 50 psi (345 kPa), because there is a possibility of gas
steel and designed to withstand specified routine pres- coming out of solution from the dielectric fluid at low
sure tests, which are based on the nominal internal sys- pressure. It is also used for sampling the oil for dissolved
tem pressure. The bodies for extra-high pressure that gas analysis (DGA), which is used as one of the diag-
require a 1000-psi routine factory test for 1 hour incor- nostic tools for evaluating the condition of the termina-
porate gussets to increase mechanical strength. Typi- tions in service.
cally, there is a 0.75-in. (2-cm) valve at the body that is
Diffusion Chamber
used when the termination is filled with dielectric fluid
The introduction of dielectric fluid at high velocity into
from an external source, not from the pipe.
the system, either during initial filling or if force cooling
Insulators is used, can eventually be damaging for top papers of
The insulator housing is made of high-strength glazed
porcelain to withstand high internal pressures during
normal and emergency operations as well as factory
routine tests. Both the inside and outside surfaces of the
insulators are glazed to provide a higher mechanical
strength of the unit.
Connectors
As described previously, connectors are provided in cop-
per, bimetallic (copper-aluminum for aluminum cables),
and aluminum. Copper connectors are pressed on to the
cable conductor using a 4-ram press for conductors
larger than 750 kcmil (380 mm2). Smaller conductors
are pressed with a single-ram press. For large aluminum
conductors above 1000 kcmil (507 mm2), a combination
of MIG welding and 4-ram pressing is used. This con-
nection method is similar to that used on pipe-type Figure 8-67 Typical connector for HPFF cable
joints with aluminum conductors. Aluminum connec- terminations (courtesy of G&W Electric Company).
tors for cable sizes below 1000 kcmil are pressed using a
4-ram press. Bimetallic connectors are recommended for
8-58
cable insulation. Diffusion chambers prevent fluid from Equipment Terminations for Pipe-type Cables
impinging directly on the cable, thus eliminating the Equipment-type terminations are also single-phase
risk. They are mainly used in high-speed circulation units, and they are installed either in a gas-insulated
and force-cooled systems. An example of a diffusion substation (GIS termination) or in a transformer (oil-
chamber is shown in Figure 8-69 with a capacitor- immersed termination). Due to the fact that the termi-
graded termination. nation is immersed either in SF6 gas or in oil, the
required external creepage distance of the insulator is
Termination Installation
considerably less than for outdoor termination, result-
Outdoor terminations can be installed either on the sta-
ing in a relatively short overall length.
tion structure or on transition poles. Figure 8-68 is an
example of a pipe-type cable termination installed on a
Similar to outdoor terminations, the paper roll is uti-
transition pole.
lized to provide stress relief at the cable-shielding edge.
Outdoor Pipe-type Cable Terminations, 230 and Internal filters and check valves facilitate oil communi-
345 kV cation between the termination and the pipe. The hol-
Terminations for 230 kV and above are distinguished low connector features a valve for vacuum application,
from lower-voltage class terminations by voltage grad- venting, and oil sampling (see Figure 8-70).
ing along the porcelain insulator, utilizing individual
“doughnut” capacitors connected in series. This method The cable metallic shield and cable pipe must be
provides a very uniform voltage grading of the external insulated from any other grounded component to allow
insulator surface, which makes insulators less suscepti- separate grounding and cathodic protection of the pipe.
ble to pollution and enables designing the terminations
with shorter insulators than if the termination was with-
out capacitors. Figure 8-69 shows a cross section of a
230 kV outdoor termination.
The bottom capacitor is connected to the test bushing

i n t h e t e r m i n a t i o n b o d y. C a p a c i t a n c e o f t h e
termination is measured at the factory and should be
checked prior to and after installation on the cable to
verify that the capacitors and all connections are sound
and have not been damaged during either transport or
installation. The capacitance of the ter mination
installed on the cable is somewhat higher due to stray
c ap a c i t a n c e s b e t we e n t h e cabl e c o nd u c t o r a nd
“doughnut” capacitors.
Figure 8-68 Pipe-type termination on pole (courtesy of Figure 8-69 230 kV capacitor graded pipe-type
G&W Electric Co.). termination (courtesy of G&W Electric Co.).
8-59
Since the cable metallic shield and pipe are connected that are at the potential of the cable grounding shield.
to the subassembly of the base plate and pipe stub, and These are very fast transients with rise times in the
the base plate rests on the mounting plate, both the nanosecond range and peak values that can reach in
base plate and mounting plate need to be isolated from excess of 50 kV. Surge voltage limiters installed across
a ny o t he r gro un de d c om po ne nt . T he s t an d- o ff pipe stub insulators might be required if such overvolt-
insulators and pipe stub insulator serve this function. ages are developing. The GIS manufacturer and end
The stand-off insulators separate the mounting plate user should participate in specifying surge voltage
from the grounded support structure, and the porcelain limiters, but they should be supplied by the cable
pipe-stub insulator insulates the base plate from the termination manufacturer.
termination body, which in turn, is connected to the
grounded GIS enclosure. The valve at the termination top is used for attachment
of a vacuum pump during oil filling. It is also used in
GIS switching transients may cause short-duration over- emergencies to bleed gasses from the termination devel-
voltages between the GIS enclosure and components oped in the dielectric fluid—e.g., after the system pres-
Figure 8-70 345 kV GIS pipe-type termination (courtesy of G&W Electric Company).
8-60
sure has been reduced below 50 psi for a prolonged mounting plate, and pipe-stub should be loosely
period of time. The valve can also be used for oil sam- installed on the termination support structure. Then the
pling for dissolved gas analysis and other dielectric tests. GIS housing with the preinstalled interface plate must
The provision of a maintenance port in the GIS housing be installed to the GIS bus flange and the support hard-
in the vicinity of the termination valve can significantly ware on the structure for the GIS housing adjusted. The
reduce the work involved in oil sampling. next step is to install the termination body with a prein-
stalled pipe-stub insulator onto the interface plate.
Dimensional requirements of the interface between Finally, the support hardware on the termination struc-
gas-insulated switchgear and pipe-type terminations ture needs to be adjusted so that the bottom flange of
are specified in IEEE 1300 Standard. Nevertheless, a the pipe stub insulator can be attached to the base plate
close working relationship between the end users, without force. Only after all hardware on all connec-
installers, and manufacturers of GIS gear and the cable tions is tightened to the required torque and all inter-
accessory is required. face flanges are marked with alignment marks can the
pipe stub be welded to the riser pipe.
All termination components between the mounting
Internal Pressure Characteristics
plate (point “B”) and the end flange of the GIS gear bus
The internal pressure characteristics of pipe-type termi-
(point “A”) are rigid and require strict alignment at the
nations are very important to the system operation since
site. It is mandatory to “dry fit” these components prior
the terminations are usually at the source of the dielec-
to welding the pipe-stub subassembly to the riser pipe,
tric fluid supply and at high and low elevations of a
as shown in Figure 8-71. First, the stand-off insulators,
Figure 8-71 “Dry-Fit” of rigid GIS and termination components (courtesy of G&W Electric Company).
8-61
cable system. The termination operating pressures in 50 psi margin of safety. Utilities also provide backup
these cases may limit the cable system operations. power supplies to the pressurizing pumps or provide
other backup pressurization systems to ensure proper
IEEE Standard 48 specifies the internal operating pres- pressures during major power outages. It may be possible
sure at which terminations can operate continuously. to provide methods to reduce the chances of pothead fail-
Standard internal pressures are 200 psi (1380 kPa) and ure if all pressure is lost by automatically injecting SF6 if
300 psi (2070 kPa). A factory withstand pressure test of the pressure drops below 25 psi (175 kPa) (Engelhardt et
2.5 times internal pressure is specified as a routine test. al. 1990).
For higher internal pressures, the standard refers to an
agreement between the customer and manufacturer. Ampacity of Pipe-type Cable Terminations
Extra-high strength terminations are available and are The maximum continuous and emergency current load-
safe to implement. Routine factory tests for extra-high ings in the termination are limited by maximum allow-
strength terminations include a 1-hour pressure test at able temperatures of the cable insulation at the
1000 psi (6900 kPa). It is essential that the pothead pres- conductor. Conductor temperature limitations have been
sure rating, riser pipe sizes, the installation of diffusion established at 85°C for normal operation and (most com-
chambers on risers, and conductor sizes for the risers monly) to 100°C and 105°C, respectively, for 300-hour
consider not only the short-term ampacity require- and 100-hour emergencies.
ments, but also future load growth, emergency operat-
ing scenarios, and the resulting possibility that the The ampacity rating of pipe-type terminations for typi-
utility might need circulation or forced cooling. cal self-cooled systems is usually higher than the cable
in-earth for equal conductor sizes, so the terminations
The allowable pressure for pipe-type cable terminations are typically not ampacity limiting. However, termina-
is determined by the strength of the porcelain insulator. tions are often ampacity limiting with forced-cooled sys-
Table 8-10 shows the internal pressure characteristics t e m s, r e q u i r i n g l a rg e r c o n d u c t o r s i z e s i n t h e
for standard- and high-strength terminations. terminations and/or supplemental cooling.
Maintaining proper pressure in the terminations is criti- The pipe-type terminations are designed to terminate
cal to cable system operation. A minimum operating not only one specific cable but rather to accept a range
pressure of 125 psi (860 kPa) at the termination is speci- of cables. Cables can vary in conductor size and mate-
fied by most utilities to avoid the chance of termination rial, in insulation type (paper or LPP), and in the thick-
failure. Some utilities install pressure switches at the base ness of the insulation wall. The external factors that
of the pothead to initiate urgent alarms or trip the circuit influence ampacity are ambient temperature, solar radi-
if pressure drops below a critical value. It is generally ation, air density, and wind. All these factors need to be
accepted that the minimum pressure that would ensure evaluated when establishing the ampacity rating of a
continuous suppression of ionization in HPFF systems is particular termination.
75 psi. A trip point could be as high as 125 psi, leaving a
EPRI sponsored three projects for ampacity evaluation
of the terminations. Two projects (EPRI 1975, 1982)
Table 8-10 Typical Operating Pressures for Pipe-Type deal with evaluation of termination ampacity in normal
Terminations operation, while the third one (EPRI 1986) deals with
the emergency rating of the termination.
Standard High Strength
psi (kPa) psi (kPa)
Nominal operating pressure 200 (1380) 300 (2070) Based on a number of experiments on real terminations
in controlled laboratory conditions, computer programs
Max. continuous operating
pressure (over 10 hours)
275 (1895) 410 (2825) were developed to calculate termination ampacity
initially in steady-state conditions and then for
(up to 10 hours)
300 (2070) 450 (3100) emergencies (i.e., an increase of current load above
maximum continuous load). Both matrix and network
Max. peak transient 400 (2680) 600 (4140)
models were evaluated, and it was found that a simpler,
Max. field (proof) test 350 (2410) 525 (3620) network model offers sufficient accuracy. This is
Factory test (1 hour) 500 (3450) 750 (5170) because the vertical heat flow in the termination is
Minimum operat- Fluid-filled 125 (860) 125 (860) dominated by the heat flow in the cable conductor.
ing pressure Gas-filled 150 (1035) 150 (1035) Vertical heat flow caused by oil convection in the
annular space around the cable and paper roll in the
termination can be disregarded and yet still achieve the
8-62
desired accuracy. The advantages of the network node N4 (in this case, heat generated by current in con-
approach arein the ease of model creation and fast ductor); B6, B7, and B8 are heat conductances; and
computation compared to other methods. dT/dX is change of temperature in time in node 4.
A simplified network model for a pipe-type 138 kV out- The output of the computer program is the list of emer-
door termination is shown in Figure 8-72. The heat con- gency currents for specified emergency periods, selected
ductance in each branch is determined from the geometry pre-emergency loads, conductor starting temperatures
of the termination. The model represents a series of con- and ambient temperatures.
centric shells with varying thermal resistances and capac-
itances, simulating an actual design structure. 8.3.4 Self-Contained Fluid-Filled Terminations
A self-contained fluid-filled (SCFF) termination con-
The computer method is based on fundamental equa- sists of a single-conductor, hollow-core cable with a
tions for the heat balance. For example, heat balance for metallic sheath, a lower metalwork, a stress-relief cone,
node N4 may be expressed by following two equations: an insulating enclosure that supports the fluid pressure,
a cable connector, and an aerial connector. An example
of an outdoor self-contained termination is shown in
dT Figure 8-73.
T 4 • S 4 − T 3 • B 6 − TA • B7 − T 5 • B8 = W 4 − H 4 • 8-1
dX
Se lf - con ta i ned te r mi nat ion s c an be o f t he o ut -
S 4 = B 6 + B 7 + B8 8-2 door/indoor, metal-enclosed GIS, and oil-immersed
transformer type, as described at Section 8.3.1. The
stress-control methods adopted are either geometrical-
Where, T3, T4, and T5 are temperatures at nodes N3, capacitive or capacitor-graded up to 500 kV and can be
N4, and N5; TA is temperature at node NA (next to installed on either paper or LPP cable insulations.
B7); H4 is heat capacitance connected with the nodal
volume at node N4; and W4 is external heat input at
Figure 8-72 Network model for thermal analysis of a 138-kV outdoor termination
(courtesy of G&W Electric Company).
8-63
Outdoor /Indoor Terminations General Description A mounting flange is used to mount the termination on
This section reviews the materials and design issues of a supporting structure and to support the cable sheath.
the major components that are most important for The mounting flange is generally made of galvanized
SCFF outdoor/indoor terminations. mild steel, but with high conductor currents (above
approximately 1000 A) and a small opening in the
Lower Metalwork
mounting plate, stainless steel (or slotted galvanized
The lower metalwork includes a wiping sleeve (end bell)
mild steel) must be used. Stand-off insulators are used
and a mounting flange (base plate).
between the structure and the mounting flange to allow
flexibility in connecting cable metallic sheath to the
A wiping sleeve is manufactured from copper alloy,
ground. Cable systems that are cross-bonded require
with an opening to allow it to slide over the cable
mounting insulators.
sheath. The sleeve is hot-tinned to provide a good
soldering surface. The wiping sleeve is attached to the Stress Cone
metallic sheath by a solder wipe, or by a mechanical The geometrical stress cone consists of an impregnated
means in the case of some aluminum-sheath cables. kraft-paper roll, carbon-black crepe tape, tinned
Reinforced wipes are required with double-lead sheaths copper-shielding braid, and crepe tape cover. Some
or with armored cables operating at pressures above termination designs have an additional stress-control
40 psi (275 kPa). A lead wedge is used to center the element, which is assembled over the stress cone. The
sheath in the wiping sleeve before the wipe is made. A paper roll is made of electrical-grade kraft paper, which
cable ground boss is commonly supplied, silver-brazed has been dried and impregnated with a dielectric fluid in
to the wiping sleeve, to provide easy grounding or a manner similar to the cable insulation. Some termina-
bonding of the cable system. tions up to 69 kV do not have the additional stress-
control element, which is assembled over the stress cone.
Paper rolls are custom-designed, utilizing a double log
slope to fit a specific cable size. 345 kV and higher
voltage terminations use two or more paper rolls in
series with a crepe tape fill-in between the rolls.
The capacitor-graded design utilizes a factory-made or

site-manufactured preimpregnated paper roll includ-
ing suitably positioned metallic foils (condenser cone)
able to uniformly distribute the stress along the insula-
tor surface.
Insulator Housing Assembly
The insulator housing assembly consists of a porcelain
insulator with sheds of suitable creepage and with a non-
magnetic upper metalwork, generally of stainless steel.
The insulating enclosure is designed to operate at fluid
pressure of 15-100 psi (105-690 kPa). A corona shield is
normally adopted in order to screen the sharp edges of
the upper metalwork and avoid corona effects. Capacitor-
graded terminations using doughnut capacitors are
factory-assembled. The termination is sealed and pres-
surized with a dielectric gas before shipment. A test
bushing is provided on the body for measuring the capac-
itance and power factor before and after installation. All
test data should be retained for future reference in the
event of problems.
Cable Connector
The cable connector, through its stem, provides a gasket
seal and mechanical attachment to the termination
Figure 8-73 SCFF 420-kV outdoor termination. upper metalwork. Copper connectors are generally
pressed on with a 4-ram press. The connector is
equipped with a fluid valve, which allows the fluid flow
8-64
to be reduced during construction and the evacuation of bonded systems is generally obtained through the use of
the termination. The termination is evacuated to 500 a sectionalizing ring attached to the end bell.
micrometers or less and then filled with a degassed
Stress Cone
dielectric fluid. After the termination has been impreg-
The construction of the stress cone is the same as
nated, the fluid valve is opened and removed from the
described earlier for outdoor terminations.
connector stem.
Insulator Housing Assembly
Aerial Connector
The insulator housing assembly consists of an epoxy
The aerial connector attaches the cable connector to the
resin insulator whose dimensions generally comply with
electrical system. If the termination connector is to be
IEC 60859 specifications (IEC 1999). In some designs,
connected to a flat or round bus, a flexible connection is
the top of the insulator incorporates a solid high-voltage
recommended. Cantilever forces should be kept to a
aluminum electrode to prevent any ingress of the pres-
minimum on any HV termination.
surized fluid, especially SF6, into the cable.
Internal Pressure Characteristics
Cable Connector
Internal pressure characteristics of SCFF terminations
In the case of open-top insulators, the connector stem
are very important to the system operation because the
provides a gasket seal and mechanical attachment of the
terminations are usually at the high and low elevations
cable to the insulator. If the insulator is blind, a plug-in
of the cable system. The operating pressures in these
connector is required in order to transmit the current
cases may limit cable system operations.
from the conductor to the high-voltage electrode. The
locking of the conductor is accomplished through
The internal pressure characteristics are determined by
spring-loaded pins.
the thickness of the copper body wiping-sleeve assem-
bly. Table 8-11 shows the internal pressure ratings for
SCFF terminations.
SF6/ Oil-Immersed Terminations General
Description
This section reviews the materials and design issues of
the major components that are most important for
SCFF SF6 and oil-immersed terminations. Figure 8-74
shows a typical SCFF SF6- immersed termination.
Lower Metalwork
Details are the same as described earlier for outdoor ter-
minations. The insulation of the cable from the equip-
ment and the mounting structure for single-point
Table 8-11 Typical Operating Pressures for Self-Contained

Terminations
High
Standard Medium Strength psi
psi (kPa) psi (kPa) (kPa)
Nominal operating pressure 15 (105) 40 (275) 100 (690)
Max. continuous operating
22 (150) 55 (380) 138 (950)
pressure (over 10 hours)
24 (165) 60 (415) 150 (1035)
(up to 10 hours)
Max. peak transient 32 (220) 80 (550) 200 (1380)
Max. field (proof) test 28 (195) 70 (485) 175 (1205)
Factory test (1 hour) 40 (275) 100 (690) 250 (1725)
Figure 8-74 SCFF SF6-immersed 420-kV termination.
8-65
Bus Bar Connector standards relevant to terminations and to the composite

A bus bar connector is used to connect the cable con- insulators are also available and are listed below:
nector or the top of the electrode (in blind top designs)
• IEC 60859 “Cable connections for gas-insulated
to the apparatus bus bar. For SF6-immersed termina-
switchgear for rated voltages of 72.5 kV and above.”
tions, the dimensions and characteristics of this connec-
tion are standardized by IEC 60859 specifications. For • IEC 60815 “Guide for the selection and dimension-
transformer terminations, the connection is flexible and ing of high voltage insulators for polluted conditions
screened by a suitable corona shield. – Part 1: Definitions, information and general princi-
ples.”
8.3.5 Industry Specifications and Standards • IEC 61462 “Composite Insulators—Hollow insula-
IEEE Standard 48-1996, “Standard Test Procedures tors for use in outdoor and indoor electrical equip-
and Requirements for Alternating-Current Cable ment—Definitions, test methods, acceptance criteria,
Terminations, 2.5 kV through 765 kV” covers all and design recommendations.”
outdoor, apparatus, and indoor terminations used on ac • IEC 62217 “Polymeric insulators for indoor and out-
cables having a laminated or extruded insulation. door use with a nominal voltage greater than
Compressed-gas-insulated terminations are not covered 1000V—General definitions, test methods, and
in this standard. acceptance criteria.”
This standard specifies the design and production tests For terminations installed in polluted environments, it is
for laminated-dielectric cable terminations assembled necessary to evaluate the long-term behavior of the
and ready for service for insulation classes up to 765 kV, polymeric housing, which despite porcelain insulators,
a n d fo r ex t r ud e d - di e l e c t ri c c abl e t e r m i n at i o n s is subjected to aging. For this reason, IEC 62217 has
assembled and ready for service for insulation classes up introduced an accelerated aging cycle of up to 5000
to 230 kV. hours duration where, in addition to the voltage, differ-
ent environmental stresses are applied to the full-size
The test sequences for design tests for both extruded termination or to a representative test insulator—i.e.,
cable and laminated cable terminations are specified. A UV radiation, salt fog, steam fog, rain, drying periods.
minimum of four samples are required for dielectric
design tests, two of which have to undergo a 30-day More test details are described in Chapter 14.
cyclic aging test at double line-to-ground voltage with
cable conductor temperature within 5°C of the cable
8.4 OTHER ACCESSORIES
maximum rated emergency operating temperature for a
period of six hours in each cycle. The partial discharge High-voltage cable systems are complex, and a number
design tests are required only for terminations for of ancillary accessories may be needed in order to allow
extruded cables, and the ionization design tests are the system to operate in a proper manner and guarantee
required only for laminated cable terminations. the requested performances.
Routine (production) tests require dielectric tests on the 8.4.1 Sheath Bonding, Sectionalizing, Link
termination insulator, and pressure and leak tests on Boxes, and Sheath Voltage Limiters
components and factory-assembled seals.
Sheath interrupts, link boxes, and sheath voltage limit-
IEEE Standard 1300-1996 (reaffirmed in 2002), “IEEE ers are important components for the bonding of
Guide for Cable Connections for Gas Insulated Substa- extruded-dielectric and SCFF cable systems. Sheath
tions” provides the basis for coordination of design, bonding alternatives and basic bonding design princi-
material supply, installation, and test procedures for the ples are discussed in detail in Chapter 10.
connection of the termination into gas-insulated substa- Sectionalizing—Metallic Screen/Shield/Sheath
tion (GIS). Interruptions
Many troubles will be encountered with the specially
IEC international standards are applicable to the trans- bonded metallic sheath arrangement if there is a lack of
mission cable terminations up to the voltage of 500 kV. attention to surge protection for the shield interruptions
The electrical test requirements are the same as those required for any special shield/sheath bonding system.
indicated for the joints in Section 8.2.5. Specific IEC Any termination of the shield represents a dielectric dis-
continuity and point of intense electrical stress if not
8-66
adequately relieved in the design. In addition, it is a A routine factory test is recommended in order to ascer-
location at which an incoming transient overvoltage will tain the integrity of the cable jacket (IEC 60229, ICEA S-
be reflected and multiplied in magnitude, possibly 108-720). Many polymers with varying density grades
doubling or more depending on design. are used as jacketing materials. Irrespective of the mate-
rial chosen, it is mandatory that maintenance testing be
The shield interruption must prevent current flow in conducted at regular intervals to ensure integrity over
both the metallic shield/sheath as well as in the semi- the life of the cable circuit. This is usually done with a dc
conductive insulation shield (which cannot carry more test potential of a few kV to perhaps 15 kV depending on
than charging current levels without significant jacket material, thickness, manner of installation, and
heating) and simultaneously provide for electrostatic sometimes service age of the circuit. To facilitate inter-
shielding of the underlying cable insulation. For the preting leakage-current results over time, it is recom-
insulation shield interruption, a thin layer of suitable mended practice, especially for cables installed in ducts
insulating material is applied over one side of the with high insulating rating, to apply a conducting surface
interrupted area, then the shield of the other side is to the outside of the jacket to serve as a ground electrode
extended over the insulating layer to overlap it, for conducting dc tests between the metallic sheath or
providing electrostatic shielding by capacitive coupling. shield and ground. Deposited, tightly adhering graphite
In prefabricated accessories, the interruption is coatings are commonly used; some manufacturers
no r ma l ly i nc or po rate d in t h e f a c t or y -m a de extrude a thin layer of semiconducting jacket material
components or placed in a low electric stress position. directly over the insulating jacket.
Link Boxes
Some designs of interruptions for extruded cables
There are as many styles and designs of link boxes as
incorporate a pad of metal-oxide-loaded, high-
there are hybrid cable configurations for special bond-
dielectric-constant polymeric material immediately over
ing. The principal function of each is to provide reason-
the insulation at the interruption to serve as a primary
ably easy access to the sheath-bonding connections for
surge suppressor backed up by the sheath voltage
inspection and removal of grounds to permit testing the
limiters connected externally.
cable’s insulating jacket for integrity. For this reason, the
link boxes are generally located in manholes or pillars.
For sheathed cables, the metallic shield is usually inter-
rupted by inserting a cast-epoxy insulating flange in the
A t l o c at i o n s wh e r e ov e r vo l t ag e p ro t e c t i o n o f
splice sleeve. For wire-shielded unsheathed cables, it is
shield/sheath interruptions is provided, the link box
customary to terminate the wires at the splice ends and
contains the sheath voltage limiters physically and elec-
to extend the metal shielding with copper mesh and
trically close to the points to be protected. Figure 8-75
overlap it along with the overlapped insulation shield. In
shows a link box of the diving-bell-lid–type, containing
special cases, similar interruptions have been incorpo-
sheath voltage limiters. Completely sealed link boxes
rated in the shields of extruded-dielectric cables at loca-
with a bolted lid are today usually adopted.
tions in the cable run itself, but this solution is
applicable only to low-insulation screen stress cables.
The link box design must comply with the following
Figure 8-4 shows a splice with a shield interrupt.
requirements:
In all cases, it is essential that the dielectric strengths, • The links inside the box must be able to carry the
and the ac, impulse, and switching surge levels of the maximum foreseen short-circuit current and to sup-
interruptions pass the required tests and be coordinated port the relevant electrodynamic forces.
with the calculated magnitudes of standing ac and tran- • The clearances between the different links must with-
sient overvoltages for the circuit in the system fault con- stand an impulse voltage equal to two times the resid-
ditions. ual voltage of the sheath voltage limiter at 10 kA 8/20
Cable Jacket (Outer Sheath) µs current plus a 25% safety factor (ELECTRA
The reliability of specially bonded systems is totally 1990a).
dependent on the integrity of the insulating jacket, which • It is essential for the link box to be watertight; for this
must permanently withstand standing and transient volt- reason, it must pass a type test consisting of water
ages without breakdown or significant current leakage.
8-67
immersion followed by a voltage withstand test In principle, three different types of SVLs can be
(ELECTRA 1981). adopted: spark gaps, nonlinear resistors in series with
spark gaps, and nonlinear resistors alone.
Disconnecting link boxes—usually without sheath volt-
age limiters, except for rare terminal configurations— Simple metal spark gaps used in earlier practice proved
are frequently used at terminal base plates and normally unreliable because, following an operation (flashover),
grounded splice locations to facilitate jacket testing. their electrodes were found to have fused together, thus
obviating the special bonding method being employed.
Connections between components and the link box are Another disadvantage is that they present delays in
often made with coaxial cables to minimize inductive their response when very steeply rising transient over-
reactance between the device and ground to optimize voltages occur.
the protective level.
Nonlinear resistors in series with a spark gap can pre-
Sheath Voltage Limiters (SVLs) vent the damage due to the power frequency current
Sheath voltage limiters (also called surge diverters) are (limited by the resistor), but their response to steeply ris-
usually classified according to their performance with ing waves is inevitably slower than that of the nonlinear
respect to: resistor alone.
• voltage at power frequency that can be withstood for
a conventional duration, generally 2 seconds (voltage Nonlinear resistors, also called metal oxide varistors
induced on the cable metallic sheath during short cir- (MOVs), offer immediate response to steeply rising
cuits) overvoltages, and, if correctly used within their service
• residual voltage appearing at its ends when a tran- limits, are subject to no deterioration under working
sient current 8/20 µs having amplitude 10 kA flows conditions. For this reason, they have effectively sup-
through it (current flowing along the cable metallic planted gapped arresters of all types. However, they
sheath as a consequence of transient overvoltages). must be protected from moisture, and their energy dissi-
Figure 8-75 Link box with sheath voltage limiters
8-68
pation capacity (in case of switching surges) must be performed at 5 kVdc; in fact, the current that they
carefully chosen. absorb in this condition is negligible.
Bonding Leads
Among metal oxide varistors, the zinc-oxide-type is
The bonding lead is the insulated conductor forming the
preferable compared to the silicone carbide type due to
connection between the metallic screen of the HV cable
its better nonlinear voltage versus current characteristic
termination or the joint and the relevant link box. The
(see Figure 8-76), smaller dimensions, and lower sensi-
bonding lead should be capable of sustaining the short-
tivity to the application of repeated transient overvolt-
circuit current of the system.
ages (Par migiani 1986). To minimize the risk of
moisture absorption, the metal oxide blocks are usually
Bonding leads interposed between the cable accessory
encapsulated in an epoxy resin casting or placed in a
(joint or termination) and the SVL inevitably reduce
suitable container. Intermediate-class, medium-voltage
somewhat the protective effect of the SVLs. To keep this
arresters, generally in the range of 3-10 kV, are used as
effect to a minimum, they should be as short as possible
SVLs (see Figure 8-77).
and of concentric type. If nonconcentric bonding leads
are used for the SVL connection, they should be laid
Reference (Ball et al. 1965; CIGRE 2005) and others
bunched together. In those conditions where the link
listed in ANSI Standard 575 give insight into selecting
boxes are very close to the joints, bundling the leads
the appropriate rating for the overvoltages associated
may not be practical, but is not needed due to the low
with individual cases.
inductance (CIGRE 2005).
Zinc oxide varistors with a discharge voltage above
The bonding lead must have the insulation level coordi-
5 kVac need not be removed from the link box during
nated with that of the sectionalizing of the joint.
the periodic integrity test of the cable jacket, generally
Figure 8-76 Voltage-current characteristic of SiC and ZnO nonlinear resistors.
Figure 8-77 SVL metal oxide surge diverters (courtesy of ABB).
8-69
REFERENCES Chu, H. et al. 1992. “Development of 138 kV Transition

Akiyama, T. et al. 1991. “Development of Prefabricated Between HPPT Cable and Extruded Cable.” IEEE/PES.
Joint for 154 kV XLPE Cables.” IEEE Trans. Vol. 6.
No. 2. CIGRE. 1994. No. 89. Accessories for HV Extruded
Cables. Types of Accessories and Terminology.
Allam, E. M., A. L. McKean, and F. A. Teti. 1986.
“Optimized PPP-Insulated Pipe-Type Cable System for CIGRE. 2001a. No. 177. Accessories for HV Cables
the Commercial Voltage Range.” IEEE Power Engineer- with Extruded Insulation.
ing Society Conference Paper. No. 86. T&D 569-8.
CIGRE. 2001b. No. 194 Construction, Laying and
Amerpohl, U. et al. 2002. “Dry Type Outdoor Termina- Installation Techniques for Extruded and Self Con-
tion with New Stress Control Management.” CIGRE tained Fluid Filled Cable Systems.
2002 Session. Paper 21-106.
CIGRE. 2002. No. 210. Interfaces in Accessories for
ANSI/IEEE. 1988. P575. “Guide for Application of Extruded HV and EHV Cables.
Sheath Bonding Methods for Single Conductor Cables
and the Calculation of Induced Voltages and Currents CIGRE. 2005a. No. 279. Maintenance of HV Cables
i n C abl e S h e at h s. ” I n s t i t u t e o f E l e c t ri c a l a n d and Accessories.
Electronics Engineers.
CIGRE. 2005b. No. 283. Special Bonding of High Volt-
Attwood, J. R. et al. 1993. “Prefabricated Accessories age Power Cables.
for Extruded Polymeric Cable at 66 kV to 150 kV.” IEE
paper. November. CIGRE 2006. No. 303. Revision of Qualification Proce-
dures for HV and EHV AC Extruded Underground
Attwood, J. R. et al. 1998. “Development of High Stress Cable Systems.
HV and EHV XLPE Cable Systems.” CIGRE 1998 Ses-
sion. Paper 21-108. Couderc, D. et al. 1996. “Development and Testing of a
800 kV PPLP Insulated Oil Filled Cable and its Acces-
Ball, E. H., E. Occhini, and G. Luoni. 1965. “Sheath sories.” CIGRE 1996 Session. Paper 21-22/04.
Overvoltages in High Voltage Cables Due to Special
Sheath Bonding Connections.” IEEE Transactions on De Ridder, E. et al. 1999. “170 kV Common Termina-
Power Apparatus and Systems. Vol. 84. tion System for Cables with Extruded Insulation.” Jica-
ble 1999. Paper A5.1.
Beker, J. et al. 1998. “XLPE Cable Systems for 400 kV
and Above.” CIGRE 1998 Session. Paper 21-106. De Ridder, E. et al. 2002. “Compact 170 kV Transition
Joint.” CIGRE 2002 Session. Paper 21-110.
Blodgett, R. B., and J. S. Engelhardt. 1975. “Preliminary
Qualification of a HPOF Pipe Cable System for Service Dejean, P. et al. 2003. “New Dry Outdoor Terminations
at 765 kV.” Proceedings of the IEEE Power Engineering for HV Extruded Cables.” Jicable 2003.
Society. Paper T 75212-6. Winter Power Meeting.
ELECTRA. 1973. No. 28. The Design of Specially
Bolza, A. et al. 2002. “Prequalification Test Experience Bonded Cable Systems.
on EHV XLPE Cable Systems.” CIGRE 2002 Session.
Paper 21-104. ELECTRA. 1976. No. 47. The Design of Specially
Bonded Cable Systems (part II).
Chatterjee, S. et al. 1995. “Development of HV and
EHV Single Piece Premoulded Joints.” Jicable 1995. ELECTRA. 1981. No. 75. Recommendations for Tests
Paper A4.4. on Anti-Corrosion Coverings of Self Contained Pres-
sure Cables and Accessories and Equipment for Spe-
Choi, M. K. et al. 1998. “Development of 400 kV XLPE cially Bonded Circuits.
Cable and Accessories in Korea.” CIGRE 1998 Session.
Paper 21-107. ELECTRA. 1983. No. 90. Forced Cooling of Cable
Joints and Terminations.
8-70
ELECTRA. 1990a. No. 128. Guide to the Protection Hahm, Y. H. et al. 1988. “Underground 161 kV Trans-
of Specially Bonded Cable Systems against Sheath mission Project in Des Moines, Iowa.” Proceedings of the
Overvoltages. IEEE PES Summer Meeting. Paper 87 SM 592-9. July.
ELECTRA. 1990b. No. 128. The Steady State Thermal Halperin, H., and K. M. Miller. 1929. “Reduction of
Behavior of Accessories for Cooled Cable Systems. Sheath Losses in Single Conductor Cables.” AIEE
Transactions. Vol. 48. April.
ELECTRA. 2004. No. 212. Thermal Rating of High
Voltage Cables. Hassler, S. P. et al. 1984a. “M.O.V. Arrester Protection
of Shield Interrupts on 138 kV Extruded Dielectric
Engelhardt J.S. et al. “Model Testing of An Improved Cables.” IEEE Transactions on Power Apparatus and
Oil Dielectric for High Pressure Oil-Filled Cable Termi- Systems. Vol. 103. November. pp. 3334-3341.
nations.” IEEE Transactions on Power Delivery. vol. 5,
no. 3, July 1990. Hassler, S. P. et al. 1984b. “Shield Interrupt Overvolt-
ages on 138 kV Extruded Dielectric Cables.” IEEE
Engelhardt, J. S. 1972. “Ampacity Considerations of Transactions on Power Apparatus and Systems. Vol. 103.
HPOF Joints.” Proceedings of the IEEE 1972 Under- November. pp. 3327-3333.
ground Transmission Conference. Paper 72-CHO 608-0-
PWR (Supplement). Hwang, S. C. et al. 1990. “The Development of 154 kV
XLPE Cable Accessories in Korea.” CIGRE 1990 Ses-
EPRI. 1975. High Ampacity Potheads. Report PB-248 sion. Paper 21.103.
342, October
IEC. 1986. Publication 60815. “Guide for the Selection
EPRI. 1982. High-Ampacity Terminations. Report no. and Dimensioning of High Voltage Insulators for Pol-
EL-2233. July. luted Conditions – Part 1: Definitions, Information and
General Principles.”
EPRI. 1984. Designer’s Handbook for Forced-Cooled
High Pressure Oil-Filled Pipe-Type Cable Systems. IEC. 1998. Publication 61462. “Composite Insulators—
Report no. EL-3624. July. Hollow Insulators for Use in Outdoor and Indoor Elec-
trical Equipment – Definitions, Test Methods, Accep-
EPRI. 1986. Emergency Ratings of Cable Terminations. tance Criteria and Design Recommendations.
Report no. EL-4293. January.
IEC. 2005. Publication 62217. “Polymeric Insulators for
EPRI. 1988. Joints and Terminations for Pipe-Type Cable Indoor and Outdoor Use with a Nominal Voltage
Insulated with PPP Tapes. Report no. EL-6039. October. Greater than 1000V – General Definitions, Test Meth-
ods and Acceptance Criteria.”
Gahungu, F. et al. 1993. “The HV Transition Joints.”
IEE 1993. IEC. 1982. Publication 60229. “Tests on Cable Over-
sheaths Which Have a Special Protective Function and
Gazzana-Priaroggia P., and G. Palandri. 1956. “Use of are Applied by Extrusion.”
Synthetic Casting Resins in the Manufacture of Very
High Voltage Cable Accessories.” CIGRE. Paris, IEC. 2001. Publication 60287. “Calculation of the Con-
France. 16th meeting. Paper 215. Vol. 11. tinuous Current Rating of Cables (100% Load Factor).”
Goldlust, J. A., and K. J. Smith. 1967. “Welded FM IEC. 1989. Publication 60853-2. “Calculation of the
Connection for Aluminum Pipe Cables.” IEEE Transac- Cyclic and Emergency Current Rating of Cables, Part 2;
tions on Power Apparatus and Systems. Vol. 86. No. 10. Cables Rated Above 36 kV.”
October. pp. 1265-1273.
IEC. 1999. Publication 60859. “Cable Connections for
Gregory, B. et al. 1991. “A Range of Transition Joints Gas-insulated Switchgear for Rated Voltages of 72.5 kV
for 33 kV–132 kV Polymeric Cables to Paper Insulated and Above.” First Edition.
Cables.” Jicable 1991.
IEC. 2001. Publication 62067. “Power Cables with
Extruded Insulation and Their Accessories for Rated
8-71
Voltages Above 150 kV (Um = 170 kV) up to 500 kV Mcllveen, E. M., R. Waldron et al. 1978. “Mechanical
(Um = 550 kV) - Test Methods and Requirements.” Effects of Load Cycling on Pipe-Type Cable.” IEEE
Transactions on Power Apparatus and Systems. Vol. 97.
IEC. 2004. Publication 60840. “Power Cables with May/June.
Extruded Insulation and Their Accessories for Rated
Voltages above 30 kV (Um = 36 kV) up to 150 kV Minemura S. et al. 1989. “500-kV Oil-filled Cable
(Um = 170 kV) - Test Methods and Requirements.” Installed on Bridges.” Proceedings of the IEEE Trans-
mission and Distribution Conference. New Orleans.
IEEE. 1996. Standard 48-1996. “IEEE Standard Test Paper no. 89TD 354-2 PWRD.
Procedures and Requirements for Alternating-Current
Cable Terminations 2.5 kV Through 765 kV.” Minemura, S. et al. 1990. “Completion of Installation of
500 kV PPLP-Insulated Self-Contained Oil Filled Cable
IEEE. 2000. Standard 404-2000. “IEEE Standard for Along Seto Ohashi Bridge for Honshu-Shikoku Inter-
Extruded and Laminated Dielectric Shielded Cable connecting Transmission Line.” Sumitomo Electric
Joints Rated 2500-500 000 V.” Technical Review. No. 29. January.
IEEE. 2002. Guide 1300-1996 (R2002). “IEEE Guide Miranda, F. J., and P. Gazzana Priaroggia. 1976. “Self-
for Cable Connections for Gas-Insulated Substations.” contained Oil-filled Cables. A Review of Progress.” Pro-
ceedings of I.E.E. (British). Vol. 123. No. 3. March.
Jerome Hydraulics Co., Inc. 1965. “Type FM Connec-
tor.” Data Bulletin no. 1565. February. Nakabasami, T. et al. 1985. “Investigation for Commer-
cial Use of 275 kV XLPE Cables and Development of
Kikuchi, K. et al. 1992. “Recent Technical Progress in Extrusion Type Molded Joint.” IEEE Trans. Vol.104.
Accessories for Extra-high Voltage XLPE Cables in No. 8.
Japan.” CIGRE 1992 Session. Paper 21-203.
Nakamura, S. et al. 2003. Development of Cold Shrink-
Kreuger, F. H. 1989. Partial Discharge Detection in able Joints for 110-230 kV XLPE Cable.” Jicable 2003.
High-Voltage Equipment. Butterworth & Co Ltd. Paper A.5.3.
Kubota, T. et al. 1994. “Development of 500 kV XLPE Nakanishi, Y. et al. 1995. “Development of Prefabri-
Cables and Accessories for Long Distance Underground cated Joint for 275 kV XLPE Cables.” IEEE Trans. on
Transmission Lines – Part II: Jointing Techniques.” Power Delivery. Vol. 10. No. 3.
IEEE/PES 1994 Winter Meeting.
Nelson, W. et al. 1978. “A Premolded Splice for 138 kV
Kunze, D. et al. 2000. “Macroscopic Internal Interfaces Extruded Dielectric Cables.” IEEE Trans. on Power
in High Voltage Cable Accessories.” CIGRE 2000 Ses- Delivery.Vol. 97. No. 6.
sion. Paper 15-203.
Occhini, E. et al. 1978. “Self Container Oil Filled Cable
Le Peurian, S. et al. 1999. “Development of Synthetic Systems for 750 and 1100 kV - Design and Tests.”
and Composite Ter minat ions for HV and EH V CIGRE 1978 Session. Paper 21-08.
Extruded Cables.” Jicable 1999. Paper A5.4.
Oesterheld, J. et al. 1992. “Optimized Design of Acces-
Le Peurian, S. et al. 1992. “Synthetic Terminations for sories for 245 kV and 400 kV XLPE cables.” CIGRE
Very High Voltage Cables.” CIGRE 1992 Session. Paper 1992 Session. Paper 21-202.
21-201.
Okada, et al. 1999. “Development of Prefabricated Joint
Leufkens, P. P. et al. 1992. “A new Generation of Acces- for 500 kV XLPE Cable.” Jicable 1999. Paper A5.2.
sories for (E)HV Extruded Cable Systems.” CIGRE
1992 Session. Paper 21-204. Parmigiani, B. 1988. “Premolded Accessories for HV
Extruded Cables.” CIGRE 1988 Session. Paper 21-08.
Luton, M. H. et al. 2004. “Entirely Synthetic Termina-
tions for High Voltage Cables – Assessment of Service Parmigiani, B. et al. 1987. “Development of a Rapid-fit-
Life.” CIGRE 2004 Session. Paper B1-109. ting Pre-molded Joint for Extruded Insulation Cables.”
Jicable 1987.
8-72
Parmigiani, B. et al. 1986. “Zinc Oxide Sheath Voltage Takeda, N. et al. 1995. “Development of 500 kV XLPE
Limiters for HV and EHV Power Cables: Field Experi- Cables and Accessories for Long Distance Underground
ence and Laboratory Tests.” IEEE Trans. on Power Transmission Lines- Part IV: Electrical Properties of
Delivery. Vol. PWRD-1. January. p. 164-170. 500 kV Extrusion Molded Joints.” IEEE/PES 1995
Summer Meeting.
Parpal J. L. et al. 1998. “Prequalification Testing of
345 kV Extruded Insulation Cable System.” CIGRE Tanaka, T. et al. 2000. “Dielectric Characteristics of
1998 Session. Paper 21-101. Interfaces in Prefabricated Joints of Extra-high Voltage
XLPE Cables.” CIGRE 2000 Session. Paper 15-206.
Peschke, E. F. et al. 1996. “A New Generation of Joints
for XLPE Insulated Extra-high Voltage Cables.” Tanaka, T. et al. 1992. “Development of Prefabricated
CIGRE 1996 Session. Paper 21-204M. Type Straight Through Joint for 275 kV XLPE Cables.”
IEEE Power Eng. Review.
Rosevear, R. D. et al. 1995. “Development of an XLPE
Insulated Cable System for 400 kV Operation.” Jicable Tsuchiya, S. et al. 1999. “Development of Block Molded
1995. Paper A.1.4. Joint for 275 kV XLPE Cable and its Application to
Actual Line.” Jicable 1999. Paper A5.7.
Roughley, T. H., J. T. Corbett et al. 1974. “Design and
Installation of a 138 kV High Pressure Van Der Wijk, G. P. et al. 1998. “Development and
Gas-Filled Pipe Cable Utilizing Segmental Aluminum Qualification of a New 400 kV Cable System with Inte-
Conductors.” IEEE Transactions grated Sensors for Diagnostics.” CIGRE 1998 Session.
on Power Apparatus and Systems. Vol. 93. No. 2. Paper 21-103.
March/April.
Vitkovitch, D. 1966. Field Analysis. London & Prince-
Saito, M. et al. 1974. “New Prefabricated Accessories for ton, NJ: Van Nostrand.
66 – 154 kV Crosslinked Polyethylenecables.” Under-
ground Transmission and Distribution Conference. Wakabayashi, H., H. Tsuji, and H. Kurihara. 1970.
“Reconsideration of Design Practice for an Oil-Filled
Schmid, M. et al. 1995. “Use of Elastomeric Material Cable Joint and the Development of a Smaller Joint.”
for Manufacturing of HV and EHV Accessories.” Jica- IEEE Transactions on Power Apparatus and Systems.
ble 1995 Paper A.3.3. vol. 89, no.6, July/August 1970.
Selsing J. 1985. ”A Versatile Computer Method for Watanabe, Y. et al. 1986. “Development of New 500 kV
Computation of Conductor Temperatures in Cable Ter- Laminated Paper Insulated Self-Contained Oil-Filled
minations and Pipe Cable Systems.” IEEE Transactions Cable and its Accessories.” IEEE Transactions, Paper 86
on Power Apparatus and Systems. vol. PAS-104, no 4, SM 390-9.
April 1985.
Weedy, B. M., and J. P. Perkins. 1967. “Steady State
Short, H. D. 1949. “A Theoretical and Practical Thermal Analysis of a 400 kV Cable
Approach to the Design of High-Voltage Cable Joints.” Through-Joint.” Proceedings of the IEEE. Vol. 114.
AIEE Transactions on Power Apparatus and Systems.
vol. 68, 1949. Willems, H. et al. 1995. “A New Generation of HV and
EHV Extruded Cable System” Jicable 1995 Paper A.1.6.
Stepniak, F. M. et al. 1987. “Effect of Aging on Pre-
molded Elastomeric Splices for 138 kV XLPE Cable.” Yagisawa, K. et al. 1990. “Development of Extrusion
IEEE Trans. Vol. 2. ‘Type Molded Joint for 275 kV and 500 kV XLPE
Cables.” Hitachi Cable Review. No.9. August.
Takaoka, M., T. Mohtai, and S. Yoshida. 1980. “Devel-
opment of Flexible Joint for Large Capacity Submarine Acknowledgments:
O.F. Cable.” IEEE Transactions on Power Apparatus and • CIGRE for the provided large technical documenta-
Systems. Vol. 99. No. 3. May/June.
tion
• Bruno Parmigiani for the important technical contri-
bution
8-73
CHAPTER 9 Hydraulic Design and

Equipment
Author: Dave Purnhagen, Underground Systems Inc. (USi)
Reviewer: Jay A. Williams, Power Delivery Consultants, Inc. (PDC)
Hydraulic or gas management systems are essential for the operation of many types of
underground high-voltage cables including high-pressure fluid-filled pipe-type, self-
contained fluid-filled, and high-pressure gas-filled pipe-type cables. The hydraulic or gas
system functions as a means to provide pressure, thus enhancing the dielectric properties
of the cables, and also as a heat transfer medium. This chapter discusses hydraulic system
theory, equipment, and application. It provides the engineer with the basic formulae to
design the hydraulic system and assess its performance. The various types of equipment
for pressurizing and cooling cable systems are also treated. Finally, a section is devoted to
system design guidelines, including such issues as reliability and environmental impact.
Dave Purnhagen has over 35 years of experience designing forced-

cooling and hydraulic equipment for underground fluid-filled cable
systems and the complex data acquisition and control systems
associated with forced-cooled cable systems. He has engineered many of
the cooling systems, including refrigerated systems, presently utilized on
electric utility pipe-type cables. He designed the Electric Power
Research Institute forced-cooling facility at Waltz Mill and as principal
investigator conducted a four-year experimental program to determine
design parameters for forced-cooled cable systems. The results were published in 1984 in
the Designer’s Handbook for Forced-Cooled High-Pressure Oil-Filled Pipe-Type Cable
Systems, which is considered the industry standard for forced-cooled calculations and
guidance for cooling system design. Mr. Purnhagen went on to develop the concept of
forced-cooled dynamic feeder rating systems and ultra-high-sensitivity leak detection
systems for high–pressure, pipe-type fluid-filled and self-contained fluid-filled cables. As
guest scientist at Brookhaven National Lab, he worked with the tracer technology
scientific staff to develop equipment and strategies for implementing perfluorocarbon
tracers for cable leak location.
Mr. Purnhagen has been active in professional organizations and has chaired
IEEE/PES/ICC Task Groups on Forced Cooling and has served as Convener of CIGRE
working group 21-08 on forced-cooled cables. He is well published with over 30 industry
publications related to underground pipe-type cable systems and has delivered many
industry presentations and seminars.
9-1
Chapter 9: Hydraulic Design and Equipment EPRI Underground Transmission Systems Reference Book
9.1 INTRODUCTION 9.2 ENGINEERING DESIGN PRINCIPLES

The pressures of dielectric fluid-filled cable systems vary
principally due to loading induced thermal cycles and to 9.2.1 Cable System Pressure Limits
seasonal environmental temperatures. The need for The low-pressure limit is determined by the dielectric
hydraulic accessories and equipment arises from the properties of the pressurized cable insulation system and
requirement to control these pressures such that the depends on the design of the cable. It has been shown
pressure-bearing capability of components is not that ac breakdown strengths of oil-impregnated paper
exceeded and that the dielectric integrity of the insula- can improve by 30-40% between pressures of 0.34 and
tion system is not compromised at low pressures. In 1.03 MPa (50 to 150 psi) (Lanfranconi et al. 1974), but
addition, hydraulic systems can be utilized to improve only 10-15% additional improvement is realized by rais-
heat transfer from the cable systems by circulation and ing the pressure another 0.69 MPa (100 psi). The high-
forced cooling. Forced-cooling equipment places addi- pressure limit is established by the pressure limitations
tional burdens on the pressure regulation equipment. of pressure-bearing components in the system such as a
The hydraulic system may also be required to identify termination or cable sheath. Transient conditions can be
and possibly control fluid losses in case of a breech of taken into consideration for the high-pressure limits if
the containment system. the limit is based on a mechanical stress.
HPFF Cable System
Chapter 9 reviews hydraulic design and equipment.
HPFF cables take advantage of the nominal improve-
Section 9.2 outlines engineering design principles.
ment in ac breakdown strength above 1.03 MPa
Section 9.3 identifies the calculations for expulsion,
(150 psi) by operating in the upper 1–1.7 MPa (200-
demand, and pressure drop due to flow. Section 9.4
250 psi) range, because the steel encasement permits this
describes pressurization equipment, and Section 9.5
without penalty. Furthermore, HPFF cables are
covers forced-cooling equipment. Section 9.6 reviews
exposed to gas during installation and operation, and
considerations for system design. Section 9.7 presents a
operating at elevated pressures ensures that all gas
worked example.
remains in solution, inhibiting ionization. It is generally
accepted that the minimum pressure that would ensure
The term “hydraulic” is loosely applied to cover both
continuous suppression of ionization in HPFF systems
liquid dielectric fluids (oil) as well as gas (nitrogen).
is 0.52 MPa (75 psi), but most utilities limit the mini-
Accessory equipment and calculating procedures for
mum operating pressure to 0.69–0.83 MPa (100–
High Pressure Fluid Filled (HPFF) pipe-type, High
120 psi). Normal operating pressure levels in the 1.4–1.7
Pressure Gas Filled (HPGF) pipe-type, and Self Con-
MPa (200-250 psi) range provide a healthy margin of
tained Fluid Filled (SCFF) cable systems are covered.
safety for system transients, pressure losses due to leaks,
and other hydraulic disturbances. Recently, normal
In Chapter 9, The International System of Units (SI) is
operation somewhat below these nominal pressures has
used, with English units in parentheses. SI units are con-
been explored to optimize forced-cooled operation. In
sistent with information in Chapter 11 and other refer-
these cases, pressures are closely monitored and con-
ences needed to complete the calculations, and English
trolled by Programmable Logic Controller (PLC)-based
units are the most common for general discussion. Pres-
systems, which provide alarming, control, and safety
sure refers to gauge pressure, whether it is in SI units
shutdown capability. The high-pressure limitation in
(Pascals) or English units (lbf/in 2 = psi), unless it is
static HPFF cable systems is usually the termination.
specified as absolute pressure.
Chapter 8, Section 8.3.3, provides a table of termination
high-pressure limits for various normal and transient
Section 11.4, “Forced Cooling” in Chapter 11 is
conditions. In a forced-cooled system, the high-pres-
required to obtain dielectric fluid properties, and the
sure limit may be piping or a component near the dis-
worked example of Chapter 11 should be utilized to
charge of the pump, not necessarily the termination.
supply the temperatures for the worked example in this
chapter. The principal additional references with which SCFF Cable System
the reader should become familiarized are: Oil Flow and SCFF cables operate under the principle that all gas is
Pressure Calculations for Pipe-Type Cable Systems eliminated from the cable insulation system during
(AIEE 1955), Oil Flow and Pressure Calculations for installation and operation. Operating pressures are
Self-Contained Oil-Filled Cable Systems (Buller et al. thereby established by the dielectric requirements of
1956), and Designer’s Handbook for Forced Cooled each voltage class system and the pressure-bearing capa-
High-Pressure Oil-Filled Pipe-Type Cable Systems bility of the cable sheath. The pressure transient during
(Purnhagen 1984). cable loading or unloading and the axial distribution of
9-2
EPRI Underground Transmission Systems Reference Book Chapter 9: Hydraulic Design and Equipment
pressures under these conditions are other important The thermal expansion for a given volume can be
considerations when establishing the nominal operating expressed by:
pressure. SCFF cables are usually classified as “low
pressure,” when nominal operating pressures may be in ∆Vtn = β n ·Vn ·(T2 n - T1n ) m3 9-1
the 14-414 kPa (2-60 psi) range, or “high pressure,” Where:
which utilize pressures that exceed 414 kPa (60 psi) and n = an integer referencing a cable component.
may be as high as 1379 kPa (200 psi). Pressure is main- ∆Vtn = change in volume due to temperature
tained on “low-pressure” SCFF cables by accumulator- change, m3.
style reservoirs, as described in Section 9.4.3, and on βn = coefficient of volumetric thermal expan-
“high-pressure” SCFF cables by an active pumping sion, 1/°C.
plant. In both cases, gas contact with the dielectric fluid T2n = temperature at condition 2, °C.
is eliminated. T1n = temperature at condition 1, °C.
HPGF Cable System
The pressure-related volumetric change of a component
HPGF cables are similar to HPFF cables, except that
can be expressed by:
nitrogen with nominal pressures in the 1.4 – 1.7 MPa
(200-250 psi) range is utilized to transmit pressure to
the insulation. The insulation is impregnated with a (
∆Vp n = α n ·Vn · P2 n - P1n ) m3 9-2
very viscous fluid, and is somewhat thicker than corre- Where:
sponding voltage class HPFF cables. The maximum n = an integer referencing a cable component.
operating voltages of HPGF cables are limited to ∆Vpn = change in volume due to pressure change,
138 kV. The advantage is that the pressurized gas dis- m3.
tributed along the length of the system absorbs the αn = compressibility factor, 1/Pa.
pressure changes due to loading and seasonal tempera- P2n = Pressure at condition 2, Pa.
ture changes without complex peripheral equipment. P1n = Pressure at condition 1, Pa.
Additionally, these cables can be installed in areas with
significant elevation differences along the route, Table 9-1 lists suggested values for the thermal expan-
because they are not limited due to extreme pressure sion coefficients and compressibility for common cable
differences that could limit a fluid-filled cable system. materials.
The minimum pressure preventing ionization in HPGF
cables is not well documented. Typical operating
Table 9-1 Volumetric Thermal Expansion Coefficients and
practice is to alarm low pressures at approximately
Compressibility Modulus for Common Cable Materials
1.24 Mpa (180 psi) and disconnect at 1.03 Mpa (150
psi). As with HPFF cables, the high-pressure limit is Volumetric Coef- Volumetric
ficient of Thermal Compressibil-
usually restricted by the terminations. Material
Expansion, β ity Modulus, α
(1/ °C) (1/Pa)
9.2.2 Basic Hydraulic Equations Copper .000039
The primary need for pressure regulation is a result of Aluminum .000052
Lead .000066 Negligible
thermal expansion and contraction of cable compo-
nents, primarily the dielectric fluid in fluid-pressurized Steel .000027
systems. The increase in pressure also alters the geome- Impregnated Paper .00028 8.28 x 10-10
try of the cable system by expanding the containment Dielectric Fluid – Mineral .00075
(i.e., pipe or sheath) and compressing internal compo- Dielectric Fluid – Poly-
.00079
butene
nents. The increase in pressure also drives the free fluid 6.91 x 10-10
Dielectric Fluid – Alkylben-
along the cable’s axis in the fluid containment section of zene
.00081
the cable. The basic equations describing these hydraulic
phenomena are given in this section.
9-3
The cable component volume, Vn (m3), is computed by: The reader is directed to Section 11.4 for a thorough
treatment of pressure drop in cable systems. The static
⎡ ⎛ D ⎞2 ⎛D ⎞ ⎤
2
hydraulic head due to an elevation change is:
Vn = Fn ·π ·⎢ N n ⎜ n ⎟ − N n −1 ⎜ n −1 ⎟ ⎥·L m3 9-3
⎣⎢ ⎝ 2 ⎠ ⎝ 2 ⎠ ⎦⎥ ∆Pelev = γ·g c ·∆ h Pa 9-6
Where: Where:
Fn = material fill of each cable component, γ = density of fluid, kg/m3.
dimensionless. gc = gravitational conversion constant, m/s2.
Dn = outer diameter of cable component, m. ∆h = elevation difference, m.
Dn-1 = inner diameter of cable component, m.
N = number of cable components, dimension- In general, a thermal expansion increases pressure in a
less small axial section of cable, creating a pressure differen-
L = length of cable section, m. tial, driving the fluid along the axis of the cable toward a
point of egress. By dividing the cable into small axial
The Fn term is the fill of the specific material within the increments, a hydrodynamic solution can be obtained
cylinder defined by D n and D n-1 . The fill (F n ) for the through system partial differential equations, combin-
metallic portion of compact segmental conductors is ing the phenomena described by Equations 9-1 through
approximately 0.84, and the impregnated paper in the 9-5. The static hydraulic head, Equation 9-6, is then
insulation cylinder is approximately 0.50. The remaining superimposed for a total solution. These equations are
oil fill for the conductor and insulation is 0.16 and 0.50, usually solved numerically and only justified if
respectively. The fill for the copper portion of a SCFF extremely accurate modeling of system volumetric flows
keystone conductor can approach unity. is required. This approach is used in high-sensitivity
leak-detection systems (Engelhardt and Purnhagen
The containment component (i.e., pipe or sheath) is 1991) and sometimes for long-length SCFF cable sys-
computed by: tems. For most applications, simplifying assumptions
2 such as ignoring the compressibility terms for reservoir
⎛D ⎞
Vcont = π ⎜ m ⎟ ·L m3 9-4
sizing or the axial pressure distribution due to thermal
⎝ 2 ⎠ expansion in HPFF cables can be made so that the
Where: equations can be used independently.
Dm = inner diameter of cable containment com-
ponent, m. 9.2.3 Thermal and Hydraulic Driving Forces
The primary thermal and hydraulic driving forces that
The frictional pressure gradient along the length of the affect cable system hydraulics are: (1) cable loading, (2)
cable due to flow can be expressed by: ambient thermal cycles, (3) elevation changes, and (4)
forced cooling. A solution to Equation 9-1 requires
∆P γ·V 2 coupling to the thermal equations needed to compute
= K f f· , Pa/m 9-5 the temperatures of cable components at the condi-
∆x 2·D h tions of interest—i.e., T1n and T2n. These equations are
Where: described in Chapter 11. Knowledge of the ambient
Kf = correction factor, dimensionless. earth temperature is also required to determine the
f = friction factor, dimensionless. overall operating temperature range for reservoir siz-
γ = density of fluid, kg/m3. ing. Finally, forced cooling can create additional
V = velocity, m/s. demands on the hydraulic system by rapidly impress-
∆x = length, m. ing reduced temperatures upon the cable system and at
Dh = hydraulic diameter, m. the same time redistributing pressures due to flow fric-
tional losses.
9-4
9.3 HYDRAULIC CALCULATIONS components. The pressure change of a HPGF cable

system can be computed by coupling the volumetric
9.3.1 Expulsion Calculation equation to the Boyle’s Law equation (Equations 9-10
The term “expulsion” is used to designate the fluid that and 9-11):
( )
is expelled from the cable system due to thermal expan-
sion and must be accepted by the reservoir. Although Vk HPGF = Vk pipe − 3· Vk cond + Vk ins m3 9-10
the term denotes positive flow from the cable, the equa-
tions are equally valid for contraction, or negative flow. V1HPGF
For the purpose of sizing reservoirs, the compressibility P2 HPGF = P1HPGF · Pa 9-11
term in Equation 9-2 is usually ignored, and only the V2 HPGF
effect of thermal expansion effects is considered. The Where:
general equation for expulsion (Vxp) is: k = an integer denoting a condition—i.e., 1 =
initial condition, 2 = final condition.
m
V xp = ∑ ∆Vt n − ∆Vt cont m3 9-7
P = absolute pressure, Pa.
n =1
9.3.2 Demand Calculations
Where:
∆Vtcont = the change in containment volume The term “demand” denotes the rate of change of fluid
volume or “flow rate” required to maintain a desired
Therefore the expulsion for HPFF cable becomes: pressure in a cable system. The peak demand is a result
of cable unloading, usually from its highest emergency
VHPFFxp = 3·(∆Vt cond + ∆Vt ins ) + ∆Vt oil - ∆Vt pipe m3 9-8 operating temperature. The demand calculation dictates
Where: the capacity of the pressure delivery system from
∆Vtcond = ∆Vtcond metal + ∆Vtcond oil reservoir to cable. The demand calculation is accom-
∆Vtins = ∆Vtins paper + ∆Vtins oil plished by computing the rate of temperature change at
small time increments ∆t, over the entire transient
The expulsion for a single-conductor SCFF cable period using the transient thermal equations of Chapter
becomes: 11 or equation 9-12. Generally, for HPFF cables, the
axial pressure gradient due to thermal expansion flows
VSCFFxp = ∆Vt oilduct + ∆Vt cond + ∆Vt insul − ∆Vt sheath m3 9-9 can be ignored and the computed temperature changes
can be directly applied to the expulsion equation,
Equation 9-7. The computation is conducted for
For reinforced or ar mored sheaths, the ar mor is
successive time increments until the maximum or peak
mechanically coupled to the sheath, preventing it from
value for demand is obtained. The demand equation in
expanding freely. Although it is possible to explicitly
general terms is then:
calculate this effect, it is suggested that the expansion
coefficient for the sheath be modified by a 0.80 multipli-
⎛ m ⎞
cation factor to account for the armor.
⎜ ∑ ∆Vt n − ∆Vt cont ⎟
Similarly, expulsion equations for any cable type, such QDemand = ⎝ n =1 ⎠ m3/s 9-12
as for three-conductor oil-filled cables can be generated dt
by summing the internal components volumetric
It should also be noted that, in actual practice, with
changes, ∆Vtn terms, then subtracting the containment
multiple cables and/or fluid-return pipes in close prox-
volumetric change, ∆Vtcont.
imity, the thermal interaction between these systems
must be taken into account when evaluating fluid
HPGF cable systems rely on pressure variations within
demand. For example, with many parallel cable installa-
the pipe to accommodate expansion and contraction of
9-5
tions, the loss of one cable circuit results in a rapid also be considered. This is computed by Equation 9-13,
increase in load in the companion circuit. as follows:
HPFF Cable System Qfc = Qcp · βoil · (Tin-Tout) m3/s 9-13
The required capacity of the pump is determined by Where:
computing the maximum demand due to current Qcp = Cooling plant flow, m3/s.
unloading, as well as the demand that can be caused by βoil = Thermal expansion coefficient of oil, 1/°C.
forced-cooling equipment. Cooling of the dielectric Tin = Inlet temperature to cooling plant, °C.
fluid, as well as the pressure redistribution due to the Tout = Outlet temperature of cooling plant, °C.
pump during startup conditions, can create the demand
caused by forced-cooling equipment. Another factor affecting demand in forced-cooled
cable systems is the transient redistribution of
The per-unit-length demand due to cable unloading pressures during the startup transient of the pump. By
(ahpff), has been computed numerically using Equation applying Equation 9-2 to the oil volume and dividing
9-12 for several common HPFF cable constructions and by the time to redistribute pressures, the following
is presented in Figure 9-1. The starting condition is expression for demand due to pressure redistribution
100°C, 300-hr emergency followed by an abrupt (Qpr) can be generated:
disconnection of load. Although the curves of Figure 9-
1 are illustrative of typical demand curves, it should be Qpr = ∆Vpoil/∆tpr m3/s 9-14
noted that there might be load scenarios that cause Where:
greater demands. ∆tpr = the time to redistribute pressures from P1
to P2.
If the cable system is forced cooled, the additional
demand caused by the cool-down transient (Qc) from a If the cable system is of uniform geometry, then average
self-cooled condition to a forced-cooled condition must system pressures can be taken for P1n and P2 n. If, for
instance, an un-cabled pipe is used in the forced-cooling
loop, then the average pressure of that volume must be
separated from the cabled pipe. If the pressure redistri-
Figure 9-1 Per-unit-length demand vs. time for several HPFF cable constructions
following a sudden disconnection of emergency load.
9-6
bution demand is negative, because average pressures Unlike HPFF cable systems the axial pressure gradient
are decreasing during the pump startup transient, then due to thermal flows in SCFF cables cannot be ignored
the Qpr term can be ignored in Equation 9-14. (see Section 9.3.3). In addition, the pressure sources
(reservoirs) may have complex volumetric – pressure
The required pump capacity (Qhpff) is then: relationships and need to be coupled to the system
Qhpff = ahpff ·L·Ka Qfc + Qpr m3/s 9-15
model. Calculating temperatures at successive time
+
increments using the appropriate transient methodolo-
Where: gies in Chapter 11 and coupling these temperatures to
L = length of cable, m. the demand, Equation 9-12 can determine SCFF fluid
ahpff = per-unit-length demand due to cable demands. The per-unit-length demand due to cable
unloading, m2/s unloading (ascff), has been computed numerically for
Ka = 1.666x10-8, conversion factor (converts several common SCFF cable constructions and is pre-
cm3/min) sented in Figure 9-2. The starting condition is 100°C,
Qpr = demand due to pressure redistribution, 300-hr emergency, followed by an abrupt disconnection
m3/s) of load. Although the curves of Figure 9-2 are illustra-
Qfc = demand due to a forced-cooled cool-down tive of typical demand curves, it should be noted that
transient, m3/s there might be load scenarios that cause greater
demands.
The pump size should be selected on the basis of
Equation 9-15 with a reasonable safety factor applied The one-dimensional demand (point location) is then:
such as 1.2-1.5. Over-sizing the pump may be counter-
Qscff = ascff · L · Ka m3/s 9-16
productive, causing rapid pump cycling of the pump
during normal operation and overfeeding a leak in case The demand (Qscff) must then be coordinated with the
of a leak. The designer should also consider the pressure profile of the cable system, axial pressure drop
likelihood of coincident events, such as a sudden discon- due to thermal flows and characteristics of the reser-
nection of load while the cooling plant is rapidly cooling voirs. This is generally accomplished by computing axial
the system, or provide the control means to prevent the pressure distribution curves using a model such as that
simultaneous occurrence of such events. described in Section 9.3.3 at each successive time inter-
SCFF Cable System val, and under a number of loading scenarios. Reservoir
The fluid demand in SCFF cables is principally a result design and size are then adjusted such that the maxi-
of the thermal transient due to current unloading. mum and minimum pressure limits are not exceeded.
Figure 9-2 Per-unit-length demand vs. time for several types of SCFF cable construction
following a sudden disconnection of emergency load.
9-7
HPGF Cable System be accomplished numerically by dividing the system into

The demand requirements for HPGF systems are not small axial segments ∆x, as indicated in Figure 9-3.
an issue, because the contraction due to cable loading
is accommodated by the compressibility of the nitro- The flow expulsions for each segment can be computed
gen medium, which is distributed along the length of using Equation 9-7 at each successive time step and
the system. summed as the flow propagates along the cable duct
axis. The pressure gradient across each segment can
9.3.3 Pressure Drop Due to Flow then be computed using Equation 9-15, where the
hydraulic diameter in this case would be the inner diam-
HPFF Cables eter of the oil duct. If the oil duct contains a spiral sup-
Pressure losses due to thermal expansion flows can usu- port tape, the inner diameter of the support tape can be
ally be ignored in HPFF cables. However, certain older used for a conservative result. So, for each time step, an
systems with small-diameter pipes (< 152 mm, 6 in.) and axial pressure gradient is then computed. This result
very viscous fluids such as high-viscosity polybutene must be coupled to the pressure volume (PV) response
and high-viscosity mineral oil (HVP, HVM) can exhibit of the reservoir system, which is treated in Section 9.4.
significant pressure drops due to thermal expan-
sion/contraction (AIEE 1955). Forced-cooled systems HPGF Cable System
and even moderate or slow circulation systems can cre- Thermally induced axial flows in HPGF systems are
ate significant pressure losses along the length of the insignificant, if not nonexistent, because the pressure
cable. The frictional pressure drop along the length of medium is evenly distributed along the length of the
the cable due to flow can be expressed by: cable. However, the pressure drops due to gas flow dur-
ing filling operations and other maintenance procedures
∆P = K f f
(γ ·V ·∆x) 2
Pa 9-17
can be of interest. Equation 9-15 can be used to calcu-
late those pressure drops, provided that the appropriate
2·Dh gas fluid properties are used at the pressure of interest.
Where: Table 9-2 provides the properties of nitrogen at atmo-
Kf = correction factor, dimensionless. spheric pressure and at 1.38 MPa (200 psi).
f = friction factor, dimensionless.
γ = density of fluid, kg/m3.
9.4 PRESSURIZATION EQUIPMENT
V = velocity m/s.
∆x = length, m.
9.4.1 Introduction
Dh = hydraulic diameter, m.
(D )
Pressurizing plants, commonly called pumping plants,
2 2
pipeID − 3.03·Dshield are employed as the pressure source for HPFF cable.
Dh = 9-18 They may also be used or required on SCFF cables,
D pipeID + 3.03·Dshield although reservoirs are generally used to furnish the
pressure source for those cables. Pumping plants are
Section 11.4 provides the necessary details, including used on SCFF cables when the lengths are long, when
fluid properties and friction factors, for completing the pressure can be applied at the end points only, and/or if
computation. The correction factor includes the addi-
tional hydraulic impedances due to cable snaking, fit-
tings, bends, joints, and diffusion chambers. Although
these can be computed discretely, Kf of 1.2 is typical for
most systems.
SCFF System
The axial pressure gradients in SCFF oil ducts due to
thermally driven flows are significant, and the transient
due to a step function unloading often dictates the over- Figure 9-3 Hydraulic model of SCFF cable system.
all hydraulic design. During unloading, the pressures in
the center of the cable system can become dangerously
Table 9-2 Properties of Nitrogen at Different Pressures
low if the pressures at the feeding ends cannot sustain
the flow demand. Upon cable loading, the pressures in Density Viscosity
Pressure
(kg/m3) (m2/s)
the center of the system can become dangerously high if
0.1 MPa 1.15
the reservoir system cannot easily accept the expulsion. 1.6x10-5
140 MPa 22.1
The determination of the axial pressure distribution can
9-8
the cable is of the high-pressure type. Underwater sures are gauge (i.e., increments above atmospheric pres-
SCFF cables and SCFF cables with severe elevation sure), unless otherwise indicated.
change may require pumping plants rather than reser-
voirs. HPGF cables do not require pumping plants or Midnight until 9 am: The loading is low, the circuit is
reservoirs. Extruded-dielectric cables require no fluid cooling, and the fluid is contracting. The pressure drops
pressurization, except for small reservoirs to accommo- to 1276 kPa (185 psig); the pump turns on and brings
date fluid expansion and contraction in some termina- the pressure to about 1689 kPa (245 psig), where it is
tion designs. held by opening the normal relief valve for the five min-
utes that the pump is set to run. At the end of five min-
9.4.2 Pressurization Plants (HPFF /SCFF Cable utes, the pressure slowly decays until the pump starts
Systems) again and the cycle repeats.
Overview 9 am to 6 pm: The loading increases, the circuit heats up,
In HPFF cable systems, pressurization plants, consist- and the fluid expands and flows back to the reservoir
ing of steel reservoirs (tanks), and a mechanical pump- tank through a cable relief valve, which clamps the pres-
ing system are utilized due to the high volume of sure to about 1758 kPa (255 psig).
dielectric fluids within the cable and/or cable pipe, large
expansion volumes, and high pressures. This permits the 6 pm to Midnight: The loading decreases, the circuit
storage of bulk amounts of fluid by relieving the high- starts cooling, and the fluid begins contracting. The
pressure expulsion fluid into a volume at pressures daily cycle is repeated.
somewhat above atmospheric. Upon demand, the
pumping system draws the low-pressure fluid from the Figure 9-4 shows fluid pressure variation for pump-
reservoir and raises its pressure, thereby supplying the operating conditions upon demand. If two circuits are
cable system at the higher required operating pressures. installed in parallel, a slow flow of 38-57 l/min (10-
This cycle is repeated indefinitely as the cable load 15 gal/min) can be established between them to smooth
cycles and system conditions change. hot spots. In this case, the pressure is held constant at
Operational Description about 1379 kPa (200 psig) by a relief valve. The plant
A typical sequence of operations for a pressurizing plant illustrated is designed to maintain a pressure range of
is shown in Figure 9-4 and is described below. All pres-
Figure 9-4 Sequence of operations of pressurized plant.
9-9
1276-1724 kPa (185-250 psig) for normal operation of In general, at least two pumps and two pressure-control
the cables. ladders are recommended, with one per cable pipe
preferred. This configuration gives the redundancy
Pumping Plant Hydraulic Design necessary for reliable unattended operation of the pres-
The pumping plant has the following hydraulic sub- surizing plants.
systems, as shown schematically in Figure 9-5a and pic-
torially in Figure 9-5b. The tank and nitrogen system, shown schematically in
Figure 9-6a and pictorially in Figure 9-6b, serves to sup-
• Tank and nitrogen system
• Fluid pressure section (pumps).
• Pressure control system (ladders).
• Outlets
Figure 9-6a Schematic of tank and nitrogen blanket

system.
Figure 9-5a Schematic of a fluid pressure and pressure

control system portion of a pressurizing plant.
Figure 9-5b Typical fluid pressure and pressure control

system portion of the pressurization plant (Courtesy of Figure 9-6b Typical tank and nitrogen blanket system
Prysmian Cables & Systems). (Courtesy of Prysmian Cables & Systems).
9-10
ply or accept the active fluid of the cable system. In tronic instrumentation, providing fluid level alarms as
addition, it can provide reserve fluid to permit contin- well as accurate level readings locally and remotely.
ued flow of fluid toward a hydraulic leak. The tank is
protected from contaminants, including moisture, by a See Section 9.4.5 for details regarding determination of
blanket of nitrogen over the fluid. Nitrogen pressure the appropriate tank volume.
varies with volume as fluid enters and leaves the tank.
Typical pressure-versus-volume curves are shown in The fluid pressure system consists of a positive displace-
Figure 9-7. Under normal operation, no nitrogen is lost ment pump typically sized from 19 to 57 l/min (5-
or added. 15 gal/min). Additionally, the hydraulic controls consist
of a safety relief valve to protect the pump in case of a
Alternately, the tank can be operated under vacuum. blockage at the pump outlet, the normal pump relief
This is necessary when pumping plants supply SCFF valve that sets the upper limit of pressure that the pump
cable systems, or it may also be an option for HPFF can deliver to the cable system, and a pressure switch
cable systems when low gas contents are desirable. and timer mounted in the pressure control section that
starts the pump on a falling pressure and keeps the
The pressure–volume characteristic is not an issue in pump running for a set time. A check valve prevents
this case. Special pumps, a low-elevation pumping fluid from the system from flowing back through the
sump, or some other means must be devised to maintain pressurizing pump into the tank. These components are
satisfactory positive pump suction pressures. shown in Figure 9-5A, B. The pressure control system
sets the upper limits of pressure during periods of
For nitrogen-based systems, the storage tank is fur- increasing cable load. The cable relief valve provides the
nished with a rupture disk and ASME-rated relief valve path for the relief of excessive pressure built up in the
to relieve over-pressures. Typical nitrogen operating cable system due to the thermal expansion of the fluid.
pressure ranges from 14 kPa (2 psi) to 103 kPa (15 psi)
gauge. A nitrogen cylinder and pressure regulator sys- Other valves can be provided in the fluid pressure sec-
tem replaces any lost nitrogen. High and low nitrogen tion:
pressure alarms indicate any abnormal pressure condi- • Reduced pressure relief valve—used in maintenance
tion. The storage tank may be equipped with partitions
work to provide a lower system pressure—for exam-
to retain fluid for use in a good cable when a pump is
ple, to make a freeze of the fluid in the cable pipe. A
using fluid from the tank to supply a leaking cable.
pressure of 414 kPa (60 psi) is common.
Fluid levels can be monitored with both visual and elec-
• Oscillation relief valve—used to provide a lower pres-
sure return path when the unit is equipped for oscilla-
tion.
• Migration control valve—used to keep the fluid level
in the tanks balanced manually when plants are used
at both ends of a line.
• Manual bypass valve—used for plant testing and set-
ting of relief valves.
The outlet section provides for the delivery of the fluid

to the cable system, as shown in Figure 9-5. Strainers
are furnished to trap contaminants from the cable sys-
tem from entering the pressurizing plant. Without the
strainer, any incoming particulates could lodge on relief
valve or check valve seats, causing the valve to weep and
often providing symptoms similar to a small leak.
A manifold to connect pressure sections to cable pipes

is made in this section. Sensors are connected after the
last valve in this system to measure pressure and
provide high-and low-pressure alarms for the cable.
Manual or automatic flow-limiting devices may be
Figure 9-7 Typical nitrogen pressure curves.
mounted around the last valve in this system to control
9-11
the amount of fluid that can be fed to a leaking cable • Flow limiting in effect
pipe. Insulating flanges are usually supplied at the • Plant access
outlet of the plant, to isolate it from the cathodic
protection voltage. • Activation of fire suppression
The pressurizing plant should be connected to the cable Figure 9-8 shows a typical control panel. The operation
system with a minimum of 5.1-cm (2-in.) diameter pipe. and control screens can display a virtual switch and light
At shallow depths of burial, in very cold country or panel mimicking an electromechanical control panel
when high viscosity fluids are used, the pipe size should with HAND/OFF/AUTO switches, indicator lights, and
be increased to 7.6 cm (3 in) diameter. Lengths up to START–STOP buttons. An example of a virtual electro-
90+ meters (300+ ft) are acceptable. mechanical control panel is shown in Figure 9-9.
Instrumentation and Controls A particularly useful screen for plant operation and
Current pressurization plant control schemes are usu- control is the “live” flow diagram, which is shown in
ally based on operation with Programmable Logic Con- Figure 9-10. The status of pumps, pressures, tempera-
trollers (PLC) and sometimes electro-mechanical relays. tures, and flows can be displayed in real time, and valve
Essentially a rugged industrial computer system, the positions and flow paths can be also displayed if the
PLC offers control logic programming flexibility, valves are equipped with limit switches. This is particu-
enhanced data acquisition capability, and live “touch larly useful for a remote user when rapid response to a
screen” control panels. The virtual control panels are cable system condition is necessary.
divided into several categories:
• Annunciator/Alarm Screens The PLC also provides the capability of remote moni-
toring and control. The virtual panels are available
• Operation and Control Screens
remotely via commercial or public networks, and are
• Data Display and Charting Screens given security codes, so that a remote user can view sys-
• Diagnostic Screens tem status, acknowledge alarms, and perform any func-
tion that the virtual control panel provides. The PLC
The annunciator screen furnishes information on the also provides for sophisticated control schemes, such as
condition of the cable system and indicates present or integration with forced-cooling equipment operation to
incipient trouble. A remote alarm is commonly pro- prevent unwanted pressure excursions.
vided. Some of the commonly displayed alarm points
are listed below: An automatic backup manual control system is also
usually provided to start the pressure pump in the event
• High nitrogen pressure in the tank that the PLC-based control system has failed. The reli-
• Low nitrogen pressure in the tank
• High tank fluid level
• Low tank fluid level
• High pressure for each cable pipe
• Low pressure for each cable pipe
• Frequent operation of each pump
The selection of alarm settings should be based on the

criticality and time to respond to an alarm. Some of the
additional alarms that can be furnished are listed below:
• Rupture disk failure on tank
• Low tank nitrogen cylinder pressure
• Pressure pump failure
• Continuous pump operation
Figure 9-8 Control panel showing PLC touch-screen
• Strainer clogged (courtesy of USi).
9-12
ability of present PLC-based pressurization plants is Leak Detection

very high and self-diagnostics capabilities can avert fail- Various levels of sophistication can be integrated into
ures at inopportune times. pressurizing plants for detecting the presence of a fluid
Figure 9-9 Virtual electromechanical control screen (courtesy of USi).
Figure 9-10 Control panel showing touch-screen with live process flow diagram (courtesy of
MAC Products, Inc.).
9-13
leak on the cable line. These systems are, in increasing • Gas-driven auxiliary pumps—to provide pressuriza-
order of sensitivity: tion in case of plant electrical service outage. Com-
• Alarm indicating low dielectric fluid level in the tank pressed air or nitrogen drives the piston-type pump.
• Active monitoring of fluid level in storage tank • A dc standby pump in case of loss of ac supply
• Frequent or continuous operation of pressurizing • Dual ac power supply

pump • Emergency generators
• Cumulative flow meter • Instrumentation—provides offsite readouts
• On-line leak detection system with high-accuracy • Remote control—required when the pumping plant
flow meters in combination with Dynamic Mass Bal- cannot be reached readily
ance Computation. (See Chapter 15, Section 15.7, • Exterior painting and finishing for appearance
“Leak Management Techniques for Fluid-Filled
Feeders.”) Standby Emergency Plants
Some utilities have portable, emergency pressurizing
At a minimum, the low-dielectric-fluid-level alarm and equipment that, along with a dielectric fluid tank truck,
pressurizing-pump-frequent-operation alarm systems
are typically provided for all pressurizing systems.
Enclosures
Pumping plants can be located indoors or outdoors.
The outdoor plant typically contains the tank, piping,
and controls in an insulated enclosure, mounted on a
steel base. The enclosure provides the environmental
control for lighting, heating, cooling, and dehumidifica-
tion. The indoor plant is mounted in a building fur-
nished by the utility. It may have either the tank and
pressure section as two parts, or the tank and pressure
section mounted on a common base. The indoor plant
may be used when a building must be furnished for
other equipment, or when ambient temperature is very
high or low. For both outdoor and indoor plants, the
entire pumping plant, including tanks, pressure delivery
section, and instrumentation and controls, is usually Figure 9-11 Skid including pumps, relief ladders, and
pressure-monitoring instrumentation (courtesy of MAC
mounted on a single skid and transported by truck. Products).
Some exceptions to a single integrated unit on a skid are

if a high-capacity vertical tank is used or if the tank(s) is
installed remotely due to space restrictions. Figure 9-11
is a photograph of an indoor skid, and Figure 9-12 is a
photograph of an outdoor plant.
Plant Auxiliaries
The following items can be added to a pumping plant to
meet particular operating
or maintenance conditions:
• Moat or sump—to contain fluid from a leaking
pump or valve, or a leak from the tank
• Level switch—to sound an alarm when fluid is
present in the containment
• Fire detection/protection
• Air conditioning Figure 9-12 Outdoor pressurizing plant (courtesy of USi).
9-14
can be used to replace a disabled plant for a short time. be spaced along the cable to keep the pressure rise or
This unit can also be used to train operators. Figure 9- drop within the rating of the reservoir and the cable
13 shows a portable pumping plant. sheath. If this cannot be achieved, the fluid duct size of
the cable may have to be increased. Reservoirs have a
Modifications for SCFF Service limited capacity to feed fluid to a leaking cable. They
Pumping plants designed for SCFF cables require a are used on cable with a pressure range of 14-414 kPa
‘gas free’ design and improved cleanliness and tightness. (2-60 psi). Typical reservoir sizes are 38-151 l (10-
The nitrogen blanket over the tank in a pipe-type cable 40 gal), and they are typically spaced at 610-1219 m
plant is replaced by a vacuum system which maintains (2000-4000 ft) intervals.
vacuum over the dielectric fluid in the tank. The return-
ing fluid from the cable can enter the tank via spray Reservoir Types
nozzles, drip trays, etc., to further enhance the degasifi- The three types of reservoirs, which are illustrated in
cation process. Stainless-steel piping may be required Figure 9-14, are commonly referred to as follows:
for certain types of fluids, and special seals and pumps • Gravity feed type (CC) reservoir
are usually required. Care must be taken so that ade-
quate pump suction pressures are maintained under all • Balanced pressure type (DC) reservoir
tank fluid levels, flow rates, and temperatures. Test • Bladder type (AC) pressure reservoir
requirements for SCFF systems are usually more strin-
Gravity Feed (CC) Reservoirs
gent than for HPFF pressurizing systems. Sophisti-
These consist of a number of cells connected in parallel
cated flow-limiting systems are often installed to limit
by a manifold enclosed in a tank that is open to the
fluid loss while maintaining the integrity of a leaking
atmosphere and partially filled with idle dielectric fluid.
cable.
The cells are filled with degasified dielectric fluid, and
the cell manifold is connected through tubing to the
9.4.3 Reservoir Systems (SCFF Cables)
cable joint or a termination. As dielectric fluid is forced
Overview into the reservoir, the idle dielectric fluid surrounding
The low volumetric requirements and moderate pres- the cells rises and indicates the quantity of working
sures of low-pressure SCFF cable systems allow the use dielectric fluid by means of a glass sight gauge mounted
of passive “bladder” type reservoirs as pressurization on the tank. The Type CC reservoir discharges its rated
sources. Reservoirs have no moving parts in the dielec- quantity of fluid with a very small change in pressure.
tric fluid side and do not permit dissolved or free gas to This characteristic makes it especially applicable at
enter the fluid. They have limited fluid capacity and may points in the profile where fluid capacity is required, but
where pressure, in addition to normal head pressure, is
undesirable. This type of reservoir is very efficient.
Through static head, it exerts an almost constant pres-
sure on the cable because its working range is over the
flat part of the curve. The Type CC reservoir is generally
used at terminal ends where it feeds directly into the ter-
mination, because it must be mounted at a suitable ele-
vation to maintain positive pressure on the cable even
during transient conditions and yet not exceed the max-
imum allowable pressure on the cable. It can be used in
manholes and connected to a cable joint, provided the
cable profile drops away from the reservoir elevation.
Balanced Pressure (DC) Reservoirs
These are similar in construction to the gravity type just
described, except that the tank enclosing the cells is
filled with gas (CO2 or N2) and closed to the atmo-
sphere. Gas pressure can be impressed upon the outside
of the cells and thence upon the fluid, having the effect
of raising the reservoir to an elevation comparable to
the CC reservoir. The type DC reservoir generally is
Figure 9-13 Portable pumping plant (courtesy of USi). used where the fluid demand of an intermediate section
of single-conductor cable requires a pressure-volume
9-15
relationship that cannot be satisfied by the other types (15 psi) when the tank is full. The capacity of the AC
of reservoirs or to control pressures when there are sig- reservoir is determined by the size of the tank and the
nificant elevation changes. The reservoirs and any neces- number of cells that it contains. This type reservoir is
sary supplementary gas tanks may be located in the used most frequently for supplying intermediate dielec-
stop-joint manholes at both ends of an intermediate tric fluid sections of three-conductor fluid-filled cable
fluid section or can be located in auxiliary manholes lines having a relatively flat profile, since with three-con-
and fed to the middle of the section. ductor cable, the fluid channels are readily available at
every joint, and feeder-type joints are not required. It
Bladder Type (AC) Reservoirs
can be used also on single-conductor lines with flat pro-
These consist of a number of gas-filled sealed cells
files, where the feeder joint required at the point of fluid
enclosed in a tank containing degasified dielectric fluid.
feed to the cable does not appreciably increase the size
The tank is closed to the atmosphere, but is connected
of the manhole over that required for a normal joint.
through tubing to the cable at some convenient point,
The AC reservoir offers the advantages that individual
such as at a joint or termination. When fluid is forced
units can be spread uniformly over an fluid section,
into the reservoir, the gas in the cells is compressed, and
reducing pressure drop to a minimum; it can be located
the pressure rises in accordance with the characteristic
in available space in normal manholes, and is conve-
curve of this type reservoir. The gas in the cells, usually
niently monitored by pressure-type relays. Bladder-type
CO2, is typically adjusted to a pressure of 6.9 kPa (1 psi)
reservoirs are typically found on cable shipping reels.
when all usable fluid is out of the tank and to 103 kPa
Figure 9-14 Types of reservoirs.
9-16
Although bladder-type (AC) reservoirs are relatively Under normal load cycling conditions, the typical oper-
inefficient, because the full pressure range cannot be uti- ating pressure range for the nitrogen blanket is 13.8 kPa
lized, they are economical. (2 psig) to 103 kPa (15 psig). An alarm is usually pro-
vided for pressures in excess 103 kPa (15 psig). The solu-
The pressure-volume characteristics of the aforemen- bility of nitrogen (% by volume) in typical dielectric
tioned reservoir types are shown in Figure 9-15. fluids is between 10 and 12% at the lower pressure and
between 17 and 22% at the higher pressure. The dis-
9.4.4 Gas Cabinets (HPGF cables) solved gas, which diffuses into the fluid in the tank and
HPGF cables do not require a pumping plant. For this then into the cable system over a long period of time, is
type of cable system, a gas cabinet is usually provided to an average of the upper and lower solubility. Dissolved
hold nitrogen supply cylinders, spare nitrogen bottles gas analyses have substantiated this with typical N2 gas
for rapid change-out, pressure gauges, and pressure contents reported in the 10-14% range. Higher operat-
alarms and contacts to trip the cable upon low pressure. ing pressures and higher dissolved gas contents run the
Older systems provide cable pressure alarms, but with risk of out-gassing and ionization at higher pressures
newer systems, provision for low-gas-cylinder-pressure than ordinarily expected. On the other hand, in the
alarms should be considered. This can prevent the situa- interest of conserving tank volume, the nitrogen pres-
tion where a low-cable-pressure alarm is initiated, and sure ranges in excess of 103 kPa (15 psig) but below
then it is discovered that none of the available supply 207 kPa (30 psig) could be considered for short time
cylinders has sufficient gas. excursions due to emergency loading.
9.4.5 Computing Reservoir Volume The tank volume may be divided into three volumes: (1)
reserve fluid volume, (2) active fluid volume, and (3)
Pressurization Plant Tanks nitrogen gas space.
It is recommended that the tank size requirement be
determined via an explicit calculation, because the The reserve fluid (Vres) consists of:
needs and requirements for individual systems vary
greatly. Knowledge of the steel tank pressure capability
• The “dead fluid,” which is the static fluid below the
top of the suction pipe that cannot flow to the pump.
is required. A typical design and test pressure for a steel
tank is 414 kPa (60 psig), and a rupture disk of this • Emergency fluid required to feed a leak. The required
magnitude is affixed to the tank as required by code. emergency fluid volume should be based on the leak
rate time, which is the estimated time to find and
repair the leak. The leak rate could be assumed to be
the maximum output of the pump or the flow rate of
a flow-limiting device.
The active fluid (Vactive) consists of:

• The maximum expulsion, as computed in Equation
9-7.
• Oscillation fluid required for oscillation—if oscilla-
tion between tanks at either end of a circuit is used to
smooth temperatures.
The maximum nitrogen pressure can then be computed

for a given tank size Vtank by Equation 9-19.
Pmax :=
(Pmin+ Patm) ⋅(Vtank − Vres) − P
Vtank − ( Vres + Vactive) atm Pa 9-19
One strategy is to determine the tank size necessary to

limit the maximum pressure to 103 kPa (15 psig) under
“normal” loading conditions, and then determine the
pressure at emergency loading conditions to ensure that
Figure 9-15 Pressure-volume characteristics of it remains below 207 kPa (30 psig).
reservoirs.
9-17
Reservoirs 9.5.2 System Components

To determine the size of the reservoir—or more appro-
Pumps
priately, the quantity and placement of reservoirs—the
The primary component in any forced-cooled system
designer must first determine the pressure-volume
is the pump. The three major classifications of pumps
requirements of the cable system resulting from analy-
are rotary, piston, and centrifugal. Both rotary and pis-
ses, as described in Section 9.3. This will involve expul-
ton pumps are “positive” displacement pumps, with
sion calculations, as well as axial pressure distributions
the characteristic of a steady-flow output with large
due to flow demands, and total pressures due to the ele-
variations in differential pressure. This offers the
vation changes. The specific pressure-volume character-
advantage of positively developing the flow desired to
istic of the reservoir, which must be obtained from the
achieve the cable rating independent of the pressure
manufacturer, must then be coupled to the cable so that
losses in the cable system. On the other hand, pres-
the cable pressures remain in range under all loading
sure transients due to full-bore startups and shut-
and unloading conditions. If pressure limits are
downs need to be controlled.
exceeded, then additional reservoirs can be added.
Centrifugal pumps are very sensitive to differential pres-
9.5 FORCED-COOLING EQUIPMENT sure with flow output rapidly degrading as differential
pressures increase beyond optimum pressure-flow
9.5.1 Introduction design point. This disadvantage requires that the
The temperatures achieved by the various cable system designer have a high degree of confidence in the hydrau-
components are determined by the magnitude and dura- lic design calculations when applying a centrifugal
tion of the losses, the thermal impedances of the cable pump. In many cases, the designer does not have control
components and the surrounding paths to relevant heat over the hydraulic design, and last-minute construction
sinks, and the temperature of the heat sinks involved. issues can result in additional hydraulic impedances.
In self-cooled cables, all of the cable losses must flow One advantage of centrifugal pumps, on the other hand,
through the surrounding backfill and soil. The sur- is less abrupt pressure disturbances during the startup
rounding soil thermal resistance could be typically one- and shutdown periods. Piston pumps have not been
half of the total thermal resistance of the cable-soil or used for forced-cooling applications due to the high
more. The tendency for moisture to migrate away from maintenance and pulsations characteristic of this type
the cable/soil interface may further limit the current of pump.
capacity of the cable system.
Rotary Pumps
Forced cooling, which involves the mechanical propul- Rotary pumps fall into several subcategories: screw,
sion of the dielectric fluid along the axis of the cable sys- gear, and vane. The balanced screw type has been found
tem, removes heat in parallel with heat flow to the to be best suited for forced-cooling service on HPFF
surrounding soil, thereby improving the current-carry- cables, since its maintenance requirements are low. The
ing capability of the cable system. HPFF cable systems screw pump consists of two intermeshing helical rotors
are amenable to forced-cooling, because the fluid vol- contained in a housing. Dielectric fluid enters one end
ume inside the pipe presents an ideal circulation path of the rotors, where it is trapped by the pockets formed
and is in close contact with the cable. Forced-cooled by the threads. The dielectric fluid is moved along to the
extruded-dielectric and SCFF systems most often use discharge port, much like a nut on a threaded screw.
closely spaced buried parallel pipes with a circulated Clearance between the rotors is important. One type of
coolant that removes heat through conduction through screw pump incorporates a set of bearings and timing
the backfill soil. There are no known forced-cooled gears to maintain this minimum clearance. A second
extruded-dielectric or SCFF cables in North America. type utilizes a hydrodynamic film formed between
This section covers forced-cooling equipment for HPFF rotors to maintain the clearance. The latter of the two is
systems only. Equations for determining cooling equip- used almost universally for forced-cooling service due to
ment heat load and power requirements can be found in its low noise characteristics and reliability. Care must be
Chapter 11 Ampacity, Section 11.4.7. taken, however, to operate within specific guidelines for
fluid viscosities, differential pressures, suction pressure,
9-18
and speed so as to ensure that the hydrodynamic film is tem hydraulic instability. This feature can accommo-
maintained at all times. Figure 9-16 is a photograph of date complex forced-cooled schemes or subsequent
rotary helical screw pumps with a capacity of approxi- design changes.
mately 1136 l/min (300 gpm).
Methods to Reject the Heat
Centrifugal Pumps Heat removed from the dielectric oil or a coolant must
Centrifugal pumps, unlike positive displacement pumps, ultimately be rejected to the environment. A few basic
exhibit a substantial reduction in flow as differential methods are discussed here.
pressure increases. In the centrifugal pump, pressure is
Air Heat Exchangers
developed by the action of centrifugal force. Oil enters
the hub of the propeller and is thrown radially to the Air-cooling is probably the most direct method of
periphery of the housing. Due to their simplicity, cen- rejecting heat. Air-cooled heat exchangers can be used
trifugal pumps have been used in several forced-cooled to directly remove heat from the dielectric fluid, or they
installations. As an advantage, they provide a relatively can be utilized to condense refrigerant if a refrigeration
“soft”-start characteristic. In addition, favorable tem- system is used for lower dielectric fluid temperature
perature-flow characteristics permit the use of centrifu- applications. The achievable temperatures are limited by
gal pumps, in some cases, without specialized flow- the maximum design air ambient temperature. Air heat
control devices; as temperature increases, both viscosity exchangers may consist of a series of finned tubes, usu-
and pressure drop decrease, allowing flow to increase. ally several hundred, suspended between box-type head-
Centrifugal pumps, however, cannot easily accommo- ers. The flow of air over the finned tubing of the air heat
date additional pressure differential. exchanger can be accomplished in a number of ways.
Natural convection creates motion of the air over the
Pump Drives finned tubing as a result of differences in air density;
The major advance in pump technology has been the however, most air heat exchangers use forced convection
advent of variable-speed drive mechanisms. The most for improved heat transfer coefficients. Fans can be
popular is a Variable Frequency Drive (VFD), which is employed in either a forced-draft or induced-draft
a system for controlling the rotational speed of an ac arrangement. In a forced-draft arrangement, fans blow
electric motor by controlling the frequency of the elec- the air over the finned tubing. Air-inlet noise levels tend
trical power supplied to the motor. VFDs usually to be higher than exit noise levels for forced-draft heat
employ controllers, which are easily interfaced to the exchangers. In induced-draft arrangements, the fans
process control system, or PLC, which allows control draw air over the finned tubing, which tends to create a
of startup speed, rate of increase in speed, target speed, more equal distribution of airflow over the face area,
and other variables. This permits low startup speeds, thus improving efficiency. The induced-draft design cre-
and limits in-rush currents and pressure transients ates higher air-exit velocities than the forced-draft
occurring in the cable system. Slow predetermined design, reducing the possibility of recirculation of hot
ramp-ups to final speed also reduce the chances of sys- air. Air-exit noise levels are generally higher than air-
inlet noise levels on induced-draft heat exchangers. Fig-
ure 9-17 shows induced-draft heat exchangers for direct
dielectric fluid cooling.
Figure 9-16 Positive displacement rotary helical screw

pumps (courtesy of USi). Figure 9-17 Induced-draft air-dielectric fluid heat
exchangers (courtesy of USi).
9-19
Cooling Towers pended inside a cylindrical vessel (the shell) between

Water can also be used to cool the dielectric fluid or headers (tube sheets). For forced-cooling applications,
condense refrigerants. One method of rejecting heat the coolant is usually in the tubes and dielectric fluid in
from the water is by use of a cooling tower, where water the shell. This exposes the dielectric fluid, which usually
is sprayed through an area of moving ambient air, which has poorer heat transfer characteristics than the cool-
evaporates a portion of the water. The temperatures ant, to a larger surface area and better flow characteris-
achievable by cooling towers are lower than air heat tic due to the placement of baffles. Also, the dielectric
exchangers because they are limited by the maximum fluid, being considerably more viscous than the coolant,
design wet-bulb temperature, which is usually lower creates less pressure drop in the shell as opposed to the
than the design ambient air temperature. Because some tubes. The most common coolant in the tubes is water
of the water is lost by evaporation, makeup water must or a water/antifreeze solution. A shell and tube evapora-
be provided to replenish the supply. The makeup water tor in a mechanical refrigeration system uses a refriger-
required is approximately 1% of the water flow-rate. ant as the coolant. In this case, a change of phase occurs
Cooling towers, like air-heat exchangers, can operate in the tubes as the refrigerant turns from liquid to gas as
using either natural convection or forced convection. heat is absorbed. The primary concern with shell and
Fan arrangements are normally induced-draft to keep tube heat exchanges is a leak occurring at the interface
the fan motors out of the water spray. Compared to air- of the tubing and the tube sheet. This could permit con-
heat exchangers of equivalent capacity, cooling towers tamination of the dielectric fluid by the coolant when
are more compact, less expensive, and less noisy. Their the pressure of the dielectric fluid is low for mainte-
major undesirable aspect is fogging. This phenomenon nance or otherwise. A double-tube-sheet design, illus-
occurs on cool days, when ambient air does not have the trated in Figure 9-18, is, therefore, recommended for
capacity to absorb the moist discharge air. This may be HPFF forced-cooling service.
especially problematic near high-voltage components in
substations. Since a cooling tower is an open system, the The double-tube-sheet design provides separation of the
water supply can become corrosive as water evaporates dielectric fluid from the coolant with an intermediate
and the mineral concentration increases. This requires chamber that allows for visual or automatic monitoring
substantial maintenance or water treatment. of the presence of fluid in the chamber. Although a leak
directly through the tube is very unlikely, double-tube
Direct Water Cooling
designs could also be employed. In addition to monitor-
If a water supply such as a reservoir, well, river, or cool-
ing of the intermediate chamber, pressure decay tests are
ing pond is available, it can be used directly to dissipate
recommended as routine maintenance to detect the
heat. The primary advantage of direct water cooling is
presence of a leak.
that the temperature of the supply is usually cooler than
ambient air, permitting lower dielectric-fluid discharge
temperatures. These supplies are likely to possess corro-
sive chemicals and minerals, which may impair the per-
formance of the heat exchanger. A water treatment
facility upstream of the heat exchanger is recommended
in these cases. A second problem is environmental
impact. Cable overheating requiring forced-cooling is
likely to occur during hot weather when water resources
are usually low. Another environmental consideration is
the addition of heat to the water source, which may
affect plant or animal life.
Heat Exchangers
A secondary coolant loop is often employed to dissi-
pate heat to the atmosphere such as water in a cooling
tower arrangement or an antifreeze coolant to an air-
coolant heat exchanger. In these cases a liquid-liquid
heat exchanger is required to remove heat from the
dielectric fluid.
Shell and Tube
The most common liquid-liquid heat exchanger is the Figure 9-18 Double-tube-sheet shell and tube heat
shell and tube type, where a bundle of tubes is sus- exchanger.
9-20
Compact Heat Exchangers should be based not only on efficiency and reliability,
Compact heat exchangers are characterized by a high but also on ease of maintenance.
surface area per unit volume, which can result in a
higher efficiency than conventional heat exchangers, in a 9.5.3 Overall Cooling Plant Design
significantly smaller volume. The most common type of
compact heat exchangers is referred to as a plate-and- System Arrangements
frame heat exchanger. These consist of a series of Forced-cooled systems must consist of a pump and a
formed plates with corrugated surfaces and integrated method to reject the heat. Various arrangements of
channels, which are bolted together with intermediate pumps, heat exchangers, coolants, and mechanical
gaskets and suspended within a frame. These have not refrigeration systems are possible. Cooling systems usu-
been used for HPFF cable forced-cooling applications, ally fall into two broad categories: air-cooled systems
but have been proposed for restricted space applications and water-cooled systems. An example of an air-cooled
or for improving the capacity of existing systems. The system is shown in Figure 9-19.
gasket material for HPFF service must be of a type com-
patible with the dielectric fluid. Hot dielectric fluid from the cable system enters the pre-
cooler (an air-cooled heat exchanger) that reduces the
Mechanical Refrigeration dielectric fluid to within 5-10C° above ambient air tem-
Mechanical refrigeration can be employed when dielec- perature. The dielectric fluid temperature is further
tric fluid temperatures below those that can be achieved reduced in the evaporator (a refrigerant-dielectric fluid
with ambient systems (air, cooling towers, or water sup- shell and tube heat exchanger) to temperatures that are
ply) are required. A refrigerant vapor-compression usually as low as ambient earth. The heat removed by
cycle is usually employed, which absorbs heat at low the refrigerant in the evaporator is rejected to the air in
temperature and rejects heat at a temperature higher the condenser, which consists of another air-cooled heat
than ambient. The phases of a vapor compression cycle exchanger. The two cooling stages are followed by the
consist of: (1) evaporation phase, where the refrigerant pump, which circulates the dielectric fluid at the desired
absorbs heat from the low-temperature cooling fluid flow rate and raises its pressure. By placing the pump
and turns from liquid to gas; (2) the compression phase, after the cooling equipment, the pressure rating of the
where the pressure and temperature of the refrigerant cooling components can be lowered, and the pump suc-
are elevated via a compressor; (3) the condensation tion pressures are lowered, exposing the mechanical
phase, where heat from the refrigerant is liberated to the rotating seals to lower pressures. A disadvantage is that
high-temperature fluid or ambient air and turns from the pump heat input raises the dielectric fluid tempera-
gas to liquid; and (4) the expansion phase, which takes ture somewhat. In some cases, placing the pump in front
the high-temperature liquid refrigerant and lowers its of the cooling components can be desirable, because the
pressure in an expansion device, permitting the evapo- heat imparted by the pump is easily dissipated by the
ration of the refrigerant at low temperature. heat exchangers. However, higher pump suction pres-
sures and exposure of the cooling components to high
The essential components of a mechanical refrigeration pressures must be dealt with. The most common
system are the evaporator, compressor, condenser and arrangement of forced-cooled systems is simply an air-
expansion valve. The compressor is the prime moving cooled heat exchanger followed by a pump.
part of a mechanical refrigeration system. Compressors,
like pumps, can be positive displacement (rotary,
piston) or centrifugal designs. Forced-cooling applica-
tions, to date, have used positive displacement compres-
sors, with the rotary screw design being the most
p o p u l a r. T h e ap p l i c at i o n u s u a l ly re q u i re s t h e
development of substantial differential pressures for
which rotary screw pumps are well suited. The
maintenance requirements of mechanical refrigeration-
based systems are higher than ambient-based systems.
Careful selection of a mechanical refrigeration system
Figure 9-19 Air-cooled refrigeration system.
9-21
An example of a water-cooled system is shown in Figure refrigeration systems in the foreground and the control
9-20. The water-cooled system is similar to the air- unit in the background.
cooled system, except that heat is removed from the
dielectric fluid or refrigerant by water or a water-based The mechanical refrigeration system cools a propylene
coolant. An additional pump, the coolant or water glycol coolant system that is connected with insulated
pump, is employed to pump from the cool-water source, piping in a cooling loop to the control unit. The control
which may be a cooling tower or one of the other unit houses the dielectric fluid circulation pumps, the
described sources. Coolant pumps are usually of the coolant circulation pumps, and primary and backup
centrifugal variety, which is ideally suited for this appli- double-tube-sheet shell and tube heat exchangers, as
cation. shown in Figure 9-22.
Environmental Impact
The advantages of a two-stage system are: (1) the pre-
cooling stage requires substantially less energy to Coolants and Refrigerants
remove heat, and (2) either stage can serve as a backup • Secondary coolants are usually water-based with the
in case of failure of one stage. Some recent applications
addition of antifreeze agent, corrosion inhibitors,
have eliminated the pre-cooler because design ambient
and stabilizers. The preferred antifreeze agent is pro-
air temperatures were high and no energy benefit could
pylene glycol, which is environmentally more benign
be realized. Designing a completely redundant mechani-
as compared to ethylene glycol. The appropriate mix-
cal refrigeration system compensated for the loss of the
ture for the regional water supply and application is
second stage as a backup in this case. Figure 9-21 is a
usually left to a water treatment expert. The chemical
photograph of a water-based coolant mechanical refrig-
status of the secondary coolant loop needs to be
eration system with primary and backup mechanical
determined on a regular basis with additional treat-
ment as required.
• Refrigerants can be classified as chlorofluorocarbons
(CFCs), hydrochlorofluorocarbons (HCFCs), and
hydrofluorocarbons (HFCs). CFCs are considered
the most harmful in depleting the ozone layer and
magnifying the greenhouse effect because they con-
tain chlorine, which is highly reactive with ozone.
They are very stable chemically, and therefore persist
in the atmosphere for long periods of time. R12 falls
into the category of CFCs. Although they contain
Figure 9-20 Water-cooled refrigeration system.
Figure 9-21 Mechanical refrigeration cooling plant Figure 9-22 Primary and backup shell and tube heat
(courtesy of USi). exchangers (courtesy of USi).
9-22
chlorine, HCFCs are less stable in the atmosphere Noise

and, therefore, have considerably less impact. R22 Specifying noise levels in the early design stages is the
falls in the classification of HCFCs. HFCs contain best procedure to control noise levels. The addition of
no chlorine and are generally considered to have no attenuation devices in the later stages of design or as an
impact on the environment. R134a is an HFC. Fed- after-thought is seldom effective. The first step in a good
eral Clean Air Act Section 606 requires the phase-out noise design is to provide for high efficiency in the fun-
of both R12 and R22. CFCs (R12) were taken out of damental design of components. High-efficiency fans
production as of January 1, 1996. R22, an HCFC, deliver a given air flow with less energy input and conse-
can still be used, but is to be phased out for new quently less noise. The same is true for all other moving
equipment by January 1, 2010 and completely phased components. The second most important factor is to
out by 2020. Although Section 606 of the Clean Air “isolate” noise. Moving or vibrating machinery should
Act allows R22 to be obtained and used for the next be isolated from structures that can act as amplifiers.
10 to 20 years, there are substantial penalties for leak- The third step is to provide sound-absorbing materials
age or venting of R22. R134a, R407c, and R410a are nearby noisy components. Reflected noise is additive
all replacement refrigerants proposed as candidates and can substantially increase the inherent noise level of
for new systems. Although R134a is the preferred a given device. Finally, treatment with attenuation
refrigerant, previous systems designed for R22 may devices such as walls and air intake/outlet baffles can be
suffer 40-60% losses in capacity. R407c, on the other applied, provided they do not restrict airflow. Given the
hand is a thermodynamic equivalent of R22, and is heat load and land usage, Table 9-3 can be used to
considered a ‘drop in’ for R22. It is not a pure refrig- approximate noise levels. A weighted noise level of
erant, but a zeotropic “blend” of refrigerants, mean- 55 dBA may be considered a daytime residential noise
ing that it behaves as several refrigerants, not as one. level, and 65 dBA a daytime light industrial noise level.
The major constituent of R407c is R134a. Leakage is
a problem, because the constituents leak at different 9.6 SYSTEM DESIGN CONSIDERATIONS
rates, thus upsetting the ratio of the blend. R407c is a
viable alternative while manufacturers gear up to pro- 9.6.1 Pressurization Plants
duce R134a-specific equipment. R410a is a 50/50
blend of R-32 and R-125 and is a higher efficiency General
R22 replacement than R407c. Although not as The availability and reliability of pressurization systems
benign as R134a, R410a has very low ozone deple- are paramount to preserving HPFF cable systems. To
tion or global warming potential. R410a is not a improve reliability, multiple pump/relief ladders within a
‘drop in’ replacement refrigerant for R22 and plant and multiple pressurization plants or reservoirs at
requires re-design of equipment. Forced-cooling separate locations on a cable system should be consid-
refrigeration equipment installed in the 1970s and ered. If one plant or pump/relief system (ladder) fails,
1980s employed R12 and R22 refrigerants. These sys- then a system must be considered to revert to the
tems have been retired. A recent application (year backup plant or ladder. A passive system, such as set-
2000) has used R407C. ting the controls of the backup plant or ladder just out-
side the limits of the primary system, is always desirable.
Energy Usage
The energy usage of cooling systems can be computed in Good system design practice, when multiple cables are
detail utilizing the Designer’s Handbook (Purnhagen pressurized, should consider hydraulic isolation
1984). For estimating purposes, the coefficient of perfor- between cable circuits. If one of the circuits develops a
mance (COP), which is the ratio of heat energy leak or the cable fails electrically, a means must be con-
extracted to the energy used, is approximately 4 for two-
stage systems. A pre-cooler alone can have a COP of 5
or 6, and a mechanical-refrigeration-only system may Table 9-3 Land usage per unit heat load based on Noise
have a COP of 3. These numbers appear to be low com- Level
pared to residential air conditioning units, etc. because
they include the energy to pump the dielectric fluid Land Usage (m2/kW)
around the cooling loop. Noise Level of Water Cooled
Air Cooled System
Equipment (dBA) System
Land Usage 65 0.2 0.4
The land usage of cooling systems can be computed in 55 0.6 1.2
detail utilizing the Designer’s Handbook (Purnhagen
1984). Table 9-3 can be used to approximate land usage.
9-23
sidered to prevent the loss of pressure or contaminated be required on long cables because of pressure drop in
fluid from propagating to the “good” circuit. the pipe.
Radial Cable Systems
The need for multiple plants or other backup options,
With a radial cable system, a single pumping plant at
such as remote auto crossover manifolds, etc. should
the hub is not recommended, because all circuits are
consider the impact of loss of cables. Only if the cable is
not critical to the system operation, where other means
of transferring the load are readily available, or where
loss of the cable does not result in high-energy replace-
ment costs, should a stand-alone plant be acceptable.
Otherwise multiple plants at separate locations should
be considered.
In addition, operator training, on a regular basis, is a

key aspect of ensuring reliable operation of the plant
and the cable system itself. The overall system design—
which includes the placement and number of pumping
plants, the assignment of total tank volume, and a
control scheme— depends on many factors, some of
which are:
• Length of cable
• Importance of keeping the line in service
• Contingency considerations
• Hydraulic considerations
- Elevation profile Figure 9-23 Hydraulic schematic—single-pipe system.
- Circulation or oscillation requirements
- Reduced pressure and flow requirements
• Service response-time considerations
• Cost
Some Basic Guidelines for System Arrangements
Single-Pipe Systems
Figure 9-23 shows a schematic of a single-pipe system.
A single cable pipe between two stations can be pressur-
ized by one pumping plant. Although this arrangement
has the advantage of lowest cost, its disadvantages are:
(1) there is no redundancy, (2) pressure cannot be sup-
plied to the remote side of a freeze, and (3) thermal
smoothing via oscillation is not possible. The preferred
system for a single-pipe is two single-pump pressurizing
plants, one at each end.
Dual Cable Pipe Systems
Figure 9-24 shows a schematic of a dual-pipe system.
Dual cable pipes between the same stations can be
pressurized by one pumping plant, with a flow-limiting
crossover at the far end. The plant can furnish fluid to
both sides of a freeze or leak, and fluid temperature can
be averaged by circulation. Again there is no
redundancy. A better, although more expensive, system
would have a pumping plant at each end, and this may Figure 9-24 Hydraulic schematic—dual-pipe system.
9-24
jeopardized by a major failure of the primary plant. tanks using 100% of the required reserve but only 50%
Pumping plants located at the remote ends of each cir- of the required expulsion fluid. This results in total tank
cuit are recommended, or a hybrid design, as shown in volumes of 125-150% of the theoretical requirement for
Figure 9-25. typical systems.
Assignment of Tank Volumes Thermal Mitigation Techniques
The total required tank volume could be computed The movement of the dielectric fluid via a slow circula-
using the equations in previous sections of this chap- tion arrangement or oscillation can be an effective
ter. If one plant is used on the system, then that plant’s means to mitigate a short hot-spot 91-305 m (300-
tank volume should be 100% of the volume computed. 1000 ft). The improvement in rating is usually due to
If more than one plant is associated with the cable sys- allowing the highest permissible conductor temperature
tem, then strategies to conserve fluid volume can be because the AEIC (AEIC 1996) specifications apply a
considered. If, for instance, two plants are installed, one 10°C penalty if there is not adequate knowledge of the
at each end of the cable system, then each plant could be thermal environment. Alternately, the equations in Sec-
100% of the required fluid volume, which provides 100% tion 11.7 of Chapter 11 can be utilized to explicitly com-
redundancy, or a total tank volume of 200% of the theo- pute the temperature rise due to circulation.
retical requirement. The sum of the individual plants
Slow Circulation
tank volumes can be reduced if the contingencies are
When a “go” and “return” exist, forming a loop, slow
carefully analyzed, but should always exceed 100% of
circulation can be introduced. The pipes could consist
the required volume, because it is virtually impossible to
of two cable pipes or a cable pipe and a fluid line, or sev-
synchronize the operation of multiple plants so that the
eral cable pipes and fluid lines. One technique is to use
tank levels of each plant are equal. Furthermore, if one
one “ladder” as the go and the second ladder as the
plant is inactive, the second plant not only needs to
return. The pressure pump on the go is set for continu-
accommodate the active fluid, but it may be required to
ous operation, and its pressure is controlled by the
service a leak. One strategy is to size the individual
pump relief valve. The cable relief on the return ladder is
set to operate at a lower pressure than the pump relief,
allowing the fluid to return to the tank, where it is
drawn by the pump, thus forming a continuous loop.
The same concept can be employed using a remote
pumping plant on parallel cables, except that the corre-
sponding cable pumps and remote cable relief valves are
set as a go and a return. This arrangement can result in
preferential fluid migration from one tank to another,
so fluid must be redistributed from time to time.
Another limitation of the “go–return” ladder concept is

that flow rates are restricted to the flow of the pressure
pump. When higher flow rates are desired, or when very
large systems are involved, it is desirable to equip the
pumping plant with a dedicated circulation pump. This
is usually incorporated between ladders and on the
cable side of the control valves. The disadvantage of this
scheme is that only one pumping plant pressure control
point is required, and it must, therefore, be able to
Figure 9-25 Hydraulic schematic—radial cable system.
accept the full active fluid volume. Multiple tanks,
therefore, should each be sized to accommodate nearly
9-25
100% of the fluid requirement. Figure 9-26 illustrates valves, either manually operated or with a motor
the common circulation schemes. actuator that can be brought into service during a leak.
A third strategy for controlling fluid spills for massive
Oscillation
leaks is a flow limiter, which controls the maximum flow
Oscillation, sometimes referred to as “shuttling,” can be
delivered, independent of pressure. These techniques
used if a loop cannot be formed (as in a single-cable
can be employed individually or in combination with
pipe installation). Two pressurization plants, one at
one another.
either end of the circuit, are required. A portion of the
tank volume, which must be incorporated in the
9.6.2 Forced-Cooled Systems
“active” fluid, must be provided for oscillation. Fluid is
then pumped from one tank to another. The process Introduction
stops when the fluid level of the active plant falls to a Many schemes are available for the employment of
predetermined set point. The process can be controlled forced cooling. The basic elements of a forced-cooled
easily with communication between plants. It is also system are: the cable pipe(s); a supply or return pipe;
possible to control an oscillation without interplant turnaround point (crossover); exit and entry points; and
communication. Oscillation is not as efficient as circula- the pump and the heat exchanger. The “cooling loop”
tion, because the fluid passes over the hot spot succes- refers to the smallest system composed of these ele-
sively. It is recommended that the oscillation slug be ments. Subsequently, most calculations refer to the
three to four times the length of the hot spot to cool it “cooling loop” as the basic entity.
effectively. Oscillation systems are also prone to tank-
tank fluid migration. Basic Configurations
There are many possibilities for integrating the basic ele-
Addressing Cable Leaks ments into cooling loops. Some examples are provided
An important function of the pressurization system is to in this section. Cable pipes are often used as the supply
maintain the integrity of the cable system, both the and return paths for the fluid. When two or more cir-
leaky system and other “good” systems, during the leak cuits are installed in parallel, a simple arrangement is
for a period of time until the leak can be repaired. If the the two-pipe system of Figure 9-27.
pumping plant is servicing more than one cable system,
then tanks are often equipped with partitions to a height The cooling plant is located near one set of terminals,
equivalent to the reserve fluid for each cable. This with the turnaround point located at the other end of
prevents the leaky cable from fully draining the fluid the circuit. All exit and entry points are shown at the tri-
that may be necessary for normal operation of the furcators, which is common when natural cooling of the
“good” cable. A reduced pressure operation capability is riser sections is sufficient for the desired rating. The
also often employed to reduce cable system pressures major advantage of this arrangement is elimination of
and leak rates, provided that the cable is taken out of the cost of an additional fluid pipe for a supply or
service. This usually involves an additional set of relief return. A major disadvantage is a lack of redundancy,
which will result in the loss of forced-cooling in the
event of failure of one cable circuit. Since forced-cooled
system capabilities are most often limited by pressure
drop, it may be necessary to divide the circuit into seg-
ments. A system employing two cooling segments and a
Figure 9-27 Basic cooling loop using two cable pipes.
Figure 9-26 Slow circulation schemes.
9-26
common fluid supply pipe to form four cooling loops is forced-cooled systems. The designer has some control
shown in Figure 9-28. over pressure drop issues with the following items
commonly addressed:
Two levels of redundancy are provided by this arrange- • Lower viscosity dielectric fluid (only effective in lami-
ment. If one cable circuit should fail, the cooling plants
nar flow)
are able to direct their full capacity toward the remain-
ing circuit after the failed circuit has been isolated by • Higher pressure rated terminations (mainly effective
appropriate valves operations. Secondly, if one cooling if termination is exposed to pump outlet pressure)
plant should fail or lose power, the remaining plant can • Larger diameter piping (cost and total fluid volume
be used to force-cool the entire system, but generally at issues)
reduced capacity. The flow path of Figure 9-28 shows • Lower temperatures (reduces flow rate but cost of
cooled fluid delivered to the cable pipe after traveling
additional cooling may be prohibitive)
the length of the section through the supply pipe. This
flow direction can be advantageous, since the high deliv-
The designer must also consider the placement and
ery pressure at the plant outlet is reduced by friction
operation of the pressurization system with respect to
losses en route to the terminals. Flow in the reverse
the forced-cooled equipment. Pressure plants should
direction, however, may be preferable. If pressure drop is
ordinarily be connected to the cable system near the ter-
not substantial, heat transfer from the return pipe to the
minations. The pressurization plants can be set at their
earth can be maximized by employing this flow direc-
normal operating pressures if a fluid supply line is used.
tion, and in some cases, the thermal requirements of the
If a fluid return line is used, or if the cooling loop con-
cooling station can be greatly reduced. When the rating
sists of two cable pipes, then the terminations will be
exceeds the riser/termination rating fluid can exit (or
subjected to high circulation pump outlet pressures. The
enter) chambers mounted on the risers, just under the
pressure plants must be then set to high pressures when
terminations, which improves the ratings of the risers
the circulation is in effect. The designer must consider
and terminations as well. Longer cable circuits may
the condition, however, when the circulation is off. The
require intermediate (between terminals) cooling station
pressures all along the system will be elevated, which
locations to achieve design levels. Environmental impact
can exceed maximum design pressures when there are
and power supply to the intermediate location are prime
substantial elevation differences. A dual-pressure range
considerations. For added reliability, it is sometimes
pumping plant, which is integrated with the cooling
desirable to hydraulically isolate adjacent cables so that
plant, may be required in this case. Multiple loops
an electrical or hydraulic failure on one circuit cannot
require further considerations, including the pressure
affect the performance of the remaining circuit(s). The
ramifications when one cooling system of a multiple
system shown in Figure 9-29 illustrates the employment
loop fails. Pressures tend to drive upwards or down-
of intermediate locations and hydraulically isolated
wards relatively rapidly with this imbalance. A control
cable circuits.
sequence, which involves rapidly shutting down all cool-
System Design Considerations ing units, is usually required.
Forced-cooling equipment must be integrated into the
cable system while remaining within the pressure and Transient thermal and hydraulic conditions can occur
thermal limitations of the cable system. The addi- under normal and emergency operation of forced-
tional pressure required to overcome frictional losses is cooled systems, which may damage the cable system.
the major consideration that the designer faces. In The fluid in a cooling loop does not accelerate instantly
fact, the pressure drop in the cooling loop limits most
Figure 9-29 Multiple-loop system with an intermediate

Figure 9-28 Two-segment, four-loop arrangement with location and complete hydraulic isolation.
common supply pipe.
9-27
when the circulation pump starts. As a result, cavitation 9.7 WORKED EXAMPLE
at the pump inlet can momentarily occur, causing low The following worked example is based on the high-
enough pressures to activate alarms, resulting in an pressure fluid-filled Worked Example #4 of Chapter 11,
automatic pump shutdown, or in severe cases, out- Ampacity. The Chapter 11 worked example provides the
gassing. To prevent this, the system should be designed required temperatures for the normal operating condi-
to accelerate slowly using either a variable-speed tions, and the methodologies in Chapter 11 can be used
controller or a flow bypass around the pump with an to compute the emergency temperatures. These temper-
automatic valve, which slowly closes over a period of 5- atures are provided in Table 9-4 below. This worked
10 minutes. A pressure transient will also occur due to a example then computes the fluid volumetric expulsion
rapid pump shutdown or pump failure. On normal and demand using Equations 9-14 and then computes
shutdown, a means to slowly reduce the flow should be tank size using Equation 9-18.
incorporated. In a multiple-loop configuration, startup
and shutdown sequencing is important. All pumps Cable Circuit Description
could be started simultaneously, but this may put a large
Circuit length: 49213 ft., 15,000 m
demand on the pressurization system due to pressure
Number of Circuits: 2
redistribution. Another solution is to sequence the
Conductor: 2500 kcmil, 1267 mm2
pumps, starting with the cooling plant closest to the
pressurization plant in control, and ramp up the system Insulation: Laminated paper polypropylene, 0.600 in.,
to an intermediate pressure and wait for stability. After 0.015 m
achieving stability, the pumps are ramped up to their Skid Wires: 2 - SS 0.100 in. x 0.200 in. x 3.00 in. lay
final speed in the same sequence. The shutdown Fluid: Low Viscosity Alkylbenzene
sequence should be the reverse of the startup sequence.
Provision must also be mad e for an unplanned Since there is no additional fluid movement due to oscil-
shutdown due to an emergency trip of a pump. Cooling lation, the active fluid can be set equal to the expulsion
should also be introduced slowly, since the initial fluid.
cooling thermal transient can be a much greater burden Vactive = VHPFFxp = 5,051 gal/ckt
on the pressurization system than the fluid demand due
to loading. Assuming a 10 gal/hr leak that must be supplied for 10
days and 600 gal.dead fluid, then:
Vreserve = 10 gal/hr x 240 hr + 600 gal = 3000 gal/ckt
Table 9-4 Hydraulic Worked Example

Conductor Insulation Fluid Pipe Reference
Base Material Copper Paper LVAB Steel Chap. 11
Dn 1.824 in 3.044in. 8.125in. na Chap. 11
Dm - - - 8.125 in Chap. 11
N 3 3 1 1
Fn 0.84 0.5 1 1
Vn-base material 16,834 gal 17,887 gal 76,737 gal - 9-3
Vn-oil 3,206 gal 17,887 gal - - 9-3
Vm - - - 132,552 gal 9-4
βn–base mat’l (1/°C) 0.000039 0.00028 0.00081 0.000027 Table 9-1
Tmin 10°C 10°C 10°C 10°C Chap. 11
Tmax (Norm) 85°C 78.8 68.9°C 65°C Chap. 11
∆Vtn (Norm) 244 gal 1,341 gal 3,662 gal 197 gal 9-1
VHPFFxp (Norm) 5,051 gal 9-8
Tmax (Emerg) 105°C 94.1° C 85.6° C 82.9° C
∆Vtn (Emerg) 309 gal 1,640 gal 4,699 gal 261 gal 9-1
VHPFFxp (Emerg) 6,387 gal 9-8
9-28
Using Equation 9-19 and setting the maximum nitrogen The demand must then be computed using Equation 9-
pressure to 15 psi (maximum recommended pressure for 15 and selecting the demand parameter ahpff from Fig-
normal operation), then: ure 9-1. Given the maximum per-unit-length demand
Pmax = 15 psig (103.4 kPa) due to cable unloading from Figure 9-1,
Patm = 14.7 psi (101.4 kPa) ahpff = 1.42 cm3/min/meter = 0.114 gal/min/kft
Pmin = 2.0 psig (13.8 kPa) Qhpff = ahpff · L = 0.114 gal/min/kft· 49.213 kft
Vtank = 14,539 gal (choose a 15,000 gal tank/ckt) = 5.61 gpm.
Therefore, two 15,000 gal tanks are required for the two Therefore, the pressure pump for each circuit should be
circuit system. For reliability purposes, the tanks should approximately 1.2 x 5.6 gpm or ~ 7 gpm.
be located at remote ends of the system.
REFERENCES
The tank, under normal conditions, is apportioned as
AEIC. 1996. Specification CG1-96. Third Edition. April
follows:
1996. Guide for Application of AEIC Maximum Insula-
Vres = 3,000 gal tion Temperatures at the Conductor for Impregnated-
Vactive = 5,051 gal Paper-Insulated Cables.
VN2 = 15,000 - 3,000 - 5,051 = 6,949 gal
AIEE. 1955. Oil Flow and Pressure Calculations for
The tank volume and pressure must then be checked for Pipe-Type Cable Systems. AIEE Committee Report.
the emergency condition—i.e., 105ºC conductor tem- Transactions of AIEE. Part 111. Vol. 74.
perature and expulsion equal to 6,387 gal. Using Equa-
tion 9-19, Buller, F. H. et al. 1956. Oil Flow and Pressure Calcula-
PTankEmerg = 21.0 psig tions for Self-Contained Oil-Filled Cable Systems.
This falls well below the maximum recommended pres- Engelhardt, J. S., and D. W. Purnhagen. 1991. Dynamic
sure under emergency transients of 30 psig. The addi- Rating and Underground Monitoring System, Feasibility
tional 1,336 gal (i.e., V H P F Fx p (Emerg) – V H P FF x p Study. Underground Systems, Inc., July. EPRI Report
(Norm)) of emergency expulsion fluid is easily accom- No. EL-7341.
modated by the available nitrogen volume.
Lanfranconi, G. M. et al. 1974. “Self-Contained Oil-
The tank, under emergency conditions, is apportioned Filled Cables for High Power Transmission in the 750 –
as follows: 1200 kV Range.” IEEE Transaction Paper T 74 202-8.
Vres = 3,000 gal
Purnhagen, D. W. 1984. Designer’s Handbook for Forced
Vactive = 6,387 gal
Cooled High-Pressure Oil-Filled Pipe-Type Cable Sys-
VN2 = 15,000 – 3,000 – 6,387 = 5,613 gal
tems. July. EPRI report no. EL-3624.
9-29
CHAPTER 10 Grounding and Cathodic

Protection
Authors: Dennis E. Johnson, POWER Engineers, Inc.
Adriano Santini, Consolidated Edison
Reviewer: John S. Rector, Black & Veatch
Designing the cable system grounding is essential to the safe and successful operation of
any cable system.
This chapter is divided into two parts: Grounding of single-core cable systems, and the
cathodic protection of pipe-type cable systems. The section on grounding of single-core
cables addresses basic design principles, various bonding methods, selection of compo-
nents, and basic design calculations. The section on cathodic protection addresses basic
design principles, various cathodic protection systems, and selection and design of a
cathodic protection system.
Related information may be found in other chapters. Chapters 4, 5, and 6 provide details
on the cable design and construction of dielectric, pipe-type, and SCFF cables. Chapter 8
covers cable system accessory design.
Dennis Johnson is a Senior Project Engineer with POWER Engineers,

Inc. (POWER). He received a BSEE degree from Brigham Young Uni-
versity (1985). He joined Arizona Public Service in 1986, where he ini-
tially was responsible for distribution system design. He later worked
for Black & Veatch from 1988 to 2001 as a design and project engineer
on numerous substation, overhead, and underground transmission
projects at voltages ranging from 69 to 345 kV. Mr. Johnson has worked
as a Senior Project Engineer in POWER’s underground transmission
design group since 2001. Mr. Johnson is a Member of the IEEE Power Engineering Soci-
ety and a Voting Member of the IEEE Insulated Conductors Committee (ICC). He is a
member of various ICC subcommittees that are developing guides and standards for
high-voltage underground cable systems. He is a registered professional engineer in the
states of California, Connecticut, Iowa, Kansas, Kentucky, Nevada, New York, Pennsyl-
vania, Texas, Vermont, and Virginia.
Adriano Santini is a Technical Specialist in Corrosion with Consoli-

dated Edison Company of New York. He holds a Bachelor’s Degree in
mechanical engineering from the City College of New York (1971) and
a Master’s Degree in engineering management from the Polytechnic
Institute of New York (1977). Since he joined Con Edison in 1972, Mr.
Santini has been involved in all corrosion control aspects of the com-
pany’s electric and gas facilities. He worked extensively in improving the
cathodic protection status of Con Edison’s pipe-type cable system with
special focus on its interaction with New York City’s grid of dc subways and commuter
railroads. As Manager of Corrosion Control, he was responsible for all corrosion engi-
neering and field testing of Con Edison’s buried and submerged metallic facilities. More
recently, Mr. Santini has been Manager of Gas Transmission Planning, and is presently
10-1
Chapter 10: Grounding and Cathodic Protection EPRI Underground Transmission Systems Reference Book
working in Transmission Feeders Engineering. Mr. San- University’s Appalachian Underground Corrosion
tini is a 35-year member of NACE International and Short Course. He has authored various technical papers
has served on and chaired several of its technical com- and made presentations at national conferences on such
mittees. He is also an accredited NACE Cathodic Pro- subjects as stray current mitigation, pipeline integrity
tection Specialist and is an instructor at West Virginia assessment, and computer applications in corrosion.
10-2
EPRI Underground Transmission Systems Reference Book Chapter 10: Grounding and Cathodic Protection
10.1 INTRODUCTION transformer action between the cable conductor and

An important part of designing any electrical system is sheath and is not a static-type voltage as is frequently
to determine the appropriate method of grounding. This and erroneously thought. ANSI/IEEE Standard 575
chapter discusses the various types of grounding sys- offers the caution that “the voltage on the insulated
tems that are typically implemented on underground sheath of a transmission cable may be considered analo-
cable systems. This chapter includes two sections: gous to the voltage on the conductor of non-shielded
grounding for single-core cables, and corrosion and secondary cables” (Ball et al. 1965, ANSI/IEEE Stan-
cathodic protection for pipe-type cable systems. dard 575-1988).
The geometric configuration of all conductors involved,

10.2 GROUNDING OF SINGLE-CORE CABLES
including parallel current-carrying circuits of any sort,
This section discusses the basic principles for grounding affects the magnitude of the induced voltage, which can
single-core transmission cable and the most common be, and usually is, different for each cable except for tre-
sheath-bonding systems now in use. Section 10.23 pro- foil and transposed geometries. Because of their geome-
vides detailed calculations and examples for calculating try and the magnetic steel pipe enclosing the three-phase
induced sheath voltages and currents. conductors, pipe-type cables are essentially three-phase
cables and are not addressed here and do not require
10.2.1 Sheath Bonding special bonding.
Large investment costs generally associated with the
installation of underground transmission circuits nor- The more commonly employed configurations of single-
mally mandate operational optimization of lines from core cables—self-contained fluid-filled (SCFF) and
an efficiency and power throughput capacity stand- extruded-dielectric, as well as the many options, hybrids,
point. Single-conductor cables and the use of low-loss, and engineering considerations—are described very
high-dielectric-strength insulating materials and thoroughly in ANSI/IEEE Standard 575-1988 and a
improved cable jackets gained popularity in the mid- comprehensive reference list is presented there
1960s. Along with their application at sub-transmission (ANSI/IEEE Standard 575-1988). Some excerpts from
and transmission voltages, there is significant interest in ANSI/IEEE Standard 575 are included here; more com-
single-conductor cables and the problems of the induced plex cases can be analyzed from materials presented in
voltages and currents associated with their use. Many of the reference. The basic theory was originally published
these problems (for example, failure of sheath insula- in 1929, first as the Master’s Degree Thesis of K.W.
tors, failure of cable jackets, and sheath corrosion) have Miller and later as an AIEE paper (Halperin and Miller
been recognized since metallic-sheathed cables were first 1929) and remains a classic reference on the subject for
used. However, increasing ampacity requirements and those with unorthodox cases and the zeal to analyze
short-circuit capabilities of modern power systems have them rigorously.
accentuated some problems, while improvements in
Sheath-Bonding Alternatives
sheath insulations have virtually eliminated others. This
Any sheath-bonding or grounding method must per-
section discusses basic principles and the selection of
form the following functions:
bonding systems.
• Limit sheath voltages.
Basic Principles
All alternating current-carrying conductors create an • Reduce or eliminate the sheath losses.
external magnetic field that induces a voltage in all • Maintain a continuous ground path to permit fault-
other conductors within its influence field, including the current return and adequate lightning and switching
shields and sheaths of any high-voltage cables. Parallel surge protection.
neutral or grounding conductors, control and communi-
cation cables, metal pipes, and other metal structures, The following is a brief discussion of four bonding
such as concrete rebar that are influenced by the mag- methods.
netic field, also may have voltages induced on them.
Multi-point Bonding
The magnetic fields and induced voltages follow basic When the cable sheaths of a three-phase circuit are
and well-known electrical engineering principles (i.e., directly interconnected (bonded) and/or grounded at
the induced voltage is directly proportional to the power multiple points, there will be a current flow in the
current flowing and to the logarithm of the spacing sheaths determined by the magnitude of the induced
between cables; and it increases linearly with the section voltage and the electrical impedance of the sheath cir-
length). Note that the induced voltage is the result of cuit. Because both the induced voltage and sheath
10-3
impedance are proportional to section length, the operation and shunt the current to ground during a
induced current is not affected by the distance between cable fault. They can be considered mandatory at all
bonding points. This induced current creates an I2R loss locations other than those where the open circuit is
in the sheath of each cable resulting in reduction of the achieved by insulating the terminal base plates with
amount of load current that can be carried for the same standoff insulators designed to withstand occasional
total allowable heating. This sheath loss for heavily flashovers. Figure 10-2 shows two approaches for single-
loaded cables can approach and, in some cases, exceed point bonding.
the conductor I2R loss. Without question, solidly
bonded and grounded sheaths represent the simplest Since the cable sheath may be exposed to contact by
solution to dealing with induced sheath voltages. This is personnel, it is common practice to specify a maximum
the most common bonding method used for under- permissible voltage during full-load operation. It is
ground distribution circuits. However, since under- understood that this voltage will be greatly exceeded
ground transmission lines are more expensive and during system transients and short circuits. In the early
typically require higher circuit ratings, different bonding days of single-core cable, the cables were non-jacketed
methods are needed. Figure 10-1 shows solidly bonded lead-sheathed cables. Because of potential corrosion
sheaths. problems, the permissible sheath voltage was limited
from 12 V to 17 V. Today, most cables are designed and
When a cable circuit is expected to carry a high load, installed with an insulating jacket, so the induced volt-
special bonding is often economically desirable as the age no longer poses a corrosion problem. This maxi-
reduction in losses will typically result in a smaller con- mum sheath voltage varies considerably between
ductor size. Two basic approaches that are commonly countries from 60 V to 600 V. In North America, the
considered for special bonding are single-point bonded trend is to limit the maximum allowable standing volt-
sheaths (open-circuited sheath operation) and cross- age during normal operation to 100-200 V. During
bonded sheaths, plus many hybrid combinations. emergency operation, the trend is to limit the standing
Another method, which was used more frequently in the voltage to less than 275 V.
past, is impedance bonding.
When a ground fault occurs on a power system, the
Single-Point Bonding
associated zero-sequence current must find a return
In a single-point bonded system, the shields of the three
path. Since a single-point bonded cable is grounded only
cables are connected and grounded only at one location
at one end, it cannot carry any of the fault current,
for each section length. At all the other points along the
except in the case of a cable fault. Therefore, an external
cable, a voltage will appear from sheath to ground, with
path must be provided. This is done by using a parallel
the maximum voltage occurring at the farthest point
ground continuity conductor (sometimes called and
from the grounded end. The sheaths must, therefore, be
earth continuity conductor). A separate cable, sized to
adequately insulated from ground. Since there is no
accommodate the expected fault current, is installed in
closed sheath circuit, a device is needed to complete the
parallel with the cable. The spacing of the conductor
circuit and protect the end of the cable from anticipated
should be as close as possible to limit the voltage rise of
transient overvoltages. These surge protective devices
the sheath to an acceptable level during a single-phase
are commonly called surge arresters, surge diverters, or
fault. The ground continuity conductor is usually insu-
surge voltage limiters (SVL) and should be considered
lated so as to avoid any corrosion risk, and is subject to
at all open-circuited shield locations. The purpose of
voltage induction from the power cables in the same way
these devices is to prevent current flow during normal
Figure 10-1 Solidly bonded sheaths. Figure 10-2 Single-point bonding.
10-4
as any other parallel conductor. To avoid circulating current return path on the sheaths independent of a par-
currents and losses in this conductor, it is preferable, allel conductor and is preferred for longer circuits.
when the power cables are not transposed, to transpose
the parallel ground continuity conductor. The voltage induction in parallel cables resulting from
balanced loads can be reduced or eliminated by trans-
Figure 10-3 shows methods that can be used for transposition, and this is particularly desirable for specially
position of the parallel cable or conductor. When the bonded cables for the reasons given above. Transposi-
power cables are laid in flat formation with wide spac- tion has the additional advantage of balancing the
ing, the parallel conductor should be between the power impedances of the three-phase cables. The transposition
cables at the position shown. If there is insufficient of heavy power cables is not generally practical except at
space between the power cables to adopt this position joint positions, however; hence, the complete transposi-
precisely, the parallel conductor should still be between tion cycle of the three phases will occupy three cable
the power cables. If the cables are touching or in trefoil, lengths. For cable circuits that consist of only one or two
the parallel conductor should be laid immediately lengths, it is not usual to transpose the power cables, but
alongside the power cables. In all cases, the parallel con- the parallel conductor, if used, may be transposed as
ductor should be transposed at the center of the section indicated in Figure 10-3.
length or route length to an identical position on the
other side of the formation. In a duct bank design where With cables not transposed, as illustrated in Figure 10-4,
the cables are arranged in an “L-shape,” the parallel it is impossible to achieve an exact balance of induced sheath
conductor should be located in position inside the voltages unless the cables are laid in trefoil. When the cable
“L-shape,” if possible, where the conductor is about the conductors are transposed at each joint position, the
same distance from all the cables. induced sheath voltages will be neutralized irrespective of
cable formation, provided the three minor sections are
identical in length. Transposition is the preferred method
Cross-Bonding for providing the best balance of sheath voltages. However,
Cross-bonding consists of dividing the circuit length into the practical difficulty of transposing large extra-high-
groups of three minor sections (“triads”), equal in length, voltage cables generally means that a majority of those lines
if possible, to have induced voltages of equal magnitude are installed not transposed.
in each minor section. By cross-connecting the sheaths in
ab, bc, and ca phase order as shown in Figure 10-4, it is It is not generally possible to divide the route length into
then possible to bond and ground all three sheaths at the exactly matched minor section lengths, nor is it always
ends of the triad without significant current flow since the possible to maintain a constant spacing of the cables
closing voltage will generally approach zero. Since the throughout the route. In continuous cross-bonded sys-
sheaths are cross-bonded together, it is normal practice to tems, it also may be desirable to have a total number of
use sheath-sectionalizing insulators at joint positions. minor sections not exactly divisible by three. In practical
This system has the advantage of a solid metallic fault systems, there is, therefore, generally some imbalance,
and it may be necessary to calculate the circulating
sheath currents that are present so as to assess their
Figure 10-3 Transposition of parallel conductors to reduce

induced voltages (ANSI/IEEE Std 585 1988). Figure 10-4 Cross-bonding.
10-5
effect on the cable rating. ANSI/IEEE Std. 575 has indi- • The impedance devices are relatively expensive since
cated that if the imbalance in minor section lengths is they must be custom designed.
15% or greater, it is considered necessary to account for • In normal operation, third harmonics may be intro-
derating effects due to the sheath circulation currents
duced into the sheath, which may cause interference
(ANSI/IEEE Standard 575-1988).
on nearby telephone lines.
Impedance Bonding • Stray direct currents may enter through the ground-
Impedance bonding consists of bonding the cable ing, which may cause saturation of the iron cores and
sheaths together through an impedance. This impedance impact the operation of the impedance devices.
may consist of simple reactors or of devices such as sat-
urable reactors and bonding transformers. In all these The primary advantage of impedance bonding is that it
methods, a certain amount of sheath current is permit- is effective in limiting the induced sheath currents
ted so as to reduce losses and sheath voltages. This regardless if the circuit segments are equal or not.
method is not commonly used today. Figure 10-5 shows
a schematic diagram of impedance bonding using a Design of Bonding Methods
bonding transformer. In the design and selection of special sheath-bonding
arrangements, consideration must be given to the fol-
During normal cable operation, the induced sheath volt- lowing aspects:
ages on the three cable sheaths are approximately equal • Cable sheaths are usually expected to be nominally at
and 120 degrees out of phase. Consequently, there is ground potential, but in a specially bonded system
very little current that flows through the cable sheaths to they may have appreciable voltages with respect to
ground via the sheath bonding transformers. Single- ground. Consideration should be given to any safety
line-to-ground fault conditions produce a zero-sequence aspects that may arise and to any limiting values of
voltage across the bonding transformer, and the fault sheath voltage that are specified.
flows to ground through relatively low impedance. The
sheath bonding transformers must be designed so that
• Complete elimination of circulating sheath currents
they will not saturate as a result of induced sheath volt- may not always be possible because of practical diffi-
ages produced by normal and short-term emergency culties in the choice of cable length and spacing. It is
operating currents. The cable sheaths also are connected then important to calculate the residual sheath cur-
to local ground by means of sheath voltage limiters. The rents and assess their effect on the cable rating.
sheath voltage limiters protect the cable sheaths and • The use of special bonding gives rise to sheath over-
sheath interrupts from transient overvoltages. voltages during system transients and faults, and the
magnitudes of those overvoltages must be consid-
Although a partial suppression of the induced sheath ered. A sheath voltage limiter will be needed, and in
voltage is obtained using this method, there are a num- all cases consideration must be given to the coordina-
ber of disadvantages that limit their application. tion of the cable jacket insulation level, bond lead
insulation level, and shield interruption insulation
• Additional manhole space is required.
level in relation to the overvoltages to which the insu-
lation will be subjected.
Figure 10-5 Typical schematic for impedance bonding (ANSI/IEEE Standard 575-1988).
10-6
• Failure of a part of the sheath insulation or of a tective relays. In single-point-bonded systems, the
sheath voltage limiter may result in large sheath cur- coupling path between fault currents and the sepa-
rents and losses and, hence, may cause overheating of rate continuity conductor may be devious and always
the cables. Consideration must, therefore, be given to inferior to the very positive coupling between fault
the duty imposed on the sheath voltage-limiting currents and cable sheaths, which represents the opti-
device and to the monitoring and maintenance of the mal return path.
complete systems in operation.
Selection of Special Bonding
Since a minimum of three minor sections is needed to
For single-conductor cable circuits carrying large cur-
form a cross-bonded system, it is not normally applica-
rents in excess of 500 A, there is no clear-cut point at
ble to cable circuits comprising only one or two lengths.
which special bonding should be introduced. The extra
For such circuits, single-point bonding is widely used.
cost of the larger conductor size cables needed for a sol-
For longer lines, multiple single-point bonding may be a
idly bonded system must be balanced against the cost of
better alternative in lieu of cross-bonding. This is espe-
the additional equipment and the maintenance cost aris-
cially the case when a spare cable is installed in addition
ing from the greater complexity of a specially bonded
to the phase cables (a fourth cable in the case of a single
system. However, if high projected loadings are
circuit or a seventh cable for a double circuit line), or
expected, special bonding is often economically desir-
where the section lengths are very unequal.
able as the reduction in losses allows an appreciably
smaller conductor size to be used. As stated in the
Although the cable sheaths of a single-point bonded
ANSI/IEEE Std. 575, special bonding should be consid-
system are generally of a cross-sectional area and con-
ered (but not necessarily adopted) at loads above about
ductivity that makes them quite capable of carrying
500 amperes per cable (ANSI/IEEE Standard 575-
short-circuit currents due to through-faults in the power
1988). If the initial loadings are lower, it may be prefera-
system, they are unable to do so because they are
ble to provide the required shield interruptions, but
grounded at one point only. A parallel ground continu-
jumper them. Installing the fairly expensive and high-
ity conductor is, therefore, required, and this adds
maintenance-cost link boxes and surge arrester hard-
appreciably to the cost of the cable system.
ware can be deferred until critical loading is
approached. The reasons for deferring the special bond-
The principal advantage of cross-bonding is that, while
ing are as follows:
induced sheath currents are inhibited during normal
• There are personnel safety considerations related to balanced load operation, the sheaths do form a continu-
high standing voltages and transient overvoltages on ous path from end to end of the cable circuit and are
the cable sheaths, joint housings, and terminator grounded at both ends. Sheath currents can, therefore,
bases that are implicit in any of the special bonding flow during ground faults, and the necessity for the par-
schemes. allel ground continuity conductor is removed. In addi-
• Adopting special bonding techniques to maximize tion to the economy achieved by the elimination of the
the rating imposes the necessity of a more stringent ground continuity conductor, the cable sheaths function
and expensive maintenance schedule. This is impor- more effectively as screening conductors during ground
tant since failure of any component of the system faults than a parallel ground continuity conductor.
used to interrupt circulating current flow, or damage Hence, the voltages induced in parallel cables are less
to the insulating jacket of the cable, can result in during ground faults in a cross-bonded system than for
undetected currents and over-heating. a similar single-point bonded system.
• The ancillary equipment for properly protecting the Because the preferred method depends on circuit length
cable jacket and interruptions from overvoltages is and interruption locations (usually at a splice), the
expensive, especially for longer circuits where cross- hardware selection and cost cannot be generalized. In
bonded sheaths are usually indicated. any event, surge voltage limiters are strongly recom-
• ANSI/IEEE Standard 575, “Parallel Ground Conti- mended at all shield interruptions. Figure 10-6 shows
nuity Conductor,” discusses the need to provide a the recommended equipment and connection for a
parallel conductor to ensure positive return of fault cross-bonded cable system.
currents to system neutral and to be sensed by pro-
10-7
Cable Bonding Components and to determine the amount of energy that SVLs must
Link boxes, surge voltage limiters, bonding cables are dissipate due to high frequency (traveling wave) tran-
important components of a bonding system. These sients. Prior to the development of computer simulation
bonding components are discussed in detail in Chapter tools for power system electrical transients, such as the
8, Section 8.4.1. Electromagnetic Transients Program, or EMTP (Dom-
mel 1994), it was common practice to make conservative
Sizing of SVLs
assumptions when specifying equipment for sheath
It is well known that relatively high sheath transient
transient overvoltage protection for cross-bonded and
voltages can occur when cable sheaths are single-point
single-point bonded cable systems. However, with the
bonded or cross-bonded to minimize induced sheath
trend to longer distances between SVLs and increased
currents (Marti et al. 1996; Itoh et al. 1950). These tran-
standing voltage limits, it is often necessary to perform
sient overvoltages are typically caused by
EMTP simulations to determine maximum overvoltages
• lightning current impulses entering an underground and SVL energy dissipation requirements accurately
cable system from overhead lines, (Emin et al. 2001).
• fault conditions, and
The following criteria should be considered when select-
• switching transients such as from line energization. ing an SVL:
SVLs are commonly used to limit the voltage of the • The SVL should be suitable for continuous operation
cable sheaths during transient overvoltage conditions; with an applied voltage equal to the sheath-standing
however, the amount of energy that the SVLs must dissi- voltage under either normal or emergency load.
pate increases with lower protective levels and increased • The SVL should be able to withstand the 60-Hz
distances between SVLs. overvoltages resulting from system faults. Caution
should be used in selection of nonlinear resistance-
The sheath overvoltages are primarily the result of two type limiters because of problems associated with
factors. The first is the mutual coupling among the 60-Hz overvoltages.
cable phases that occurs when the magnetic flux created
• With nonlinear resistance-type SVLs, a maximum
by one cable links the other cables in the same trench.
duration should be specified; this is normally twice
The second transient overvoltage phenomenon is
the maximum fault clearing time for the system to
caused by the discontinuity to traveling waves presented
allow for reclosing.
by the sheath interrupts. Traveling waves created by
lightning, switching, and faults reflect at the sheath • When calculating 60-Hz voltages appearing across
interrupts, resulting in superposition of the incident SVLs, allowance should be made for the SVLs that
and reflected waves. are star or delta connected.
• Spark gaps and surge arresters should be able to
While it is possible to estimate the power frequency withstand impulse currents for the same duration as
component of fault-current-initiated sheath overvolt- specified for the impulse requirements for the main
ages using methods described in the other sections of lightning (surge) arresters on the system.
this guide, more sophisticated calculation tools are
required to accurately model transient sheath voltages A nonlinear resistance-type limiter should be able to
absorb, without damage, the energy dissipated due to
switching, including switching associated with a fault
external to the cable circuit. Experience and calculations
indicate that the energy dissipated in the nonlinear resis-
tances due to switching is not an important design crite-
rion for normal cross-bonded circuits. However, for long
single-point bonded circuits or lengths of single-point
bonded cable that terminates long circuits, the switching
surge energy may be important, and calculations should
be made for these cases.
10.2.2 Structure Grounding

Direct-embedded steel structures and structures with
Figure 10-6 Sheath connections for cross-bonding. drilled-pier foundations where the foundation’s steel
10-8
reinforcing bars are electrically bonded to the structure system. These voltages are important both from a safety
generally constitute effective and acceptable ground. A standpoint and from the need to calculate the voltages
separate ground conductor should be provided for seen by various components of the system.
bonding cable accessories such as arresters, link boxes,
Induced Voltages
and terminations. This separate ground conductor
Of primary interest are voltages induced in the sheaths
should be bonded to the structure ground and the
of single-conductor cables by positive- and zero-
parallel ground continuity conductor. Some utilities do
sequence current. The sheath standing voltage, which is
not run a separate ground conductor from the shield wire
the term that is commonly used to describe the voltage-
to the structure ground, thus using the inherent low surge
to-ground induced on the sheath by full-load current, is
impedance of the steel structure. Using the steel structure
a function of the positive-sequence current. Calcula-
as the ground for the cable system is not recommended
tions are required for both positive- and zero-sequence
currents to determine fault-induced sheath voltages.
It is also important to consider the grounding for each
The general procedure for doing this is discussed in the
manhole. A ground ring should be provided inside the
following paragraphs; however, readers are referred to
manhole and connected to ground rods. If suitable
several technical papers on this subject for more specific
grounding cannot be achieved with ground rods, ground
details (CIGRE ANSI/IEEE Std 575 1988, CIGRE
wires should be placed in the trench or under the man-
Working Group 1990, Kuwahara and Doench 1963).
hole to achieve a suitable ground.
The problem of induced sheath voltages is obviously
10.2.3 Design Calculations
minimized by solidly bonding the cable sheath to ground
Possible faults in power systems, both internal and at each termination and joint. Even though there is an
external to a cable circuit, affect the design of cable appreciable voltage created in the cable sheath by the
systems from several aspects. First, the cable must be high sheath currents at locations other than the ground
designed so that the cable insulation will not be point, they are usually well below the short-term ac-
damaged by the heat produced by fault current flowing withstand voltage of the cable jacket. However, solidly
in the conductor or the cable shield/ sheath. Secondly, bonded sheaths are often avoided to minimize the
fault current will induce high voltages on a single- steady-state energy losses from the induced sheath
conductor cable sheath that may damage sheath voltage currents. This sometimes results in the use of single-
limiters, shield interrupts, or puncture the cable jacket. point bonding for short cable lengths such as substation
If sheath voltage limiters are used to protect the shield “getaways” or “dips” at airport runway crossings.
interrupts and cable jacket, they must be designed to
absorb the energy they dissipate while limiting the In order to calculate the induced sheath voltage, it is
sheath overvoltage. important to determine the associated impedances.
Chapter 16 discusses the methods for calculating cable
Special sheath bonding methods have been devised to impedances. Figure 10-7 shows a general three-phase
reduce or eliminate the losses in a cable system caused cable circuit, which is the basis for matrix Equation 10-1
by induced sheath currents. While these special bonding to relate the series voltage drop along the cable conduc-
methods reduce sheath losses, they can result in a signif- tors to all of the currents flowing in the high-voltage
icant sheath-to-ground voltage at various points in the conductors and the cable sheaths. Equation 10-1 was
Figure 10-7 Cable circuit for positive- and zero-sequence impedance calculations.
10-9
taken from Chapter 16, Equation 16-21. This figure As an alternative to the above calculation procedure,
shows the sheaths bonded at both ends. ANSI/IEEE Std. 575 gives the following equations for
calculating the open circuit sheath voltages for cables in
⎡Ea ⎤ ⎡ Zaa Zab Zac Zas1 Zas 2 Zas 3 ⎤ ⎡Ia ⎤ a flat configuration. These equations are for a, b, c rota-
⎢ ⎥ ⎢ ⎥⎢ ⎥ tion with phase b as the reference phase.
⎢Eb ⎥ ⎢ Zab Zbb Zbc Zbs1 Zbs 2 Zbs 3 ⎥ ⎢Ib ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎛ 1 S
⎜ − ln
3 4S ⎞
⎟
Es1 = jωIb(6.4 x 10-5) ⎜ 2 dsm + j 2 ln dsm ⎟ V/1000 ft 10-3
⎢Ec ⎥ ⎢ Zac Zbc Zcc Zcs1 Zcs 2 Zcs 3 ⎥ ⎢Ic ⎥ ⎝ ⎠
⎢ ⎥=⎢ ⎥⎢ ⎥ 10-1
⎢ 0 ⎥ ⎢ Zas1 Zbs1 Zcs1 Zss Zs1s 2 Zs1s 3 ⎥ ⎢Is1 ⎥ Es2 = jωIb(6.4 x 10-5) log e
2S
V/1000 ft 10-4
⎢ ⎥ ⎢ ⎥⎢ ⎥ dsm
⎢ 0 ⎥ ⎢ Zas 2 Zbs 2 Zcs 2 Zs1s 2 Zss Zs 2 s 3⎥ ⎢Is2 ⎥ ⎛ 1 S 3 4S ⎞
⎢ ⎥ ⎢ ⎥⎢ ⎥ Es3 = jωIb(6.4 x 10-5) ⎜ − log e
⎜ 2
⎝ dsm
+j
2
ln ⎟
dsm ⎟⎠
V/1000 ft 10-5
⎢ 0 ⎥ ⎢ Zas 3
⎣ ⎦ ⎣ Zbs 3 Zcs 3 Zs1s 3 Zs 2 s 3 Zss ⎥⎦ ⎢⎣Is3 ⎥⎦ Where:
S = axial spacing between adjacent cables, in.
Since the objective of the calculations is to determine dsm = mean sheath diameter, in.
the series voltage drops along the high-voltage ω = angular frequency, radians 2πf.
conductors in terms of positive- or zero-sequence
current, the cables may be assumed to be connected to Figure 10-8 shows the open-circuit sheath voltages for
ground at one end to simplify the equations. Sheath cables in flat and triangular formations.
voltages at the left end of the cables are given by the
following matrix equation.
⎡Es1 ⎤ ⎡ Zas Zbs1 Zcs1 ⎤ ⎡Ia ⎤

⎢Es2 ⎥ = ⎢ Zas 2 Zbs Zcs 2 ⎥ ⎢Ib ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥ 10-2
⎢⎣Es3 ⎥⎦ ⎢⎣ Zas 3 Zbs 3 Zcs ⎥⎦ ⎢⎣Ic ⎥⎦

Where:
ES1, ES2, E S3 are the voltages induced at the open
end of the sheaths when the ground is removed at
the right end of Figure 10-7.
A common procedure is to assume 1 ampere of either

positive-or zero-sequence current. The voltages
calculated by the above equations will give the induced
sheath voltages in volts per thousand feet per ampere of
positive- or zero-sequence fault current. Figure 10-8 Induced sheath voltage as a function of ratio of
separation to cable diameter.
10-10
For zero-sequence currents, the sheath voltages with sheath is single-point grounded. Calculate the voltage at
respect to a neutral conductor laid parallel to the high- the open end of the sheath for 10,000 amperes of posi-
voltage cables (Figure 10-9) are tive- and zero-sequence current. Other parameters
needed for calculating the sequence impedances are:
⎡ 2
2Sag ⎤
E s1 = I o ⎢ R g + jω(6.4 x10 ) loge
-5
⎥ V/1000 ft 10-6
1000 kcmil; 23 µΩ/ft
⎣⎢ d sm GMR g ⎦⎥ Conductor size; resistancea:
Sheath size; resistance: 100 mils, lead; 142 µΩ/ft
Earth electrical resistivity: 100 Ω-m
⎡ Sag Sbg ⎤
E s2 = I o ⎢ R g + jω(6.4 x10-5 ) log e ⎥ V/1000 ft 10-7 Insulation thicknessa: 0.850 in.
⎣⎢ GMR gSab ⎦⎥ Geometric mean radius of conductor: 0.445 in.
Conductor diametera: 1.060 in. (compact round)
a. Values per AElC CS7-87

⎡ SagScg ⎤
E s3 = I o ⎢ R g + jω(6.4 x10 -5 ) log e ⎥ V/1000 ft 10-8
⎢⎣ GMR g Sac ⎥⎦ The geometric mean distance between the conductors is:
Where:
Io = zero sequence current GMDφ = 3 9 x 9 x18 = 11.34 in.
Sag, Sbg, Scg=axial spacings between the centers of
cables a, b, and c and the ground conduc- The equivalent depth of the earth return current is:
tor, in.
dsm = mean sheath or shield diameter, in. De = 3346 100 = 33500 in.
Rg = resistance of the ground conductor, Ω/1000
ft.
The radius from the center of the cable conductor to the
GMRg= geometric mean radius of the ground con-
middle of the sheath is:
ductor, in.
1.06
Example Calculation for Induced Sheath Voltages
rsm = + 0.85 + 0.05 = 1.43 in.
Assume a 138-kV XLPE cable installed in a horizontal 2
configuration with phase spacing of 9 in. between the
centers of the cables. The cable is a 1000-kcmil, compact The induced sheath voltages could be calculated using
round aluminum conductor and a 100-mil thick lead either matrix Equation 10-2 or Equations 10-3 through
sheath. Assume the cable is 1000 ft in length and the 10-8. The following calculations use the for mer
approach. The first step is to calculate the mutual
impedances for the circuit. Chapter 16 discusses the
method for calculating the mutual impedances. Using
the following equation from Chapter 16 for mutual
impedance between a conductor and its concentric
sheath with earth return, the expression is
De
Zas1 = 0.01807 + j 0.0529 log 10 Ω/1000 ft 10-9
rsm
The conductor-to-sheath mutual impedance for phase

“a” with earth return is
33500
Zas = 0.01807 + j 0.0529 log 10
1.43
Figure 10-9 Induced sheath voltage with neutral
conductors. = 0.01807 + j 0.2311 = 0.2318∠85.53 Ω/1000 ft
10-11
The mutual impedance between sheath “a” and conduc- sequence currents for cable circuits with multi-section
tor “b” is cross bonding.
33500 10.3 CORROSION AND CATHODIC

Zbs1 = 0.01807 + j 0.0529 log 10
9 PROTECTION
This section reviews the methods and procedures associ-
= 0.01807 + j 0.1889 = 0.1897∠84.53 Ω/ 1000ft ated with the prevention of pipe and sheath corrosion
on transmission cables. Primary consideration will be
The mutual impedance between sheath “a” and conduc- given to the steel pipe used with pipe-type cable.
tor “c” is
10.3.1 Corrosion Theory
33500 It is a common observation that after a steel pipe has
Zas = 0.01807 + j 0.0529 log 10 been buried for some time, a material forms on its sur-
18
face that is unlike the pipe itself or any of the material
= 0.01807 + j 0.17295 = 0.1739∠84.0 Ω/1000 ft that is in contact with its surface. It is important to note
that some of the steel is “missing” as indicated by the
The next step is to calculate the voltage induced in the a- pipe’s pitted surface. Since this new material was not
phase sheath by balanced three-phase currents in each there before, it is reasonable to assume that it was pro-
of the three phases (positive- sequence current). duced by a chemical reaction between the steel and its
environment. What actually takes place is that some
Es1 = ZasIa + Zbs1Ib+Zcs1Ic
parts of the steel give up electrons and react with the
Es1 = 137 + j 501.9 = 520 V/l000 ft positively charged hydroxyl ions in the soil to form the
aforementioned material, commonly known as “rust.”
The voltage induced in the a-phase sheath due to zero- The electrons travel to a separate area of the steel, are
sequence currents would be discharged into the soil, and join with positively charged
hydrogen ions to form hydrogen gas. The area of the
Es1 = ZasIo + Zbs1Io + Zcs1Io
= (0.2318∠85.53) (10,000∠0) + (0.1897∠84.53) steel where the rust forms is called the anode and the
(10,000∠0) area of the steel that discharges electrons into the soil is
+ (0.1739∠84.04) (10,000∠0) called the cathode. Note that the steel at the cathode
plays no part in the reaction other than to transfer elec-
Es1 = 542 +j 5929 = 5,954∠84.77 V/1000 ft trons from the anode to the soil.
If an SVL is placed at the open end of the sheath to limit This process is called “corrosion,” and in order for it to
overvoltages produced by switching surges and light- occur, four components must be present, which together
ning, it must be able to absorb the heating energy caused form the “corrosion cell” as shown in Figure 10-10.
by the induced ac voltage for fault conditions. Thus, the • Anode—The part of the metal that corrodes; it gives
sheath voltage limiters for the cable in the above exam- up electrons and reacts with ions in the soil.
ple would have to be able to withstand approximately
6000 V for the maximum fault-clearing time without
damage. Reference (CIGRE Working Group 1990) dis-
cusses additional design requirements for sheath over-
voltage protectors for cross-bonded cable systems. For
cables with cross-bonding, the induced sheath voltages
for balanced three-phase faults are typically higher than
for single line-to-ground faults. Single line-to-ground
fault voltages for cross-bonded cables are similar to
cables with single-point bonding because the voltages
induced by the zero-sequence currents in each transpo-
sition section are all in phase. The method shown above
can be used to calculate the voltages induced by zero-
sequence currents. Reference (Kuwahara and Doench
1963) derives rather complex equations and approxi-
mate formulas for the voltages induced by positive- Figure 10-10 Metal corrodes at anode (courtesy
Consolidated Edison Company of New York, Inc.).
10-12
• Cathode—It receives electrons from the anode and metals in order of their propensity to corrode results in
discharges them into the soil; it is the part of the a list called the Galvanic Series, a portion of which is
metal that does not corrode. shown in Figure 10-11.
• Electrolyte—The soil in which both the anode and
Active metals such as magnesium or zinc are at the top
cathode are buried. It is the ionic connection between
of the list and metals such as gold or platinum are at the
the two.
bottom. Connecting metals at the top of the list with
• Return path—The metallic connection between the those at the bottom and placing them both in an
anode and cathode through which electrons flow. electrolyte would create a corrosion cell where the active
metal becomes the anode and the passive metal becomes
In addition to these, there must be a potential difference the cathode. A corrosion cell made from two metals that
between the anode and cathode that will cause the flow are closer to each other in the Galvanic Series will be
of electrons. This difference in potential can develop less active than one made from two metals that are
because of chemical differences between the anode and further apart.
cathode or it can develop because of chemical differ-
ences within the soil. Potentials resulting from such con- This potential difference can exist between different
ditions cause “galvanic corrosion.” areas of the same steel pipe because the manufacturing
process does not yield a completely homogeneous
A second, perhaps less common but certainly more metal. Each area of the pipe may contain more or less
aggressive, cause of corrosion occurs when the difference iron, more or less carbon, and so on. In fact, this steel
in potential between the anode and cathode is caused by pipe can be thought of as being made up of many differ-
electrical interference by outside sources. This is called ent metals all connected to each other and at different
“electrolytic corrosion” or “stray current” corrosion. potentials than each other. The result is in not just one
Galvanic Corrosion corrosion cell but many such cells all corroding the sur-
This “self-generated” corrosion activity is the most face of the same pipe. Although a galvanic cell that
common cause of corrosion-related failures of pipelines exists on the surface of the same metal is weak, it can
and other metal structures. The following are the corro- still produce considerable damage, especially if other
sion cell components that are formed when buried steel factors affect it. Factors that contribute to the activity
pipes corrode: of corrosion cells include the following:
• Anode—Since the corrosion products are visible on • Temperature—Raising the temperature of the metals
the pipe, some part of the pipe surface must be the or the electrolyte will increase the reactions at the
anode. anode and cathode. As the temperature is increased
to the point of drying the electrolyte, the corrosion
• Cathode—This is not so apparent, but since the pipe activity will actually decrease.
is not metallically connected to anything else, the
cathode must be some other part of the same pipe. • Stress—Areas of metals that are under stress are
anodic to areas of the same metal that are not
• Electrolyte—This is the earth which surrounds the stressed and, therefore, will corrode faster.
buried pipe.
• Oxygen content—A metal that is in contact with soils
• Return path—The metallic connection between the with different oxygen content will develop anodic and
anode and cathode is the pipe itself.
What about the potential difference that is required

between the anode and cathode to cause current flow
and corrosion to take place, where does it come from?
The answer lies in the fact that metals are not found in
nature in the state in which they are commonly used. In
their natural state and to a varying degree, metals exist
as compounds with other elements such as oxygen, sul-
fur, and carbon. Depending on the stability of these
compounds (ores), more or less energy is required to
separate each metal from them. Those metals that
require the most energy will more readily corrode or Figure 10-11 Position of common metals on the galvanic
return to their natural state; those metals that require series (courtesy Consolidated Edison Company of New
the least energy are less likely to corrode. Arranging York, Inc.).
10-13
cathodic areas. The anodic areas will be the ones in The problem occurs because the tracks are supported by
contact with soils with low oxygen concentration. wooden or concrete ties that are not effective isolators;
• Concrete—Metals that may be partially buried in therefore, the track is essentially grounded. This allows
concrete and partially in soil will develop anodic and the ground to become a parallel conductor to the rail
cathodic areas. The areas of the metal in contact with for the dc current to return to the transit substation.
the soils will be anodic to the areas in the concrete. This current “strays” off its intended path, hence the
name “stray current.” The amount of current that leaves
• Cinders—Soils that contain cinders are corrosive to the rails and travels through the ground is a function of
metals buried in it. Each cinder that comes in contact the rail to ground resistance. During the last several
with the metal will create a localized corrosion cell years, rail operators have made great efforts to maxi-
with the cinder as the cathode and the surrounding miz e this resistance, resulting in corresponding
metal as the anode. decreases in stray current levels. However, considering
Electrolytic (Stray Current) Corrosion the many miles of rails that comprise dc subways and
Dc currents that travel through the soil can affect the commuter railroads in large cities, it is unlikely that
corrosion “status” of metallic structures in their path. stray current can be completely eliminated. To do so
Although these currents can originate from various would require a separate, isolated conductor to return
sources, the most common are current to the substation.
• dc railroads and Just as the ground is a good conductor for these stray
• cathodic protection rectifiers. currents, so is any metallic facility that is buried in it.
Pipe-type cables are very susceptible to stray current
Dynamic Interference interference because the pipe is electrically continuous
Currents from dc railroads are referred to as dynamic from end to end, making its longitudinal resistance very
stray currents because their magnitude changes continu- low. Therefore, if such cables are located within a stray
ously depending on the movement of the train or trains. current area, they also become parallel conductors to
Dc railroads are powered by current from transit sub- the rail for returning current to its substation source.
stations that energize the “third” rail located along each
track as shown in Figure 10-12. The train’s motors The stray current process sets up two corrosion cells as
receive this current via a “shoe” that is attached to the shown in Figure 10-13.
train and slides along the third rail. The circuit is com-
pleted as the current flows through the train’s wheels The first cell is in the area where the current is dis-
onto and along the tracks back to the substation. charged from the rails into the ground and is picked up
Figure 10-12 Stray currents from dc railroads “interfere”

with buried pipelines (courtesy Consolidated Edison
Company of New York, Inc.). Figure 10-13 Corrosion cells set up by stray currents
(courtesy Consolidated Edison Company of New York,
Inc.).
10-14
by the pipe. Here the corrosion occurs on the bottom their localized nature makes them more difficult to
surface of the rails (anode), but its affect is generally not notice during routine corrosion surveys.
severe since the surface area is large. The pipe that picks
up the current in this area is the cathode of this corro- 10.3.2 Corrosion Coordinating Committees
sion cell, and it does not corrode. The second corrosion To resolve both dynamic and static interference issues,
cell is set up where the current tries to return to its committees have been organized in most urban areas to
source, near transit substations. Here, stray currents are facilitate interaction and communication among com-
discharged from the pipe surface back into the ground; panies that may install and operate or may be affected
the effect, however, is much worse since the stray cur- by systems such as dc railroads or impressed current
rents are discharged from very small areas on coated rectifiers. These committees are sometimes referred to as
pipes. The corrosion is concentrated in these areas and corrosion coordinating committees or electrolysis com-
can easily result in the penetration of the pipe wall. mittees. They usually have regularly scheduled meetings
Static Interference at which individuals discuss and attempt to resolve spe-
Stray currents also can emanate from nonvarying dc cific problems in a mutually satisfactory way. A list of
sources. The most common of these are cathodic protec- the various committees in the United States can be
tion rectifiers whose function, as will be discussed later, obtained from NACE International, a professional
is to prevent corrosion on other underground facilities. organization of corrosion engineers headquartered in
Unlike dc currents from railroads whose flow into the Houston, Texas.
ground is unintended, cathodic protection rectifiers
impress dc current into the ground deliberately to pro- 10.3.3 Corrosion Prevention
tect these facilities. Just as in the previous case, as they Cathodic protection is the most widely accepted and
travel through the ground, these currents can be picked most effective method of preventing corrosion on buried
up and discharged by other metallic structures in the pipes. Ironically, this method uses the concept of the
area. At locations where these structures receive current corrosion cell as the basis for corrosion prevention. As
from the neighboring cathodic protection system, they was discussed before, for corrosion to take place, each of
also receive some degree of protection. Corrosion the four components of the corrosion cell must be
occurs at the locations where this current is discharged present. One of these is the cathode, the component that
back into the ground as shown on Figure 10-14. does not corrode. The idea of cathodic protection is sim-
ply to create a corrosion cell whose cathode is the pipe
Although the levels of these currents are typically much that is being “protected.” Such a corrosion cell can be
lower than those from railroads, their damaging effect is created in two ways:
continuous since their source is constant. In addition, 1. Connect the steel pipe to be protected with another
metal that is higher on the galvanic series. This will
create a potential difference between the two metals
such that current will be discharged from the anodic
metal into the electrolyte and be picked up by the
pipe. Such a system is called a galvanic cathodic pro-
tection system. The typical galvanic cathodic protec-
tion system that is applied to buried steel pipes uses
magnesium as the anode.
2. Connect the pipe to be protected to another metal
and apply the required potential between the two
with an outside source. Such a source is a rectifier
whose positive terminal is connected to the “expend-
able” metal, making it the anode. Its negative termi-
nal is connected to the pipe, making it the cathode
and, therefore, “protecting” it. This method is called
an impressed current system and is used when pro-
tecting larger or longer buried structures where more
Figure 10-14 Interference by cathodic protection rectifiers current is required.
(courtesy Consolidated Edison Company of New York,
Inc.).
10-15
Theoretically, this is all there is to cathodic protection; a voltmeter and is placed in the soil at grade above the
however, its realistic application requires at least two pipe at a number of locations along the pipe. The poten-
more steps. The first is to minimize the area of the pipe tial reading at each location indicates the corrosion or
that is in contact with the electrolyte (soil). This is done protective condition of the pipe at that location.
by coating the pipe with a good quality coating of high
dielectric strength. Since the coating effectively sepa- An unprotected, coated steel pipe will have a potential
rates the structure’s surface from the electrolyte, the to a copper-copper sulfate electrode of -0.5 to -0.7 V.
anode – either impressed current or galvanic – can dis- When cathodic protection is applied, the potential
charge much less current to achieve the same level of becomes more negative. The most common criterion for
protection. Second, ensure the pipe to be protected is indicating adequate cathodic protection is -0.85 V to a
electrically isolated from any other metal structure to copper-copper sulfate reference electrode. When taking
which it is mechanically connected. The idea is to potential readings, it is important to ensure that the
increase the resistance of the pipe with respect to ground electrode makes contact directly with the soil. If these
to achieve cathodic protection with the least current readings are measured on pipes that are installed under
applied. This leads to the question of how much current pavement, an attempt can be made to take these read-
is enough? ings by placing the reference cell on the pavement. Such
readings, however, can be erroneous; therefore, they
10.3.4 Cathodic Protection Criteria must be verified periodically by comparison to measure-
When a cathodic protection system is energized, the ments taken with the cell in contact with the earth under
potential of the pipe with respect to the surrounding soil the pavement through bar-holes made for that purpose.
is changed. By measuring the pipe-to-soil potential, it These readings also can be compared with readings
can be ascertained whether the resultant potential meets taken at nearby test stations where direct contact with
the criteria for protection. The method used for measur- soil can be made.
ing pipe-to-soil potential is shown in Figure 10-15.
Metallic contact to the pipe is made through a previ- 10.3.5 Applying Cathodic Protection to Pipe-type
ously installed test station, or contact can be made Cables
directly to the pipe at an aboveground location. The As any other buried steel pipe, pipe-type cables are sub-
most commonly used reference electrode is a copper- ject to corrosion attack unless they are cathodically pro-
copper sulfate half cell. It is rugged and stable under tected. As previously stated, conventional cathodic
field conditions and is easy to maintain. It consists protection requires the application of dc current by
essentially of a rod or tube of pure copper immersed in a either a galvanic or impressed current anode and by
saturated solution of copper sulfate. Contact to the soil maximizing the pipe to ground resistance by the use of
is made through a porous plug at the bottom. This ref- dielectric coatings and isolation from other grounded
erence electrode is connected to the negative terminal of structures. Unlike other buried pipe, however, pipe-type
cables have characteristics that make the application of
cathodic protection more difficult. All of these charac-
teristics are related to the absolute requirement for pipe-
type cables to be connected to substation ground mats
for ac fault current protection.
Pipe-type cables, for example, must be electrically con-

tinuous from end to end, to allow fault currents to flow
unimpeded. Therefore cathodic protection of pipe-type
cables must be applied along the entire pipe length and
cannot be applied to individual sections that are electri-
cally isolated from each other, as is often done on gas or
other pipelines.
The grounding requirement also means that the pipe-

type cable must be connected to a large copper ground-
Figure 10-15 Measurement of pipe-to-soil potential using ing grid. This creates a very large and aggressive dissim-
a high resistance voltmeter and Cu-CuSO4 reference cell
(courtesy Consolidated Edison Company of New York, ilar-metals corrosion cell with the pipe as the anode and
Inc.). the copper grid as the cathode.
10-16
Finally, connecting the pipe-type cable directly to sta- ground path. This in turn can only be done by maximiz-
tion ground makes the ground mat part of the structure ing the pipe-to-ground resistance by coating the pipe
to be protected. This condition is not only unnecessary, with a coating of high dielectric strength. By maximiz-
because of the relative passivity of copper, it is also det- ing the ground path resistance, only a small fraction of
rimental to cathodic protection of the feeder pipe, the rectifier current is allowed to be discharged to the
because it lowers its pipe to ground resistance. ground mat. To review the corrosion cell that is created
by this cathodic protection method, the station ground
Because of these characteristics, the cathodic protection mat is the anode and, therefore, corrodes. The earth
system employed must include on appropriate ground- around the pipe and the ground mat is the electrolyte,
ing method that is able to overcome these limitations. the pipe is the cathode that does not corrode, and the
rectifier is both the return path as well as the source of
As described below, various grounding methods have the potential between the anode and cathode.
been developed and improved over the years.
Although this method has been used for many years, it
Resistor-Rectifier has several problems that have made it ineffective. The
In this method, a 0.004-ohm resistor capable of carrying first is that it generates the minimum required potential
the anticipated fault current is connected between the difference at the substation end of the pipe and it
pipe and the substation ground mat. A 100-ampere rec- assumes that the high initial coating resistance prevents
tifier is then connected across the resistor with the posi- this potential from attenuating. This may be true for
tive terminal connected to the station ground. As shown very short feeders, but as the length increases, the pipe-
in Figure 10-16, the current from the rectifier circulates to-soil potentials of the pipe fall quickly below criteria.
through two parallel paths. In addition, coating resistance only deteriorates over
time; therefore, the current requirements increase. This
The first path is the resistor bar and the second is the increase can only be met with larger rectifiers with pro-
ground between the ground mat and the pipe. The portional increases to current flow through the resistor
amount of current in each path is inversely proportional and current discharge to ground, further corroding the
to their resistance. Since the object is to impress a nega- ground mat. Finally, depending on the relative position
tive voltage of approximately 0.4 V between the pipe of each substation, stray currents can be picked up and
and ground, virtually all of the rectifier current should discharged not only by the pipe but by the ground mats
be circulating through the resistor bar. This can only be themselves, with additional detrimental effects to the
accomplished by maximizing the resistance of the discharging mat.
Figure 10-16 The rectifier changes the dc potential of the pipe, and the
“resistor” conducts ac fault current to ground (courtesy Consolidated
Edison Company of New York, Inc.).
10-17
Polarization Cell trons at the cathode react with hydrogen ions in the elec-
In the above section, the dc current required to protect a trolyte to form hydrogen, which “coats” or “polarizes”
well-coated pipe is a small fraction of the rectifier’s out- the surface of these plates, impeding further current
put. The rest of its current is shunted through the flow. When the negative plates become fully polarized,
0.004-ohm resistor, whose only and rare function is to the current flow through the cell is virtually eliminated.
discharge ac faults to ground. If the resistor could be The polarization cell is, in effect, a chemical insulating
replaced with a device that could similarly conduct very joint. The current output of the rectifier is, therefore,
short duration ac faults to ground, yet continuously reduced by the amount that was previously shunted
provide dc isolation to the pipe, the high-current recti- through the resistor while the current that was previ-
fier could be replaced with a small dc current source. ously discharged from the ground mat remains the
More importantly, this dc current source could be same. This current reduction makes it possible to com-
placed outside the substation and avoid the corrosive pletely eliminate the two large rectifiers (one at each end
effect of dc current discharge from the substation of the feeder) and replace them with much smaller units
ground mat. One such device is the polarization cell. that are installed along the feeder’s route. These “field
rectifiers” discharge protective current into the ground
The polarization cell consists of two electrodes each via “anode groundbeds” that are separate from the sub-
connected to multiple stainless steel or nickel plates. The station ground mats, thereby, completely eliminating
plates are immersed in an alkaline electrolyte as shown their corrosion issues.
in Figure 10-17. The typical ac rating for such cells is
55 kA for three cycles. Although polarization cells are good substitutes for
resistors-rectifiers as a method of ac grounding and dc
The electrolyte consists of a 30% solution of potassium isolation, there are some things to consider when using
hydroxide or a 15% solution of sodium hydroxide. A them. It is important to remember that the polarization
layer of mineral oil is placed on the electrolyte and acts film that forms on the negative plates is tenuous and can
as a seal to prevent evaporation that would change the break down. One of the reasons is the fact that each
concentration of the electrolyte. It also separates the pipe-type cable has some ac current that is induced on
electrolyte from the atmosphere, thereby, preventing its the pipe because of the load unbalance on the three
reaction with atmospheric gasses such as carbon diox- phases. This current flow, although small, flows continu-
ide. The cell resembles a battery, but, unlike a battery, it ously to ground, passes through the cell, and breaks
has no charge of its own. When it is connected in place down the polarization film. This, in turn, causes an
of the resistor and the rectifier is energized, dc current increase of dc current flow through the cell, which means
flows through it turning it into a corrosion cell. Plates more corrosion and deterioration of the positive plates.
that discharge current into the electrolyte become
anodic and corrode while plates that pick up current This polarization breakdown also can be caused if the
become cathodic and do not corrode. The excess elec- applied current causes the dc pipe-to-ground potential
to reach the voltage blocking capability of the cell,
approximately 1 volt. At this point, any further increase
in current applied does not result in a corresponding
increase in potential, rather it causes a breakdown in the
polarization film and increased flow of corrosive dc cur-
rent through the cell.
This situation can happen in two ways. First, depending

on the location of the current source (rectifier) along the
feeder and on the relative pipe-to-ground resistance of
different areas of the feeder, the pipe to soil potential at
one end is often higher or lower than at the opposite
end. In fact, the potential at one end may be below pro-
tective criteria while at the other end the feeder may be
well above criteria. It is apparent that if a rectifier’s out-
put is raised to correct the problem at one end, the
Figure 10-17 Components of a polarization cell (courtesy polarization cell’s blocking voltage at the other end can
Consolidated Edison Company of New York, Inc.).
be reached, at which point the polarization film will
10-18
break down. The second way that this can happen is in ity, which, of course, means that part of normal cell
situations where the substation ground mat is in a stray maintenance includes cleaning and inspection of the
current pickup area. Here, stray currents are added to plates for any sign of deterioration.
the rectifier current and can cause the potential across
Solid-State Isolator
the cell to easily reach the limit and break down the
Although the polarization cells are effective dc isolation
polarization film.
devices for cathodic protection of pipe type cables, their
chemical characteristics make them high-maintenance
Note that in both of these cases, the pipe’s potential
pieces of equipment. The maintenance requirements
remains above criteria; therefore, the problem that is
make their operation costly and their reliability some-
affecting the cell may remain unnoticed unless the cur-
what tenuous. For these reasons, a new device was devel-
rent flow through the cell is monitored periodically and
oped to take its place, remedying some of the problems
is verified to be minimal. If not, corrosion of the cell’s
of polarization cells.
positive plates can reduce their surface area and limit
their ac current-carrying capabilities during a fault.
The device, shown on Figure 10-19, is called an isolator-
surge protector (ISP), and it is a solid state device com-
To take care of the limited blocking voltage issue that
posed of three main components: capacitors, inductors,
can occur with polarization cells (and because the pipe
and thyristors. The ISP, like the polarization cell,
to ground potential at each end of the pipe-type cable
replaces the resistor as the connection between the pipe
cannot be determined in advance), it is necessary to
type cable and ground. The ISP’s capacitors can block
design cell cabinets as shown in Figure 10-18. These
up to 12-V dc, a value that makes it possible for one ISP
cabinets can contain more than one cell that can be con-
to be more than adequate for any situation. If this volt-
nected in series, thereby producing a blocking voltage
age should exceed 12 V, the gate circuit would sense this
that is higher than the pipe-to-ground voltage.
increase and would open one of the two thyristors and
effectively short out the ISP and allow dc current to
Finally, polarization cells need to be maintained to
flow. During this time, cathodic protection levels would
insure that they retain their chemical properties. The
drop until the dc voltage returns to a level below the
electrolyte’s specific gravity must be kept at its proper
nominal 12 V, at which point cathodic isolation would
concentration, and the recommended amount of min-
be restored.
eral oil must be kept floating on the electrolyte. In addi-
tion, the sediment that collects at the bottom of the
container must be periodically removed. It is important
to remember that even under normal operating condi-
tions, the dc current flow through the cell is never zero;
therefore, some corrosion of the positive plates takes
place. The sediment is the result of this corrosion activ-
Figure 10-18 Cabinet installed near pothead structure Figure 10-19 The ISP is mounted on the pothead
designed to accommodate up to three polarization cells structure (courtesy Consolidated Edison Company of
(courtesy Consolidated Edison Company of New York, New York, Inc.).
Inc.).
10-19
The flow of ac current, generated by the unbalanced tested and maintained, cathodic protection levels may
load on the three conductors, also has the effect of be compromised.
reducing the voltage that the ISP can block during nor-
mal operations. However, the reduction is rarely enough Voltage-breakdown gaps are used less often than the
to be of concern considering the 12-V starting point. A other grounding methods described. A survey of 26
more important consideration in choosing the right ISP power companies in 1972 concerning corrosion control
is its steady-state current rating with respect to the ac practices (NACE 1972) indicated that only four compa-
unbalance current that is likely to flow through it. If this nies used them on their pipe-type cables.
current exceeds the steady-state current rating of the
ISP, the gate circuit will recognize it as a fault, close the Voltage-breakdown gaps are not standard equipment
thyristors, and short out the ISP. The cathodic protec- for the power industry. Rather, each company usually
tion levels will then fall below criteria and, since the designs and builds its own. For example, in 1977, Com-
unbalance current is continuous, the thyristors may stay monwealth Edison used the equipment shown in Figure
closed for long periods of time. 10-20. This particular gap was designed to break down
at voltages as low as 35 V.
Real ac faults ironically pose fewer problems because of
their short duration. When a fault occurs, it raises the The gap is described (Hymes 1977) as “an aluminum
voltage across the capacitors higher than 12 V. The gate oxide film cutout disc, identical to the type used with
circuit senses this voltage rise and alternately opens each series street lamps.” The disc is backed by copper anvils
of the two thyristors every half cycle, thereby discharg- held against the disc by spring pressure. This enables the
ing the fault current to ground. When the fault clears, installation to conduct the high current generated dur-
the thyristors reset to their open position. ing a fault. This particular voltage gap has two
500-microfarad capacitors connected across the gap to
When lightning strikes in or near substations, the high- prevent the inadvertent breakdown problems that had
voltage surges may pass through the ISP. In such cases, been experienced and that were attributed to switching
the fast voltage rise across the inductor is such that the surges. These capacitors were not entirely successful in
surge protector is placed in operation and diverts most solving that problem, since (Hymes 1977) other means
of the surge current to ground. were being considered.
Field Rectifiers
The ISP is an effective device for this application. It also
As was previously discussed, when the grounding resis-
is a favorite of substation operators since it requires no
tor, which was installed at each substation end of each
onsite maintenance. To ensure that its components are
pipe-type cable, is replaced with one or more polariza-
in good working order, ISPs need to be tested periodi-
tion cells or with an ISP, the associated rectifiers can be
cally and after the feeder experiences a fault. These tests
moved away from the substation and placed (in the
are done using special computer-controlled testing
field) along its route. Doing so eliminates the corrosion
devices developed for this purpose. If any of its compo-
problem that these rectifiers can cause on the substation
nents are found to be damaged, the unit must be sent
ground mats. In addition, installing these rectifiers away
back to the manufacturer for repair.
Voltage-Breakdown Gap
Another cathodic-protection-compatible device that has
been used to “connect” pipe-type cables to ground is
called the Voltage-Breakdown Gap. This device, some-
times referred to as a Spark-Gap, is designed to break
down at a given voltage when a fault occurs and to con-
duct the resulting fault current to ground. It is inserted
in the grounding connection at the station ends of the
pipe-type cable, in effect keeping the pipe and its
cathodic protection system “disconnected” from the
ground mat during normal cable operation. The break-
down in the gap that occurs during a fault causes the
cathodic protection levels to drop because of the direct
connection between the pipe and the ground mat. The Figure 10-20 Example of a voltage-breakdown gap as
gap has no means of resetting itself, so unless they are used by Commonwealth Edison (courtesy Consolidated
10-20
from the substation ends of the feeder results in better available land such as park areas, grassy areas near
current distribution and reduced current density. For highway ramps, and river banks.
these reasons, fewer rectifiers are typically required for
each application. An impressed current system is comprised of an ac cur-
rent source as input to a rectifier. The rectifier’s dc out-
Note: Although the cathodic protection system can be put is connected between the pipe (negative) and a
either galvanic or impressed current, generally pipe-type number of anodes (positive) that are buried in the
cables are protected by impressed current systems. The ground near the rectifier. The rectifier dc current is dis-
main reason is that they need to be protected as a whole charged into the ground by the anodes, travels through
because they cannot be sectionalized into small, isolated the ground, and is picked up by the pipe surface. Once
pipe sections where galvanic systems would be more again, this is nothing more than a corrosion cell whose
appropriate. If galvanic cathodic protection systems cathode is the pipe to be protected. The rectifier voltage
were to be used on these cables, the magnesium anode and current outputs are based on the requirement for
clusters would need to be installed at relatively short each application. The current output is determined by
intervals of 200 to 300 ft. This is because their relatively the current required to protect the pipe. The current
low driving voltage of approximately 1.5 V makes them required is, in turn, based on the total pipe-to-ground
ineffective for much greater distances, especially as the resistance (ohms), which is equal to the unit-coating
coating resistance decreases over time. In addition, gal- resistance (ohms-square foot) divided by the total sur-
vanic anodes also could be a problem for stray currents face area of the pipe (square feet). The rectifier’s voltage
where they could actually increase stray current pickup. output is determined by the anode ground-bed resis-
In any case, galvanic anodes cannot be used in place of tance, which is a function of the soil resistivity and the
the rectifier in resistor-rectifier grounding systems type, number, size, and configuration of the anodes that
because they would not be able to overcome the gal- are buried in that soil.
vanic effect of the substation’s copper ground mat and
the stainless steel (SS) riser pipes. Anode materials for impressed-current systems are cho-
sen based on their ability to discharge current at low
Impressed current systems are comprised of a rectifier rates of consumption. The most often-used materials
and a number of anodes. Ideally, such a system should are graphite and high-silicon cast iron. Anodes made of
be located midway along the cable’s route or should be platinized titanium or niobium also are quite common.
evenly spaced if more than one is required as shown in The common graphite and high-silicon cast iron anodes
Figure 10-21. Realistically, since pipe-type cables are are 5-ft (1.5-m)-long cylindrical rods. The cast iron
located within city environments, the location of these anodes are supplied in a variety of diameters from 1.5-3
impressed current systems is more likely dictated by in. (3.8-7.6 cm), while most graphite anodes are 3 in. in
diameter as shown in Figure 10-22.
Figure 10-21 Field rectifiers are installed along the cable’s

route, outside substations (courtesy Consolidated Edison Figure 10-22 High-silicon cast iron anodes prepackaged
Company of New York, Inc.). in coke breeze (courtesy Consolidated Edison Company
of New York, Inc.).
10-21
These anodes can be installed either vertically or hori- Similarly, the coating resistance should be as high as
zontally. They are installed in groups referred to as possible to minimize the current required from the recti-
ground beds. Each anode is backfilled with a specially fier. This will not only keep the operational cost of the
prepared coke breeze. The purpose of this backfill is to rectifier low, but more importantly will minimize static
lower the anode-to-earth resistance and to provide a interference with other structures. The units have step-
uniform environment for the anode, a condition that down transformers with tap settings so that the output
makes for uniform anode consumption. voltage can be adjusted.
The anodes are manufactured with attached wires. In a Most rectifier units are contained in housings that pro-
typical ground-bed installation, these wires are con- vide convection cooling. In corrosive atmospheres or in
nected to a header cable, which, in turn, is connected to hazardous locations, rectifier units can be oil-immersed
the positive terminal of the rectifier. It is worth noting and provided with explosion-proof fittings. Rectifiers
that although the anodes are the intended current dis- are available with automatic controls that can keep the
charge points, any other metal that is connected to the current output constant in reaction to changes to
positive terminal of the rectifier and is buried will be a ground-bed resistance. These controls also can be set to
point of discharge, including the header cable itself. vary the output of the rectifiers in response to changes
Corrosion of the positive header cable is, in fact, the in pipe-to-soil potential on the pipe. Although this fea-
most common cause of failure in these impressed cur- ture can be useful in the short run, changes in pipe-to-
rent systems. This is why, as a general rule, the insula- soil potentials should be investigated in depth as to their
tion on the header cable should be the best available, cause and not just addressed by adjustments to the out-
and extra care should be taken to avoid damaging it put current of the rectifier. The most useful option that
during installation. can be purchased with rectifiers is a remote monitoring
system. These systems make it possible to monitor the
Rectifiers used for these applications can be obtained in output of the rectifier often, if not continuously. These
a large variety of output voltage and current ratings. monitors also can be wired to permanent reference cells
Standard units with current ratings of less than to continuously monitor the actual pipe-to-soil poten-
10 amperes to more than 250 amperes and voltage tial of the pipe and transmit or record the readings for
ratings of 10 V to more than 100 V are listed in the specific time intervals. Such data can be very useful to
catalogs of the manufacturers. Most units are single- insure that the rectifiers operate properly and that
phase, with the bridge connected as shown in Figure cathodic protection of the pipes is not interrupted.
10-23, although larger capacity units are often three-
Galvanic Protection Systems—Magnesium vs. Zinc
phase units.
Anodes
Although the cathodic protection system can be either
It is good practice to design the system so that the
galvanic or impressed current, generally pipe-type cables
ground-bed resistance is as low as practical. This will
are protected by impressed current systems. The main
minimize the output voltage from the rectifier needed to
reason is that they need to be protected as a whole,
obtain the required current discharge from the anodes.
because they cannot be sectionalized into small, isolated
pipe sections where galvanic systems would be more
appropriate. If galvanic cathodic protection systems
were to be used on these cables, the magnesium anode
clusters would need to be installed at relatively short
intervals of 200 to 300 ft. This is because their relatively
low driving voltage of approximately 1.0-1.5 V make
them ineffective for much greater distances, especially
when soil resistivity and current requirements are
considered. In addition, galvanic anodes could also be a
problem in stray current areas, where they could actually
increase stray current pickup. In any case, galvanic
anodes cannot be used in place of the rectifier in resistor-
Figure 10-23 Schematic of typical cathodic protection rectifier grounding systems because they would not be
rectifier (courtesy Consolidated Edison Company of New able to overcome the galvanic effect of the substation’s
York, Inc.). copper ground mat and the substation riser pipes.
10-22
If galvanic anodes are determined to be appropriate for formly. It lowers the anode-to-earth resistance, thereby
a particular application, the following guidelines should increasing the current. It retains moisture, which main-
be used in determining which galvanic metal to use. tains a lower resistance. It also acts as a depolarizing
agent and prevents the anode from going passive.
The protective current generated by galvanic anodes
depends upon the inherent voltage that exists between 10.3.6 Stray Current Mitigation
the anodes and the structure, as well as the resistivity of In cases where pipe-type cable is affected by static stray
the soil in which the anodes are installed. Galvanic currents that emanate from other cathodic protection
anodes consume themselves in the process of generating systems or other steady-state sources, the solution is to
the protective current. This is why the system is often communicate the problem to all parties involved and try
referred to as a sacrificial anode system. The rate of to rectify the situation to everyone’s satisfaction.
consumption depends upon the amount of current gen-
erated as well as the material of the anode. For example, The solution could entail moving the offending rectifier
the theoretical consumption rate of zinc is 23.5 lbs to a different location. It also could include bonding the
(10.7 kg) per ampere-year, while that of magnesium is pipe-type cable to the rectifier’s negative terminal, in
8.7 lbs (4 kg) per ampere-year. In actual practice, not all effect making it part of the protected structure. If these
of the metal that is consumed goes to generate cathodic
protection current. Some of the metal is consumed in
self-corrosion. Zinc is approximately 90% efficient, and
magnesium is approximately 50% efficient. Therefore
the actual corrosion rates for these metals are 26 lbs
(11.8 kg) of zinc per ampere-year of protective current
and 17 lbs (7.7 kg) of magnesium per ampere-year.
Magnesium is more widely used than zinc because of its

higher driving voltage. The solution potential of
magnesium is 1.55 V to a copper-copper sulfate
magnesium reference electrode, while that of zinc is
1.10 V. Therefore, the use of zinc has been limited to
very low-resistivity soils and to seawater applications.
Zinc is rarely effective in soils with resistivities of more
than 1500 ohm-cm unless the pipe is very well-coated
and the current requirement for protection is small. Figure 10-24 Distributed magnesium anodes (courtesy
Magnesium, on the other hand, can be used quite Consolidated Edison Company of New York, Inc.).
effectively in soils of 5000 ohm-cm, and on well-coated
pipe, can be used in soils of 10,000 ohm-cm or more.
Galvanic anodes are buried directly in the soil at inter-

vals along the pipeline, as shown in Figure 10-24. The
distance away from the pipe is usually 3-5 ft (1-1.5 m).
The anodes are supplied prepackaged in a cloth bag in a
specially prepared backfill mixture and are equipped
with a connecting wire connected directly to the pipe.
The anodes may also be connected to each other with a
header cable, which, in turn, is connected to the pipe, as
shown in Figure 10-25. Very often a shunt is inserted in
the connecting cable so that the current output can be
read. The shunt is housed in a test box, as shown in Fig-
ure 10-26.
The backfill mixture most often used is 70% gypsum,

Figure 10-25 Distributed galvanic anodes with header
25% bentonite, and 5% sodium sulfate. This backfill cable (courtesy Consolidated Edison Company of New
serves a number of purposes. It provides a uniform envi- York, Inc.).
ronment for the anode so that the anode corrodes uni-
10-23
steps cannot be agreed on, as a last resort another recti- Figure 10-29 is a typical graph and shows the when the
fier could be installed on the cable to counter the effect pipe-to-negative bus potential moves in the positive
of the offending rectifier. This should not be necessary if direction, so does the pipe-to-soil potential. From this it
all parties are cooperative. is reasonable to infer that, since the pipe-to-negative bus
potential is the driving force for current flow between
Where dynamic stray currents are present, pipe-type the pipe and the substation, it also is the cause for pipe-
cables are particularly affected because of their length to-soil potential becoming more positive as a result of
and electrical continuity. As previously mentioned, current discharge. The degree of correlation between the
dynamic stray currents result from the operations of dc two readings can be calculated or can be deduced from
transit and commuter railroads traveling to and from as the shape of the X-Y graph. Good correlation results
well as within metropolitan areas. This is the same envi- are in a thin straight line, while a wide line indicates less
ronment where pipe-type cables are found. correlation. In cities with many subways, commuter
The areas of the cable that pick up or discharge current

are readily visible on its potential profile.
As shown on Figure 10-27, the pickup areas are negative

and occur over large areas, while discharge areas are
positive and are more concentrated. To correct the situ-
ation, it is necessary to first find what is causing the
problem. It is often the case that when the cable is in the
proximity of transit substations, the cause is obvious.
However, best practices dictate that interference testing
be done in cooperation with transit personnel so that
the proper solution can be agreed on.
Interference testing consists of measuring the pipe-to-

soil potential of the pipe simultaneously with the poten-
tial between the pipe and the negative output terminal
(bus) of the transit substation as shown on Figure 10-28. Figure 10-27 Pipe-to-soil potential profile measured along
the cable’s route (courtesy Consolidated Edison Company
of New York, Inc.).
By plotting an X-Y graph of the two measurements, con-
clusions can be made as to the cause of the interference.
Figure 10-28 Interference tests are performed in discharge

Figure 10-26 Galvanic anode cluster installed with areas (courtesy Consolidated Edison Company of New
current measuring shunt (courtesy Consolidated Edison York, Inc.).
Company of New York, Inc.).
10-24
trains, and crisscrossing pipe-type cables, interference metallic path was added, the slope became negative.
tests rarely yield graphs with good correlation. This is From this, the following can be deduced:
because the cable being tested is almost always affected • A metallic connection between the pipe and the nega-
by stray currents from different substations, not just the
tive bus will keep the pipe from discharging stray cur-
one where the interference tests are being performed.
rent into the soil.
Having established that stray currents are being dis- • The resistance of the metallic path must be low
charged from the pipe into the ground so that they can enough that the slope of the X-Y graph of pipe to soil
return to their substation source, the solution is to pre- vs. current flow is negative.
vent this current discharge by providing it with a better
return path to the substation. The most obvious way to
do this is to install a cable connecting the pipe and the
negative bus of the transit substation as shown in Figure
10-30. Substituting this metallic path for the electrolytic
earth path would prevent current discharge into the soil
and, therefore, prevent corrosion of the pipe surface.
The effect of this connection can be demonstrated by

again recording the pipe-to-soil potential simulta-
neously with the current flow on the cable. Figure 10-31
shows the X-Y plot of these readings, and it shows that
the pipe-to-soil potential becomes more negative as cur-
rent in the bond cable increases. By comparing the two
Figure 10-30 Cable connection is made between pipe-
graphs, it is apparent that when the current was
type cable and transit substation (courtesy Consolidated
returned through the electrolytic (more resistive) path, Edison Company of New York, Inc.).
the slope of the X-Y graph was positive, but when a
Figure 10-31 X-Y graph of PS vs. drain current indicates

“overcompensation” (courtesy Consolidated Edison
Figure 10-29 X-Y plot indicates correlation between PS Company of New York, Inc.).
and P-negative bus readings (courtesy Consolidated
10-25
The metallic connection between the pipe and the tran- This differs drastically with the only other method that
sit substation’s negative bus is appropriately called a has been used to counter the effects of stray currents,
stray current drain bond, and its optimum resistance is the interference rectifier. This is an impressed-current
that which results in an X-Y plot with a vertical slope. A system (as previously described) whose function is to
drain bond of proper resistance will allow stray current change the potential of the pipe in the negative direction
to return to its source without any change in the pipe- by the amount that the stray currents change it in the
to-soil potential. The value of drain bond resistance is a positive direction. The problem is that the rectifier out-
function of the pipe-to-ground resistance and pipe-to- put is constant and must be designed to counter the
negative bus resistance, two values that are difficult to maximum interference level even though this only
measure. Even if they were measured, the results would occurs during a very small portion of the day. The effect
be approximate, at best. Therefore, the best way to is that most of its current output does not benefit the
determine the resistance of a drain bond is to do so pipe and may cause damage to neighboring pipes
empirically. Again, in cooperation with transit person- through static interference as previously discussed.
nel, a temporary drain bond is installed in series with an
adjustable resistor. X-Y plots are generated for various The second point is that it is incumbent on the operator
resistance values until the vertical slope is achieved as of the pipe-type cable to survey the drain’s performance
shown in Figure 10-32. A drain having this resistance is often to insure that the levels of current and pipe-to-soil
then installed permanently. The installation is com- potential remain relatively constant. Any significant
pleted with the addition of a diode to the circuit. The change usually is the result of decreased pipe-to-ground
diode allows current to flow only from the pipe to the resistance, which can be the result of accidental contacts
substation. It is required to prevent the drain bond itself between the pipe and other buried metallic structures.
from becoming a path for stray current flow onto the These must be located and cleared as soon as possible.
pipe as shown in Figure 10-33.
10.3.7 Pipe to Soil Potential Surveys
There are two important points to remember when After the cathodic protection system is energized, it is
installing a stray current drain bond. The first is that it necessary to evaluate its effectiveness. This is done by
is the most effective way to mitigate the effects of stray performing a potential survey over the entire length of
current interference because it operates automatically. the pipe-type cable. This survey consists of measuring
Its operation is controlled by the stray current activity pipe-to-soil potential at every test station installed along
itself, which changes drastically at different times during the route of the cable, typically near joint manholes.
the day or night. It only functions when stray currents These test stations consist of a small metal box that, in
are present and is completely passive when they are not. most cases, is mounted flush with the pavement. In
some cases, where the environment permits, these test
Figure 10-33 A drain bond is comprised of the cable, the

resistor, and the diode (courtesy Consolidated Edison
Figure 10-32 A drain bond of proper resistance will keep Company of New York, Inc.).
the pipe-to-soil potential unchanged at any level of drain
current (courtesy Consolidated Edison Company of New
York, Inc.).
10-26
stations can be mounted above grade. In either case, back to the basic corrosion cell, it is apparent that by
they contain a wire that is connected to the pipe below coating the cathode, the pipe is isolated from the electro-
and have an open bottom through which the reference lyte, thereby preventing current flow and stopping the
cell can make contact with the earth. Readings taken at corrosion process. In fact, if the cathode (the pipe) were
these test stations only reflect the potential of a small perfectly coated and were to remain so forever, nothing
pipe section below the test station; however, they are further would need to be done. There would be no need
representative of the potentials over the entire surface of for cathodic protection. Unfortunately, this is never the
the pipe. This is a reasonable assumption if the condi- case. Protective coatings on pipe-type cables are sub-
tion of the coating and the type of backfill are generally jected to more severe conditions than coatings on other
uniform and if the protective current source (rectifier- pipelines for the following reasons:
ground bed) is some distance away from the pipe. How- • Cables often operate at elevated temperatures. There-
ever, this may not be the case, and it is the purpose of a
fore, the coatings used on pipe-type cables must be
qualified person to determine if a pipe-to-soil profile
able to withstand greater heat than is usually the case
along the route of the cable is, in fact, representative of
for underground pipelines.
the cathodic protection status of the entire pipe surface.
Such a person may call for additional readings to be • Coatings must possess thermal characteristics that
taken between test stations to get a better picture of the permit them to dissipate that heat to the surround-
situation. The additional readings that may be called for ing soil.
may range from several to many. The ultimate potential • Coatings are often subjected to the momentarily high
survey is called a close interval survey (CIS) whose electrical potentials resulting from switching surges
objective is to actually measure the potential of the and fault currents. The coating must have sufficient
entire pipe surface. Such a survey is one of the most use- dielectric strength to resist puncturing.
ful tools in finding cathodic protection problems on any
pipeline, ranging from stray current interference over Four basic types of coatings have been used successfully
large areas to finding coating damage. on pipe-type cables (NACE 1972). These are asphalt
mastic, coal-tar enamel, polyethylene, and polypropy-
A CIS survey is performed by taking successive pipe-to- lene. Thin film or fusion-bonded epoxy coatings have
soil potential measurements with the reference electrode not received acceptance for pipe-type cables because of
placed on the soil above the pipeline at closely spaced their relatively low electrical strength in resisting punc-
intervals. The generally accepted interval length is 3 ft to ture from switching or circuit failure. The most common
5ft. These measurements are taken with cathodic protec- coatings used today are the plastic coatings.
tion applied to the pipe. As the reference electrode
advances along the pipeline, the potential readings are Asphalt mastic and coal tar are hot-applied coatings
substantially the same at each successive location if the that are bonded to a pipe whose surface has been
coating is intact. If there is a holiday in the coating in cleaned (grit or shot-blasted) and primed (NACE 1977).
the vicinity of a given reference electrode placement, the The asphalt mastic most commonly used for pipe-type
value of the potential reading at that location will be less cable is an air-blown asphalt mixed with a sand aggre-
negative than the readings along the remainder of the gate and a mineral filler, reinforced with glass fiber
pipeline. This also is true if the potential is being (NACE 1976). This coating, applied to a thickness of
affected by stray current discharge, either over a large 0.5 in. (1.3 cm) is marketed under the trade name
area or localized over a short distance. In urban areas Somastic. However, it is susceptible to brittleness at low
where the pipe-type cable may be buried under pave- temperatures. Somastic was the most widely used coat-
ment, performing such surveys is more difficult and the ing for pipe-type cable through the early 1970s and has
result may be confusing. However, if proper steps are the longest history, dating from a Detroit Edison instal-
taken during data collection and if the data are properly lation in 1941 (Sinclair 1956).
interpreted, these surveys can be very useful even in
paved areas. Coal-tar enamel also has had a long history of
successful use, particularly in the general pipeline
10.3.8 Coatings industry. On most pipelines, coal-tar enamel was usually
Although ISPs, polarization cells (PCs), rectifiers, and applied in accordance with American Water Works
stray-current bonds are necessary components of a Association Standard C203 (AWWA 1986), which
good cathodic protection system, none of these could required a thickness of 0.094 in. (0.24 cm). However,
effectively protect a pipe if it were not for a high-quality many operators of pipe-type cables specified greater
dielectric coating being applied to its surface. Referring thicknesses, varying from 0.125-0.1875 in. (0.31-
10-27
0.48 cm) (NACE 1972). Most coal-tar enamel coatings range of -21 to 82°C. Plastic coatings are susceptible to
were reinforced with fiberglass and had an outer wrap of deterioration resulting from ultraviolet radiation when
felt and kraft paper. stored outdoors. The addition of carbon black to the
resin increases the resistance of the coating to the ultra-
As is the case with asphalt mastics, coal-tar enamels violet radiation.
also are susceptible to low-temperature brittleness.
There are two tests that are performed to insure that
Plastic coatings, both polyethylene and polypropylene, pipe coating is in the best possible condition after it is
are referred to generically as extruded polyolefin-resin installed and backfilled. The first test is called a holiday
coating systems (NACE 1985). There are two basic detector (jeep) test. This test uses an instrument to
methods used for extruding the resin onto the pipe. In apply high-voltage dc across the coating over its entire
one method, the resin coating is crosshead-extruded on surface. For coatings such as extruded polyolefin with a
the pipe in the shape of a continuous sleeve over a hot- thickness of 60 mils or more, the voltage applied is 18–
applied mastic adhesive. In the other method, the resin 20 kV. This test is intended to find small pinholes that
coating is side-extruded and spirally wrapped and fused are missed by visual inspection. If any such “holidays”
as a seamless coating over a butyl rubber adhesive. Fig- are found, they are repaired before the pipe is lowered
ure 10-34 shows a sandblasted pipe as it rotates under into the trench. The pipe sections are welded together,
two extruding machines. The first wraps a film of butyl after which the joints are cleaned and field-coated, typi-
rubber on the pipe surface that is immediately covered cally with cold-applied tapes or shrink sleeves. The
with a film of polyethylene from the second extruding coated joints, and sometimes the entire pipe, are again
machine. The coating is then cooled by water nozzles jeeped immediately prior to backfill.
that extend from the blue header as the coated pipe
moves to the right. After backfill, a second test is performed to ensure that
the coating was not damaged during backfilling. This
Although 40 mils (1 mm) is indicated as the standard test, called the coating resistance test, is performed by
sheath thickness for pipe sizes of 6-in. (15-cm) diameter applying a known amount of current to the pipe section
or more, most operators have specified a 60-mil and measuring the pipe-to-soil potential of the pipe
(1.5-mm) sheath thickness. The high-density polyethyl- before and after current application. The change in
ene coating is available in thicknesses from 50-250 mils potential divided by the current gives the pipe-to-
(1.3-6.4 mm). ground resistance of the pipe section. The surface area
of the buried pipe is calculated and multiplied by the
These coatings do not have the low-temperature limita- pipe-to-ground resistance. The result is a unit resistance
tions of asphalt and coal tar. The polyethylene coating value for the coating, which, if not damaged, can easily
has a continuous operating range of -40 to 74°C (or as reach 2,000,000 ohms-square feet. Although some com-
high as 82°C, depending upon the manufacturer). panies specify such high values of coating resistance,
Polypropylene is said to have a continuous operating each company should set its own value of coating resis-
tance that they are willing to accept.
10.3.9 Disbonded Coating

Coatings are essential components of any cathodic pro-
tection system; however, they have one characteristic
that not only can render such systems ineffective but can
actually contribute to the corrosion of the pipe surface.
To be effective, coatings must be well bonded to the pipe
surface when first applied and they must remain well
bonded. When coating is completely missing, such as
might be the case when the coating is damaged during
third-party excavations, the exposed surface of the pipe,
although in contact with the soil, is protected by the
direct pickup of cathodic protection current. However,
in cases of disbondment, the coating is not missing; it is
Figure 10-34 Butyl rubber adhesive is side-extruded over simply not adhered to the surface of the pipe. Therefore,
pipe surface, and polyethylene is extruded over the water can penetrate under it and support galvanic cor-
adhesive (courtesy Consolidated Edison Company of New
York, Inc.). rosion of the pipe surface as shown in Figure 10-35.
10-28
This condition is one that cannot be overcome by the • Third-party excavations: It often happens that third-
application of cathodic protection because the coating party contractors do not adhere to accepted con-
prevents cathodic protection current from reaching the struction practices. They don’t make use of the one-
pipe surface. Pipe-to-soil potential readings that are call system. They expose and work around these
measured annually over these feeders do not reveal these pipes with little regard to the coating, and backfill
areas of disbondment because they are not part of the these excavations with anything available: rocks,
“circuit” through which the applied cathodic protection bricks, chunks of concrete, trash, and other debris.
current flows. The wet area under a section of dis-
bonded coating can be thought of as an environment There are two other factors that may contribute to coat-
that is separate from the backfill in which the rest of the ing disbondment. The first is temperature, and the sec-
pipe is buried. ond is excessive levels of cathodic protection. Intuitively,
it is understood the elevated temperatures at which pipe-
There are several projects being undertaken to better type cables operate should affect the coating adhesion
understand this problem and to find a solution for it, over time. Although temperature is a contributing fac-
but none have been found to date. Field observations of tor to other primary causes, there is no evidence that the
corrosion problems have resulted in some understand- operational temperature of pipe-type cables by itself
ing of why coatings disbond. The causes can be placed causes coatings to disbond. If it did, the problem would
in three groups: be much larger and much more widespread.
• Coating application: It has been observed that dis-
bonding problems have occurred at field joints where As for high levels of cathodic protection—referred to in
two pipe sections are welded together. This is not sur- the previous discussion of corrosion cells—when a
prising, since regardless of the type of coating used, corrosion cell is formed, hydrogen is generated at the
coating application at field locations is not of the cathode. It is thought, therefore, that since the protected
same quality as it is in the controlled environment of pipe is the cathode of a corrosion cell, the hydrogen that
a factory as shown in Figure 10-36. is generated at its surface may migrate under the
coating, forcing it to separate from the steel. This
• Mechanical damage: This generally results from poor condition cannot be discounted; however, it is most
construction practices such as not maintaining likely to occur at the edges of a pre-existing coating fault
proper clearances between the pipe and adjacent and generally doesn’t extend too far under the coating
structures. Often, pieces of micarta, wood planks, or from that point.
similar materials have been used as substitutes for
several inches of clean backfill to separate pipe-type 10.3.10 Atmospheric Corrosion
cables from other underground structures such as
The term “atmospheric corrosion” is applied to corro-
concrete ducts, metal conduits, foundation walls, sub-
sion that occurs on metal structures that are not buried
way tunnels, or even protection plates that are used
when depth-of-cover is shallow. The result is that the
weight of the pipe on these boards or the vibration
from vehicle traffic over these pipes or subway trains
under these pipes cracks the coating, allowing water
to penetrate under it.
Figure 10-35 Missing vs. disbonded coating (courtesy

Consolidated Edison Company of New York, Inc.). Figure 10-36 Corrosion under disbonded, field-applied
coal-tar coating (courtesy Consolidated Edison Company
of New York, Inc.).
10-29
or submerged. It is galvanic in nature in that it is not backfilling the manhole with moist sand as shown in
caused by any interference from outside sources; rather, Figure 10-38. This would allow cathodic protection cur-
it is caused by the metal’s reaction with its environment, rent to reach the pipe inside the manhole.
which is the moisture that is contained in the “atmo-
sphere” such as humidity and rain. Pipe-type cables can This step should only be considered after the pipe coat-
be susceptible to this type of deterioration in areas ing has been inspected for any sign of disbondment and
where they are not buried. These include inside man- repaired if necessary.
holes and tunnels, hanging from bridges over highway
or rivers, and, under certain circumstances, inside cas- 10.3.11 Casings
ings. Unlike buried pipes, ambient temperature also is Although they facilitate pipe installation and provide
an issue in these cases because the temperature extremes them with a certain measure of mechanical protection,
between winter and summer cause the pipe and the casings can and often do affect the level of cathodic pro-
coating to expand and contract at different rates, result- tection that these pipes receive. When a pipe is installed
ing in the coating cracking and disbonding. Alternating inside a casing, it is imperative that the pipe be kept
humid, wet, and dry conditions in these areas are ideal electrically isolated from the casing by means of appro-
for continued corrosion attack of the pipe surface and priate skids that are attached to the pipe before it is
often result in failures as show in Figure 10-37. inserted into the casing. Immediately after the pipe is
fully inserted, the resistance between the pipe and the
These failures cannot be prevented by conventional casing must be measured to verify that isolation
cathodic protection methods because cathodic protec- between the pipe and casing has been maintained. The
tion, by definition, involves the pickup of dc current by annular space at both ends of the casing must then be
the protected surface directly from the electrolyte that sealed to prevent moisture from entering the casing.
surrounds it. Therefore, only structures that are buried
or submerged can be cathodically protected. Since pipes Not withstanding these initial precautions, the following
on bridges and generally in manholes are not buried and conditions can occur over time that will affect the corro-
cannot pick up this current, they cannot be protected in sion status of the pipe.
this manner.
• The casing remains sealed, keeping the annular space
For pipes on bridges or in tunnels, the standard method clean and dry. Under this idyllic condition, corrosion
to prevent atmospheric corrosion is to thoroughly will not take place since there is no electrolyte in con-
inspect them and insure that the coating has not deterio- tact with the surface of the pipe.
rated or become disbonded from the pipe surface. If • The casing seals fail, but the casing becomes only
such conditions are found, the coating must be partially filled such that the water does not come into
removed, the pipe surface inspected, repaired (if neces- contact with the surface of the pipe. This condition
sary), and recoated. This process must be repeated peri- results in a humid environment inside the casing that
odically to insure that corrosion does not progress to the will support atmospheric corrosion of the pipe sur-
point of failure. The same applies for pipes in manholes;
however, there is another step that can be considered:
Figure 10-37 Atmospheric corrosion of pipe surface in Figure 10-38 Cathodic protection current can reach pipe
joint manhole (courtesy Consolidated Edison Company of surface inside a backfilled manhole (courtesy
New York, Inc.). Consolidated Edison Company of New York, Inc.).
10-30
face. The only defense against this condition is to Despite this lower average corrosion rate, lead can
ensure that the pipe is well coated when first installed become very reactive in poorly aerated soils as well as in
since, unlike pipes in manholes or on bridges, it can- soils high in organic acids. Like steel pipe, lead-sheathed
not be regularly inspected. cable is vulnerable to dissimilar metal corrosion when
connected to copper grounds. It also is susceptible to
In both of these cases, a metallic contact between the differential oxygen attack, particularly when installed in
pipe and the casing will not affect the pipe inside the conduit ducts that become infiltrated with mud and that
casing but will affect the cathodic protection levels of are not properly drained.
the buried pipe outside the casing because the bare cas-
ing acts as a large coating fault on the pipe. Lead is an amphoteric metal (it is capable of reacting
• The casing becomes filled with water, thereby fully or with either an acid or an alkali). Thus, unlike steel, lead
is susceptible to severe corrosion in highly alkaline envi-
partially submerging the pipe. If the pipe is under
ronments (Uhlig 1969). For this reason, lead-sheathed
cathodic protection, this situation is actually the
cables are particularly vulnerable to corrosion from
most favorable one. Here the cathodic protection cur-
stray dc currents. At locations where current leaves the
rent can reach the pipe by traveling through the water
cable, the sheath is made anodic, and the lead can
in the annular space and preventing corrosion of its
corrode at a rate of 75 lbs (34 kg) per ampere-year. At
surface. This also would apply if ampacity consider-
locations where the cable collects current, the sheath is
ations require that the casing be filled with grout. The
made cathodic, and the environment is made more
grout is an electrolytic conductor and would carry
alkaline. An accumulation of highly alkaline products
the cathodic protection current from the casing to the
can generate corrosion of the sheath. Cathodic
pipe surface. However, care should be taken not to
corrosion of lead is indicated by the presence of red or
leave voids in the grout that could create galvanic
yellow corrosion products.
cells, as was discussed previously.
• The casing is full of water and in metallic contact Because of this vulnerability, almost all electric utilities
with the pipe. The effect is that the cathodic protec- with lead-sheathed cables in metropolitan areas at the
tion current that is picked up by the surface of the time when street railway operation was at its height
casing travels through two separate paths on its way (1940s and 1950s) conducted tests and performed sur-
to the pipe surface: the metallic path created by the veys to determine the corrosion conditions. A 1954
contact and the electrolytic path created by the water. report summarizing the practices of 22 companies
The amount of current in each of these parallel paths issued by the National Association of Corrosion Engi-
is inversely proportional to their resistance. As the neers (NACE 1954) indicated that most companies
resistance of the metallic path decreases, the amount relied on potential measurements for that determina-
of protective current that reaches the pipe surface via tion. The major concern was to find locations where
the electrolytic path (water) also decreases. anodic conditions existed as indicated by positive cable
potentials. Some companies also were concerned about
10.3.12 Lead-Sheathed Cable locations where the cables were highly cathodic as indi-
Although lead is an active metal in the galvanic series, it cated by highly negative potentials.
becomes passive in many corrosive environments
because of the formation of relatively insoluble lead Corrosion control practice at that time consisted
compounds (Uhlig 1985). This passivity probably mainly of measures to mitigate the adverse effects of
accounts for the reasonably good corrosion experience the stray current. In the process, many cables in stray-
with lead-sheathed cable usage in the electric power current areas benefited by receiving inadvertent
industry. On average, lead placed underground corrodes cathodic protection. But with the decline in the
less than steel. A series of field tests was conducted by operation of street railway systems, these benefits also
the U.S. National Bureau of Standards (now the declined, and many miles of bare, lead-sheathed cable
National Institute of Standards and Technology) were consequently exposed to galvanic corrosion
between 1910 and 1950 in almost all types of soil conditions. Some companies instituted cathodic
throughout the United States. These tests showed that protection programs that have been successful in
the average corrosion rate of steel in 44 soils was preventing this galvanic corrosion. But the use of
approximately 9 times the average corrosion rate of lead cathodic protection for cables by the power industry, in
in 21 soils (Romanoff 1957). general, has been limited. General practice has been to
rely on jacketed cable for replacement of corroded cable
and for use in most new installation work. Protective
10-31
jacketing for cable is available in either PVC or polyeth- Despite these difficulties, some companies have reported
ylene extruded over the lead sheath. Jacket thicknesses success with cathodic protection programs as simple as
range from 80-110 mils (2.0-2.8 mm). Some suppliers installing a magnesium anode at each manhole. Other
call for a reduction in the thickness of the lead sheath c o m p a n i e s h ave i n s t a l l e d m o r e s o p h i s t i c at e d ,
when a jacket is specified. Others cover the jacket with impressed-current systems. These latter systems often
a thin layer of bituminous compound. use strings of high-silicon, cast-iron anodes installed in
spare ducts of conduit runs. The most common criterion
The limited use of cathodic protection is probably for determining cathodic protection calls for a cable
because cathodic protection of bare lead-sheathed cable potential of -0.70 V measured to a copper-copper sul-
is inherently more complicated than is the case for fate reference electrode. Another popular criterion is a
coated steel pipe. First, more current is required. Sec- voltage shift of 0.25 V. To avoid the possibility of
ond, cables are often installed in conduit rather than cathodic corrosion, the cable potential should not be
directly buried; therefore, proper distribution of that more negative than –1.5 V to a copper-copper sulfate
current is more difficult. Third, the risk of cathodic cor- reference electrode (NACE 1990).
rosion exists because of the amphoteric nature of lead.
10-32
REFERENCES Halperin, H. and K. W. Miller. 1929. “Reduction of

ANSI/IEEE Standard 575-1988. 1987. The Institute of Sheath Losses in Single Conductor Cables.” AIEE
Electrical and Electronics Engineers. “IEEE Guide for Transactions. Vol. 48. April.
the Application of Sheath-Bonding Methods for Single-
Conductor Cables and the Calculation of Induced Volt- Hassler, S. P., et al. 1984a. “M.O.V. Arrester Protection
ages and Currents in Cable Sheaths.” New York. of Shield Interrupts on 138 kV Extruded Dielectric
Cables.” IEEE Transactions on Power Apparatus and
AWWA. 1986. American Water Works Association. Systems. Vol. 103, pp. 3334-3341. November.
“Coal-Tar Protective Coatings and Linings for Steel
Water Pipelines—Enamel and Tape—Hot-Applied.” Hassler, S. P., et al. 1984b. “Shield Interrupt Overvolt-
Standard C203-86. ages on 138 kV Extruded Dielectric Cables.” IEEE
Trans actions on Power Apparatus and Systems. Vol. 103.
Ball, E. H., E. Occhini, and G. Luonim. 1965. “Sheath pp. 3327-3333. November.
Overvoltages in High Voltage Cables Due to Special
Sheath Bonding Connections.” IEEE Transactions on Hymes, R. W. 1977 “Pipe-Type Cable—Construction,
Power Apparatus and Systems. Vol. 84. Coating and Operating Experience.” Materials Perfor-
mance. Vol. 16. No. 4. pp. 32-37.
CIGRE. 1982a. IEC Publication 229. “Tests on Cable
Oversheaths Which have a Special Protective Function ICEA. 1969. Insulated Cable Engineers Association,
and are Applied by Extrusion.” Geneva. Inc. “Short-Circuit Characteristics of Insulated Cable-
1969.” ICEA Publication P-32-382. Massachusetts.
CIGRE. 1982b. IEC Publication 287. Second Edition.
“Calculation of the Continuous Current Rating of Itoh, Y., N. Nagaoka, and A. Ametiani. 1950. “Tran-
Cables (100% Load Factor).” Geneva. sient Analysis of a Cross-bonded Cable System Under-
neath a Bridge.” IEEE Transactions on Power Delivery.
CIGRE. 1989. IEC Publication 853-2. First Edition. Vol. 5. No. 2. April.
“Calculation of the Cyclic and Emergency Current Rat-
ing of Cables, Part 2: Cables Rated above 36 kV.” Kuwahara, K. and C. Doench. 1963. “Evaluation of
Geneva. Power Frequency Sheath Currents and Voltages in Sin-
gle Conductor Cables for Various Sheath Bonding
CIGRE. 1990. Working Group 07 of CIGRE Study Methods.” IEEE Transactions on Power Apparatus and
Committee 21. “Guide to the Protection of Specially Systems. Vol. PAS 82.
Bonded Systems Against Sheath Overvoltages.” Electra.
No. 128, pp. 48-61. Marti, L., T. E. Granger, and A. Morched. 1996.
“Sheath Over-voltages in HV Underground Cables.”
CIGRE. 2004. CIGRE Technical Brochure No. 250. CIGRE Paper 33-201.
Working Group B1.19. “General Guidelines for the
Integration of a New Underground Cable System in the NACE. 1954. National Association of Corrosion Engi-
Network.” August. neers. “Tests and Surveys for Lead-Sheathed Cables in
the Utilities Industry.” Publication 4B254.
CIGRE. 2005. CIGRE Technical Brochure No. 283.
Working Group B1.18. “Special Bonding of High Volt- NACE. 1972. National Association of Corrosion Engi-
age Power Cables.” October. neers. “Survey of Corrosion Control Practices for Pipe-
Type Cables in the Utility Industry.” Publication
Dommel, H. W. 1994. “The EMTP Theory Book, Sec- 10A172.
ond Edition.” The University of British Columbia.
NACE. 1976. National Association of Corrosion Engi-
Emin, Z., P. K. Basak, a nd C. Ferguson. 2 001. neers. “Extruded Asphalt Mastic Type Protective Coat-
“Simulation Studied to Improve Design for Mid-life ings for Underground Pipelines.” Standard RP0276-76.
275 kV Cable Refurbishment.” Paper submitted to
IEEE. August. NACE. 1977. National Association of Corrosion Engi-
neers. “Application of Organic Coatings to the External
Surfaces of Steel Pipe for Underground Service.” Stan-
dard RP0275-7.
10-33
NACE. 1985. National Association of Corrosion Sinclair, W. A. 1956. “Corrosion Control Practices for
Engineers. “Extruded Polyolefin Resin Coating Systems Pipe-Type Cables on the Detroit Edison System.” Cor-
for Underground or Submerged Pipe.” Standard rosion. Vol. 12. No. 8. pp. 31-34.
RP0185-85.
Uhlig, H. H., Ed. 1969. Corrosion Handbook. New
NACE. 1990. National Association of Corrosion Engi- York: John Wiley & Sons. Chapter by R. M. Bums.
neers. “Criteria and Test Methods for Cathodic Protec- “Corrosion of Lead and Lead-Alloy Cable Sheathing.”
t i o n o f L e a d S h e at h C ab l e. ” P ro p o s e d NAC E
Recommended Practice. Houston, Texas. Uhlig, H. H. 1985. Corrosion and Corrosion Control.
Third Edition. New York: John Wiley & Sons. p. 358.
Romanoff, M. 1957. Underground Corrosion. U.S.
National Bureau of Standards, Circular 579.
10-34
CHAPTER 11 Ampacity
Authors: Earle C. (Rusty) Bascom, III, Power Delivery Consultants, Inc. (PDC)
Anthony Ernst, Underground Systems Inc.(USi)
Reviewers: David Purnhagen, Underground Systems Inc.(USi)
Brian Gregory, Cable Consulting International Limited
Rachel I. Mosier, Power Delivery Consultants, Inc. (PDC)
Ampacity calculations form the heart of transmission-cable design since the ultimate
purpose of a cable circuit is to transfer the desired amount of power. This chapter
describes the calculation procedures, reviews the principles for transient ampacity calcu-
lations, discusses forced cooling, and briefly covers the procedures for ampacity audits
and uprating.
Earle C. (Rusty) Bascom, III is a senior engineer with Power Delivery

Consultants, Inc. He holds an A.S. (1987) in engineering science from
Hudson Valley Community College, Troy, New York, a B.S. (1989), and
M.E. (1990) in electric power engineering from Rensselaer Polytechnic
Institute, Troy, New York, and an M.B.A. (1993) from the State Univer-
sity of New York at Albany. Mr. Bascom began his career in 1990 work-
ing with underground cable systems for Power Technologies Inc. before
joining Power Delivery Consultants, Inc. as a senior engineer in 1999.
There, he performs engineering studies and design involving advanced ampacity, system
specifications, uprating, magnetic fields, forced cooling, and rapid circulation calcula-
tions, and economic evaluations of underground and submarine cable systems. Mr. Bas-
com is a Senior Member of the IEEE, Power Engineering Society and Standards
Association, a Voting Member of the Insulated Conductors Committee, and a registered
professional engineer in New York and Florida. He is the author of more than 30 techni-
cal papers and publications and the holder of one patent.
11-1
Chapter 11: Ampacity EPRI Underground Transmission Systems Reference Book
11.1 INTRODUCTION Section 11.2 describes an approach for calculating

steady-state ampacity that is focused on the use of the
11.1.1 Ampacity IEC method—generally recognized as providing better
The ampacity, or current rating, of a cable is one of the accuracy for ac losses and being easier to follow—but
most important concepts for the cable engineer to accounts for cyclic loading by using the “loss factor”
understand, since its calculation encompasses many approach from Neher-McGrath. The section defines
aspects of the cable construction and installation. equations, gives ranges of applicable data for all param-
Although the calculation approach is conceptually sim- eters used in ampacity calculations and provides sensi-
ple—the solution of the thermal equivalent of Ohm’s tivity analyses for various installation and design
law—properly selecting and calculating the values for variables.
each component of the equivalent thermal circuit
requires a great deal of thought and careful planning. Transient ampacity calculations are complex. Section 11.3
reviews the principles for transient calculation procedures
Cable ampacity is principally determined based on the based on the IEC-80853 approach and includes the basic
allowable conductor current that does not exceed insu- equations for the most commonly used procedures.
lation temperature limits when installed under a partic-
ular set of conditions and energized to a selected Forced cooling is a specialized topic, mainly applied to
voltage. Other sections of this book discuss the details pipe-type cables, that represents a way to increase the
of voltage stress and the intricacies of selecting an insu- ampacity of both new and existing cable circuits. Sec-
lation material. This chapter focuses on the temperature tion 11.4 reviews forced cooling and its application to
limit that may be applied to a given insulation type— pipe-type cables, and gives simplified equations for cal-
effectively determining the conductor temperature and culating forced-cooled ampacities.
current— based on the following parameters:
Section 11.5 briefly reviews procedures for ampacity
• Peak current and load-cycle shape audits and uprating, Section 11.6 provides worked
• Conductor size, material, and construction examples, and Section 11.7 lists the extensive nomencla-
• Dielectric loss in the insulation ture used throughout this chapter.
• Current-dependent losses in the conductor, metallic It is important to note that other system components or
shields (screens), sheath, and pipe (for pipe-type sys- load-flow conditions could limit the current-carrying
tems) capacity of the transmission circuit to a value lower
• Thermal resistances of insulation, sheaths and cover- than the ampacity of the cable. Other sections of this
ings, filling medium, conduit or pipe wall (including book provide system implications for installing cables
pipe coating), and the various earth layers with other transmission equipment.
• Thermal capacitances of these components of the
thermal circuit 11.2 STEADY-STATE AMPACITY
• Mutual-heating effects of other cables and other heat CALCULATIONS
sources
11.2.1 Calculation Principles
• Ambient earth temperature
The cable ampacity is the solution to the heat transfer
problem of heat leaving the cable at a rate which allows
Both steady-state and emergency ampacities depend
the conductor temperature to reach but not exceed the
upon these variables, although emergency ampacities
maximum allowable temperature of the insulation mate-
have a much greater dependency on the thermal capaci-
rial immediately adjacent to the conductor. The heat is
tances of each component of the thermal circuit.
generated from ohmic (I2R) losses in the conductor,
shield/sheath, and other metallic layers (e.g., pipe for
Many ampacity calculations are based on a 1957 paper
pipe-type cables), along with dielectric losses generated
by J. H. Neher and M. H. McGrath (Neher and
in the insulation. The “steady-state” ampacity is the rat-
McGrath 1957). Later work by CIGRE documented an
ing of the cable under normal loading, with normal load
ampacity procedure in the International Electrotechni-
shapes, over a long period of time. Ampacities are stated
cal Commission (IEC) standard (IEC-60287 and IEC-
in terms of peak current, but it must be recognized that
60853), which provides a step-wise approach to calculat-
the allowable peak currents are affected by load shape,
ing ampacity based upon cable construction. The two
as discussed in Section 11.2.2.
calculation approaches—Neher-McGrath and IEC—
give similar results, although their treatment of daily
load cycles is different.
11-2
EPRI Underground Transmission Systems Reference Book Chapter 11: Ampacity
Heat generated in the cable is removed by thermal con- In this circuit:

duction to ambient earth and, ultimately, air. Engineers • Heat (watts) corresponds to electrical current
familiar with Ohm’s law know that electrical current (amperes)
flowing through a resistance produces a voltage drop (or
rise) according to the following relationship: • Thermal resistance (thermal ohm-meters) corre-
sponds to electrical resistance (ohm)
ΔVoltage = Current ⋅ R ac • Temperature drop (C°) corresponds to voltage drop
(volts)
An analogous relationship may be used to describe ther-
mal conduction where heat flowing through a thermal Thermal resistance is the product of a material property,
resistance produces a temperature drop (or rise) accord- thermal resistivity, times a geometric term that accounts
ing to the following: for the dimensions and shape of the component. Resis-
tivity is defined in units of C°-m/W (or K-m/W),
ΔTemperature = Heat ⋅ R Thermal although soil thermal properties are often listed in units
of C°-cm/W; values in the latter units are 100 times the
former. For a material with a resistivity of 1C°-m/W, a
This basic concept is extended to model heat emanating
heat flow of 1 W/m2 will cause a 1C° temperature rise
out of a buried cable through the various cable layers,
when flowing through a thickness of 1 m. Section 11.2
trench backfill, and native earth. Throughout this chap-
provides a further discussion of thermal resistivity.
ter, “thermal” variables (thermal resistance, thermal
capacitance, thermal resistivity, etc.) are shown with an
Ampacity calculations are done on a unit cable length of
overline (e.g., R Thermal ). This is to distinguish them 1 m. A “thermal ohm-meter” (TOM) (C°-m/W) is
defined, as shown in Figure 11-2, to be the thermal
from “ac” variables where no overline is used. The con-
resistance that causes a 1C° temperature rise of the con-
vention used for a temperature change is to show these
ductor, when there is 1 W per meter of heat generated in
values with units of “Celsius degrees” (C°), while tem-
the conductor.
peratures are shown in “degrees Celsius” (°C).
In its simplest form, an ampacity calculation requires
Calculating cable ampacity requires determining the
calculating losses and the temperature rise due to those
peak current that will give an allowable conductor tem-
losses flowing through the various thermal resistances.
perature rise above ambient temperature for the cable
construction, installation, and design operating condi-
The procedure is described in detail in Section 11.2.5
tions. The calculation procedure is basically the solution
and is summarized as follows:
of the thermal equivalent of Ohm’s law, as shown in Fig-
ure 11-1. 1. Select a cable construction and conductor size that
are anticipated to meet the ampacity requirements
for the expected installation conditions.
2. Refer to industry standards (IEC, AEIC) or manu-
facturer’s limits for the maximum allowable conduc-
Figure 11-1 Basic thermal circuit. Figure 11-2 Definition of thermal ohm-meter.
11-3
tor temperature for the cable insulation material. To verify that rated current can be calculated by Equa-
Determine the temperature rise over ambient earth tion 11-1, consider the following dimensional analysis:
temperature that will give this value.
⎛ Ohm ⎞⎛ C° − m ⎞
3. Calculate dielectric loss. ( )
R ac ⋅ Σ Qi R thermali = ⎜ ⎟⎜
⎝ m ⎠⎝ Watt ⎠
⎟
4. Calculate the electrical resistances of each current-
⎛ volt / ampere ⎞ ⎛ C° − m ⎞
carrying component of the system for the expected =⎜ ⎟⎜ ⎟
operating temperature of that component. ⎝ m ⎠ ⎝ volt − amperes ⎠
5. Calculate the thermal resistance of each component ⎛ C° ⎞
=⎜ 2 ⎟
of the system, including the earth. ⎝ amperes ⎠ 11-2
6. Calculate the temperature rise due to dielectric loss
flowing through the thermal resistances, and subtract and,
that number from the total available temperature rise.
ΔT
7. Solve the Ohm’s law equivalent of the circuit of Fig- I=
ure 11-1, to determine the ampacity that achieves the
allowable temperature rise.
(
R ac ⋅ Σ Qi R Thermal i )
8. If calculated ampacity is too high or too low, make C°
adjustments in conductor size and/or installation = = amperes
parameters, and repeat these steps as necessary to
C° / amperes 2
11-3
achieve the desired ampacity.
So dimensionally, the ampacity solution is consistent
The solution of the thermal circuit shown in Figure 11-1 with the units of the various elements in the equivalent
is summarized for ohmic (I 2 R) losses in the ampacity thermal circuit.
procedure of Equation 11-1.
The total thermal circuit for an extruded transmission
ΔT = Wi ⋅ R Thermali cable, shown for an installation in conduit (common in
the United States and increasingly used internationally),
Wi = I 2 R ac Qi is illustrated as an electrical analog in Figure 11-3. The
thermal equivalent for a pipe-type cable is shown in Fig-
ΔT
I= ure 11-4. The thermal circuits of Figure 11-3 and 11-4
(
R ac ⋅ Σ Qi R Thermal i ) 11-1
show thermal capacitances for each component of the
cable system. All of these thermal capacitances are eval-
Where uated for transient or emergency calculations. For
ΔT = allowable temperature rise, C°. steady-state calculations with a continuous, non-varying
Wi = losses developed in the conductor, W/m. magnitude of load current (100% load factor), thermal
capacitance is excluded from the thermal circuit. This
R Thermal i = thermal resistance of various “i-th" rating is sometimes referred to as the “continuous” or
layers, C°-m/W (thermal ohm-meter [C°- “sustained” rating in many parts of the world. In the
m/W]). United States, the practice is to allow for the cyclic varia-
Rac = electrical resistance, ohm/m. tion of the load current (i.e., a load/loss factor of less
Qi = the ratio of losses in the “ith” layer plus the than 100%), which usually allows for a smaller conduc-
ratio of losses in all inner layers (including tor size to be selected. This “cyclic” rating can be devel-
the conductor) to that of the losses of only oped using a “cyclic rating factor” from IEC 60853 or
the conductor. using the “loss factor” method from the Neher-McGrath
paper. In the Ohm’s Law analogy, a change in applied
The summation sign, Σ, denotes that we sum up the voltage (temperature) does not instantly change the volt-
various products of ac and thermal resistances from the age across an electrical capacitor, and so for the purpose
equivalent thermal circuit that will be discussed below; of calculation under cyclic variation, the electrical (ther-
this notation is for simplicity and is not numerically mal) resistance of the earth may be considered to have
correct. been reduced.
In Figures 11-3 and 11-4, the thermal resistances and

heat sources are shown, analogous to electrical resis-
tances and current sources in a conventional Ohm’s law
circuit. The temperature of the conductor, Tc, is deter-
11-4
mined by adding up the temperature rises above the Components of the Cable Thermal Circuit
ambient earth temperature, Te a r t h , as heat passes Each of these components is summarized below and dis-
through the various thermal resistances. The compo- cussed in detail later in this chapter.
nents of the thermal circuit represent one cable, includ-
Temperatures and Temperature Rise
ing the thermal resistance as seen from one cable, which
Temperatures of interest include conductor tempera-
is mathematically reduced to allow for the mutual heat-
ture, cable/earth interface temperature, and ambient
ing among adjacent cables. The earth component of the
earth temperature.
thermal circuit is actually a thermal composite R-C net-
wo rk , wh i c h fo r s i m p l i c i t y h a s b e e n s h ow n a s
To calculate the ampacity, the conductor temperature is
ΣR n ⋅ C n . determined for a given current and checked against the
maximum allowable conductor temperature based on
Qi can be found from Figures 11-3 and 11-4. For exam- the insulation material. For the above thermal circuits,
ple, QS = (WC + WS) / WC and QP = (WC + WS + WP) / the temperature rise above ambient is found using
WC. QS and Q P are used to relate all of the heat losses Equation 11-4, where the ac losses (I2R) and dielectric
heat losses (Wd) pass through the various thermal resis-
experienced by the various cable layers back to the con-
ductor losses. So, for example, the heat losses flowing tances to give a temperature rise (Equation 11-4 is a
generalization):
through the jacket thermal resistance, R Jacket , in Figure
11-3 are WC + WS or numerically equivalent to WC • QS. ( )
Tmax = I 2 R ac + Wd × R thermal + Tambient
ΔT = Tmax − ΔTd − Tambient = I 2 R ac × R thermal

11-4
Figure 11-3 Equivalent thermal circuit for extruded-dielectric and self-contained

fluid-filled cable types.
Figure 11-4 Equivalent thermal circuit for pipe-type cables.
11-5
Typical ampacity calculations assume that the voltage mal capacitance of surrounding soil on steady-state rat-
remains constant, so dielectric losses are fixed. Then, ings of buried cables is addressed by looking at
the only unknown in Equation 11-4 is the current, as “attainment” factors that evaluate the temperature
follows (as a convention, thermal quantities are shown response of the surrounding soil in response to changes
with an over-line to distinguish these values from electri- of heat output. This will be discussed further in Section
cal quantities). The term, Qi, is introduced to properly 11.3.6.
account for the ac losses at the various, “ith” layers of
Special Situations for Ampacity Calculations
the thermal circuit in Equation 11-5, which are summed
Chapter 11 deals principally with buried cable systems,
together to give the total effect of temperature rise
which account for the vast majority of transmission
between the conductor and ambient earth:
cables. Cables in air—for example, on bridges, riser sec-
tions, etc.—are treated separately in Section 11.2.6. This
ΔT
I=
( ) amperes 11-5 section also discusses related topics such as cables in air-
R ac ⋅ Σ Q i R Thermali filled tunnels.
Losses Trenchless installations (horizontal directional drilling,

Losses consist of ohmic (I2R) losses in the conductor, pipe jacking, etc.) are discussed in Section 11.2.7, and
dielectric losses in the insulation, and losses due to cir- the derating effects of cables crossing external heat
culating and eddy currents in the shield/sheath. Addi- sources are discussed in Section 11.2.8.
tional eddy current and hysteresis losses are generated
when extruded and self-contained cables are put in cas- Typically associated with system studies, fault current
ings and in pipe-type cables. Submarine cables may have capability calculations are specialized methods for
additional losses due to the armor. determining ratings during very short duration events
Thermal Resistances lasting less than a second. The intent is to determine
Thermal resistances for single-conductor cables what current levels can be sustained by the cable before
(extruded, self-contained) consist of the cable insula- a circuit breaker or other circuit protection has time to
tion, jacket, annular space between the cable and inside operate. These sorts of calculations are much different
of the conduit, the duct material (when used), and the than typical emergency or transient ratings that last sev-
earth, which includes the thermal backfill (concrete, eral minutes to many hours and are described in Section
Fluidized Thermal Backfill™ [FTB]) around the cables 11.2.9. The background and application of fault current
along with the native soil. There are also thermal resis- calculations are discussed in Chapter 16.
tances to account for mutual heating among cable
phases and adjacent circuits. Submarine cable ratings generally follow the same
approaches as conventional buried cables, although
Pipe-type cables have thermal resistances in the insula- there are additional considerations from the standpoint
tion, pressurization fluid (liquid or gas) in the pipe, of soil types that are typically found on sea beds, at the
pipe-coating material, and the earth, including backfill, bottom of lakes and in rivers, and in the land-fall
native soil, and mutual heating components. Section regions. Also, high-voltage direct current (HVDC) cable
11.4 discusses additional thermal resistances that systems are not affected by dielectric losses, but may be
develop when fluid circulation or forced cooling are limited by temperatures other than the conductor tem-
used on pipe-type cables. perature. A discussion of the installation considerations
and general design parameters, including ampacity, for
For all cable types, the thermal resistances of metallic submarine cables is contained in Chapter 7.
layers (shield/screen, sheath, armor, pipe, etc.) are Approaches
neglected because the thermal conductivity of these lay- Virtually all steady-state ampacity calculations are
ers is much larger as compared to the nonmetallic lay- based on the Neher-McGrath procedure given in the
ers. For this reason, these thermal resistances do not classic paper, “The Calculation of Temperature Rise
appear in the equivalent thermal circuits used for cable and Load Capability of Cable Systems” (Neher and
ampacity. McGrath 1957). The procedure has been refined and
Thermal Capacitances characterized based on cable system types by CIGRE in
Each of the components in the cable system has a ther- the IEC Publication 60287; this latter document is now
mal capacitance that is important when evaluating situ- the predominant approach used throughout the world,
ations where the load will be changing, particularly for both in hand calculations and by computer programs.
short-term emergency ampacities. The effect of the ther-
11-6
The procedure to calculate ampacity is relatively simple 345 kV) for calculating dielectric losses. System voltage,
to follow and has been verified to be accurate by many particularly for the extra-high-voltage cables, has impli-
tests. The procedure used for ampacity analyses details cations for charging current and system stability. This
each of the terms—thermal resistance, thermal capaci- topic is discussed in great detail in Chapter 16.
tance, ac resistance, etc.—described above. The IEC-
60287 procedure is the basis for presentation of most Design peak current levels should be chosen carefully.
material in this chapter. However, the Neher-McGrath Choosing an unrealistically high current—for example,
paper is good background, particularly when referring to match the never-attained thermal rating of an over-
to hand calculations done on older cable systems in the head line— can result in selecting an unnecessarily large
United States where U.S. units (inches, feet, etc.) are conductor size, adding forced cooling, or installing an
employed. For the calculation procedure described here, additional circuit.
the nomenclature and symbols used in the Neher-
Expected Load Shape
McGrath paper are followed, along with the daily loss
The large thermal mass and long thermal time constant
factor, though the equations are from IEC-60287 and
of the cable system permits an ampacity that is deter-
use metric units to be consistent with worldwide practice.
mined more by the average daily load rather than the
Definition of Terms peak load. As described in Section 11.2.2, the earth
Symbols and terms used in this chapter are defined in thermal circuit is modified to account for the effect of
Section 11.7. These definitions are consistent with those the daily load shape.
of the Neher-McGrath paper using metric values. “T”
will be used to denote temperatures, “R” for ac resis- The earth's temperature response to a time-varying heat
tances, “ R ” for thermal resistances, and “W” for heat input can be represented by a series of thermal R-C sec-
(watts). For reference, the equivalent IEC symbols are tions. The inner sections close to the heat source have a
“θ” for temperature or temperature rise, “W” for heat, short time constant. They represent the “close in” por-
“T” for thermal resistance, and “Q” for thermal capaci- tion of the earth with relatively small mass (or volume)
tance. of earth. Sections farther from the heat source have pro-
gressively larger thermal mass and longer time con-
11.2.2 Utility System Data stants. The temperature rises of the inner sections react
System conditions for cable operation must be carefully to short time changes in heat input (i.e., respond to
defined to accurately determine the required conductor high-frequency load changes), while the sections further
size, and to provide cable ampacities for all operating out react to long-time changes in heat input (i.e.,
conditions. The most important system conditions are respond to lower frequency load changes). For cyclic
system voltage, expected peak current, expected daily loads of a 24-hour-time period, the earth thermal circuit
load shape, loading during contingencies and preload can be represented by a two-zone steady-state model. In
levels, and reactive power flow. Note that although the an inner zone, the temperature rise is proportional to
term, “cable loading,” is commonly used instead of the the peak joule (I2R) losses. In the outer zone, the tem-
more proper term, “peak current,” current through the perature rise is proportional to the average joule losses.
cable may have little relationship to system load. On The diameter of the circular boundary between the two
some tie lines, for example, cable current is inversely zones, defined to be DX, can be related to the length of
proportional to system load. The focus of this chapter is the load cycle and the earth thermal diffusivity or ther-
the current-carrying capacity of cable circuits, rather mal capacity. This is the method of accounting for the
than the system load flow characteristics. effect of a cyclic load adopted in the Neher-McGrath
paper and used in this chapter.
System Voltage and Expected Peak Current
System voltage is seldom a variable, although in some Many system planners are familiar with a “load factor”
cases the power transfer requirements may dictate that relates the peak load to the average daily load. In
selecting a higher voltage level. In the interest of conser- cable systems, we are interested in heat output—a func-
vatism, many engineers select maximum system voltage tion of I2R—so we use the daily “loss factor,” which is
rather than nominal voltage (e.g., 362 kV rather than
11-7
essentially the load factor of the losses as defined by the Other Methods for Modeling Load Shape
following equations: Other approaches can be applied to account for load
cycles in cable ratings. The IEC method in its basic form
24 assumes that the load remains constant (e.g., a loss fac-
∑ Ii tor of 100%), but ratings by this method can be modi-
Load Factor = i =1 fied by a companion document, IEC-60853 (IEC 1989),
24 ⋅ I max to account for the periodic load variations, particularly
when the load is not closely sinusoidal. The IEC
24 approach to account for load variations does not easily
∑ I i2 lend itself to hand calculations and generally gives close
Loss Factor = i =1 agreement to the Neher-McGrath method under typical
24 ⋅ I 2max 11-6
loading patterns.
Where: Longer load cycles have been modeled using a Fast Fou-
Ii are hourly load values (amperes). rier Transformer (FFT) method (Thomann et al. 1991)
Imax is the maximum hourly load value during the and are used today in some computer programs. The
cycle (amperes). approach uses a FFT to convert a load period in time
domain to frequency domain, where a steady-state tem-
With the increasing availability of load information perature calculation is done at each of the frequencies.
from utility data acquisition systems, loss factors can be The calculated temperatures in frequency domain are
calculated directly using historical load records. In passed through an inverse FFT to give a continuous
many cases, though, only the daily load factor is known. function that shows variations in temperature with
The loss factor can be estimated from the load factor respect to time. A basic assumption for this method is
using an empirical relationship. One such relationship is that the analysis period is part of an ever-repeating
as follows: cycle.
Loss Factor = 0.3(Load Factor ) + 0.7(Load Factor )2 Contingency Criteria and Cable Preload
11-7
Ampacity calculations must consider contingency crite-
(Distribution cables sometimes follow a similar relation- ria as well as steady-state conditions. The cable design
ship, Loss Factor = 0.2 (Load Factor) + 0.8 (Load Fac- engineer should determine the expected contingency
tor)2, but it is important to note that these load factor- loading and duration, and frequency of occurrence of
to-loss factor relationships are specific to a given system the contingency. Typically, ampacity requirements are
and should not necessarily be applied arbitrarily.) The given for transient loading (e.g., 15 or 20 minutes),
loss factor versus load factor relationship based on short-term emergencies (2-4 hours), long-term emergen-
Equation 11-7 is shown graphically in Figure 11-5 (note cies (8-24 hours), and emergency repair time (300 hours
that the loss factor is lower than the load factor for a to one month or longer).
given load shape).
The heat-storage capacity of the cable and the immedi-
ately surrounding earth are important for calculating
transient and emergency ratings. This capacity depends
upon temperatures prior to the contingency. A cable
operating at one-half rated load will have current-
dependent losses roughly one-quarter as great as that at
full-load losses; the temperature rise due to current-
dependent losses will, therefore, be one-quarter as great
as that at full-load temperatures. Many utilities have
preload criteria assumed to be a percentage of rated cur-
rent—typically 50 to 75%. Conductor temperature is
calculated based upon preload conditions, and that tem-
perature then forms the basis for calculating transient
and emergency ratings. The design engineer should,
therefore, determine expected preload prior to emer-
Figure 11-5 Loss factor vs. load factor for a typical
gency conditions. Section 11.3 describes how to calcu-
transmission cable. late ampacities for transient and emergency loadings.
11-8
Cable preload also has implications for fault current stress levels in the insulation, the insulation thickness is
capability calculations. For these types of calculations, it designated mostly by the system voltage and is some-
is common to assume full rated load and the corre- what standardized for most cable voltage classes.
sponding rated temperatures on the cable prior to the
initiation of the fault. This will be discussed in greater The current-carrying capacity—the subject of this chap-
detail in Section 11.2.9. ter—is dependent on many cable system characteristics.
These data must be obtained to use in the calculations:
Reactive Power Flow
The power requirements should be precisely defined. • Conductor material, size, construction, and proper-
Ampacity calculations give a total current in the con- ties
ductors that will achieve the maximum allowable con- • Conductor semiconducting shield material, thick-
ductor temperature. That total current is the phasor ness, and properties
sum of real current and reactive current, as shown in
• Insulation material, thickness, and properties
Equation 11-8 and Figure 11-6.
• Insulation semiconducting shield material, thickness,
I total = I 2real + I 2react or I real = I 2total − I 2react and properties
MVA = MW 2 + MVAR 2 or MW = MVA 2 − MVAR 2

• Metallic insulation shield type (wire or tape) and con-
struction
11-8
• Sheath material, thickness, properties, and electrical
Two sources of reactive power flow in the cable may connection
reduce the allowable real power flow: charging current • Jacket material, thickness, and properties (single-con-
generated by the cable itself, and through-flow of reac- ductor cables)
tive power. Charging current can be calculated readily,
and is treated in Section 11.2.5, Calculation Procedures. • Conduit properties and size (unless direct buried)
Through-flow of reactive power is a function of system
conditions, and should be evaluated using a load-flow Additionally for pipe-type cables:
program as described in Chapter 16. The cable engineer • Pipe-filling medium
and system planner may have to resolve the total current
• Pipe size
or MVA rating (thermal) of the cable with the reactive
power flow that is a function of the overall power sys- • Pipe-coating material, thickness, properties
tem. As a separate step, the cable engineer and system
planner should determine if the required real power Cable data are discussed in detail in Chapters 4 through
transfer (MW) will be met based upon reactive power 7. Many of the items have only a minor effect on ampac-
flows and system compensation. Refer to the example at ity. Components with larger effects are conductor mate-
the end of this chapter for determining real power trans- rial and size, insulation material and thickness, sheath
fer from a calculated total current rating. material and properties, and the sheath-bonding
method. A brief summary of the effect of each item on
11.2.3 Cable Data cable ampacity is given below.
Ampacity calculations should begin by selecting a cable Conductor Material, Size, Construction, and
that is expected to meet the power transfer requirements Properties
while supporting the necessary line-to-ground voltage The materials, size, construction, and properties of the
for the selected insulation type. Depending on permitted conductor are important parameters that determine the
ohmic losses in the conductor. Stranding and conductor
treatment affect ac losses, as described in Section 11.2.5.
Figure 11-7 shows the effect of conductor size and mate-
rial on the ampacity of a 230-kV extruded-dielectric
cable, with two circuits arranged vertically in a duct
bank.
Conductor Semiconducting Shield Material,
Thickness, and Properties
Conductor semiconducting shielding has a very minor
effect on ohmic losses. There is an effect on thermal
Figure 11-6 Real, reactive, and total current. resistance, which is normally taken into account by
11-9
Insulation Semiconducting Shield Material,

Thickness, and Properties
Differences in insulation semiconducting shielding have
a minor effect on losses for all cable types. Thermal
resistance of the semiconductive shielding should be
taken into account, generally by considering the shield-
ing as part of the insulation material.
Shield and Sheath Material, Thickness, Properties,
and Electrical Connection
Ampacity is affected by inductively-induced circulating
currents flowing in the metallic shield and sheath in
extruded-dielectric and self-contained cables, as well as
in armor wires (if present). Eddy current losses also
Figure 11-7 Ampacity of 230-kV extruded-dielectric develop in the shields of all cable that use continuous
(XLPE) cable as a function of conductor size and
material. metallic layers (e.g., extruded lead, longitudinally-taped
foil, corrugated metallic sheaths, etc.), including in pipe-
including the shield material as part of the insulation type cables. As described in Section 11.2.5, if single-con-
thickness for thermal resistance calculations. ductor cable sheaths are grounded at only one point or
if they are perfectly cross-bonded, eddy-current losses
Insulation Material, Thickness, and Properties are induced, which give a small reduction in cable
Insulation has a major effect on cable ampacity. As ampacity. If the sheaths are grounded at more than one
described in Chapters 4 through 7, insulation material point or if there are different minor-section lengths in
determines the allowable conductor operating tempera- cross bonding, induced currents in the sheath result in
ture. The dielectric constant and dissipation factor of ohmic losses that are sometimes almost as large as those
the insulation, and its dimensions, will determine the in the conductor. The magnitude of the induced cur-
dielectric losses. Insulation dimensions and thermal rents and shield and sheath losses are a function of the
resistivity determine the insulation thermal resistance electrical resistivity of the shield and sheath. Figure 11-9
for a given conductor size. The impact of insulation shows the difference in ampacity for an extruded-dielec-
material selection is most pronounced on self-contained tric cable with single-point-grounded and with multiple-
fluid-filled or pipe-type cables when deciding between point-grounded sheaths.
kraft paper and laminated paper-polypropylene insula-
tion materials. Figure 11-8 shows cable ampacity for The topic of sheath bonding is discussed in greater
conventional impregnated-paper insulation and for lam- detail in Chapter 12.
inated paper-polypropylene insulation as a function of
conductor size for a pair of 345-kV pipe-type cables.
Figure 11-8 Ampacity of 345-kV pipe-type cable as a function of insulation

material and conductor size.
11-10
Jacket Material, Thickness, and Properties losses are increased. Metallic conduits and steel cable
The jacketing over single-conductor cables presents an pipes have a negligible thermal resistance. Significant
additional thermal resistance that must be considered circulating current losses may be generated in stainless
when an ampacity calculation is performed. steel conduits used for some single conductor cables
because the conduits are usually grounded at one or
Cable to Duct or Pipe both ends. Riser pipes between a trifurcator and termi-
When cables are installed in conduits, heat generated in nation in pipe-type cables can also have circulating cur-
the cable must pass through the air in a duct system. Since rent losses if the grounding is not properly managed.
air has a relatively high thermal resistance to heat as com-
pared to most other materials, sometimes conduits are Corrosion Coating Material, Thickness, and
filled with water or a grout material to reduce the thermal Properties
resistance and increase ampacity by a few percent. Corrosion coatings for casing pipes or on the outside of
pipe-type cables give an additional thermal resistance,
A fluid-filled pipe-type cable will have an ampacity but the effect is fairly small.
about 2% greater than a nitrogen gas-filled, pipe-type
cable because of the better heat transfer through the liq- 11.2.4 Installation Data
uid. Cables designed for gas-filled systems also require a Installation conditions have a substantial impact on
slightly greater insulation thickness, so there is an added cable ampacity, and the user often can change installa-
thermal resistance from the insulation. Fluid boundary- tion conditions to optimize ampacity. Major installation
layer resistances are important in circulation and factors affecting ampacity include the following:
forced-cooling calculations, as described in Section 11.4.
• Burial depth of cable
• Spacing between circuits, and among phases
Conduit Material, Thickness, and Properties
In addition to the thermal resistance from the cable sur- • Thermal resistivity of native backfill and controlled
face to the conduit, the presence of a nonmetallic duct backfill (see Section 11.2.6)
introduces an additional thermal resistance in the cir- • Dimensions of trench and controlled backfill
cuit, though it does not affect losses. When using • Presence of nearby heat sources
trenchless installation methods, this can be a significant
factor when considering the generally thicker conduit • Ambient earth temperature
materials and the use of a casing. Plastic casing pipes
introduce an additional thermal resistance layer. General comments on the effects of cable installation
conditions on ampacity are given in the following sub-
The presence of a metallic conduit or casing pipe sections.
(including the pipe of a pipe-type cable) results in signif- Burial Depth
icant hysteresis losses due to magnetically-induced cur- All heat generated by the cable must eventually reach
rents. In addition, the conductor, shield, and sheath the surface of the earth. A greater depth of burial cre-
Figure 11-9 Effect of sheath grounding method on single-conductor

cable ampacity.
11-11
ates a longer heat flow path and greater thermal resis- increasing thermal resistivity). So, although a shallow
tance, and results in lower ampacity, even though earth burial depth reduces thermal resistance, other installa-
temperature may decrease and moisture levels may tion factors must be considered when evaluating ampac-
increase with depth. Figure 11-10 shows the effect of ity and placement of cables.
burial depth on the ampacity of a pair of 230-kV
Spacing Between Circuits and Among Phases
crosslinked polyethylene cables with 1200 mm2 (2368
Spacing between circuits, and among phases for single-
kcmil) conductors. Ambient earth temperature is
phase cables, determines the heating effect of one cable
assumed to drop for every 30 cm (12 in.) of increased
on the others (see Figure 11-11). Increasing the spacing
burial depth, which would be reasonable for northern
gives greater ampacities, although for single-conductor
latitudes during summer months. A typical burial depth
cables, ampacity for the close trefoil configuration may
for a trenched installation is about 1.1 m (42 in.).
be higher because of better cancellation of magnetic
fields and lower sheath losses when sheaths are multiple-
Barring other factors, it is generally better to have a
point grounded. For this reason, multipoint bonded
more shallow burial depth to minimize the earth ther-
cables (mostly at distribution voltages) are placed in a
mal resistance. However, ambient soil temperatures are
single conduit or in close-trefoil (triangular) configura-
often greater near the surface (potentially reducing
tion to minimize circulating current losses.
ampacity), and soils tend to be drier above the water
table (potentially reducing soil moisture content and
Figure 11-10 Effect of burial depth on cable ampacity of 230-kV XLPE cables.
Figure 11-11 Effect of circuit spacing on ampacity of 230-kV extruded-dielectric cables.
11-12
Thermal Resistivity of Native Backfill describes soil characteristics, thermal testing, and types
Since all heat generated in self-cooled cables must flow of backfill materials.
through the earth, the thermal resistivity of native back-
Trench Dimensions
fill strongly affects cable ampacity. Except for conductor
Increasing the trench dimensions beyond minimum
size, most of the cable construction is fixed based upon
requirements permits the use of additional controlled
the voltage class of the cable. The native soil outside the
backfill, which increases cable ampacity. Utilities some-
cable trench is also fixed (and represents a large part of
times widen the trench at specific locations of high
the total thermal resistance), but must be factored into
native soil thermal resistivity to preserve circuit ampac-
the ampacity calculations. Higher soil thermal resistivity
ity or mitigate a hot spot.
results in a higher thermal resistance to heat leaving the
cable and lower ampacity. Presence of Nearby Heat Sources
Heat sources within a distance of about 4 m (12 ft) of
Soils may also have great variability in soil conditions typical trenched installations raise the temperature of
over a few meters of cable route. Therefore, it is impor- the earth, reduce the allowable flow of heat from the
tant to find the worst conditions along the cable route— transmission cable, and thereby reduce the cable ampac-
“weakest link” in the chain—and base ratings on that ity. For interfering heat sources such as other transmis-
limiting location. Figure 11-12 shows the impact of soil sion cables, distribution circuits, or steam mains, an
thermal resistivity on ampacity. Soils with a thermal interference temperature rise is calculated as described
resistivity that is low in natural (in situ) conditions may in the “Thermal Resistances” portion of Section 11.2.5
dry out substantially when exposed to cable heating. and in Section 11.2.8.
Good quality thermal backfills show relatively minor
changes in thermal resistivity as moisture content is Ambient Earth Temperature
reduced. Although the surface of the earth is the ultimate sink for
heat generated from the power cable, heat flows to the
Controlled Backfill ambient earth for the purposes of ampacity calcula-
Backfilling the trench with a material having controlled tions. Ambient earth temperature varies substantially
thermal resistivity is often an effective way to increase throughout the world, as discussed in Chapter 12. Val-
cable ampacity. Figure 11-12 also shows the effect of ues for ambient temperature for areas around the world
adding a controlled backfill with a resistivity of 0.5C°- are given in IEC 60287-3-1 and later in this chapter in
m/W in the trench, this being surrounded by native soil Tables 11-2 and 11-3. Like native soil thermal resistivity,
having resistivities shown on the axis of abscissas. The there is little control over the ambient soil temperature,
figure also illustrates the improvement in ampacity that but it must be factored into the cable ratings. One con-
might be achieved by adding additional thermal backfill sideration about the ambient temperature is the burial
up to the top of the trench, a technique sometimes used depth. Greater temperature extremes are experienced
in areas with particularly poor native soil conditions. close to the surface, so summer ambient temperatures
Using a good quality thermal backfill (thermal sand, are more significant at shallow depths. Also, surface
Fluidized Thermal Backfill™ [FTB]) can improve the coverings have an impact on ratings. Soil ambient tem-
allowable current-carrying capacity. Chapter 12
Figure 11-12 Effect of soil resistivity on ampacity of 230-kV XLPE cable duct bank.
11-13
peratures may be 3-5°C warmer below asphalt than Temperatures

other areas because of the increased solar absorptivity
Conductor (Insulation) Temperature
of the surface.
The maximum operating temperature of a conductor is
limited by the temperature performance of the insula-
11.2.5 Calculation Procedures
tion material. Extruded materials—polyethylene,
Significant activity in the late 1940s and early 1950s crosslinked polyethylene, and ethylene-propylene rub-
resulted in a Symposium on Temperature Rise of Cables ber—can operate up to a certain temperature where the
and the publishing of many papers on cable ampacity materials may soften or otherwise experience a degrada-
(Neher 1949). This activity culminated in the publication in dielectric performance or because thermal
tion of the Neher-McGrath paper, “Calculation of Tem- expansion may damage the shielding layers or sheath.
perature Rise and Load Capability of Cable Systems”
(Neher and McGrath 1957), which is the basis for Paper-based insulations demonstrate a relationship
almost all steady-state ampacity calculations, including between temperatures and aging rate. For example, the
those in IEC 60287. The calculation procedures in this aging rate of impregnated paper doubles with every
section follow the IEC 60287 approach, with roots in 8-10C° increase in temperature. Maximum conductor
Neher-McGrath including the use of the loss-factor temperatures for paper-insulated cables are based upon
method to account for daily load cycles. The interpreta- a 40-year life at maximum operating temperature, plus a
tion in this chapter is intended to simplify rating calcu- limited number of excursions to higher, emergency oper-
lations for the major transmission cable types, although ating temperatures (see Section 11.3).
IEC 60287 and Neher-McGrath continue to be good
references. Typical insulation temperature limits are summarized in
Table 11-1. Utilities following AEIC recommendations
The thermal circuit for an extruded underground cable
system, shown originally in Figure 11-3, is reproduced
here as Figure 11-13. This section addresses tempera-
tures, losses, and thermal resistances of the cable sys-
tem, thermal resistances of the earth system, and
solving the thermal circuit to give rated ampacity. All
ampacity calculations are on a per-conductor basis,
even for pipe-type cables where there are three conduc-
tors in the pipe.
A typical trench cross section for an extruded-dielectric

cable system is shown in Figure 11-14, and a typical
pipe-type circuit is shown in Figure 11-15 (trefoil cables
on the left and cradled cables on the right). Nomencla-
ture used in this section is shown in these two figures,
and definitions of these and other terms are provided at
the end of this chapter, along with detailed worked
examples. Figure 11-14 Extruded-dielectric cable system
installation.
Figure 11-13 Electrical analog of cable thermal circuit.

Figure 11-15 Pipe-type cable system installation.
11-14
may reduce these maximum-operating temperatures by ambient soil temperatures in recent years due to place-
10 C° when the thermal environment in which they are ment of cables below pavement and generally higher
installed is not well documented or understood. ambient temperatures seen in recent years.
Longer-duration emergencies are often considered to “Ambient” soil temperature is the temperature at the
account for the typically longer repair times encoun- burial depth of the cable in the absence of any non-
tered with transmission cables. Some utilities permit native heat sources. These temperatures are usually
emergencies up to 720 hours (30 days) for this purpose established during a route thermal survey (described in
on the basis that the emergency calculations usually Chapter 12). Typical values at a 1 m (3.3 ft) depth for
consider the temperature response to a step function representative cities are given in Tables 11-2 and 11-3.
and that the ultimate emergency temperature is not
reached until the end of the emergency rating period.
Ambient Earth Temperature
The other “given” temperature is the ambient earth
temperature, which varies substantially depending on
geographic location and time of year. Summertime tem-
peratures generally decrease with depth, and wintertime
temperatures generally increase with depth down to 5-6
m (15-20 ft). Below that depth, ambient earth tempera-
ture is reasonably constant until much greater depths.
The variation in ambient earth temperature is linked to
the variation in the ambient air temperature over the
year and, within 60 cm (2 ft) of the surface, the solar
heating. Figure 11-16, taken from a paper (Neher 1949)
shows this effect for the Philadelphia, Pennsylvania
area. Many utilities have adopted somewhat higher
Table 11-1 Allowable Conductor Temperature for Different

Insulation Materials
Maximum Normal
Insulation (Continuous) Maximum Emergency
Material Temperature Temperature
105-130°Ca, 72 hours
Crosslinked continuousb
90°C
Polyethylene 105°C- peak at end
of 100-300 hours Figure 11-16 Ambient earth temperature as a function of
105°C peak at end: up depth for Philadelphia.
Impregnated to 100 hours
85°Cc
Paper 100°C peak at end:
100-300 hours
Table 11-2 U.S. Ambient Soil Temperatures at Typical
105°C peak at end:
Laminated Paper up to 100 hours Installation Depthsa
85°Cc
Polypropylene 100°C peak at end: Maximum Maximum
100-300 hours Location Summer Winter
Ethylene-Propy-
90°C 130°C °C °C
lene Rubber
Atlanta 25 20
Linear
Boston 22 18
Low-Density 75°C 90°C
Polyethylene Chicago 22 18
Denver 22 17
a. Depends on shield construction and agreement from manu-
facturer. Honolulu 30 30
b. Based on Association of Edison Illuminating Companies Miami 30 25
CS-7 specification. However, the unpublished draft CS-6 New York 27 18
specification limits the temperature of HV transmission Palo Alto 22 20
cable to 105°C. San Juan, PR 31 29
c. Outside the United States, paper-insulated cables are per-
mitted to operate up to 90°C. a. E.g., 1.1 m (42 in.).
11-15
Table 11-3 International Ambient Soil Temperatures (from Where:

IEC 60287-3-1) TAverage =Average annual air temperature, °C.
Maximum Maximum
A = Difference between maximum and mini-
Location Summer Winter mum air temperatures, C°.
°C °C x = Depth below the earth's surface, in.
Australia 25 18 a = Thermal diffusivity, in.2/hr (2.75 from
Austria 20 0 Neher-McGrath).
Canada 20 -5 t0 = Length of period, hours (8760 hours
Finland 15 0 equals one year).
France 20 10 t = Time since maximum air temperature
Germany 20 0 occurred, hr.
Italy 20 5
Japan 25 15 The temperature is only an estimation because it could
Netherlands 20 5 be influenced by surface conditions (including surface
17 (south) 0 (south) cover that might influence the solar absorptivity of the
Norway
13 (north) -5 (north) ground), localized geography and weather.
Poland 15 5
Temperatures of Outer Cable Layers
Singapore 30 25
South Africa 22 15
Temperatures of the cable shield, sheath, and, if present,
Sweden 15 0
pipe or duct may be calculated once the thermal circuit
Switzerland 20 -10
of Figure 11-13 is solved. This is an iterative process,
United Kingdom 15 15
given that the electrical resistances of the shield and/or
sheath are temperature dependent, and the temperatures
will not be known precisely prior to calculating the resis-
Ambient earth temperatures can often be obtained from tance of these components. For typical calculations, an
local soil conservation services. Thermal probes, described assumed value of the shield/sheath temperature of 10-
in Chapter 12, indicate ambient earth temperature at the 15C° below the rated conductor temperature is a rea-
time and depth of testing, and the utility can install a sim- sonable approximation for normal ampacity.
ple thermocouple or thermister probe that will permit
year-round measurement. The reader should be aware Depending on the soil conditions, some cable ratings are
that ambient soil temperatures tend to lag behind ambient limited by the earth interface temperature (most often
air temperature by 4-6 weeks in the higher latitudes. For for direct-buried cables), rather than the conductor tem-
example, the highest ambient air temperatures in the perature, because of a concern about moisture migra-
northern hemisphere generally occur in late July or early tion and soil drying. These conditions could increase the
August, but the peak ambient soil temperatures are not soil thermal resistivity, although a good-quality thermal
present until early to mid-September. backfill generally mitigates these concerns. A typical
earth interface temperature for large single-conductor
Within the top several meters of the earth, the ambient cables, cables in conduit, or pipe-type cables is on the
soil temperature varies as a function of a variety of order of 50-65°C (smaller diameter cables, e.g., distribu-
parameters, but most significantly due to seasonal variation voltages, might use a lower value because there is a
tions in ambient air temperature and soil depth. When greater heat flux density at a given temperature). At ele-
the ambient soil temperature is not known, it may be vated temperatures, moisture may migrate away from
estimated (Williams and Gold 2003) by knowing the the cable, which causes the earth thermal resistance to
average annual air temperature, time of year, and burial become greater, thus raising the temperature, which
depth by evaluating an equation of the form: causes further drying, and so forth until this thermal
π
− x⋅
⋅ cos⎛⎜ 2tπt − x a⋅πt ⎞⎟
a⋅t 0 11-9
T( t , x ) = TAverage + A ⋅ e
⎝ o 0⎠
11-16
runaway leads to cable overheating and eventual failure. The dielectric heat loss, Wd, can be found from the fol-
Thermal runaway is generally more of a problem with lowing equation:
higher-temperature cable systems and with cables having
a high heat flux, such as direct buried self-contained Wd = 2πf ⋅ C ⋅ E 02 ⋅ tan δ
cables. Chapter 12 addresses this concern in more detail
2πf ( 2π ⋅ ε r ⋅ ε 0 ) E 02 tan δ
and discusses the concept of heat flux, which may be a =
better indication of the tendency of the soil to dry. ⎛ D ⎞
ln ⎜ insulation ⎟
Electrical Resistance and Losses ⎜D ⎟
Dielectric losses and charging-current losses are basi-
⎝ cond _ shield ⎠
cally independent of circuit loading, while current-pro- 2πf ⋅ ε r ⋅ E 02 tan δ ⋅10−9
duced losses vary as the square of circuit loading. Since ≈ (W / m)
⎛ Dinsulation ⎞
the voltage can reasonably be treated as a constant, the 18ln ⎜ ⎟⎟
⎜D
temperature rise due to dielectric loss is calculated and ⎝ cond _ shield ⎠ 11-11
removed from the total allowable temperature rise, as
shown in Equation 11-10. Where:
C = Capacitance, F/m.
ΔTC = TC – Ta – ΔTd C°11-10 11-10 f = power frequency, Hz.
Where: εo = permittivity of free space =
TC = allowable conductor temperature, °C. 8.854 x 10-12 F/m
Ta = ambient earth temperature, °C. εr = relative permittivity of the insulation
ΔTC = conductor temperature rise due to current- (specific inductive capacitance or
dependent losses, C°. dielectric constant), dimensionless.
ΔTd = conductor temperature rise due to dielectric tan δ = insulation dissipation factor, dimen-
losses, C°. sionless.
E0 = line to ground voltage applied across
These losses are discussed in more detail below. the insulation, V.
Dielectric Losses Dinsulation = diameter over the insulation (mm)
Losses are generated in insulating materials that are Dcond_shield = diameter over the conductor shield
subjected to an alternating electrical field as the mole- (mm)
cules attempt to realign themselves each time the field
reverses. Food is heated in a microwave oven by the Since the dielectric losses vary as the square of operat-
same principle—except the fields in microwave ovens ing voltage, cable operation at maximum normal system
alternate at 25 MHz rather than at 50 or 60 Hz as in a voltage (e.g., 242 kV), rather than nominal voltage (e.g.,
transmission line. 230 kV), can increase the losses by as much as 10%.
Actual operating voltage should be used if known; oth-
Dielectric loss, Wd, is directly proportional to the square erwise, the conservative approach is to use the maximum
of the voltage and to the dielectric constant and dissipa- normal system voltage.
tion factor of the insulating material, and is affected by
cable dimensions as shown in Equation 11-11. For The dielectric constant of a material, εr, also known as
crosslinked polyethylene cables up to 138 kV and Specific Inductive Capacity (SIC) or relative permittiv-
impregnated-paper and EPR insulations up to 69 kV, ity, is the ratio of the capacitance of a pair of electrodes
dielectric losses are generally negligible. Because the separated by that material, compared to the capacitance
losses are so great in EPR insulations, they are generally with vacuum as the dielectric—i.e., it determines the
not used for cables above 115 kV. At 230 kV and higher capacitance of a given electrode system. The dielectric
voltages, laminated paper-polypropylene insulation is constant is essentially independent of temperature. Typ-
considered for new self-contained or pipe-type cable ical values are given in Table 11-4.
installations because the losses are about 20% of similar
kraft paper-insulated cables. The dissipation factor of the insulating material is equal
to the tangent of the loss angle of the material, and is
Dielectric heating comes from charging and discharging often called “tan δ”. For small loss angles, usually the
the insulating dielectric at 50 or 60 times per second. case for insulating materials, the dissipation factor,
tan δ, is essentially the same as the dielectric power fac-
tor “cos δ”. Dissipation factor is, therefore, often called
“power factor.” Dissipation factor, tan δ, or power fac-
11-17
tor, cos δ , refer to the real (in-phase) component of each second). The current flow returns through the cable
losses in the insulation, as shown in Figure 11-17. shield. Charging current heats the conductor and shield,
and can be significant for long, high-voltage cables. This
Some typical insulation parameters are listed in Table is sometimes a concern for systems engineers when eval-
11-4. Dissipation factors tend to increase with tempera- uating the stability of an overall power system. The cur-
ture, particularly for paper-based insulations. rent flows any time that the circuit is energized, so
charging-current losses should properly be considered in
Both the dielectric constant and the dissipation factor of the same manner as dielectric losses. The distribution of
a type of an insulating material can be controlled some- charging current along the conductor is nonlinear, being
what during manufacture. Paper or laminated paper- a maximum at the cable end supplying the charging cur-
polypropylene insulations can be controlled by selecting rent and a minimum at the remote end, though often the
a lower-density grade of kraft paper washed with deion- charging current may be supplied from both ends of the
ized water and using a higher percentage of polypropy- cable. Conservative calculations take the hottest position
lene relative to the paper. Modern XLPE cables at and assume that all the charging current is supplied for
transmission class voltages generally use unfilled grades the entire cable length by one end.
of homopolmer polyethylene. The dielectric properties
of crosslinked polyethylene are temperature dependent, Charging current in the conductor is a reactive current
particularly for some grades of copolymer polyethylene. in that it leads phase current by 90°. Dielectric loss in
Tan δ for impregnated-paper insulation is dependent on the insulation is supplied by the real (in-phase) compo-
temperature; it is generally lowest at 75-85°C, as dis- nent of charging current. Dielectric losses are generated
cussed in Chapters 4-7. Dissipation factor is often on a per-meter basis and increase the radial heat gener-
expressed in percent—i.e., a tan δ of 0.0005 is equivalent ated in the cable, ultimately reducing normal ampacity.
to 0.05%. The dielectric loss factor, which is the product Charging current flows all the time that the cable is
of dielectric constant and dissipation factor, is often energized, and the heat that it generates in the conduc-
used as a figure of merit for a particular insulation and tor ultimately limits the allowable real power transfer.
should, of course, be as low as possible. Further discussion is given in Chapter 16.
Charging-Current Losses
Voltages were lower and lines were shorter when the
A cable is a long capacitor, and energy must be supplied
original Neher-McGrath paper was published in 1957,
through the conductor by current flow to charge and dis-
so charging current was a smaller proportion of load
charge the capacitor at power frequency (50 or 60 times
current than for modern high-voltage cable systems. For
this reason, charging current was not separated from
load current and was assumed to have the same load
factor—i.e., to have the same cyclic pattern as the load
current. The proper approach is to determine the maxi-
mum charging current and temperature rise due to that
current at 100% load factor, and subtract that rise from
the total permitted rise. Unfortunately, this is an itera-
tive procedure that requires the solution of the entire
electrical circuit of Figure 11-13. This procedure is rec-
ommended if the circuit load factor is less than 50% and
if voltage is 230 kV or higher. As will be illustrated in
the examples at the end of this chapter, the difference in
Figure 11-17 Dissipation factor and power factor approaches can be a few percent of the rating. If this
for a dielectric material.
approach is followed, the user should be sure that the
Table 11-4 Typical Insulation Material Dielectric Constants and Dissipation Factors
Dielectric Constant Dissipation Factor
(εr) (tan δ)
Insulation Material Range Typical Range Typical
Crosslinked Polyethylene 2.1-2.3 2.3 0.001-0.005 0.001
Impregnated Paper, Fluid-Filled 3.3-3.7 3.5 0.0025-0.0045 0.0025
Impregnated Paper, Gas-Filled 3.3-3.7 3.5 0.003-0.004 0.003
Laminated Paper-Polypropylene 2.7-2.9 2.7 0.0007-0.0010 0.0007
Ethylene-Propylene-Rubber 2.5-4.0 3.0 0.002-0.08 0.005
11-18
daily load factor for current-dependent losses supplied Dc Resistance

by the planning department does not already include Dc resistance can be calculated using Equation 11-14
the effect of charging current. Equation 11-12 shows the using data from Table 11-5. The resistance is a function
calculation of charging current, Ic. Calculation of cable of the cross-sectional area of the material and is defined
capacitance, C, is shown in Equation 11-13. by the following relationship:
2πf ⋅ ε r ⋅ E 0 ⋅ 10 −9 ρ
I C = 2πf ⋅ C ⋅ E 0 = R dc20 = (ohm/m) 11-14
Area ⋅ 10 −6
(A/m)
⎛ D insulation ⎞
18 ln⎜ ⎟ Where:
⎜ D cond _ shield ⎟ ρ = electrical resistivity of metal in ohm- m at
⎝ ⎠ 20°C
11-12
Area = cross-sectional area of metal in mm2.
−9
ε r ⋅ 10
C=
⎛ D insulation ⎞ (F/m) The dc resistance can be adjusted to other temperatures
18 ln⎜ ⎟
11-13
using the following relationship:
⎜ D cond _ shield ⎟
⎝ ⎠ ⎡ T−τ ⎤
Where: R dcT = R dc 20 ⋅ ⎢ (ohm/m)
⎣ 20 − τ ⎥⎦
11-15
f = frequency, Hz.
C = capacitance, F/m. Where:
E0 = line-to-neutral voltage, V. τ = inferred temperature of zero resistance
εr = relative permittivity of the insulation (spe- from Table 11-5, °C.
cific inductive capacitance, dielectric con- T = temperature of the conductor, °C.
stant), dimensionless.
Dinsulation = diameter over the insulation (mm) The resistances must be adjusted to the temperatures at
D cond_shield = diameter over the conductor shield which the materials will operate. For example, the con-
(mm) ductor resistance would typically be adjusted to 90°C
for a crosslinked polyethylene cable. The shield and
Although electrical resistances are commonly stated in sheath resistances, and the skid wire on a pipe-type
microhms/m, charging current is sometimes stated in cable, should also be adjusted to their operating temper-
A/km or A/1000 ft because the quantity is generally ature. This may be done iteratively after performing an
small. The effect of charging current is discussed further initial ampacity calculation or by assuming some typical
in Chapter 16, and in a published reference (IEEE temperature drop across the insulation (10-25C°) below
1990). the conductor temperature.
Current-Dependent Losses Some typical values of electrical resistivity are listed in
Ohmic (I2R) losses arise from the flow of current Table 11-5, along with their temperature correction
through an electrical resistance, such as conductor cur- coefficients.
rents and induced currents in short-circuited cable
sheaths or other parallel conductors having a return
path. In addition, for ac operation, skin and proximity Table 11-5 Typical Electrical Resistivities and Temperature
effects increase the losses in conductors, and eddy-cur- Factors for Common Cable Metals
rent losses are introduced in cable shields and sheaths. Electrical Resistivity Inferred Temperature
Those losses are reflected back to conductor losses by (ρ) of Zero Resistance,
calculating an increment in ac resistance at the conduc- Material Ohm-meters at 20°C τ, °C
tor and shield. If the cables are installed in steel pipe or 2.8264 x 10-8
casing, eddy-current and hysteresis losses are also Aluminum (conductor) -228.1
2.84 x 10-8 (sheath)
reflected back to the conductor. Losses for each compo-
Brass 6.317 x 10-8 -912.0
nent of the cable system are addressed below.
Bronze 3.5 x 10-8 -564.0
Conductor Losses Copper 1.7241 x 10-8 -234.5
Conductor losses equal I2Rac, where Rac is conductor ac Lead 21.4 x 10-8 -236.0
resistance. Conductor resistance consists of three com- Stainless Steel 70 x 10-8 n.a.
ponents: dc resistance, increases due to skin effect, and Zinc 6.633 x 10-8 -218.7
increases due to proximity effect.
11-19
Note that the variation of stainless steel is minimal with lent conductor cross section in square millimeters; values
respect to temperature over the typical operating range are shown at 25°C. (Note that the conversion from
of a cable. square millimeters to kcmil is 1.9735, so a 1000 mm2 con-
ductor has an equivalent area of 1973.5 kcmil.) These
A factor 1.025 (2.5%) is typically applied to the dc resis- resistances may be corrected to temperatures other than
tance to consider the strand-lay factor relative to the 25°C using Equation 11-15 above or the following:
resistance of the equivalent size of solid conductor. The
strand-lay factor increases for conductor sizes larger Temperature Correction for Copper:
than 1000 mm2 (~2000 kcmil). RdcT = Rdc25 (1 + 0.00385 (T-25))
Some U.S. specifications (e.g., AEIC) tabulate dc resis- Temperature Correction for Aluminum:
tance as a function of conductor size, material, and con- RdcT = Rdc25 (1 + 0.00395 (T-25))
struction. These tabulations include the stranding effect
for the conductors. Table 11-6 provides the resistance of The IEC has standard conductor sizes listed in Publica-
standard AEIC conductor sizes in kcmil with the equiva- tion 60228. The IEC does not distinguish conductor
Table 11-6 Dc Resistance of Copper and Aluminum Conductors by AEIC Size (kcmil)
Resistance at 25°C, ohm x 10-6/m
Size of Conductors Copper Aluminum
Round Segmental Hollow Round Segmental Hollow
kcmil mm2 Conductors Conductors Core Conductors Conductors Core
167.8 (3/0) 85.02 210.63 210.63 210.63 344.49 344.49 —
211.6 (4/0) 107.2 166.99 166.99 166.99 274.28 274.28 —
250 127 141.40 141.40 141.40 232.28 232.28 —

300 152 118.11 118.11 118.11 193.57 193.57 —
350 177 101.05 101.05 101.05 165.68 165.68 —
400 203 88.58 88.58 88.58 145.01 145.01 —
450 228 78.74 78.74 78.74 130.91 130.91 —
500 253 70.87 70.87 70.87 — 116.14 —
550 279 64.30 64.30 64.30 105.64 105.64 —
600 304 59.06 59.06 59.06 96.78 96.78 —
650 329 54.46 54.46 54.46 89.24 89.24 —
700 355 50.52 50.52 50.52 83.01 83.01 —
750 380 47.24 47.24 47.57 77.43 77.43 78.08
800 405 44.29 44.29 44.62 72.51 — 73.16
900 456 39.37 39.37 39.70 64.63 — 65.29
1,000 507 35.43 35.43 35.76 58.07 58.07 58.73
1,250 633 28.31 28.44 28.58 46.59 46.59 46.92
1,500 760 23.59 23.72 23.82 38.71 39.04 39.04
1,750 887 20.24 20.31 20.44 33.14 33.46 33.46
2,000 1,013 17.68 17.78 17.88 29.04 29.17 29.20
2,250 1,140 15.88 15.81 15.88 26.05 25.95 —
2,500 1,267 14.30 14.24 14.30 23.46 23.36 —
2,750 1,393 12.99 12.93 12.89 20.93 20.80 21.33
3,000 1,520 11.91 11.84 11.91 19.52 19.42 19.52
3,250 1,647 11.09 10.96 — 17.85 17.65 —
3,500 1,776 10.30 10.17 10.30 16.90 16.70 16.90
3,750 1,900 9.61 9.48 — 15.49 15.26 —
4,000 2,027 9.02 8.89 8.89 14.80 14.60 14.76
11-20
resistances by conductor construction type (e.g., round, ratio”—is a function of the conductor type and con-
segmental, etc.). The standard conductor values are struction.
listed in Table 11-7.
The conductor current density and distribution are
The reader should note that the nominal conductor influenced by several factors:
resistance may differ from the maximum allowable con- • The dependence of interwire resistance on conductor
ductor resistance. The actual resistance of the conduc-
compaction and whether it is impregnated with
tor may be somewhat greater than the nominal value, so
dielectric fluid or is dry, as in an XLPE cable
this should be factored into design and as-built ampac-
ity calculations. • The use of a stranded concentric, a stranded segmen-
tal, or a hollow-core conductor
AC Skin and Proximity Effects
There is unequal current distribution in an ac current-
• The dc resistivity of the metal, the higher resistivity of
carrying conductor due to the magnetic field induced by aluminum being beneficial for current distribution.
the flow of electrical current. As a result, losses in the
conductor are affected by self and mutual inductance; The effect of the particular conductor type and con-
the self-inductance causes the current to concentrate struction is quantified by the conductor skin effect cor-
near the conductor surface—called “conductor skin rection factor, kS, and proximity effect correction factor,
effect”—and the magnetic field of neighboring conduc- kP. The reference unity values are those of a solid cylin-
tors affects the distribution of current across the con- drical conductor, which for large cross-sectional areas
ductor—called “conductor proximity effect.” When gives the greatest (worst) ac-dc ratio. The accurate mea-
determining the total losses from the cable conductor, surement of the ac-dc resistance ratio is difficult and has
these two parameters must be determined. The impact resulted in few published values being available. Early
of these effects on the dc resistance—called the “ac-dc published values are mainly for conductors impregnated
Table 11-7 Dc Resistance of Class 2 Stranded Copper and Aluminum Conductors by IEC Size (mm2)
Resistance at 20°C, ohm x 10-6 / m
Size of Conductors Copper

Aluminum or Alumi-
Plain Wires Metal-Coated Wires num Alloy
mm2 kcmil Conductors Conductors Conductors
50 98.7 387.0 391.0 641.0
70 138.1 268.0 270.0 443.0
95 187.5 193.0 195.0 320.0
120 236.8 153.0 154.0 253.0
150 296.0 124.0 126.0 206.0
185 365.1 99.1 100.0 164.0
240 473.6 75.4 76.2 125.0
300 592.1 60.1 60.7 100.0
400 789.4 47.0 47.5 77.8
500 986.8 36.6 36.9 60.5
630 1243.3 28.3 28.6 46.9
800 1578.8 22.1 22.4 36.7
1,000 1973.5 17.6 17.7 29.1
1,200 2368.2 15.1 15.1 24.7
1,400 2762.9 12.9 12.9 21.2
1,600 3157.6 11.3 11.3 18.6
1,800 3552.3 10.1 10.1 16.5
2,000 3947.0 9.0 9.0 14.9
2,500 4933.8 7.2 7.2 12.7
11-21
with dielectric fluid, as used in HPFF and SCFF cables Where:

(examples are listed in Table 11-8, including values from Area = is the cross-sectional area of the conductor,
the Neher-McGrath paper). The recent availability of mm2
XLPE cables with large stranded segmental conductors di = is the inner diameter of the conductor (e.g.,
has shown that k s and k p values are higher than for the central duct diameter), mm.
fluid-filled conductors. As a result, CIGRE Technical d'c = is the outer diameter of an equivalent solid
Brochure 272-220 (CIGRE 2005) recommends that, for conductor having the same central duct
large segmental conductors of 1600 mm2 (3158 kcmil), diameter, mm.
the skin and proximity effect factors should be mea-
sured when the cables are qualification tested; if mea- Where specific values for aluminum conductors do not
surements are unavailable, the values from Table 11-9 exist, the copper values may be used as conservative esti-
should be used. mates. Segmenting is important in reducing conductor
ac losses for large conductor siz es. A 1200 mm 2
During conductor selection, the construction of the (2368 kcmil) four-segment copper conductor has an
conductor may be specified to optimize the losses in the ac/dc resistance ratio of about 1.08. That ratio would be
conductor. Some manufacturers have considered a greater than 1.25 if the conductor had a standard con-
higher number of segments (7 or 9) in the conductor to centric construction. The ac/dc resistance ratio is less for
reduce skin and proximity effects. Values of the conduc- aluminum conductors of equivalent size because alumi-
tor skin effect correction factor and proximity effect num is characterized by a lower conductivity, and the
correction factor (also called the “transverse conductiv- oxide on individual strands gives some segregation of
ity factor”) are listed in Tables 11-8 and 11-9. currents. Some manufacturers offer five- and six-seg-
ment conductors, while others put a cupric oxide or
For hollow-core conductors, the skin effect factor, kS, enamel coating on individual conductor strands to
may be calculated using the following equations: improve the ac/dc ratios (reduce the skin and proximity
effects). The values of kS and kP in Table 11-9 provide a
4 ⋅ Area relative comparison of the impact on ac resistance
d'c = + di2
π 11-16
ratios.
2 An approximation of a Bessel Function is used in

⎛ d ' −d ⎞⎛ d'c +2 ⋅d i ⎞
k S = ⎜⎜ c i ⎟⎟⎜⎜ ⎟⎟ Neher-McGrath and IEC methods to calculate skin and
⎝ d'c +d i ⎠⎝ d 'c +d i ⎠ 11-17
proximity effect factors. The increment in ac resistance
from skin effect is a function of conductor size and con-
Table 11-8 Skin and Proximity Effect Factors for Various struction, but typically adds 5-12% to the dc resistance
Conductor Types of the conductor. The ac resistance increment for con-
Conductor Type kS kP ductor skin effect, YCS, can be found from Equations
Concentric round, dry, Cu 1.0 1.0 11-18 and 11-19, where kS is the conductor skin effect
Concentric round, in oil, Cu 1.0 0.8 factor based on the conductor construction and XS is
Compact round, in oil, Cu 1.0 0.6 between 0 and 2.8. ELECTRA No. 272 (CIGRE 2005)
Compact 4-segment, dry, Cu 0.435 0.6
Compact 4-segment, in oil, Cu 0.435 0.37 Table 11-9 Skin and Proximity Effect Factors for Large
Compact 4-segment, in oil, 0.35 (trefoil), 0.46 Segmental Conductors
0.39
pipe-type, Cu (cradled) Conductor Type kS kP
0.29 (trefoil), 0.36 Bidirectional stranding of copper
Compact segmental, Al 0.35 0.8 0.37
(cradled) bare wires
Compact 5-segment 0.435 0.37 Unidirectional stranding of copper
0.62 0.37
See bare wires
Hollow, stranded
below 0.33 (Cu), Interlayer insulated copper bare
0.5 0.37
Hollow core, 6-segment, in oil, 0.19 (Al, trefoil or wires
0.39
Cu square), Unidirectional copper oxidized
Hollow core, 6-segment, Al, 0.27 (Al, flat or vertical) (cupric oxide) coated conductor 0.35 0.2
0.26 strands
in oil
Hollow core, 4-segment, in oil 0.435 0.37 Enamel-coated conductor strands 0.25 0.15
11-22
notes that some large conductors now fall outside this distorted and resistances are increased. Neher-McGrath
range and gives an improved formulae for YCS. had proposed a value of 1.7 to account for incremental
losses when cables are installed in steel pipes, but later
2
⎛ 8πfk S ⋅ 10 − 7 ⎞ work by Katz (Katz et al. n.d.) suggested a value of 1.5
XS = ⎜
4 ⎟ would be more accurate.) The ac/dc ratio for the cable is
⎜ R dcT ⎟ the ratio of the ac resistance of the conductor, including
⎝ ⎠
skin and proximity effects, to that of the conductor’s dc
11-18
resistance alone.
4
XS
YCS = Shield, Sheath, and Skid wire Resistance
4
192 + 0.8 ⋅ X S Current in the conductor induces voltages in parallel con-
11-19 ductors, including the outer metallic layers of the cable.
Depending on the cable type, there are various metallic
When multiple (typically multiples of 3 phases) ac layers outside of the cable insulation that are affected this
current-carrying conductors are brought together, the way. A summary of typical layers is as follows:
magnetic field from each conductor causes the current
in adjacent conductors to shift to the near side of the Extruded: concentric stranded copper or aluminum
conductor, which increases the effective resistance; this wires, copper or aluminum tapes, extruded
is the proximity effect. The proximity effect is a function lead, corrugated copper, corrugated alumi-
of cable phase spacing, and can add 2-5% to the dc resis- num or corrugated stainless steel, possibly
tance of the conductor for typical spacings. When multi- armor for submarine installations.
ple circuits are involved, the proximity effect is
conservatively determined using the nearest cable phase Pipe-Type: helically applied metallic tape(s), helically
spacing. The ac resistance increment for conductor applied skid wire(s).
proximity effect, YCP, can be found from Equation 11-20
(see below), where kP is the proximity effect factor based SCFF: extruded lead alloy or corrugated alumi-
on the conductor construction. num sheath (corrugated copper has been
used very rarely), possibly with additional
The conductor resistance, including skin and proximity copper wires or tapes, and perhaps with
effects, can be calculated using Equations 11-21 and armor wires for submarine cables.
11-22:
For cables in air or soil: If the shield on an extruded or self-contained cable is
R ac = R dcT ⋅ (1 + YCS + YCP ) , ohm/m 11-21 grounded at more than one point, the induced voltage
causes current flow in the shield, which results in ohmic
For cables in a steel pipe or casing:
losses. The circulating current greatly reduces eddy cur-
R ac = R dcT ⋅ (1 + 1.5 ⋅ (YCS + YCP )) , ohm/m rent losses. However, losses in metallic shields/sheaths of
11-22 single-conductor cables from circulating currents can be
substantial and may result in as much as a 35% reduc-
(For cables in a steel pipe or casing, conductor ac resis- tion in cable ampacity, as compared to a single-point
tances are calculated for cables as if they were in air. If bonded shield/sheath or one that has perfect cross-
the cables are installed in a ferromagnetic casing, con- bonding.
duit, or pipe, as for HPFF cables, the magnetic field is
2
⎛ 8πfk P ⋅10−7 ⎞
XP 4
=⎜ ⎟
⎝ R dcT ⎠
⎛ ⎞
⎛ 4
⎞ ⎛ DConductor ⎞
2
⎜ ⎛D ⎞
2
⎟
XP 1.18
YCP = ⎜ ⋅
4 ⎟ ⎜ ⎟⎟ ⋅ ⎜ 0.312 ⋅ ⎜ Conductor ⎟⎟ + ⎟
⎜ ⎜ ⎜ ⎟
⎝ 192 + 0.8 ⋅ X P ⎠ ⎝ d phase ⎠ ⎝ d phase ⎠ XP4
+ 0.27 ⎟
⎜ 192 + 0.8 ⋅ X P 4
⎝ ⎠ 11-20
Where:
DConductor = outside diameter of conductor, mm
dphase = axial spacing between adjacent cables, mm
11-23
As an alternative, many short (< 800 m [< 2600 ft]) DInner = is the inner diameter of the shield layer,
transmission cable circuits are single-point bonded mm.
(shield/sheath directly grounded at only one location, so Area = is the cross-sectional area of the shield
circulating currents cannot develop). For longer circuits, layer, mm2.
the cable shield/sheath connections are transposed at a ρ = is the electrical resistivity of the metallic
third of the circuit length or major section length; when layer, ohm-m.
the difference in lengths is less than 15%, the net circu-
lating current is very small, so the ratings approach If a longitudinally taped foil shield or sheath is used, the
those of similar single-point bonded cables. The impact calculations should factor in the additional foil material
of larger discrepancies in the three minor section lengths used where the foil is overlapped, usually about 15 mm
should be considered for ampacity. The topic of (0.6 in.).
shield/sheath bonding is discussed in Chapter 10, and
the calculation of circulating currents is discussed in The following equation can be used for a helically-
greater detail in Chapter 16. applied metallic layer—e.g., a helically applied tape
intercalated with an insulating tape (or a tape where the
Unlike single-core cables, pipe-type cables do not suffer lay length is greater than the tape width), skid wire, or
a significant reduction in ampacity from the shield/skid armor. However, if the taped shield is not intercalated
wire being continuously bonded, because these layers with an insulating tape, Equation 11-24 for a non-taped
are generally high resistance, and the close phase spac- layer (or tube) should be used.
ing better cancels magnetic fields. The shield/skid wire
losses in pipe-type cables are generally less than 5%. 2
ρ ⎛ π ⋅ D Layer ⎞
R dc20 = ⋅ 1+ ⎜ ⎟ (ohm/m)
N ⋅ Area ⋅ 10 − 6 ⎜ Lay Layer ⎟
Regardless of the cable type, the resistance of the metal- ⎝ ⎠
lic layers outside the insulation of each cable phase must 11-25
be evaluated. These resistance values would then be Where:
adjusted for the temperature of the respective layer and ρ = is the electrical resistivity of layer, ohm/m.
appropriately applied to calculate eddy current and cir- DLayer = is the average diameter of the helical layer,
culating current loss increments. mm.
LayLayer = is the distance along the cable for one turn
Calculation of Shield, Sheath, and Skid wire Resistances of the tape or wire (e.g., the “lay”), mm.
The calculation of the dc resistance of the shield, sheath, Area = is the cross-sectional area of a single metal-
or skid wire (for the remainder of this section, these will lic tape, wire or skid wire, mm2.
all be referred to as “shields”), Rs, is given in Equations N = is the number of shield wires, shield tapes,
11-23 to 11-25. Electrical resistivity, ρ, is obtained from or skid wires.
Table 11-5, and the calculated resistance is then adjusted
to the expected operating temperature of the component For wire-strand shields, the lay of the wires depends on
using Equation 11-15. Equations for specific shield con- the manufacturer. A typical value for the lay is approxi-
structions are listed below in Equations 11-24 and 11-25. mately seven times the wire shield diameter (e.g., if the
cable diameter at the wire shield is 90 mm, then the wire
The electrical resistance, Rs, for a non-taped (tube) lay would be approximately 630 mm). For metallic taped
layer, such as an extruded or welded metallic sheath shields, the tape width is often about 22 mm (0.875 in.),
(e.g., the lead sheath on an extruded or self-contained with the tape lapped by 3 mm (0.125 in.), so the lay is
cable) can be calculated by knowing the cross-sectional 19 mm (0.75 in.).
area (Equation 11-23) of the layer and then using the
equation for resistance (Equation 11-24): For taped shields, the area is the width of the tape times
Area =
( 2
π D Outer )
− D 2Inner (mm2) 11-23
the thickness of the tape. For a skid wire, the area is:
4 π ⋅ D1 ⋅ D 2
Area = (mm2) 11-26
ρ 4
R dc 20 = (ohm/m) 11-24 Where:
Area ⋅ 10 − 6 D1 = height of the skid wire (usually half the
Where: diameter), mm.
DOuter = is the outer diameter of the shield layer, D2 = width of the skid wire, mm.
mm.
11-24
As an example, a 2.54 mm high by 5.08 mm wide (0.1 in. These currents also exist in pipe-type cables, but are
x 0.2 in.) “D”-shaped skid wire would have an area of generally very small because of close phase spacing and
10.13 mm2 (0.0157 in.2 or about 20 kcmil). high-resistance shields and skid wires.
For a stranded shield, the area of a single strand must be Calculation of the circulating current increment requires
calculated using the following: determining the average mutual reactance between adja-
cent cable phases and depends on the configuration of
2
π ⋅ D strand the phases. Equations 11-29 to 11-31 should be applied
Area = (mm2) 11-27 appropriately depending on the cable configuration that
4 most closely matches the indicated cable positions. S,
Where:
the center-to-center phase spacing, and, DSM, the mean
Dstrand = diameter of a single wire shield strand,
diameter of the shield/sheath/skid wire layer, are used in
mm.
the following:
When different metals are involved for the shield layers,
For cables in equilateral (trefoil or triangular) configu-
the resistance of each layer should be adjusted sepa-
ration:
rately to the expected operating temperature of the
shield. Typically, this is 10-25C° lower than rated con-
⎛ 2⋅S ⎞
ductor temperature. Finally, the total resistance for all X m = 4πf ⋅ 10 − 7 ⋅ ln⎜⎜ ⎟⎟ (ohm/m) 11-29
of the shield layers is determined by taking the resis- ⎝ D SM ⎠
tances electrically in parallel, as follows:
For cables in flat/vertical configuration:
1 1 1 1
= + + .... +
R S R s1 R s 2 R sn ⎛ 2.52 ⋅ S ⎞
11-28 X m = 4πf ⋅ 10 − 7 ⋅ ln⎜⎜ ⎟⎟ (ohm/m) 11-30
D
⎝ SM ⎠
The shield resistance calculated in Equation 11-28 is not
used directly. It is used to calculate the two components For cables in cradled configuration:
for ac shield losses: an increment, YCC, due to circulat-
ing currents in the shield, and an increment, YEC, due to ⎛ 2 .3 ⋅ S ⎞
eddy currents in the shield. These loss increments are a X m = 4πf ⋅ 10 − 7 ⋅ ln⎜⎜ ⎟⎟ (ohm/m) 11-31
function of the resistance, mutual reactance (described ⎝ D SM ⎠
below), and sheath bonding mode. Where:
f = power frequency, Hz.
Mutual Reactance and Shield Loss Increments for S = axial spacing between nearest phases, mm.
Circulating Currents DSM = mean diameter of the metallic shield,
Circulating currents exist when a cable metallic shield, sheath and/or skid wire, mm.
sheath, neutral, or skidwire is grounded and bonded at
both ends (e.g., multipoint bonding—see Chapter 10). Equations 11-29 to 11-31 are taken from IEC 60287 and
The resulting induced current is a function of the resis- assume that the cable phases are transposed; equations
tance and spacing among phases according to Lenz’s for nontransposed cables are also given in IEC. A
law. Circulating currents exist in single-core cables (e.g., sophisticated matrix solution for the nontransposed
extruded-dielectric or self-contained cable types), when case is discussed in Chapter 16.
the metallic sheath is grounded at both ends or on cross-
bonded systems, where the minor section lengths are not Since the goal is to calculate the conductor current, the
equal in length. Circulating-current losses can be large shield current is not calculated separately. Instead, the
for single-conductor cables with spacings greater than incremental increase in losses from shield circulating
100 mm (4 in.) and may reduce ampacity by 35%. currents can be found and reflected back as an increased
Because of these large losses, single-conductor transmis- conductor resistance by using Equation 11-32.
sion cables are typically configured with open-circuited
sheaths unless sheath voltages become excessive. In this RS 1
case, the sheaths are often cross-bonded. For long YCC = ⋅
R ac 2
⎛R ⎞
1 + ⎜⎜ S
underwater crossings, where cross-bonding is not feasi-
⎟⎟
ble, short-circuited sheaths are the only option.
⎝ Xm ⎠
11-32
11-25
Where: For cables in triangular (trefoil) formation:

YCC = incremental increase in losses from shield
2
circulating currents, dimensionless ⎛ m 2 ⎞ ⎛ DSM ⎞
RS = total shield/sheath/skid wire resistance, YSe0 = 3 ⋅ ⎜ ⋅
2 ⎟ ⎜ ⎟
ohm/m. ⎝ 1+ m ⎠ ⎝ 2 ⋅S ⎠
Rac = ac resistance of conductor, ohm/m. 0.92⋅m +1.66
⎛D ⎞
Xm = mutual reactance, ohm/m. YSe1 = (1.14 ⋅ m 2.45
+ 0.33) ⋅ ⎜ SM ⎟
⎝ 2 ⋅S ⎠ 11-35
For cross-bonded cable systems, many ampacity calcu-
lations assume perfect cross-bonding and no separate Then, the increment in losses for eddy currents may be
ground continuity conductors such that there is no net calculated:
circulating current in the cable shields. However, practi-
cal placement of joints in actual installations often 8π 2 f
means there is some discrepancy among the minor sec- X ec =
tion lengths of a cross-bonded cable system. In these ρShield ⋅ 10 7
11-36
cases, a net circulating current could develop that gener- 1.74
ates sufficient heating to cause a derating effect. Usually, ⎛t
g s = 1 + ⎜⎜ Shield
⎞
⎟⎟ (
⋅ X ec ⋅ D S ⋅ 10 − 3 − 1.6 )
a difference—longest to shortest—in minor section ⎝ DS ⎠
lengths of less than 20% allows the circulating currents 11-37
to be neglected. In situations where there is a greater
⎛R ⎞ ⎛ (X ⋅ t )4 ⎞
discrepancy in the lengths or where one or more ground YEC = ⎜⎜ S' ⎟⎟ ⋅ ⎜ g s ⋅ YSe0 ⋅ (1 + YSe1 ) + ec Shield ⎟
continuity conductors is used, a sophisticated solution ⎝ R ac ⎠ ⎝⎜ 12 ⋅ 1012 ⎟
⎠
method as described in Chapter 16 is needed to find the
11-38
circulating current and determine the loss increment. Where:
Shield Loss Increments for Eddy Currents RS ' = resistance of only the layer(s) where eddy
Eddy-current losses are principally of interest for single- currents may occur, ohm/m.
conductor cables with open-circuit sheaths or nearly Rac = ac resistance of conductor including skin
equal minor section lengths on cross-bonded cables; and proximity effects, ohm/m.
they are often neglected for cables in steel casings or tShield = thickness of the shield or sheath layer for
cable pipes. Eddy-current losses occur when a continu- eddy current loss calculations, mm.
ous concentric metallic layer exists around the cable DS = diameter over cable sheath, mm. (For cor-
core (e.g., a corrugated or extruded metal sheath or lon- rugated sheaths, use the mean outside
gitudinally taped metallic shield, but not stranded diameter: ½ (diameter over the crests +
shields). The mechanics for calculating the eddy-current diameter over the troughs) + sheath thick-
losses, borrowed from IEC 60287, are somewhat oner- ness.)
ous but not particularly complicated. The equations to DSM = mean diameter of the metallic shield,
sheath, and/or skid wire, mm.
perform these calculations were derived empirically and ρ shield =Electrical resistivity of shield at operating
are listed below. For additional background, please see
IEC 60287. temperature, ohm-m.
m = calculation coefficient.
2πf ⋅ 10 −7 gS = calculation coefficient.
m= Yse0 = calculation coefficient.
R S' 11-33 Yse1 = calculation coefficient.
For cables in flat or vertical formation (center cable in Other variables were defined earlier. Eddy-current losses
the group): for cables in cradled configuration (e.g., pipe-type) can
be neglected.
2
⎛ m 2 ⎞ ⎛ DSM ⎞
YSe0 = 6⋅⎜ ⋅
2 ⎟ ⎜ ⎟
⎝ 1+ m ⎠ ⎝ 2 ⋅S ⎠
1.4⋅m + 0.7
⎛D ⎞
YSe1 = 0.86 ⋅ m 3.08
⋅ ⎜ SM ⎟
⎝ 2 ⋅S ⎠ 11-34
11-26
AC Resistance Including Circulating Current and Eddy- For cradled configurations:

current Increments
The heat generated by the conductor is calculated using 0.0438 ⋅ D Skidwire + 0.0226 ⋅ ID Pipe
Equation 11-39 and includes the conductor skin effect YPL =
and conductor proximity effect: R ac ⋅ 10 6
11-44
WConductor = I 2 R dcT (1 + YCS + YCP ) = I 2 R ac W/m
11-39 For trefoil configurations:
When calculating ampacity, the losses in the metallic 0.115 ⋅ D Skidwire − 0.01485 ⋅ ID Pipe
YPL =
R ac ⋅ 10 6
shield, sheath, and/or skid wire are assumed to be
directly proportional to the conductor current. As such,
the ampacity equations are arranged in a manner to 11-45
allow for the solution of the conductor current in terms Where:
of all the losses from the conductor, shield/sheath/skid DSkidwire = diameter over cable skid wires, mm.
wire, and (for pipe-type systems) the cable pipe. A simi- IDPipe = Inner diameter of the cable pipe, mm.
lar relationship may also be developed for cables in cas-
ings or that have metallic armor. The losses in the The Equations 11-44 and 11-45 are for 60 Hz. If the
shield/sheath/skid wire include the circulating current power system is 50 Hz, the values of YP should be multi-
and eddy-current effects: plied by 0.76.
WShield = I 2 R ac (YCC + YEC ) W/m 11-40 The equations are empirical because the actual pipe
losses are dependent on many factors including the
magnetic permeability and resistivity. The permeability
The total losses applied to cable layers in the equivalent
is a function of the current (magnetic field) and point of
thermal circuit outside the shield include both the con-
saturation on the pipe’s B-H curve; these characteristics
ductor losses and shield losses. A loss ratio, QS, can be
are usually not known by the cable engineer for a spe-
calculated for each layer as follows:
cific line pipe. The thickness does not typically affect the
WConductor + WShield pipe losses (and does not appear in the equation above)
QS = = 1 + YCC + YEC because practical pipe thicknesses are beyond the skin
WConductor depth at power frequency.
11-41
The incremental pipe losses are injected into the thermal
If the cables are installed in ferromagnetic conduits, cas- circuit at the pipe portion of the thermal circuit:
ings, or pipes, the losses are:
WPipe = I 2 R ac (YPL ) (W/m)
WConductor = I 2 R dcT (1 + 1.5(YCS + YCP ))
11-46
W/m
11-42 Losses calculated from Equation 11-46 are on a per-
WShield = I 2 R ac (1.5 )( YSC + YEC ) conductor basis, for each of the three conductors in a
pipe-type cable. This is necessary, since the calculated
= I 2 R dcT (1 + 1.5 ( YCS + YCP ) ) ( YSC + YEC ) W / m cable ampacity is for each conductor, even for a pipe-
type system. Pipe losses can account for as much as one-
11-43
Where: third of the total ohmic losses for pipe-type cable. (Katz
RdcT = dc resistance of the cable conductor at et al. 1978) and (EPRI 1982) give further discussions of
temperature, T (usually rated), ohm/m. pipe losses and calculation procedures.
The loss ratio would be calculated accordingly. Note that the pipe-loss increment, YPL, is not multiplied
by the 1.5 factor. Note also that the pipe-loss increment
Increments for Pipe Losses is several percent greater for cables in cradled configura-
Pipe cables have additional losses from eddy current and tion than triangular configuration, because the cradled
hysteresis heating in the cable pipe as a result of the ac configuration places the cable phases in closer proximity
cables within the pipe. Empirical equations are used for to the cable pipe and has a greater average spacing (geo-
calculating incremental pipe losses in pipe-type cables metric mean radius) among phases. Pipe-type cables are
using the following equations: cradled most of the time, unless the ratio of the inside
diameter to cable diameter is less than 2.5. Therefore,
11-27
cradled configuration should generally be assumed, using the appropriate geometric factors and the mate-
unless the cables are known to lay in close trefoil config- rial’s thermal resistivity.
uration.
Thermal Resistances of Insulation, Cable Coverings, and
Duct
In a similar manner as for the shield/sheath/skid wire,
The calculation of thermal resistance is an extension of
the loss ratio for pipe losses, QP, may also be calculated:
Figure 11-18, except that most cable system compo-
WConductor + WShield + WPipe nents, being cylindrical, follow a relationship like the
QP = following:
WConductor
ρ DOuter dr
= 1 + 1.5 ( YCC + YEC ) + YPL
2π ∫DInner r
11-47
R=
Total Ac Resistance, AC/DC Ratio, and Loss Ratios (Qs, Wp)
ρ
Total ac conductor resistance equals the dc resistance = ( ln(DOuter ) − ln(D Inner ) )
plus the sum of the ac increments, as described by the 2π
equations earlier in this section. The ac/dc ratio of a ρ ⎛ ⎞
cable system is defined as the total overall ac resistance = ln ⎜ D Outer ⎟ (C° - m / W)
divided by the dc resistance of the conductor, for all 2 π ⎝ D Inner ⎠
11-48
components at their normal operating temperature. Where
DOuter is the outer diameter of the layer in milli-
Some ampacity calculations tabulate the value Q s, as meters, and
described above, which is the ratio of losses produced by DInner is the inner diameter of the layer in milli-
the shield and conductor, to that of only the conductor. meters. A bar is used over the thermal
Similarly for pipe-type cables, a value, Q p, is used to resistance terms to avoid confusion with
describe the ratio of losses in the pipe, shield, and con- electrical resistances.
ductor to those only in the conductor, including “1.5”
factor. These values will be discussed in later sections of Equation 11-48 above applies to the cable insulation,
this chapter. cable jackets and coverings, conduits, and pipe coatings.
When more than one cable is inside the layer, such as the
Thermal Resistances
case with pipe-type cables (n = 3), a factor of “n” repre-
The electrical analog of the cable thermal circuit is given
sents the number of cables producing heat within that
in Figure 11-3. The previous section of this chapter
diameter; this permits one cable only to be modeled
addressed loss (ac) calculations; the present section
inside the pipe.
describes thermal resistance calculations. The flow of
losses through these resistances gives the temperature at
ρ ⎛D ⎞
each of the nodes in the equivalent thermal circuit of R= ⋅ n ⋅ ln⎜⎜ Outer ⎟⎟ (C°-m/W) 11-49
Figure 11-3. 2π ⎝ D Inner ⎠
Thermal resistivities are given in units of C°-m/W (or Depending on the layer involved, Table 11-10 summa-
K-m/W), though units of C°-cm/W (a factor of 100 rizes what parameters to use for DInner and DOuter. Ther-
times) are sometimes used with soil thermal resistivity
test results (e.g., a value of 120C°-cm/W is equivalent to
1.20 C°-m/W). The units describe that a 1-W heat flow
through a cube 1 m on each side, having a thermal resis-
tivity of 1C°-m/W, will cause a 1C° rise across the cube,
as shown in Figure 11-18. The symbol for thermal resis-
tivity is ρ, the lowercase Greek letter “rho” with a bar
over the variable to distinguish it from an electrical
resistivity. One often hears the term “the rho of the
soil,” which is simply the thermal resistivity in C°-m/W
or the inverse of thermal conductivity.
Thermal resistivity is a property of the material, as is

electrical resistivity. Thermal resistance is calculated Figure 11-18 Definition of thermal resistivity.
11-28
mal resistances for cable to conduit or pipe and the Table 11-11 Thermal Resistivities of Common Cable
earth components will be discussed in later sections. Materials
Range Typical
Some of these parameters are illustrated in Figure 11-19. Material (C°-m/W) (C°-m/W)
6.0 (HPFF),
Impregnated Paper 5.0-6.0
The thermal resistivities of common cable materials are 5.0 (SCFF)
listed in Table 11-11. 6.0 (HPFF),
Laminated Paper-Polypropylene 5.5-6.5
5.5 (SCFF)
Crosslinked Polyethylene 3.5-4.0 3.5
Conductor and insulation semiconducting shields are
Ethylene-Propylene-Rubber 4.5-5.0 4.5
considered part of the insulation for most thermal resis-
Somastic 1.0 1.0
tance calculations. Conductor losses flow through the
PVC 4.0-4.5 4.0
entire insulation, and the dielectric losses are taken as
Neoprene 3.8-5.8 4.0
flowing through half of the insulation thermal resistance
Epoxy Resin 0.7-4.45 1.0
Water-Swellable Tapes 6.0-15.0 7.5
Thermoplastic Pipe Coating 3.5-4.5 4.0
Table 11-10 Calculation Parameters for Thermal Resistance
Layers
for calculation purposes. Rigorous evaluation of inte-
Cable Layer or
Component DInner (mm) DOuter (mm) grated dielectric losses and heat flow has shown that this
Diameter over non-
approach is correct for homogeneous insulation struc-
Conductor outer tures.
Insulation metallic insulation
diameter, Dc
shield, Di
Metallic Metallic shields and sheaths, steel casings and pipes, and
Diameter over cable
Jacket shield/sheath outer
diameter, Ds
jacket, Dj metallic conduits have negligible thermal resistances.
Conduit inner Conduit outer Cable to Conduit (or Pipe) Thermal Resistance
Conduit
diameter, Ddi diameter, Ddo
Equation 11-50 solves the empirical Neher-McGrath-
Pipe outer diameter, Diameter over pipe developed equation for the cable-to-conduit thermal
Pipe Coating
Dpo coating, Dpc
resistance for the major transmission cable applications.
Plastic Casing Casing inner Casing outer
(HDDa, etc.) diameter diameter A mean duct air temperature of 75°C or pipe-filling (liq-
uid or gas) of 60°C is usually assumed, although the
a. “HDD” is an abbreviation for horizontal directional
value may be iterated when doing detailed ampacity cal-
drilling.
culations. The thermal resistance between a cable and
the inside of a conduit or pipe can be determined using
the following, with values from Table 11-12:
n ⋅U
R= (C ° − m / W )
1 + 0.1(V + Y ⋅ Tmean ) ⋅ Dcable
11-50
Where:
Tmean = mean temperature of the duct air, or nitro-
gen gas or dielectric liquid in the pipe, °C.
Dcable = outer diameter of the cable, mm, when one
cable is installed in the conduit, or is equal
to 2.15 x outer diameter of the cable, mm,
when three cores are installed inside a
common conduit or pipe.
n = number of cables within the conduit or
pipe.
U, V, and Y as defined in Table 11-12.
Dcable for single-core cables is the diameter over the cable

jacket. Dcable for pipe-type cables is the diameter of the
Figure 11-19 Dimensions for thermal cable underneath the skid wires. The values from Table
resistances of cable layers and coverings. 11-12 are empirical; specific values for PVC or PE ducts
11-29
Table 11-12 Pipe and Duct Thermal Resistance Constants − Wρ ∞ dr − Wρ

(ln(∞) − ln(d' ) ) (C°)
2π ∫d ' r
ΔTP − = =
Configuration U V Y
2π
Fiber Duct (PVC or PE) in Con- 11-52
5.2 0.91 0.010
crete
Asbestos Cement in Concrete 5.2 1.1 0.011 The combined effect of the two temperature rises is the
Pipe-Type, HPGF 0.95 0.46 0.0021 following:
Pipe-Type, HPFF 0.26 0.0 0.0026
Earthenware Ducts 1.87 0.28 0.0036 ΔTP = ΔTP + + ΔTP −
Wρ
were not developed by Neher-McGrath, so values for = ( ln(d ') − ln(d) )
fiber duct are typically used. 2π
Wρ ⎛ d ' ⎞
Additional information on thermal resistance from = ln ⎜ ⎟
cable to duct or pipe wall is given in Appendix I of the
2π ⎝ d ⎠ 11-53
Neher-McGrath paper and in References (Buller and
Neher 1953) and (Morris and Burrell 1954). For self- Where d’ equals d, the total temperature rise is zero,
cooled, pipe-type cable circuits, the temperature drop which means that there is an isotherm halfway between
through the fluid to the pipe is calculated to be about the locations of +W and –W; for ampacity calculations,
6% of the total thermal resistance. For forced-cooled this is assumed to correspond to the earth’s surface (or,
cables, where heat is removed by the cable fluid, the in the case of submarine cables, the interface between
thermal resistances of the fluid films at the cable shield the seabed and the water).
and pipe wall are also important; see Section 11.4.
Earth Thermal Resistance By substituting d' = (h + x ) 2 + y 2 ,
The cable thermal resistances describe components of
the thermal circuit out to the interface with the sur- d'
rounding soil. For direct-buried cables, this is the pipe d = (h − x ) 2 + y 2 , and k = , the following
d
coating (on pipe-type cables) or the jacket on single- equation can be developed:
conductor cables. Single-core cables installed in con-
duits have an earth interface at the outside of the con- 2 2
⎡ k2 +1 ⎤ 2 ⎡ 2k ⎤
duit. At that point, the earth portion of the thermal
⎢x − 2 h⎥ + y = ⎢ h⎥
circuit starts. ⎢⎣ k − 1 ⎥⎦ ⎣k2 −1 ⎦
11-54
The thermal resistance of the earth portion of the circuit
often accounts for more than half of the total thermal By rearranging this equation and completing the square,
resistance and is the most variable of the resistances. values of r and L may be assigned to portions of Equa-
Modeling of the earth thermal circuit is quite complex. tion 11-54:
The foundation of the calculation method used for 2
2k
ampacity calculations is to consider the temperature rise r= h and L = k + 1 h 11-55
at a point, P, due to two heat sources that are equal but k2 −1 k2 −1
opposite located in an infinite medium of uniform ther-
mal resistivity, as shown in Figure 11-20. If the thermal
resistivity is assumed to be uniform and is unaffected by
temperature, then superposition can be used to determine
the temperature effect of each heat source at Point P.
The heat sources, +W and –W, can be thought of as infi-

nitely long, for which the temperature rise for one unit
length (e.g., 1 m) at point P is defined by the following
equations:
+ Wρ ∞ dr + Wρ
(ln(∞) − ln(d) ) (C°)
2π ∫d r
ΔTP + = =
2π Figure 11-20 Fundamental configuration for the
11-51 earth thermal resistance calculation.
11-30
Solving these two equations in terms of k gives: diameter pipes when (2L/Dearth) ≤ 10, in which the earth
surface is no longer isothermal. For example, a 345-kV
L + L2 − r 2 pipe-type cable in a 273 mm (10.75 in.) cable pipe at a
k= burial depth of 0.8 m (32 in.) (CIGRE 2003) has a ratio
r 11-56 of about 6:1. A combined equation is given in CIGRE
TB 218 using a heat transfer coefficient β and a parame-
Then, substituting back into Equation 11-53: ter h1, to allow for convection in air at the soil surface.
For a soil thermal resistivity of 1.2 C°-m/W, β is typi-
Wρ ⎛ d ' ⎞ cally ≤0.6 (C°-m2/W).
ΔTP = ln ⎜ ⎟
2π ⎝ d ⎠
β
h1 = 2(L + )
Wρ ⎛ 2L + 4L − Dearth ⎞
2 2
⎜ ⎟ ρ
= ln
2π ⎜ Dearth ⎟
⎝ ⎠ 11-57
⎢ln ⎜ 1 ( 1 earth ) ⎟⎥
⎡ ⎛ 2h + 4h 2 − D 2 ⎞ ⎤
⎢ ⎜ Dearth ⎟⎟ ⎥
Dividing both sides by the heat, W, gives the thermal ρ ⎢ ⎜⎝ ⎠⎥
resistance to ambient earth of a single cable of diameter, R earth =
Dearth, installed at depth, L, below ground: 2π ⎢ ⎡ h 2 +S 2 ⎤
⎥
⎢ ⎥
⎢ +2 ⋅ ln ⎢ ⎥
1 P
⎥
⎛ 2 2 ⎞ ⎢⎣ SP ⎥⎦
ρ ⎜ 2L + 4L − D earth ⎟ ⎣⎢ ⎦⎥
R earth = ln⎜ ⎟
11-59
2π ⎜ D earth ⎟ Where:
⎝ ⎠ 11-58 R earth = thermal resistance of earth plus mutual
This basic approach considers that the earth surface and heating and surface transfer (C°-m/W).
cable (or conduit or pipe coating) surface are isotherms, ρ = thermal resistivity of earth, C°-m/W.
and that the heat sources below ground do not interrupt
the flow of heat of adjacent heat sources, often referred
β = heat transfer coefficient at surface, C°-
to as the Kennelly Hypothesis, which is given in the m2/W.
paper, “The Temperature Rise of Buried Cables and h1 = computational parameter, mm.
Pipes” (Neher 1949). In addition to the adaptations of Dearth = diameter over conduit, pipe or cable jack-
the Kennelly formulas, the Neher paper discusses effects ets, mm.
of superposition of heat flows, the effect of solar heating L = depth from surface to center of cable, con-
and heat from the center of the earth, and the variation duit or pipe, mm.
of temperature with depth of burial. These topics are SP = axial spacing between cable, conduit or
discussed in other sections of this chapter. pipe centers, mm.
Equation 11-58 represents the normal case for a buried Certain modifications to the basic approach for normal
cable in which the surface of the earth is taken to be an burial depth ratios described above are used to consider
isotherm, this being valid when the ratio (2L/Dearth) ≥ mutual heating from multiple cable phases (or pipes)
10; for example, a 6 in. (150 mm) diameter cable at a installed in the same trench, the temperature averaging
depth of 39 in. (1000 mm) has a ratio of 13:1; shallower effects of the daily loss factor, and presences of special
burial depths are modeled using the same method, but backfill placed in the cable trench. These topics are dis-
the validity of the results may be questioned. Neher cussed later. The generalized form of the earth thermal
developed an addition to the classical Kennelly equation resistance equation that will be used in this chapter is as
method to allow for cases of shallow burial or large- follows:
ρ ⎛ ⎛ D ⎞ ⎛ ⎛ 2 ⋅ L + 4 ⋅ L2 − D 2 ⎞ ⎞⎞
R earth = ⎜
⋅ n ⋅ ln ⎜ x
⎟ + LF ⋅ ⎜ ln ⎜⎜
x
⎟ + ln ( F ) ⎟ ⎟ [C°-m/W]
2π ⎜ ⎝ Dearth ⎠ ⎜ Dx ⎟ ⎟⎟
⎝ ⎝ ⎝ ⎠ ⎠⎠
11-60
Where:
L = centerline burial depth of the center of the cable, pipe, or conduit, mm.
Dearth = diameter of the earth interface (cable OD, pipe OD, or conduit OD), mm.
Dx = diameter beyond which the effect of average daily losses applies (see below), mm.
ρ = the thermal resistivity of the soil in contact with the cable, C°-m/W.
11-31
n = number of cables within diameter, Dearth. Within diameter, Dx, temperatures fluctuate with daily
LF = the 24-hour loss factor for the load on the load shape; beyond Dx, only the average daily losses are
cables, per-unit. seen, and temperature fluctuations are small. In addi-
F = geometric factor representing mutual heat- tion to the thermal diffusivity, diameter, Dx, is affected
ing between multiple cables (this will be by the duration of the load cycle. Most rating calcula-
discussed further below). tions, including those discussed in this chapter, consider
a daily (24-hour) load cycle.
Typically the value, n, is 1 for extruded and self-con-
tained transmission cables and 3 for pipe-type cables. An empirical equation relates the thermal diffusivity
and load cycle duration (24 hours) to diameter D x
Earth Thermal Resistivity
according to the following equation:
Earth thermal resistivity can vary substantially over the
length of the circuit and with time, due to changes in
D x = 1.02 α soil ⋅ 24 (mm) 11-61
moisture content. Values range from below 0.40 to over
4.0C°-m/W. Earth thermal resistivity is discussed in
detail in Chapter 12. Calculation procedures to account Where αsoil is the thermal diffusivity in mm 2 /hr. This
for controlled backfill or a concrete envelope around value may also be expressed in m2/second (by dividing
ducts are given later in this section, but note that resis- values in mm2/hr by 3.6 x 109), which is useful for emer-
tivity of material—native soil or controlled backfill— gency rating calculations by IEC 60853.
directly in contact with the cables, conduits, or pipes
should be used in Equation 11-60. Table 11-13 summa- Unlike soil thermal resistivity, the thermal diffusivity is
rizes some typical values of soils and backfills that are not a parameter that is measured during either in situ or
used for the earth portion of the thermal circuit. laboratory soil testing. Another empirical relationship is
often used to relate thermal resistivity to thermal diffu-
Diameter Dx sivity:
Buried cable systems have very large thermal capaci-
tance; this thermal capacitance means that temperature 6.71 × 10 4
α soil = (mm2/hr)
(ρ native ⋅100)
changes from changing loads diminish with increased 11-62
0.8
distance away from the cable. Beyond a short distance
from the cables, temperature rise is a function of average Where
daily heat input from the cable, rather than peak input. is the native soil thermal resistivity in
The cable conductor, however, reaches maximum values
ρ native C°-m/W. For a thermal diffusivity of 1775
more quickly, and its temperature is a function of peak mm 2 /hr (2.75 in. 2 /hr), diameter, D x , is
rather than average load. There is a transition from approximately 210 mm (8.3 in.). Many
peak to average effect at some intermediate point. Dx hand calculations use this value, although
defines the diameter of the transition from peak to aver- most computer programs calculate the
age, and is a function of how easily heat moves into the diameter D x based on the soil thermal
soil, called the thermal diffusivity, αsoil. Diffusivity is the resistivity.
ratio of the thermal conductivity (inverse of thermal
resistivity, ρ) to the thermal capacity, C, of the soil. Note that for some larger conduits and many pipe-type
cables, diameter Dx may be smaller than the actual outer
Table 11-13 Typical Soil Thermal Resistivity Values
diameter of the cable pipe. Although this may seem
Thermal Resistivity counter intuitive, the basic form of the Kennelly
5% Moisture 0% Moisture Hypothesis assumes that the cables are “line” (zero
Soil Type (C°-m/W) (C°-m/W)
diameter) heat sources, so diameter Dx remains consis-
Fluidized Thermal Backfill 0.4 0.75
tent with the approach.
Concrete 0.6 0.8
Stone Screenings 0.4 1.0 Loss Factor
Thermal Sand 0.5 1.0 From the cable conductor out to diameter, Dx, the ther-
Uniform Sand 0.7 2.0 mal resistances of the various cable and soil layers expe-
Clay 1.0 2.5 rience the peak heat output from the cable. At distances
Lake Bottom 1.0 (50% moisture) >3.0 of Dx and further away from the cable, the soil thermal
Seabed (typical)
0.7 (100%
>3.0
resistance experiences the average, rather than the peak,
moisture) daily heat output. To account for this effect, the daily
Highly Organic Soil >3.0 >6.0 loss factor is used, which relates the peak losses to aver-
age losses. The loss factor (and related load factor) was
11-32
discussed at the beginning of this chapter. Beyond diam- depth and mutual heating on the conductor tempera-
eter, Dx, the earth thermal resistance is multiplied by tures for three equally-loaded 230-kV XLPE cable cir-
loss factor. cuits in a duct bank.
Mutual-Heating Factor
A basic assumption of ampacity calculations is that the
The presence of other heat sources, resulting in mutual
load shape is sinusoidal; for most circuits, this is a rea-
heating, reduces the ampacity of a cable system; for typ-
sonable assumption. A reference (Morris and Burrell
ical trenched installations, the area of concern is within
1954) gives a good description of the effects of daily
about 4 m (13 ft) of the trench. Cable ratings for deeper
load shapes. Loss factor applies only to current-depen-
cable installations, including directional drilling, should
dent losses. Dielectric losses are present all the time—
consider an even wider range where mutual heating
assuming that the cable is always energized—and there-
could be important. For ampacity, we will focus on
fore have a unity (1.0) loss factor.
mutual heating from nearby equally-loaded cables.
Depth of Burial Unequal loading and other types of heat sources are dis-
Depth of burial, L, is to the centerline of the cable for cussed in the next section.
which rating calculations are being performed. The cen-
terline depth—to the center of a direct buried cable, cen- Nearby equally loaded cables are evaluated by deter-
ter of the conduit for duct installations, or center of the mining an effective increase in thermal resistance for the
pipe for pipe-type—should be used for the depth of cable being analyzed. Using the thermal equivalent of
burial. Ohm’s law, the calculation determines the temperature
rise caused by the equally loaded cables on to the “hot-
Note that because of asymmetry in the cable positions, test” cable. For simplicity, many hand ampacity calcula-
usually one of the cables runs at a higher temperature tions also assume that the temperature is the same for
than the others, and this cable should be used for rating all cables, so that both the loading and resistance, and
calculations. The operating temperature of a cable for a consequentially, the heat output, are the same for all
given load is affected by the depth and relative position cables. This is a conservative assumption.
to other cables (or heat sources) in or near the trench, as
will be discussed in the next section. With the assumption that the earth’s surface is an iso-
therm, a “method of images” is used to model the heat
Sometimes the hottest cable is self-evident by inspection leaving each cable and its mutual heating effects on the
of the trench cross section (e.g., the middle cable when other cables. Two examples, a two-circuit pipe-type
three cable phases are placed in a flat configuration). cable and the three phases of a single-core cable, are
However, other installations may not be as obvious. shown in Figure 11-22.
When calculating ratings by hand, it may be necessary
to evaluate the rating based on more than one phase The mutual heating effect is evaluated knowing the dis-
position; most cable rating programs take this into con- tance from a given cable to the image of an adjacent
sideration. Figure 11-21 illustrates the impact of burial cable divided by the actual distance separating the
Figure 11-22 Illustration of mutual heating by method of

Figure 11-21 Conductor temperatures as a function of images.
duct bank position.
11-33
cables; the product of all these ratios is used to find the These values are shown graphically in Figures 11-14 and
mutual heating factor, F, as follows: 11-15.
⎛ d' ⎞ ⎛ d' ⎞ ⎛ d' ⎞ Once the mutual heating factor is known, the thermal
F = ⎜⎜ 12 ⎟⎟ ⋅ ⎜⎜ 13 ⎟⎟...⎜⎜ 1N ⎟⎟ resistance representing mutual heating can be evaluated
⎝ d12 ⎠ ⎝ d13 ⎠ ⎝ d1N ⎠ 11-63 using the following equation:
Where,
N = number of cables, conduits or pipes in the ρ
group. R mutual = ⋅ n ⋅ LF ⋅ ln (F) (C°-m/W) 11-68
d12’ = distance from cable 1 to image of cable 2,

2π
mm. Where n is the number of energized cables within the ref-
d12 = distance from cable 1 to cable 2, mm. erence location. For direct-buried cables or conduit
d13’ = distance from cable 1 to image of cable 3, installations of transmission cables, n equals one. For a
mm. pipe-type cable, n equals three. If the cables are installed
d13 = distance from cable 1 to cable 3, mm. in special backfill, the thermal resistivity used to calcu-
d1N’ = distance from cable 1 to image of cable n, late the mutual heating thermal resistance should be
mm. that of the backfill.
d1N = distance from cable 1 to cable n, mm.
Interference Temperature Rise
There are “n-1” terms in the mutual heating factor cal- The mutual-heating terms developed above assume that
culation—e.g., there are two terms for a single three- all cables are equally loaded. There are many conditions
phase extruded-dielectric circuit; there are no terms (no where the external heat source has a different heat input,
mutual heating) for a single pipe-type circuit (one pipe). such as steam mains or distribution duct banks, result-
ing in an interference temperature rise, ΔTint. The tem-
The x-y coordinate pairs are usually available for each perature rise due to each of the external heat sources is
cable, duct, or pipe in the trench. When this is the case, subtracted from the total allowable conductor tempera-
the generalized equations to calculate the distances are: ture rise in the same way that temperature rise due to
dielectric loss is subtracted. Interference temperature
rise for each source is calculated from Equation 11-69:
d1n ' = (x n − x1 )2 + (y n + y1 )2 11-64 ΔTint = Wint ⋅ R int (C°) 11-69
Wint, the heat input from the extraneous heat source is

d1n = (x n − x1 )
2
+ (y n − y1 )2
11-65
the average daily heat input. This could be heat from a
steam main, a distribution duct bank, or another
unequally loaded transmission circuit. The value of Wint
The F factor has also been solved for certain common for nearby cables is given in Equation 11-70.
geometries. For a single-circuit, one-core cable in flat
configuration, the following equation can be used:
( )
Wint = (n int ) I int 2 ⋅ R ac − int ⋅ (LFint ) + Wd int (W/m)
4L2 11-70
F= +1 Where:
Sp 2 Iint = conductor current in other interfering heat
11-66 sources, amperes.
Rac-int = ac resistance of other heat source, ohm/m.
For two pipe-type cables buried at the same depth, the LFint = loss factor of other heat source, dimen-
following equation applies: sionless.
Wdint = dielectric loss in other heat source, W/m.
4L2 nint = number of cables at the location of the
F= +1 interfering heat source, dimensionless.
Sc 2
11-67
Where: For a steam pipe or another heat source where the watt
L = burial depth of the cables, mm. output is not known, but a temperature, perhaps from a
Sp = phase spacing, mm. thermocouple, can be determined, the heat output from
Sc = circuit spacing, mm.
11-34
that source can be calculated using the thermal equiva- the effective thermal resistivity to the ground surface, as
lent of Ohm’s law in Equation 11-71: most heat flows in the region above the cables. Earth
thermal resistance should be calculated for the thermal
(Tsource − Tambient ) (W/m) resistivity of the concrete or controlled backfill, which is
Wint = 11-71
R source,earth typically a rectangular envelope around the cables, as
shown in Figure 11-23. Then a correction to the earth
Where:
thermal resistance is made to adjust for the different ther-
Tsource = surface temperature of external heat
mal resistivity outside the trench. If the soil outside the
source, °C.
trench has a higher thermal resistivity, the correction
Tambient= ambient soil temperature “remote” from
increases the total thermal resistance. If the native soil
any heat sources, °C.
has a lower earth thermal resistivity, as might be the case
R source, earth = thermal resistance to ambient earth when Bentonite-based drilling mud surrounds the con-
for external heat source, C°-m/W.
duits in a horizontal directionally drilled installation,
then the correction will decrease the total earth thermal
The effective thermal resistance to the external heat
resistance.
source is calculated from Equation 11-72 for the appro-
priate distance between the cable being rated and the
The method of calculating the correction to the native
external heat source (similar to that of mutual heating).
soil thermal resistance when the backfill envelope is
ρ rectangular with centerline burial depth, Lb, is to calcu-
R int = ⋅ n ⋅ ln(Fint ) (C°-m/W) 11-72 late the diameter, Db, of an equivalent circular envelope
2π with an isothermal circle. These parameters are illus-
Where:
trated in Figure 11-23.
Fint = mutual heating factor between the external
heat source and cable, dimensionless
R int = thermal resistance between an external heat The circumscribing circle diameter, Db, is found using
the following equation:
source and a cable, C°-m/W
⎡ ⎤
⎛ 1 ⎞⎛ x ⎞⎛ 4 x ⎞ ⎛ y2 ⎞
The cable engineer should check the effect of the exter- D b = exp ⎢⎜ ⎟⎜⎜ ⎟⎟⎜⎜ − ⎟⎟ ln⎜1 + ⎟ + ln (x )⎥ (mm)
nal heat source on all of the cables, as discussed in the ⎢⎝ 2 ⎠⎝ y ⎠⎝ π y ⎠ ⎜ x2
⎟ ⎥
⎣ ⎝ ⎠ ⎦
mutual heating section above. If the external heat source 11-73
is a cable circuit, it will also experience a temperature Where:
rise from the transmission circuit(s) for which the rating x = short dimension of the backfill envelope,
is being calculated and may require a reduction in rat- mm.
ing. If the external heat source is in the same special y = long dimension of the backfill envelope,
backfill envelope as the cable being rated, a correction mm.
factor should be applied as discussed in the next section.
The same procedure may be used to determine the ther-
Effect of Controlled Backfill and Concrete Envelope Around
Duct Bank mal resistivity correction when the backfill envelope is
Up to this point, the thermal resistances representing circular, as would be the case for a directionally drilled
heat flow to the ground surface and mutual heating tem- installation; Db is the diameter of the bore hole or casing.
perature rise have used a value for the trench backfill
material in direct contact with the cables, conduits, or
pipes. However, in actual installations, the thermal resis-
tivity of the native soil outside of the trench is typically
different, usually greater, and this must be factored into
the ampacity calculations.
For normal trenched cable installations, utilities fre-

quently place backfill with a controlled thermal resistiv-
ity in transmission-cable trenches—e.g., concrete
encasement around the conduits, thermal sand, or Fluid-
ized Thermal Backfill™ (FTB) around the cables, con- Figure 11-23 Trench backfill width and height, envelope
duits or pipes. These materials usually have a lower diameter Db, and backfill depth, Lb.
thermal resistivity than the native soil, thereby lowering
11-35
Then, knowing the centerline depth, Lb, of the backfill The procedure described above is valid as long as the
envelope, a geometric factor for the backfill can be cal- ratio of the long dimension of special backfill to the
culated: short dimension (y/x) of the special backfill is 3 or less,
which is generally the case. However, there are situations
⎛ 2 ⋅ L + 4 ⋅ L 2 − D2 ⎞ where the backfill dimensions are particularly narrow or
⎜ b b b ⎟
G b = ln⎜ ⎟ (dimensionless) short—for example, when a duct bank is rolled flat to be
⎜ Db ⎟ installed in the roadbed across a bridge—in which case,
⎝ ⎠ the long-short ratio falls outside of the 3:1 limitation.
11-74
To address these configurations, a paper (El-Kady and
The thermal resistance correction for native soil with Horrocks 1985) was developed using finite element
different thermal resistivity, ρ native , than the backfill analyses that tabulated extended values for the Gb factor
thermal resistivity, ρ backfill , can be calculated using the in considering these unusual duct bank configurations.
above geometric correction factor as shown in Equation Note, however, that the table of values was developed
11-75. using a single heat source in the middle of the backfill
envelope and does not consider the relative position of
the heat source within the backfill. Instead of using
Equation 11-75 above, the values may be picked from
Table 11-14. Note that, when selecting Gb values from
ρnative − ρbackfill
R correction = ⋅ n ⋅ N ⋅ LF ⋅ G b [C° − m / Watt]
2π 11-75
Where:
n = number of cables per conduit, location, or pipe.
N = number of occupied conduits or locations within the backfill envelope.
ρ native = native soil thermal resistivity, C°-m/W.
ρ backfill = backfill envelope thermal resistivity, C°-m/W.
Table 11-14 Extended Geometric Values (Gb) for Backfills
Lb/h (backfill depth / height)
0.6 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 20.0
0.05 0.08 0.32 0.39 0.59 0.77 0.93 1.08 1.21 1.34 1.45 1.56 1.67 1.77 1.87 1.96 2.05 2.14 2.23 2.31 2.47
0.1 0.10 0.36 0.65 0.94 1.18 1.39 1.57 1.72 1.87 2.00 2.13 2.25 2.37 2.47 2.57 2.66 2.76 2.85 2.94 3.12
0.2 0.14 0.45 1.00 1.37 1.68 1.93 2.12 2.28 2.39 2.53 2.66 2.79 2.90 3.01 3.12 3.21 3.31 3.41 3.51 3.69
0.3 0.18 0.56 1.26 1.68 2.02 2.29 2.48 2.63 2.75 2.89 3.02 3.15 3.27 3.38 3.49 3.59 3.69 3.80 3.89 4.08
0.4 0.22 0.68 1.43 1.86 2.19 2.45 2.66 2.80 2.95 3.09 3.22 3.35 3.47 3.58 3.69 3.79 3.88 3.99 4.08 4.27
0.5 0.25 0.81 1.51 1.92 2.21 2.46 2.67 2.83 2.99 3.13 3.25 3.38 3.50 3.61 3.71 3.81 3.91 4.01 4.11 4.29
0.6 0.29 0.90 1.62 2.04 2.34 2.60 2.81 2.98 3.15 3.29 3.42 3.55 3.68 3.80 3.91 4.02 4.13 4.24 4.35 4.56
0.7 0.32 0.97 1.71 2.14 2.44 2.70 2.92 3.10 3.27 3.43 3.57 3.72 3.86 3.99 4.12 4.24 4.37 4.49 4.62 4.86
0.8 0.35 1.04 1.81 2.26 2.58 2.87 3.12 3.34 3.55 3.74 3.92 4.11 4.29 4.47 4.64 4.81 5.00 5.19 5.39 5.79
0.9 0.39 1.11 1.90 2.39 2.74 3.07 3.37 3.64 3.91 4.16 4.40 4.65 4.90 5.15 5.39 5.63 5.89 6.14 6.41 6.94
h/w (height / width)
1.0 0.42 1.17 2.00 2.52 2.93 3.30 3.65 3.98 4.31 4.63 4.95 5.27 5.59 5.90 6.21 6.52 6.82 7.15 7.47 8.10
1.2 0.47 1.24 2.06 2.58 2.98 3.35 3.70 4.03 4.36 4.68 5.00 5.32 5.64 5.95 6.26 6.57 6.87 7.20 7.52 8.15
1.4 0.52 1.31 2.12 2.64 3.03 3.40 3.75 4.08 4.41 4.73 5.05 5.37 5.69 6.00 6.31 6.62 6.92 7.25 7.57 8.20
1.6 0.56 1.37 2.18 2.70 3.10 3.47 3.82 4.15 4.48 4.81 5.14 5.46 5.78 6.09 6.40 6.71 7.03 7.34 7.66 8.29
1.8 0.60 1.43 2.24 2.76 3.17 3.55 3.91 4.24 4.58 4.92 5.26 5.59 5.92 6.24 6.56 6.87 7.19 7.52 7.85 8.50
2.0 0.64 1.48 2.31 2.83 3.25 3.64 4.01 4.36 4.72 5.07 5.43 5.78 6.12 6.45 6.78 7.11 7.45 7.79 8.13 8.82
2.2 0.67 1.52 2.39 2.90 3.35 3.77 4.17 4.55 4.94 5.32 5.71 6.09 6.47 6.84 7.21 7.58 7.96 8.33 8.71 9.46
2.4 0.70 1.56 2.46 2.98 3.44 3.89 4.32 4.74 5.16 5.58 6.00 6.42 6.83 7.24 7.65 8.05 8.46 8.87 9.28 10.11
2.6 0.73 1.59 2.53 3.05 3.54 4.02 4.49 4.94 5.39 5.84 6.29 6.74 7.19 7.63 8.08 8.52 8.97 9.41 9.86 10.75
2.8 0.76 1.62 2.60 3.13 3.65 4.15 4.65 5.13 5.62 6.10 6.58 7.06 7.55 8.03 8.51 8.99 9.47 9.96 10.40 11.41
3.0 0.79 1.64 2.66 3.20 3.74 4.28 4.81 5.33 5.85 6.37 6.88 7.40 7.92 8.43 8.95 9.47 9.99 10.51 11.00 12.06
3.2 0.82 1.67 2.72 3.27 3.84 4.41 4.97 5.53 6.08 6.63 7.18 7.73 8.29 8.84 9.39 9.95 10.50 11.06 11.60 12.72
3.4 0.84 1.70 2.77 3.35 3.95 4.55 5.14 5.73 6.32 6.90 7.48 8.07 8.66 9.25 9.84 10.43 11.02 11.61 12.20 13.38
3.6 0.86 1.72 2.81 3.42 4.05 4.68 5.31 5.94 6.56 7.20 7.79 8.41 9.04 9.66 10.29 10.92 11.54 12.17 12.70 14.04
3.8 0.80 1.75 2.85 3.49 4.16 4.82 5.48 6.14 6.80 7.45 8.10 8.76 9.42 10.08 10.74 11.41 12.07 12.73 13.30 14.71
4.0 0.90 1.77 2.89 3.56 4.26 4.96 5.66 6.35 7.04 7.73 8.42 9.11 9.81 10.50 11.20 11.90 12.60 13.29 13.90 15.38
4.5 0.94 1.83 2.96 3.74 4.53 5.31 6.10 6.88 7.66 8.44 9.22 10.00 10.79 11.57 12.35 13.14 13.93 14.71 15.50 17.08
5.0 0.97 1.88 3.00 3.91 4.79 5.67 6.55 7.42 8.29 9.17 10.04 10.90 11.79 12.66 13.53 14.40 15.28 16.15 17.00 18.79
11-36
the table for cases where the long-short backfill dimen- perature rise is strictly available for current-dependent
sion is less than 3:1, there may be a slight discrepancy in heat. If the allowable conductor temperature is 90°C,
the resulting ampacity. ambient earth temperature is 25°C, and the temperature
rise due to dielectric and charging current loss is calcu-
Calculating Allowable Current
lated to be 6°C, then the rise available for current-
As derived in Equation 11-76, the allowable current for
dependent losses is 90 – 25 – 6 = 59C°. Separating the
a cable system, I, is calculated as:
two components permits a straightforward calculation
of allowable conductor current. After the dielectric
ΔT
I=
( )
losses and each of the electrical and thermal resistances
R ac ⋅ Σ Q i R Thermali are calculated, as shown earlier in this chapter, their val-
11-76 ues may be arranged, as shown in Figure 11-24, which
Where:
clearly shows the components of the thermal circuit.
ΔT = allowable conductor temperature rise, C°.
Temperature Rise Due to Dielectric Losses
R Thermal i = thermal resistance of the “i-th” layer The ac and thermal resistances complete the equivalent
in the equivalent thermal circuit,
C°-m/W. thermal circuit, so that ampacity can be calculated. The
Rac = ac resistance of the conductor, ohm/m. temperature rise due to dielectric losses, ΔTd, is calcu-
lated using the appropriate equation below. Distributed
Once an inventory has been taken of all losses and ther- dielectric losses are generated throughout the insulation,
mal resistances, calculating the ampacity of the cable but are accurately represented by injecting the total loss,
requires solving the electrical analog given in Figure Wd, at the center of the insulation thermal resistance, as
11-3. The allowable temperature drop from the conduc- shown in Figure 11-24.
tor to the remote earth heat sink is divided by the appro-
priate thermal resistances to give the amount of heat Because the voltage on an energized cable system is
that can be generated. That heat generation is divided assumed to be constant, the thermal resistances used for
into two components, non-current-dependent and cur- dielectric temperature rise are calculated in the same
rent-dependent. The temperature drop due to non-cur- manner as for current-dependent losses, except that the
rent-dependent loss (i.e., dielectric loss plus charging loss factor is 100% (1.0 per unit). Therefore, it is neces-
losses, if sufficiently high) is removed from the total sary to consider the thermal resistance values used for
allowable temperature drop. Also, any interference tem- dielectric temperature separately from those used for
perature rise, ΔTint, is also deducted from the allowable current-dependent losses, which are a function of the
conductor temperature rise. The remainder of the tem- daily loss factor.
Figure 11-24 Heat flow through thermal resistances for extruded and self-contained
cables (left) and pipe-type cables (right).
11-37
The temperature rise caused by dielectric heating can be

determined as follows. R cable − to − pipe = thermal resistance between
cable surface and pipe, C°-m/W.
For extruded and self-contained cables in conduit:
R pipe coating = pipe-covering thermal resistance,
⎛ R insulation + R jacket
1
2
⎞ C°-m/W.
⎜ ⎟ Wd = insulation dielectric losses, W/m.
ΔTd = Wd ⎜ + R cable− to −duct + R duct ⎟ [C°]
⎜ ⎟
⎜ + R earth + R mutual + R correction ⎟ Available Temperature Rise for Current-Dependent Losses
⎝ ⎠
11-77
The available temperature rise for current-dependent
losses, ΔTC, is shown in Equation 11-80.
For direct-buried extruded and self-contained cables:
ΔTC = TC − Tambient − ΔTd − ΔTint erference C°
⎛ 12 R insulation + R jacket ⎞
ΔTd = Wd ⎜ ⎟ [C°] 11-80
⎜ +R + R + R ⎟ Where:
⎝ earth mutual correction ⎠ TC = allowable conductor temperature,
11-78 °C.
For pipe-type cables:
Tambient = ambient earth temperature, °C.
ΔTd = conductor temperature rise due to
⎛ 12 R insulation + R cable− to − pipe ⎞ dielectric and charging-current
⎜ ⎟ losses, C°.
ΔTd = Wd ⎜ + R pipe coating + R earth ⎟ [C°] 11-79 ΔTinterference = temperature rise due to extraneous
⎜ ⎟ heat source, C°.
⎜ + R mutual + R correction ⎟
⎝ ⎠ Summation of Electrical and Thermal Resistances
Where:
R insulation = insulation thermal resistance, The sum of the products of electrical resistance and all
C°-m/W. of the thermal resistances to ambient earth,
R jacket = jacket thermal resistance, ( )

R ac ⋅ Σ Q i ⋅ R thermal , is calculated as shown in Equa-
tion 11-81 for single-core (extruded, self-contained)
C°-m/W.
cables, and as shown in Equation 11-82 for pipe-type
R cable − to − duct = thermal resistance between cable cables. Each of the terms is shown in Figure 11-24.
surface (jacket) and conduit
⎛ R insulation + ⎞
(duct), C°-m/W. ⎜ ⎟
R duct ⎜ ⎛ R jacket + ⎞⎟
= duct thermal resistance, C°-m/W. ⎜ ⎜ ⎟⎟
⎜ ⎜R ⎟⎟
R earth = earth thermal resistance (calcu- ⎜ cable − to − duct + ⎟
⎜ ⎟
lated with 100% loss factor), R ac × ⎜ ⎜ R duct + ⎟⎟
C°-m/W. ⎜ QS × ⎜ ⎟⎟
⎜ ⎜ R earth + ⎟⎟
R mutual = thermal resistance to account for
⎜ ⎟⎟
⎜
multiple cables within the trench ⎜ ⎜ R mutual + ⎟⎟
(calculated with 100% loss fac- ⎜ ⎜ ⎟⎟
tor), C°-m/W. ⎝ ⎝ R correction ⎠⎠
R correction = correction factor for controlled ( )
= R ac ⋅ Σ Q i R Thermal i [C° / ampere 2 ]
11-81
backfill or concrete–encased duct
(calculated with 100% loss fac-
tor), C°-m/W.
11-38
be calculated from Equation 11-84. This equation

assumes that charging current has the same load factor
⎛ R insulation + Q S × R cable − to − pipe +⎞ as load current.
⎜ ⎟
⎜ ⎛ R pipe coating + ⎞ ⎟
⎜ ⎜ ⎟ ⎟ I total = I real 2 + I reactive 2 (A) 11-84
R ac × ⎜ ⎜R + ⎟ ⎟
⎜ Q P × ⎜ earth ⎟ ⎟ For this evaluation of charging-current effects, Ireactive is
⎜ ⎜ R mutual + ⎟ ⎟ set equal to Ic max. If there is a known through-flow of
⎜ ⎜ ⎟ ⎟ reactive power, that value would be included in Ireactive as
⎜ ⎜R ⎟ ⎟
⎝ ⎝ correction ⎠ ⎠ well.
( )
11-82
Cable circuits are always limited by thermal constraints
(as compared to overhead lines that may have thermal,
Calculating Allowable Current voltage, or stability limits). The maximum length of ac
The allowable current, I, to give the specified conductor cable circuits between shunt compensation reactors is
temperature is calculated from Equation 11-76, which is limited by the charging current since charging current
repeated here as 11-83 where all terms were defined increases proportionally with length. As the length of
above. the underground line is increased, a point is reached
where the total charging current equals the cable
ΔT ampacity. This occurs at what is called the “critical
I=
( )
(amperes) 11-83
length” of the cables, where no real power may be trans-
R ac ⋅ Σ Q i R Thermal i
ferred without causing the conductor to exceed rated
Effect of Charging Current temperature. The critical length can be calculated from
Equation 11-83 shows the calculation of total current the following equation:
that gives the allowable conductor temperature rise.
I
That current consists of the phasor sum of real current Critical Length = (m) 11-85
and reactive charging current. Equation 11-12 gave the IC
calculation of charging current, Ic, per unit length. Max- Where:
imum charging current, Ic max occurs at the end of the I = normal ampacity, amperes.
line where charging current exits, and is equal to per- IC = charging current, amperes/m.
unit charging current times line length. If shunt com-
pensation or system conditions force the charging cur- The maximum feasible line length must be significantly
rent to flow equally from each end of the cable, less than the critical length in order to transmit reason-
maximum charging current, Ic max, is equal to the per- able amounts of real power. However, the concept of
unit charging current times one-half the line length. critical line length quantifies the absolute maximum
This is illustrated in Figure 11-25. lengths between shunt compensation that can be
achieved for different types of underground cables.
In ampacity calculations, charging current is considered Table 11-15 shows critical line lengths based on typical
for the last cable meter where the maximum heat is gen- insulation thicknesses and parameters.
erated. For voltages less than 230 kV and for daily load
factors greater than 50%, the allowable real current may
Figure 11-25 Charging current magnitude profile.
11-39
Table 11-15 Approximate Critical Lengths for Underground For direct-buried extruded and self-contained cables:
Cable Circuits with Typical Insulation Thicknesses
Laminated ⎛ R insulation + ⎞
Crosslinked Paper- Ethylene- ⎜ ⎟
Poly- Polypropy- Propylene- ΔTIc = Ic max 2 ⋅ R ac ⎜ ⎛ R jacket + R earth ⎞ ⎟ [C°]
⎜ QS ⋅ ⎜ ⎟⎟
Voltage ethylene Kraft Paper lene Rubber
(kV) Critical Lengths, km (miles) ⎜ ⎜ +R + R ⎟⎟
69 262 (163) 101 (63) n.a. 193 (120) ⎝ ⎝ mutual correction ⎠ ⎠
115 209 (130) 77 (48) n.a. 132 (82) 11-89

138 193 (120) 72 (45) n.a. 113 (70)
161 183 (114) 67 (42) 84 (52) n.a. For pipe-type cables:
230 130 (81) 48 (30) 58 (36) n.a.
345-400 85 (53) 32 (20) 47 (29) n.a.
⎛ ⋅R insulation + QS ⋅ R cable− to− pipe + ⎞
⎜ ⎟
500 76 (47) 27 (17) 37 (23) n.a.
ΔTIc = Ic max 2 ⋅ R ac ⎜ ⎛ R pipe coating + R earth + ⎞ ⎟ [C°]
⎜ QP ⋅ ⎜ ⎟ ⎟
The calculations above assume a 3158 kcmil (1600 mm2) ⎜ ⎜R + R ⎟ ⎟
segmental copper conductor. Laminated paper-polypro- ⎝ ⎝ mutual correction ⎠ ⎠
pylene is not generally used for voltages below 161 kV, 11-90
and the relatively high dielectric constant and dissipa-
tion factor for EPR insulation limits the application of This temperature rise is subtracted from the allowable
these cables to 138 kV or below. conductor temperature in the same manner as dielectric
losses shown earlier in Equation 11-80, and the allow-
For voltages 230 kV and greater and for load factors less able real current is calculated as shown in Equation
than 50%, Equation 11-84 will give ampacities that are a 11-83. The worked examples in the appendix at the back
few percent too high for long lines. Charging current has of this chapter illustrate the differences between this
a 100% load factor, present whenever the line is ener- approach and that discussed in the Neher-McGrath
gized, and the temperature rise is, therefore, higher than paper.
if it had the same load factor as the cable. The preferred
approach in this case is to evaluate the temperature rise Solving Equation 11-83 gives the peak allowable current
due to charging-current losses in the same manner as is for the stated cable design, installation conditions, and
done for dielectric losses. operating conditions. If the calculated ampacity is too
low or too high, the user must change one of the design
Peak charging current losses are calculated for extruded conditions, such as conductor size, and repeat the calcu-
and self-contained cables and pipe-type cables using lation. Eventually the desired ampacity is reached, and
Equations 11-86 and 11-87, respectively: no further iterations are required.
WIc max = I c max 2 × R ac ⋅ Q S (W/m) 11-86 Iterative solutions of equations of the type presented in
2 this section are ideally suited for computer program-
WIc max = I c max × R ac ⋅ Q P (W/m) 11-87 ming. Many users have developed programs to solve the
Where: ampacity equations. Most such programs follow Neher-
Ic max = maximum charging current, amperes. McGrath with enhancements from IEC 60287, and
result in a calculated ampacity as developed in this
The temperature rise due to charging current is calcu- chapter. Other programs have adapted the basic proce-
lated as shown in Equations 11-88, 11-89, and 11-90 for dures but solve for conductor temperature for a stated
the last meter of cable in the circuit, remembering to cal- cable loading; rating calculations are done by iterating
culate earth thermal resistances with 100% loss factor. the assumed load until rated temperature is achieved.
Some solution methods utilize finite element modeling
For extruded and self-contained cables in conduit: to perform a true two- or three-dimensional heat trans-
fer solution; details of these numerical techniques are
⎛ R insulation + ⎞ beyond the scope of this book.
⎜ ⎟
⎜ ⎛ R jacket + R cable − to − duct ⎞ ⎟ 11.2.6 Cables in Air and Tunnels
ΔTIc = Ic max 2 ⋅ R ac ⎜ ⎜ ⎟ ⎟ [C°]
⎜ QS ⋅ ⎜ + R duct + R earth ⎟⎟ Approach
⎜ ⎜ ⎟⎟ Transmission cables may be installed above ground for
⎜ ⎜ + R mutual + R correction ⎟ ⎟
⎝ ⎝ ⎠⎠ bridge crossings, tunnel installations, or riser sections.
All electrical resistances are the same as those calculated
11-88
11-40
in Section 11.2.5, and cable thermal resistances are the Table 11-16 Constants for Cables-in-Air Calculations (Black
same out to De, the diameter at the start of the external Surfaces)
portion of the circuit. Installation Z E g Mode
Single cable, conduit,
0.21 3.94 0.60
≥0.3 De away from
The principal differences for in-air cables are the follow- or pipe solid surface
ing: Two horizontal cables,
conduits, or pipes, 0.29 2.35 0.50
≥0.5 De away from
• Solar radiation provides heat input. The traditional touching solid surface
buried-cable calculation procedure does not consider Three cables, con- ≥0.5 De away from
0.96 1.25 0.20
this heat input. duits, or pipes in trefoil solid surface
• Heat transfer by conduction, the predominant mode Three horizontal
cables, conduits, or 0.62 1.95 0.25
≥0.5 De away from
for buried cables, is negligible. pipes touching solid surface
• Heat transfer for cables in air is by free or forced con- Two vertical cables,
≥0.5 De away from
conduits, or pipes 1.42 0.86 0.25
vection and by radiation. touching solid surface
Thermal Resistance, Cable to Air Two vertical cables,

≥0.5 De away from
In general, the ampacity of an otherwise-identical cable spaced solid surface
circuit will be greater when installed in air rather than in Three vertical cables,
the ground. The basic equation for the environment conduits, or pipes 1.61 0.42 0.20
≥1.0 De away from
touching solid surface
(outside the cable) thermal resistance is defined by
Equation 11-91: Three vertical cables,
≥0.5 De away from
spaced solid surface
1000
R air = [C°-m/W] Single cable, conduit, Attached to solid
⎛ Z ⎞ 1.69 0.63 0.25
π ⋅ De ⎜ ⎟
+ E ΔTsurface − to − ambient
4 or pipe surface
⎜ (D )g ⎟ Three cables, Attached to solid
⎝ e ⎠ onduits, or pipes
0.94 0.79 0.20
surface
11-91
Where: Note that the constants above should be applied only when
De = diameter of cable, conduit or pipe to envi- diameter, De, is not greater than 150 mm (5.9 in) for cables
spaced from a surface and 80 mm (3.1 in) for cables in direct
ronment, mm.
contact with a surface.
Z, g, E = constants from Table 11-16 based on the
installation mode.
ΔTsurface-to-ambient The surface-to-ambient air temperature rise is not a
= cable, conduit or pipe surface temperature straightforward value to determine because of the
rise above ambient air, C°. nature of convection and radiation heat loss. The heat
balance equation at the surface of the cable must satisfy
the following:
Wac + Wd = Wconvection + Wradiation − Wsolar

11-92
Where:
⎛ ⎞
⎜ 1 ⎟
Wradiation = 5.69 ⋅10 −8
(T
surface
4
− Tambient air 4
) π ⋅ De (10 ) ⎜ 1 1 ⎟ (W / m)
−3
⎜ + −1 ⎟
⎝e ε ⎠
( Tsurface − Tambient air )
4
−3
Wconvection = 8.523 ⋅ π ⋅10 ⋅ De ⋅ 3 (W / m)
Tsurface + Tambient air
2
ε = emissivity of the cable, conduit, or pipe surface.
2
⎛ D ⎞
e = ⎜1 − e ⎟ , dimensionless, with d equal to theclosest cable spacing.
⎝ επd ⎠
11-41
The radiation and convection heat loss equations are ΔTsurface-to-ambient

from (CIGRE 1985b) and account for the reradiation of = temperature difference between the cable
cables in close proximity in air. surface and air, C°.
Because of the fourth-order terms in the above series of The change in air temperature in the tunnel may also be
equations, the temperature rise from the surface to of interest when trying to size air-cooling equipment.
ambient air must be found iteratively. For a given load The tunnel air temperature change as it passes down the
condition, the temperature difference from the conduc- tunnel provides an indication of the heat being absorbed
tor to the cable, conduit, or pipe surface is determined by the tunnel air. The air mass flow through the tunnel
using the equivalent thermal circuit discussed earlier in is defined by the following equation:
this chapter, excluding the earth, mutual heating, and
backfill correction. This is then compared to an •
assumed value for the surface temperature using the m = ρ⋅A⋅υ 11-95
convection and radiation equations above. If the heat Where:
balance is satisfied, then the correct value of cable sur- •
face temperature has been found, and the air thermal m = air mass flow rate, kg/s.
resistance can be calculated. ρ = air density, kg/m3.
A = free area within the tunnel, m2.
The air thermal resistance could also be calculated using
an equation of this form: The heat removed by the air as it flows through the tun-
nel can then be found from the following equation:
Tsurface − Tambient air
R air = •
Wradiation + Wconvection 11-93 q = m⋅ CP ⋅ ( TOUT − TIN ) W
11-96
Most utilities know the maximum ambient air tempera- Where:
ture on their system. In the absence of specific data, CP = air specific heat, kJ/kg-C°.
ambient air temperature is usually assumed to be 40°C. TOUT = air temperature leaving the tunnel, °C.
A first approximation of the temperature difference TIN = air temperature entering the tunnel, °C.
between the ambient air and cable, conduit, or pipe sur- Solar Heat Input
face would be 60°C. The absorptivity, a, determines the amount of heat
absorbed by the cable surface from the sun, and ε is the
Note that loss factor does not enter in the thermal resis- amount of heat radiated from the cable to its surround-
tance equation for cables in air. This is because time ings. Typical values of the factors absorptivity, a, and
constants for cables in air are only a few hours, as com- emissivity, ε, in relation to unity values for perfect black
pared to many tens of hours for buried cables. body absorbers and radiators are given in Table 11-17.
The calculation procedures in ANSI/IEEE Std. 738- Many utilities have specific data for solar heat input,
1986, “IEEE Standard for Calculation of Bare Over- Qsun, for their location. If such data is unavailable, val-
head Conductor Temperature and Ampacity Under ues of 1000 W/m2 (93 W/ft2) for cables in a horizontal
Steady-State Conditions,” (ANSI/IEEE 1986) provides
useful information on calculating in-air ratings. Table 11-17 Absorptivity and Emissivity of Cable-Covering
Materials
A consideration for cables installed in tunnels is that Absorptivity, a Emissivity, ε
there may be forced-air ventilation longitudinally along Black plastic 0.45-0.8 0.8-0.95
the tunnel. The convective heat loss is a function of the Bitumen/jute serving 0.8 0.95
longitudinal airflow velocity within the tunnel accord- Polychloroprene 0.8 0.9
ing to an equation of the form: PVC 0.6 0.9
ΔTsurface - to - ambient PE 0.4 0.9

WConvection = Yellow plastic 0.1-0.4 0.8-0.95
1 Lead, weathered 0.6 0.63
7.6(υ ⋅ D e )0.65
11-94
Aluminum, weathered * 0.3-0.45
Stainless steel * 0.14-0.38
Where:
υ = air flow velocity within the tunnel, m/s. * No data available. Use emissivity data.
11-42
position and 700 W/m2 (65 W/ft2) for cables in a vertical The equation for cables in conduit is included because
position may be used. The calculation of solar heat single-core cables are occasionally installed in duct for
input, Wsun, is given in Equation 11-97. bridge and tunnel installations.
a ⋅ Qsun ⋅ De Allowable Conductor Temperature Rise

Wsun = W/m The allowable conductor temperature rise, ΔT, is calcu-
1000 11-97 lated as shown in Equation 11-102.
Where:
a = absorptivity, dimensionless. ΔT = TC − Tambient − ΔTd − ΔTsun 11-102
Qsun = solar heat input, W/m2.
De = diameter at start of external portion of Summation of Electrical and Thermal Resistances
thermal circuit, mm. The sum of products of electrical and thermal resis-
tances to ambient air is calculated in a similar manner
The temperature rise due to solar heat input is calcu- as for buried cables, replacing the earth thermal resis-
lated from Equation 11-97 and is subtracted from the tances with the air thermal resistance, as follows:
total allowable conductor temperature rise of Equation
⎛ R insulation + ⎞
11-80, in the same manner as dielectric losses. ⎜ ⎟
⎜ ⎛ R jacket + ⎞⎟
⎛R ⎞ ⎜ ⎜ ⎟⎟
ΔTsun = Wsun ⎜⎜ air ⎟
⎟ ⎜R ⎟⎟
R ac × ⎜ cable − to − duct + ⎟
⎝ n ⎠ 11-98 ⎜ QS × ⎜ ⎟
⎜ ⎜ R duct + ⎟⎟
Where n is the number of cables within diameter D e ⎜ ⎜ ⎟⎟
⎜ ⎜R ⎟⎟
(e.g., 1 for most single-core transmission cables, and 3 ⎝ ⎝ air ⎠⎠
for pipe-type cables).
Solar input is zero for tunnels, bridges, and other shel-

( )
11-103
tered installations.
Calculating Allowable Current ⎛ R insulation + Q S × R cable − to − pipe +⎞
⎜ ⎟
Calculation procedures are the same as those for buried
R ac × ⎜ ⎛ ⎞ ⎟
cables, with the thermal resistances for earth, mutual ⎜ Q × ⎜ R pipe coating + ⎟ ⎟
heating, and backfill correction being replaced by the ⎜ P ⎜R ⎟ ⎟
cable-to-air thermal resistance above. ⎝ ⎝ air ⎠ ⎠
Temperature Rise Due to Dielectric Losses
The temperature rise due to dielectric loss, Td, is calcu-
( )
11-104
lated as shown in Equations 11-99, 11-100, and 11-101. Allowable Current
Allowable current, I, to give the specified conductor
For extruded and self-contained cables in conduit: temperature is calculated as shown in Equation 11-83,
which is repeated below.
⎛ 12 R insulation + R jacket ⎞
ΔTd = Wd ⎜ ⎟ [°C]
⎜ +R ⎟ ΔT
⎝ cable − to −duct + R + R air ⎠ I= (amperes)
( )
duct
11-99
R ac ⋅ Σ Qi R Thermal i
For direct-buried extruded and self-contained cables:

11.2.7 Ampacity for Trenchless Installations
ΔTd = Wd ( 1
2 )
R insulation + R jacket + R air [°C]
11-100
The extensive topic of “trenchless” installations—hori-
zontal directional drilling (HDD), pipe jacking, micro-
tunneling, etc.—is covered in other portions of this book
For pipe-type cables: (see Chapter 12). This section discusses the impact on
cable ratings from installing cables using these methods.
⎛ 12 R insulation + R cable − to − pipe ⎞
ΔTd = Wd ⎜ ⎟ [°C] Implications for Ampacity
⎜ +R + R ⎟ The mechanics of calculating ampacity for trenchless
⎝ pipe coating air ⎠ 11-101 installations are no different than for installations that
11-43
utilize conventional open trenching; the Kennelly equa- sonal variation in temperature is generally much smaller
tions apply just the same. The following sections discuss than that for trenched installations. Below a depth of 6 m
the issues that should be considered for ampacity of (20 ft) or so, the ambient soil temperature approaches the
cables installed by trenchless methods. average annual air temperature for the area.
Earth Thermal Resistance for Trenchless Installations In summer months, this means that the ambient soil
Instead of a burial depth in the top few meters of earth, temperature for the trenchless sections is cooler than for
the burial depth may be tens of meters below the conventional trenched sections. However, during winter
ground, and the earth thermal resistance calculation months, the inverse is true, and the ambient soil temper-
must accurately consider this parameter. ature at the greater depths is warmer. This should be
factored into summer and winter ratings.
The thermal resistance calculation that corrects for
backfill is somewhat different with trenchless installa- Use of a Casing
tions. Generally, trenchless installations have a circular, One of the most important decisions regarding trench-
rather than rectangular, backfill envelope, so that the less cable projects is the selection of a casing. This is an
diameter of the backfill, Db, is known directly based on important issue for both civil and electrical engineering
the borehole size or the size of the casing that is used. of the project. One must decide if a casing is necessary,
and if so, decide what type of casing to use. The installa-
In addition to the circular backfill envelope, the backfill tion design details of deciding on use of a casing and
material more often has a greater thermal resistivity what type are covered in Chapter 12. This section sum-
than the native soil in trenchless installations. Many marizes the implications for ampacity.
times, the trenchless installation methods are in highly
compacted soil below fill materials and below the water If a steel casing is selected (12.5-25.0 mm, 0.5-1.0 in. in
table; sometimes the boring may be through rock. As a thickness), the heat transfer through the casing is not a
result, the native soil thermal resistivity outside of the factor because of the high thermal conductivity of the
bore is relatively good (low)—often below 0.9C°-m/W. If steel. Also, there is an opportunity to put a low thermal
a bundle of conduits or cable pipe(s) is installed without resistivity grout between the inner cable conduits and
a casing, Bentonite-based drilling mud is left around the the casing to improve ampacity. Frequently, the limita-
cables after the bore is completed; dry Bentonite can tion on ampacity is the section closest to the surface but
have a very high thermal resistivity—in excess of below where it is practical to excavate, because these
2.5C°-m/W—and, when combined with water in a drill- areas may be above the water table or consist of fill or
ing mud, is generally no better than 1.5C°-m/W (pure organic content, making the thermal resistivity worse.
water has a ther mal resistivity of approximately With extruded or self-contained cables, a steel casing
1.65C°-m/W). Cables installed within a casing may have results in additional losses from hysteresis and eddy-cur-
the interstitial space between the conduits or pipes and rent losses in the steel pipe. This additional heating may
the casing filled with pure water or a fluidized grout. unacceptably reduce the ampacity. For pipe-type cables,
However, because of pump pressure concerns when a steel outer casing does not present an ampacity prob-
grouting, the grout material usually has a relatively high lem because the magnetic field external to the cable
water content, so the thermal resistivity is usually no pipes is so low.
better than 0.9C°-m/W. In either situation—casing or
no casing—the thermal resistance correction for native With a plastic casing, there is no magnetic heating that
soil materials may be negative, since the native soil ther- can reduce the cable ratings. However, the relatively high
mal resistivity could be less than that of the drilling thermal resistivity of the plastic casing—typically
mud, water, or grout. 3.5-4.5C°-m/W—combined with the thick casing wall (50-
75 mm, 2-3 in.) means that a very high thermal resistance
The impact of deep burial depths on the earth thermal is located in a region of relatively high heat flux near the
resistance tends to diminish for typical trenchless cables. Often this higher thermal resistance added by the
depths. Although the burial depth may be high, the plastic casing offsets the magnetic heating effect of a steel
ambient temperature is often better, because surface casing, so the ampacity is only slightly better with the
solar effects are not present, and the thermal resistivity plastic casing over using steel. From an ampacity stand-
is generally better. point, there is no particular reason to select a plastic cas-
ing with pipe-type cables, although there may be
Ambient Temperature
installation issues that need to be considered.
As with conventional trenched installations, the allowable
conductor temperature rise is constrained by ambient Number of Bores and Mutual Heating
temperature. The most significant difference in the ambi- As was discussed in the trenched section, mutual heat-
ent temperature for trenchless installations is that the sea- ing effects from parallel cable circuits affect the rating of
11-44
cables. The zone of influence for mutual heating Use of water to grout the interstitial space between the
expands significantly for directionally drilled installa- inner ducts or pipes and the casing is better than leaving
tions by virtue of the deeper installation depth. Typi- the space filled with air. Air has a thermal resistivity of
cally, heat sources with horizontal spacing that is within approximately 45C°-m/W, while water is 1.65C°-m/W.
four times the burial depth should be investigated as Using water also affords the possibility that the inner
being a possible source of mutual heating. duct or pipe bundle could be pulled out at some future
time and replaced, though the removal/replacement is
This has an impact when installing multiple circuits by rarely considered for practical reasons.
directional drilling. Although a larger, more risky bore
might be attempted, installing multiple sets of cables in 11.2.8 Ampacity Implications for Crossing
a single boring has significant cost savings. However, External Heat Sources
mutual heating effects will be high. The alternative is to Invariably, underground transmission cables are faced
install two smaller bores, thereby reducing mutual heat- with the possibility of crossing other utilities, including
ing but increasing civil costs. A factor to consider is the underground transmission or distribution circuits and
easement and right-of-way constraints that may exist for steam lines. The interference temperature rise concept
two bores. discussed earlier in this chapter is a conservative
Shield Losses approach for considering these crossings. It assumes
The installation of extruded or self-contained cables in a that the external heat source is parallel to the transmis-
shared casing reduces the induced voltage, the circulat- sion cable, with no longitudinal heat flow.
ing shield currents, and the heat losses in a solidly
grounded system. Separate bores are usually widely In some circumstances, this conservative approach
spaced, because of the drilling tolerance, and if these are shows an unacceptable derating effect. The IEC has a
each used to house a single cable phase, a significant section of 60287 (Part 3) that covers the topic of cross-
increase in circulating current losses results. Although ing cables. The basic approach to these calculations is
the use of a specially bonded cable scheme in this situa- detailed below and outlined as follows:
tion eliminates circulating losses, the induced voltages 1. Select a center “reference” (usually the center) in the
per meter length remain high, and must be within the group of cables causing the derating. The horizontal
design limits for sheath voltage limiters at the adjacent distance, Sh (e.g., phase, conduit, or pipe spacing) is
joints, and section-length and width-percentage toler- used to determine mutual heating for other cables in
ances in a cross-bonded scheme. the derating group.
Grout Considerations 2. Separate contributions of interference temperature
Grouting is an option for cable installations that utilize a rise will be calculated for each cable. Evaluate the
casing. The grout generally consists of a mixture of sand, interference temperature rise of the crossing cable(s)
water, and a fluidizer such as Bentonite or fly ash. For on the circuit of interest using the equations
ampacity purposes, the cable engineer would like to have described below.
as much sand in the grout mixture as possible to mini- 3. Calculate the normal rating of each installation using
mize the thermal resistivity. The trenchless installation the procedures of this chapter, neglecting the impact
contractor is mainly concerned about pumping pressures, of the crossing. The load on the cables causing an
which can be reduced by lowering the sand content and interference temperature rise is assumed to be fixed.
increasing the water and fluidizer content. Also, the
pumping pressures from grouting, possibly combined Equations 11-105 and 11-106 define the interference
with the heat of hydration from cement in the grout, may temperature rise for oblique cable (or other heat source)
cause plastic conduits inside the casing to partially col- crossings for, respectively, the primary interference
lapse, affecting the ability to install the cables. source at depth L and other interference sources that
have a horizontal displacement from the primary cable
of Sh. at depth L. The cables cross at an angle β.
Wh ⋅ρsoil ⋅ e γ⋅Δz − 1 N − v⋅γ⋅Δz ⎛ ( L + L h )2 + ( v ⋅ Δz ⋅ cos β )2 ⎞

ΔTinf =
4π
∑ e ⋅ ln ⎜
⎜ ( L − L )2 + ( v ⋅ Δz ⋅ cos β )2 ⎟
⎟
v =1
⎝ h ⎠ 11-105
11-45
⎛ ⎛⎛ Sh ⎞ ⎞ ⎞
2
⎜ ( L + L h ) + ⎜ ⎜ v ⋅ Δz +
2
cos β ⎟ ⎟
Wh ⋅ρsoil ⋅ e γ⋅Δz − 1 N − v⋅γ⋅Δz ⎜ ⎝ ⎝ cos β ⎟⎠ ⎠ ⎟
ΔTinf =
4π
∑ e ⋅ ln ⎜ 2 ⎟
v =1
⎜ ( L − L )2 + ⎛ ⎛ v ⋅ Δz + Sh ⎞ cos β ⎞ ⎟
⎜⎜ ⎟ ⎟
⎜ cos β ⎟⎠
h
⎝ ⎝⎝ ⎠ ⎠
11-106
Where:
β = crossing angle of heat source, ° (degrees)
(0° is perpendicular, 90° is parallel).
Wh = heat output from cable, conduit, or pipe, W/m.
γ = axial thermal resistances of conductor, C°-m/W.
Δz = incremental distance over which mutual heating effect is assessed; “N” such increments are evaluated
from 1 to “v”, m.
L = burial depth of primary interference cable, m.
Lh = burial depth of crossing cable, m.
Sh = horizontal displacement of other cables grouped with the crossing cables, m.
v = incremental counter for evaluating the temperatures.
Δz should typically be set to 0.1 m (about 4 in.) unless the power systems, there is also a cable damage limit or
cables and heat source(s) being evaluated are small in fault-current capability that must be considered. This is
diameter and very close together, in which case a smaller principally of interest for the metallic shield (screen) and
step should be used. The range over which calculations sheath of extruded and self-contained cables; there is
should be evaluated is normally 6-10 m (20-35 ft), so the seldom a concern about fault-current capability for a
value of N would be on the order of 60 (6 m/0.1 m). The pipe-type cable, because the steel cable pipe has a tre-
above procedure is ideally set up for use with a computer mendous capacity to carry fault current, and the shields
program, but approximate calculations may be deter- of pipe-type cables are continuously grounded.
mined by using relatively large step sizes (0.5 m), so that
the summation term can be evaluated by hand. A com- T h e c ro s s - s e c t i o n a l a r e a o f t h e c abl e m e t a l l i c
puter spreadsheet might also be used. shield/sheath must be selected to carry fault current until
the fault is cleared by a circuit breaker, without exceeding
The derating effect can then be calculated as follows: the short-circuit temperature rating of the insulation.
Table 11-18 lists the short-circuit temperature rating for
ΔTinf transmission cable insulation and jacket materials.
Derating = 1 − ×100 [%]
TC − Tambient − ΔTd
Despite the use of polyethylene and PVC as cable jacket
11-107 materials, most manufacturers permit a maximum fault-
current temperature of 200-250°C because of conserva-
A similar set of equations could be used to evaluate the tive aspects of these calculations.
impact of the transmission cables on the distribution
duct bank (or any set of crossing cables), with the calcu-
lations iterated until the assumed rating used to calcu-
late Wh matches the calculated rating.
Figure 11-26 illustrates a possible scenario where an

extruded transmission cable circuit in a duct bank
crosses distribution cables, and the figure identifies the
parameters used in the calculation equations. The distri-
bution circuit is assumed to cause a derating effect on
the transmission circuit.
11.2.9 Fault Current Calculations

In addition to the normal or emergency ampacity values Figure 11-26 Derating of crossing cables.
that are used in consideration of load flow within the
11-46
Equation 11-108 is used to calculate the shield/sheath Table 11-19 Constants for Calculating Required Cross-
cross-sectional area, Area, that is required to prevent Sectional Area for Short-Circuit Currents
overheating. The formula assumes adiabatic conditions
during the fault; where no heat is conducted away from Material k τ (°C)
the conductor during the time of the short circuit and Copper 50237 -234.5
that all of the energy produced by I2 R losses must be Aluminum 21144 -228.1
stored by the thermal capacitance of the conductor Lead 1641 -236.0
material. The same approach may be used to calculate
the fault-current capability of a cable conductor. tion where the primary breaker fails to clear the short-
circuit current. A typical “stuck-breaker” clearing time
k ⎡ T2 − τ ⎤ is 15 cycles (at 60 Hz), or 0.25 s. Some utilities use very
I sc = Area ln ⎢ ⎥A conservative values of 30 cycles (at 60 Hz) or 0.5 s.
t ⎣ T1 − τ ⎦
11-108 Multimetal Shield and Sheaths
Where:
The approach described above is valid for a single metal-
Isc = maximum short-circuit current, A.
lic shield or sheath layer or situations where the metal
k = constant for conductor or shield/sheath
used for the shield and sheath is the same (e.g., a copper
material, dimensionless.
foil laminate used with copper shield wires). However,
Area = shield cross-sectional area, mm2.
some cable constructions utilize two different metals
t = time of short circuit, seconds.
such as copper shield wires in parallel with corrugated
τ = inferred temperature of zero resistance, °C.
stainless steel sheath or aluminum foil laminate.
T1 = pre-fault temperature of metal, °C.
T2 = allowable short-circuit temperature of the
For foil laminate sheaths, CIGRE Technical Brochure
insulation material, °C.
272 (CIGRE 2005) recommends that the shield wire con-
ductor should alone be counted to carry short-circuit
Values for τ and k for Equation 11-108 are listed in
current and that the presence of a parallel thin metallic
Table 11-19. For transmission cables, the time, t, is nor-
foil water barrier be ignored. In the case of a parallel
mally assumed to be the backup-breaker fault-clearing
thick metallic sheath (for example, of a welded or
time. That is, the worst case is assumed to be the situa-
extruded construction), then both the shield conductor
and sheath should be counted to share the current. The
Table 11-18 Short-Circuit Design Temperatures for Jacket apportion of current to the sheath layers should be based
and Insulation Materials on the magnitudes of their parallel resistances, not on
Short-Circuit Design their impedances because this gives a safer design.
Material Temperature (°C)
LDPE 150 When two different metals are used for the shield and
XLPE 250 sheath, small time steps must be evaluated as the resis-
PVC 150 tance of the materials changes during the fault, taking
EPR 250 the resistance of the two metals in parallel during each
Impregnated paper 150-160 time step and adjusting the current division between the
Laminated paper-polypropylene 150-160 materials (see Equation 11-109).
⎛ ⎛ T(t − 1) − τ1 ⎞ ⎛ T(t − 1) − τ2 ⎞ ⎞
⎜ R1 ⎜ ⎟ R2 ⎜ ⎟ ⎟
2 ⎛ Δt ⎞ ⎜ ⎝ −τ1 ⎠ ⎝ −τ2 ⎠ ⎟
T(t) = T(t − 1) + Isc ⎜ ⎟
⎝ H ⎠ ⎜ ⎛ T(t − 1) − τ1 ⎞ ⎛ T(t − 1) − τ2 ⎞ ⎟
⎜⎜ R1 ⎜ ⎟ + R 2⎜ ⎟ ⎟⎟
⎝ ⎝ −τ1 ⎠ ⎝ −τ2 ⎠⎠ 11-109
Where:
T(t-1) = shield/sheath temperature from last time step, °C.
Isc = short-circuit current, amperes.
R1 = resistance of first metal at 0°C, ohm/m.
R2 = resistance of second metal at 0°C, ohm/m.
τ1 = inferred temperature of zero resistance of first metal, °C.
τ2 = inferred temperature of zero resistance of second metal, °C.
H = total heat quantity to raise temperatures of both metals 1 C°, kg-J/C°-m3.
Δτ = time step, s.
11-47
H is calculated from the products of the volumetric spe- For cables installed in free air under bridges, on risers or
cific heat capacities and the volumes of the shield and on underground-to-overhead transition poles, a thermal
sheath. The time step should be on the order of 1 milli- transient conduction model is used out to the cable-air
second to get good results. If the final temperature at the interface; the convection and radiation heat loss (and
end of the calculation exceeds the fault–duty-rated tem- solar heat absorption, when present) are modeled as for
perature, the cross-sectional area of one or the other of normal steady-state ampacity.
the metals should be increased, or the fault duration
further limited. 11.3.3 Considerations for Distributed vs. Lumped
Thermal Capacitance
11.3 TRANSIENT RATING AND TEMPERATURE Although the thermal capacitance is distributed
CALCULATIONS throughout the cable system and surroundings as shown
Although the understanding of transient temperature in Figure 11-27, the conduction model used usually
rise and the simpler of the solution methods are reason- lumps the capacitance at discrete points (nodes) in the
able to understand and to do by hand calculations, the thermal circuit. However, a rigorous analytical solution
more accurate of the calculation methods are complex of the heat conduction equation exists for the case of a
numerical methods that are solved using computer pro- step-change in load, which assumes conduction in a uni-
grams. The reader may find useful references to industry form medium with distributed thermal capacitance and
standard methods in IEC 60853 Parts 1, 2 and 3; CIGRE resistance. This analytical solution is used for the tem-
ELECTRA 24, 44, 87, and 145; and Anders 1997. perature transient of the earth portion of the thermal
circuit for the case of a step-change in load.
11.3.1 Application
Computer programs can model the multisection “dis-
Transient calculations for power transmission cables are
tributed” network and are readily available. However, a
performed to solve for the following:
simplified “lumped” method combined with an expo-
• Temperature response to load fluctuation as a func- nential integral is published in industry standards (IEC
tion of time 60853) and is available for hand calculations.
• Maximum load for a specified time period for given
initial and final temperatures 11.3.4 Selection of Solution Method
• Pressure variation of the dielectric fluid with change Two methods are discussed here. The first may be
of load, or maximum loading as limited by pressure applied to calculating the temperature response to load
considerations. variations for any wave shape. A multisection thermal
resistance-capacitance (R – C) network represents the
This section will not cover hydraulic transient calcula- cable(s), and surrounding duct or pipe (if present) and
tions for pipe-type and self-contained cables. earth. The second method may be used to obtain the
temperature response to a step-load change or series of
11.3.2 Methodology step-load changes. This method uses a simplified (R –
C) network for the cable and an analytic solution for the
Transient-loading calculations take the thermal inertia
earth portion of the thermal circuit.
or capacitance of the cable system and its environment
into consideration, while steady-state calculations
Although both methods are generally easier to apply
ignore this thermal capacitance and consider only the
when implemented in a computer program, the multi-
thermal resistance. For cyclic loading, the steady-state
section R – C network requires a computer model, while
calculations are modified to take thermal capacitance of
the simplified (lumped) thermal model could be applied
the earth into account by multiplying the resistance
by hand calculations.
beyond diameter Dx by a daily loss factor, as described
in Section 11.2. Both transient and steady-state methods
utilize thermal conduction models.
Convection and radiation heat transfer such as at the

cable/duct air interface are converted to an equivalent
thermal resistance. Since this equivalent thermal resis-
tance is not a constant, a nonlinear resistance-capaci-
tance (R – C) model results.
Figure 11-27 Distributed thermal capacitances.
11-48
11.3.5 Multisection R-C Network Method The thermal capacitance of each section is the volume
Figure 11-28 shows a typical resistance-capacitance net- of material times the specific heat, which is given by:
work representation of an extruded cable installed in a
π ⋅ Cp n ⎛D 2 D 2 ⎞
conduit. Networks for self-contained or pipe-type cables
Cn = ⎜ n − n −1 ⎟ [J/C°-m] 11-110
are of similar form. The nodes of the network represent 4 ⎜ Nn N n −1 ⎟⎠
the conductor, several diameters through the insulation, ⎝
the metallic shield tape, jacket, bulk air temperature, Where:
conduit, and several points through the earth. Cpn = specific heat of section n, J/C°-m3.
Dn = diameter over section n, m.
Various schemes are possible for apportioning the ther- Dn-1 = diameter under section n, m.
mal resistance and capacitance of insulation and earth Nn = number of conductors inside diameter Dn.
in the thermal circuit. In the case shown in Figure
11-28, the insulation has been subdivided into n sections Table 11-20 lists the bulk (specific) thermal capacities of
using a constant diameter ratio so that each section has materials typically used in transmission cable systems.
the same thermal resistance. Then each section is
divided into two equal subsections, and the thermal The earth thermal resistance is divided in a similar fash-
capacitance of each major section is lumped at the mid- ion. If transients involving load variations of long dura-
point to form a “T” section, as shown in Figure 11-29. Table 11-20 Thermal Capacities for Cable Materials
Material Cp (J/C°-m3)
The earth is divided into m sections in the same fashion as
Copper 3.45 x 106
the insulation. Two additional sections are included for
Aluminum 2.5 x 106
the air space in the conduit (or dielectric fluid in the pipe).
Oil 1.7 x 106
The result is a network of n + m + 3 sections, each Impregnated paper 2.0 x 106
having a resistance and capacitance as shown in Figure Lead 1.45 x106
11-28. The first and last sections have thermal resistance Bronze 3.4 x 106
of Rins / (2n ) , and the rest have Rins / n . The number of Brass 3.0 x 106
sections of insulation to be employed depends on the Steel 3.8 x 106
thickness of the insulation and the period of interest, Nitrogen @ 200 psi (1380 kPa) 0.017 x106
but in general, 4-10 sections are used (CIGRE 1983). PVC 1.7 x 106
Fewer sections might be used when considering only Polyethylene 2.4 x 106
long-time transients and/or thin-walled (e.g., distribu- Rubber and rubber-like materials 2.0 x 106
tion) cables. For very short transients or extra-high- Textile servings 2.0 x 106
voltage cables, the accuracy can be improved by increas- Anticorrosion protection on pipes 2.0 x 106
ing the number of insulation sections. Earth, concrete 2.0 to 2.5 x 106
Duct materials (fiber, polyethylene, etc.) 2.0 x 106
Figure 11-28 Equivalent thermal circuit of an extruded cable in conduit.
Figure 11-29 “T” Section.
11-49
tion are to be considered, at least 10 sections are required adjusted at each time-step based on the temperatures
to accurately model the temperature rise versus time. calculated for the previous time step.
Fewer sections may be employed if only short-time tran-
sients (less than 24 hours) are to be calculated. Note that In cases where ampacity (constant load for time t) is
the product of thermal resistance and thermal capaci- required corresponding to a starting temperature, Tφ,
tance for each section has the dimension of time. The and final temperature, Tmax, at least two repetitions of
appropriate heat inputs are then applied at the conduc- the above step-by-step procedure are required. An initial
tor, insulation, shield/sheath, and duct nodes, in propor- estimate of load is input to the model. Then after the
tion to the load (or square of voltage for dielectric loss). specified time, the calculated conductor temperature,
TC, is used to obtain a second or “educated” estimate of
This network represents one isolated cable. The effect of the load according to Equations 11-112 to 11-115.
a nearby cable or three-phase circuit on the first circuit or
cable can be accounted for by another, similar network This load is input to the model for time t, and the process
with one of its earth nodes representing the temperature is repeated. Since the model is nonlinear, several repeti-
rise at cable 1 due to the loading of cable 2. Then at each tions may be necessary to converge to the required answer.
time step, the temperature rise due to circuit 2 is added to
the temperature of each node in the circuit 1 model. If Tmax − Tφ + AF ( Tφ − ΔTd − Tambient )
two three-phase circuits or three spaced cables have iden- Wf = [W / m]
R total ⋅ AF
tical characteristics and losses, a simplification involving
the use of one network may be employed. 11-112
TC − Tφ
At each node, the rate of temperature increase is: AF = [dim ensionless]
R total ( Wf − Wo )
dT δT Wn ( Tn +1 − Tn ) ( Tn −1 − Tn )
11-113
≡ = + + (C° / s)
{ }
N
dt δt Cn Cn ⋅ R n Cn ⋅ R n −1 R total = ∑ R n [C° - m / W]
n =1
11-111 11-114
Where:
Wf
Wn = heat input to the node n, W/m. It =
Tn = temperatures at node n, °C. R ac
11-115
R n = thermal resistance between Tn+1 and Tn, Where:
Tmax = final conductor temperature, °C.
C°-m/W.
Tφ = initial conductor temperature, °C.
Cn = thermal capacitance at node n, J/C°-m.
Wo = initial watt loss, W/m.
Wf = final watt loss, W/m.
The differential equations for dT/dt are approximated by
the difference equations for δT/δt. The equations (1 for R total = total thermal resistance from conductor to
ambient for constant load, C°-m/W.
each node) are solved by a numerical integration tech-
AF = transient attainment factor, dimensionless.
nique. The simplest such technique involves choosing a
Rac = conductor ac resistance at Tmax, ohm/m.
suitably small time-step (δt) and multiplying the value of
It = transient ampacity, amperes.
by δT/δt by δt.
The calculation of the attainment factor, AF, is
The initial temperatures at each node are calculated
described below.
based on initial loading and a steady-state calculation as
described in Section 11.2. Then, for each time-step and
11.3.6 Composite R-C/Exponential Integral
for each node, the temperature increase, δT, is calculated
Method
and added to the previous temperature.
A quicker, closed-form solution for ampacity can be
At each time-step, the heat inputs are adjusted accord- obtained by utilizing a two- or three-section R-C model
ing to the load variation and the preceding estimate of for the cable system and an analytic expression, the
conductor, insulation, and shield temperature. This exponential integral function, which describes the tem-
allows refinement of the temperature-dependent dielec- perature response of the earth to a step-change in load.
tric loss and electrical resistance (along with hydraulic The R-C network describes the temperature rise of the
transients). Also, the temperature-dependant cable-to- cable conductor over a fixed cable-earth interface tem-
conduit or cable-to-pipe thermal resistance may be perature (the internal portion of the thermal circuit)—
e.g., the cable temperature transient response. The expo-
11-50
nential integral function describes the temperature Where:

response of the earth interface for constant heat input Wi = heat output of a particular cable, pipe, or
(the external portion of the cable thermal circuit). The conduit including shield and sheath losses,
coupling between the internal and external portions of W/m.
the thermal circuit is approximated by calculating sepa- t = time (duration) of the response, s.
rate “attainment factors” for each section. These attain- a, b = internal cable time constants based on the
ment factors are the ratios of transient to steady-state R-C network, 1/s.
temperature rise. The cable (internal) temperature Ra, Rb= thermal resistances from the equivalent
response to a given load after time, t, is given by Equa- R-C network during the time period being
tion 11-116: evaluated, C°-m/W.
( )
ΔTcable (t) = Wi R a (1 − e − at ) + R b (1 − e − bt ) [ C°] The assignment of values to the thermal R-C network in
Figure 11-28 is based upon the cable construction.
11-116 Tables 11-21, 11-22, 11-23, and 11-24 summarize the
basic equations that are necessary to calculate the R-C
Table 11-21 Thermal R-C Network Parameters for Extruded and Self-Contained Cables, Direct Buried
Parameter Short-Time Emergency Long-Time Emergency
RA 1R R insulation
2 insulation
RB 1R + Q S ⋅ R jacket Q S ⋅ R jacket
2 insulation
⎛ D insulation ⎞ Not applicable

C i1
Cp insulation π ⋅ D conductor ⎜⎜ ⎟⎟
4
⎝ − D conductor ⎠
Ci2
Cp insulation π ⋅ D insulation ⎜⎜ ⎟⎟
4
CA C conductor + p ⋅ C i1 C conductor + p ⋅ C insulation
CB pC i 2 + (1 − p)C i1 + (1 − p )C insulation +
⎡ C shield / sheath + p' C jacket ⎤
⎢(1 − p )C i 2 + ⎥×
⎢⎣ QS ⎥⎦ C shield / sheath + p' C jacket
2 QS
⎛ Q S R jacket ⎞
⎜ ⎟
⎜⎜ 1 ⎟⎟
R + Q S R jacket
⎝ 2 insulation ⎠
p 1 1 1 1
− −
D insulation ⎛ D insulation ⎞ D insulation ⎛ D insulation ⎞
2
ln ⎜⎜ ⎟⎟ − 1 2 ln ⎜⎜ ⎟⎟ − 1
D conductor ⎝ D conductor ⎠ D conductor
⎝ D conductor ⎠
p' 1 1 1 1
− −
D jacket ⎛ D ⎞
2 D jacket ⎛ D ⎞
2
2 ln jacket 2 ln jacket
D shield ⎜⎜ D ⎟ −1
⎟ D shield ⎜⎜ D ⎟ −1
⎟
⎝ shield ⎠ ⎝ shield ⎠
11-51
Table 11-22 Thermal R-C Network Parameters for Extruded and Self-Contained Cables, in Conduit
RA 1
2 R insulation R insulation
RB ⎛ R jacket + ⎞ ⎛ R jacket + ⎞
⎜ ⎟ ⎜ ⎟
1
2 R insulation + Q S ⋅ ⎜ R cable − to − conduit + ⎟ QS ⋅ ⎜ R cable-to-conduit +⎟
⎜ ⎟ ⎜ ⎟
⎜ R conduit ⎟ ⎜ R conduit ⎟
⎝ ⎠ ⎝ ⎠

Ci1
4

Ci2
4
CB (1 − p ) Ci1 + 0.3 ⋅ Ci2 (1 − p ) Cinsulation +

Cshield / sheath + C jacket + 0.5 ⋅ Cconduit
QS
p 1 1 1 1
− −
2
ln ⎜⎜ ⎟⎟ − 1 2 ln ⎜⎜ ⎟⎟ − 1
⎝ D conductor ⎠
11-52
Table 11-23 Thermal R-C Network Parameters for HPFF Cables

RA 1R R insulation + 12 QS R Cable − to − Pipe

2 insulation
1R + Q S R Cable − to − Pipe +
RB 1
QS R Cable − to − Pipe + Q P R Pipe Coating
2 insulation 2
Q P R Pipe Coating
C i1 ⎛ D insulation ⎞
Cp insulation π ⋅ D conductor ⎜⎜ ⎟⎟ Not applicable
4
Ci2 ⎛ D insulation ⎞
Cp insulation π ⋅ D insulation ⎜⎜ ⎟⎟ Not applicable
4
CB (1 − p ) Ci1 + Ci2 Cdielectric liquid

(1 − p ) Cinsulation +
QS
p 1 1 1 1
− −
2
ln ⎜⎜ ⎟⎟ − 1 2 ln ⎜⎜ ⎟⎟ − 1
⎝ D conductor ⎠
11-53
Table 11-24 Thermal R-C Network Parameters for HPGF Cables
RA 1R R insulation
2 insulation
RB 1R + Q S R Cable − to − Pipe + Q S R Cable − to − Pipe + Q P R Pipe Coating

2 insulation
Q P R Pipe Coating
C i1 ⎛D ⎞ Not applicable
Cp insulation π ⋅ D conductor ⎜⎜ insulation ⎟⎟
4
Ci2 ⎛D ⎞ Not applicable

Cp insulation π ⋅ D insulation ⎜⎜ insulation ⎟⎟
4
CB (1 − p )C i1 + 0.3 ⋅ C i2 C dielectric liquid

(1 − p )C insulation + +
QS
0.5 ⋅ C pipe
QP
p 1 1 1 1
− −
2
ln ⎜⎜ ⎟⎟ − 1 2 ln ⎜⎜ ⎟⎟ − 1
⎝ D conductor ⎠
network parameters. In all of these tables, R ’s are ther- The values needed to calculate the cable transient tem-
mal resistances in C°-m/W, C ’s are thermal capaci- perature response are developed from the following
tances in J/C°-m, Q S is the ratio of losses in the series of equations:
conductor and shield/sheath (or conductor and
shield/skid wire for pipe-type cables) to those of only the
conductor, and QP is the ratio of losses in the conductor,
M= 1
2 (C (R
A A )
+ R B + CB ⋅ R B (s) ) 11-117
shield, skid wire, and pipe to those of only the conduc-
tor (for pipe-type cables). N = CA ⋅ R A ⋅ CB ⋅ R B (s ) 2
11-118
The duration of the emergency is determined by multi- M + M2 − N

a= (1/ s)
plying the cable thermal resistance by the cable thermal N 11-119
capacitance. If the emergency of interest is longer than
one-third of this time, then the emergency is considered a M − M2 − N
“long-time” emergency. If the emergency duration is less b= (1/ s)
than one-third of this time, the emergency is a “short- N 11-120
1 ⎛ 1 ⎞
time” emergency. There is some overlap between the cal-
()
culation approaches, so for emergency durations close to
one-third, 1 R core ⋅ C core should be evaluated using
Ra = ⎜
a − b ⎝ CA
(
− b R A + R B ⎟ (C° − m / W) )
3 ⎠
both methods and the more conservative value used as 11-121
the rating.
11-54
( )
grals in Equation 11-124 evaluate the mutual heating
R b = R A + R B − R a (C° − m / W) contributions after time, t, has passed.
11-122
Complete Temperature Transient and Computing
The cable surface attainment factor, assuming per-unit Emergency Ratings
current, can be calculated from Equation 11-123: Once the cable and earth (environment) temperature
transients are known, the complete temperature
ΔTcable (t)
α= (dim ensionless) response to a step-change in load may be computed
(
Wi R A + R B ) 11-123
using Equation 11-125.
ΔTtransient ( t ) = ΔTcable ( t ) + α ⋅ ΔTearth ( t ) [C°]

Calculating Attainment Factors 11-125
Reference (VanWormer 1955) gives equations for a
three-section R-C network for the internal portion, with The temperature of the conductor at the end of the tran-
numerical methods providing the required solution for sient (emergency) is the temperature rise from Equation
the roots of a cubic equation. 11-125 added to the temperature of the cable prior to
the emergency. The current assumed during the tran-
A two-section network is not valid for all possible transient would be adjusted until this temperature equals the
sient durations, so two different networks must be evalu- allowable emergency temperature for the cable system
ated—one for times less than a critical time, typically corresponding to the emergency duration.
about 1 hr, and one for longer-duration emergencies. The
solutions to these networks will cross at the critical time, Short-Time Transients—Cautionary Note
though the results are usually close. This requires that Caution should be exercised in applying the above
both networks be evaluated for each time, and the larger methodology for very short-time transients where high-
of the resulting transient thermal resistances selected. load currents are involved. For instance, at the end of a
15-min transient, the temperature could be increasing
Calculation of the internal (cable) temperature transient very rapidly, especially with currents calculated based
(using the R-C network) and the environmental temper- on low starting temperatures. Though there is no “ther-
ature transient (using attainment factors) was developed mal momentum” to the rising temperature, application
further and published in IEC 60853 (IEC 1989). of the calculated ampacity for longer than the time
assumed in the calculation may result in subjecting the
The temperature response to a step load may be calcu- cables to excessive temperatures with consequent loss of
lated using Equation 11-124 (see below). life or damage. Excessive heating may also result from
errors in the assumed parameters, such as ambient tem
In Equation 11-124, the first two exponential integrals perature, earth thermal resistivity, etc.
evaluate the “self ” temperature response to a single
cable, conduit or pipe. The last two exponential inte- IEC 60287 states that the accuracy of the equations for
determining temperature have not been verified for
emergency times of less than 10 min.
ρs ⎡ ⎛ Dearth
2
⎞ ⎛ L2 ⎞ N ⎡ ⎛ d i, j2 ⎞ ⎛ d 'i, j2 ⎞ ⎤ ⎤
ΔTearth (t) = Wi ⎢−Ei ⎜ − ⎟ + Ei ⎜ − ⎟ + ∑ ⎢ − Ei ⎜⎜ − ⎟⎟ + Ei ⎜⎜ − ⎟⎟ ⎥ ⎥
4π ⎢ ⎝ 16δt ⎠ ⎝ δt ⎠ j= 2 ⎣⎢ ⎝ 4δt ⎠ ⎝ 4δt ⎠ ⎦⎥ ⎦⎥
⎣
11-124
Where:
Wi = heat output of each cable (assuming equal loading), W/m.
“Ei” = notation for exponential integral.
Dearth = earth diameter for the cable, conduit, or pipe, m.
L = burial depth for the cable being evaluated (usually the hottest cable in the group), m.
t = time (duration) of the response, s.
δ = thermal diffusivity, m2/s.
ρ soil = soil thermal resistivity, C°-m/W.
di,j = distance between cable, conduit or pipe “i” to cable, conduit or pipe “j”, m.
d’i,j = distance between cable, conduit or pipe “i” to image of cable, conduit or pipe “j”, m.
11-55
11.3.7 Special Topics mization would require high confidence in ther-

The topics described below are quite specialized, and mal/hydraulic design parameters.
may be evaluated using the listed references.
This prompted a series of investigations. Full-scale tests
• Pressure transients may be calculated based on tem- were conducted by Commonwealth Edison (Williams et
peratures determined as described in this chapter. al. 1971) and Power Technologies. Inc. (Williams et al.
Equations for pressure variation will be found in Ref- 1976b), and model studies were conducted by the Mas-
erence (Mattews and Malburg 1977). sachusetts Institute of Technology (Slutz et al. 1975;
• Temperature transients of three-conductor cables are Koci et al. 1977) and the University of Illinois (Chato
discussed in Reference (Wollaston 1949). A single- 1977; Chato et al. 1980). Bulk power projects by Tokyo
conductor cable equivalent is developed. Electric also prompted investigations in Japan (Wan-
• Forced-cooled cable transient methodology is out- tanabe 1974; Wantanabe et al. 1975). The model studies
lined in Reference (CIGRE 1986). predicted as much as a ten-fold increase in friction fac-
tor due to cable snaking and relatively poor heat trans-
• Transients in splices and other accessories involve a fer from cable or pipe to filling fluid in laminar flow.
two-dimensional model and are discussed in Refer- Also, distortions of the thermal field within the cable
ence (Weedy 1988). insulation due to preferential cooling of one surface
(nonisothermal shield) would create a higher-than-
11.4 FORCED COOLING anticipated conductor temperature rise above the tem-
perature of the filling fluid (Glicksman et al. 1978). This
11.4.1 Introduction effect, sometimes referred to as an apparent increase in
This section presents calculation procedures for deter- insulation thermal resistivity, was also reported by ear-
mining steady-state forced-cooled ratings, temperatures, lier full-scale investigations in Japan (Furukawa Electric
and pressure loss for pipe-type cables. Guidelines are 1969) and the United States (Williams et al. 1976b).
also provided for the application of cooling equipment
and accessories to achieve an integrated system design. Subsequently, EPRI and the U.S. Energy Research and
These procedures are presented in simplified form and Development Administration co-sponsored the con-
may be used for planning studies or preliminary engi- struction of a highly instrumented, full-scale research
neering estimates. The user is encouraged to refer to the facility at Waltz Mill, Pennsylvania to supplement and
Designer’s Handbook for Forced-Cooled High-Pressure confirm the results of the earlier investigations and
Oil-Filled Pipe-Type Cable Systems (EPRI 1984) for a determine the applicability of such data to full-scale sys-
thorough treatment of the subject. Also, the CIGRE tems (Williams et al. 1976a). This project was completed
document, “The Calculation of Continuous Rating for i n 1 9 7 7 . T h e e n s u i n g E P R I r e s e a r c h p rog r a m
Forced-Cooled High Pressure Oil-Filled Pipe-Type (Burghardt et al. 1983) supported the high friction-fac-
Cables” (CIGRE 1987), is suggested for those only tor predictions under snaking, but found them to be
interested in thermal ratings. (Forced-cooling calcula- highly localized with increases in overall pressure drop
tions in the United States and overseas are performed in only 5-10% for typical systems. Cable-to-fluid and pipe-
SI units, and the Designers Handbook uses only those to-fluid film resistances, under forced cooling, were gen-
units. Section 11.4, therefore, uses only SI units.) erally found to be lower than self-cooled (uncirculated)
film resistances. However, for long, straight cable
Forced cooling has been applied in the United States lengths of uniform geometry, an increase in the film
since the first documented installation was commis- resistances can occur due to the formation of a thermal
sioned at the Public Service Electric & Gas Co., boundary layer. These thermal resistances are on the
Sewaren, New Jersey generating station in 1948. System same order of magnitude as those that occur in self-
design engineers were faced with many basic uncertain- cooled cables. Minor increases in apparent cable insula-
ties, which were primarily in the areas of thermal resis- tion thermal resistivities were observed during the Waltz
tances between cable or pipe and fluid and the pressure Mill experimentation, which supported earlier theoreti-
drop due to flow in the cable pipe (friction factor). Early cal studies predicting that the relatively high thermal
applications tended to use moderate levels of cooling on conductance, copper shield/brass skid wire construc-
short tie lines, and uncertainties in these parameters had tion would tend to maintain shield isothermality.
minimal impact or were easily compensated for by over-
design. In the 1970s, forced cooling was perceived as a The full-scale experiment indicated that heat transfer
means to substantially upgrade the capacity of major was much better than previously assumed on the periph-
transmission lines of considerable length. Design opti- ery of the cables not directly exposed to fluid flow. This
limited increases in apparent thermal resistivities to only
11-56
10-12% for worst-case constructions, utilizing relatively turbulent regime. The factor used to account for loss
poor thermal conductance stainless-steel shields and imbalances observed at Waltz Mill has been removed,
skid wires. Furthermore, at very high Reynolds numbers since this factor has not been substantiated by more
(high turbulence), all surfaces of the cables are equally generalized and comprehensive studies. Should a thor-
cooled, virtually eliminating the apparent increase in ough study of the imbalanced ac losses ensue, it could
thermal resistivity. In an effort to quantify losses, a calo- be more legitimately applied. A more generalized equa-
rimetric method was devised, which revealed that there tion for calculating allowable current under forced-cool-
is a difference in ac losses in each cable phase even ing (Equation 11-142) has also been substituted.
though the phase currents are equal. This difference
could be up to a 10% increase in loss on the center phase The basic cooling loop is depicted in Figure 11-30. It is
of cables in cradled configuration. It should be pointed the smallest entity containing coolant supply and return
out that this phenomenon was observed for the Waltz paths, the pump, and heat exchanger over which conser-
Mill cable/pipe geometry, which had a low cable-diame- vation of mass and energy is maintained. Hence, the cal-
ter-to-pipe-diameter ratio, and may not necessarily culation procedure refers to the “cooling loop.” Many
apply equally to all types of cable systems. Although cooling loops can be employed, and a variety of flow
this phenomenon is not peculiar to forced-cooled cable, paths can be utilized as long as temperatures and pres-
the imbalance in losses is more apparent at higher sures fall within the required boundaries. Figure 11-31
forced-cooled ac currents, which can cause heating of shows a more complex cooling system.
one phase relative to the average.
11.4.2 Units, Nomenclature, and Conventions
The results of these investigations were correlated, and The forced-cooled calculations adhere to the SI conven-
generalized equations were presented in the Designer’s tion, which will provide a 1:1 correlation with the equa-
Handbook (EPRI 1984). This section represents a tions and methodologies presented in the Designer’s
digest of the calculation portion of the Designer’s Handbook (EPRI 1984). All equations and parameters
Handbook. Complicated cable configuration and henceforth are in terms of SI units. An over-bar conven-
degree-of-snaking estimating techniques have been elim- tion is used to distinguish thermal resistance ( R ) from
i
inated, and a simplified friction-factor chart (for pres- electrical resistances ( R ).
sure-drop calculations), based on a rational distribution ac
of geometries, has been provided. 11.4.3 Calculating Losses
The losses, with the exception of the daily loss factor, are
There are several noted changes in this book, as com-
computed using equations presented for self-cooled
pared to the 1992 edition. Since the combination of the
HPFF cables given previously in this chapter (Section
aforementioned factors can lead to very pessimistic
11.2.5). The closed-form solutions given in this section
results, this edition of the reference book has adopted
require constant loss parameters over the length of the
several changes. A Reynolds number dependent equa-
cooling section. Therefore, it is recommended that the loss
tion for applying the non-isothermal shield effect has
parameters be fixed at the limiting temperature condition.
been included. As the development of a thermal bound-
ary layer is only likely in laminar flow, the impact of the
The effect of the daily load cycle and the introduction of
development of the thermal boundary layer has been
the loss factor are treated somewhat differently in the
removed from the fluid film resistance equations in the
forced-cooled equations. The forced-cooled equations
Figure 11-30 Basic cooling loop with cable pipes as the fluid supply and return path.
11-57
Figure 11-31 Multiple cooling loops.
were developed on the basis that 100% attainment is D = inside pipe diameter, m.
achieved between mid-fluid and conductor at peak
loading, while heat flow from mid-fluid outward is Area A is the cross-sectional flow area:
related to some function of the daily load factor. The
A = π ⎡( D/2 ) -3 ( d/2 ) ⎤ m
2 2
following equation for the forced-cooled loss factor,
LFfc, was empirically derived, comparing the results of ⎣ ⎦ 11-128
the steady-state solution to a dynamic model for a num- Fluid Properties
ber of cases. Table 11-25 presents a compilation of properties for
some of the commonly used fluids in pipe-type cable
LFfc =0.8 ( f ) +0.2 ( f ) dimensionless
2
11-126 systems for design and cooling-fluid selection purposes.

Where:
f = load factor The volumetric coefficient of thermal expansion, β, can
be taken as a constant equal to 0.00075 per C° over
The similarity of this equation to that used for distribu- common temperature ranges.
tion cables is by coincidence.
Calculating Thermal-Hydraulic Parameters and
Dimensionless Quantities
11.4.4 Thermal-Hydraulic Characterization
The following parameters and dimensionless quantities
The following sections deal with calculating the dimen- must be determined to compute the liquid film resis-
sionless quantities that will, in turn, be used to calculate tances and pressure loss due to friction. These should be
fluid-film resistance and friction factors. expressed as a function of temperature using the equa-
tions for the liquid properties if iterative procedures are
Geometric Parameters
used.
Diameter d is the fictitious diameter over the cable cal-
culated by taking the cross-sectional area of the cable
V is the mean velocity
over the shielding tape plus the cross-sectional area
taken by the skid wires. For typical cable construction: V = Q/A m/s 11-129
d = 1.01 x (dia. over shield), double-entry skid
wires (m). Re is the Reynolds Number
d = 1.005 x (dia. over shield), single-entry skid V Dh
wires (m). Re = dim ensionless
υ 11-130
A hydraulic diameter, Dh, is calculated to give an equiv-
alent flow diameter for the three cables in the pipe: Pr is the Prandtl Number
D 2 -3d 2 τc p υ
Dh = m Pr = dimensionless
D+3d 11-127
k 11-131
Where:
d = effective cable diameter over the shield, m.
11-58
Table 11-25 Fluid Properties

Fluid Constant τ cp k υ
TDB A -0.533 3.50 -7.0 E-5 -4.6430
B 862 1800 0.153 2.0135
LVM A -0.6300 4.25 -7.0 E-5 -5.5000
B 911 1730 0.124 2.1420
LVA B1 A -0.6200 2.50 -7.0 E-5 -5.8770
B 880 1750 0.113 2.2585
LVP A -0.6300 4.00 -7.0 E-5 -5.5400
B 846 2050 0.123 2.2770
LVA B2 A -0.6200 2.50 -6.1150
No Data
B 880 1750 2.3245
HVP A -0.6132 4.00 -7.0 E-5 -6.5270
B 862 1890 0.112 2.5026
HVM A -0.6300 3.50 -7.5200
No Data
B 941 1750 2.5739
Definition of Fluid Acronyms
TDB—Tridecylbenzene
LVP—Low-Viscosity Polybutene
LVM—Low-Viscosity Mineral Oil
HVP—High-Viscosity Polybutene
LVA B1—Low-Viscosity Alkylbenzene
HVM—High-Viscosity Mineral Oil
LVA B2—Low-Viscosity Alkylbenzene
Equations for determining thermal properties of fluids which vary with temperature
Density τ = AT +B (kg/m3)
Specific heat cp = AT + B (W-sec/kg-C°)
Thermal conductivity k = AT + B (W/C°-m)
Kinematic viscosity υ = (exp(exp(A x 10-3 x T + B))) x 10-8 (m2/sec)
T = fluid temperature in °C.
Gr is the Grashof Number g = 9.8 m/s2 = acceleration due to gravity.

β = volumetric coefficient of thermal expan-
g β ΔT D3h sion per C°.
Gr = dimensionless ΔΤ
υ2 11-132
= temperature difference between shield or
Where: pipe and mid-fluid, C°.
Q = flow rate, m3/s (Note that m3/s = GPM x Calculating Cable Component Thermal Resistances
6.309 x 10-5.).
A = cross sectional flow area, m2. Insulation Resistance— Ri
V = velocity, m/s.
Dh = hydraulic diameter, m. The insulation resistance calculated in Equation 11-48
υ = kinematic viscosity C°-m/W. must be corrected to account for a nonisothermal shield
τ = density kg/m3. due to forced fluid circulation. Table 11-26 lists correc-
cp = specific heat, W-sec/kg-C°. tion factors (Zi) for common cable constructions.
k = thermal conductivity, W/C°-m.
⎣ ( )
R icorr = ⎡1 + 1.05 ⋅ ( exp ( −0.0002 Re ) − 0.05 ) ⋅ ( Zi − 1) ⎤ ⋅ R insulation
⎦ 11-133
Where:
R i corr =corrected insulation resistance, C°-m/W.
Zi = correction factor from Table 11-26.
R insulation = calculated insulation resistance from Equation 11-49, C°-m/W.
11-59
Table 11-26 Insulation Resistance Correction Factors (Zi)

Conductor Insulation Thickness (cm/in.)
mm2 kcmil Shield Mat’l. Skid wire Mat’l. 0.508/0.2 1.016/0.4 1.524/0.6 2.032/0.8 2.54/1.0
507 1000 Copper Brass 1.01 1.00 1.00 1.00 1.00
760 1500 1.02 1.01 1.00 1.00 1.00
1013 2000 1.03 1.02 1.01 1.01 1.01
1267 2500 1.03 1.02 1.02 1.01 1.01
1520 3000 1.04 1.03 1.02 1.02 1.02
507 1000 Copper Stainless Steel 1.02 1.02 1.01 1.01 1.01
760 1500 1.03 1.02 1.02 1.02 1.02
1013 2000 1.04 1.03 1.02 1.02 1.02
1267 2500 1.05 1.03 1.03 1.02 1.02
1520 3000 1.05 1.04 1.03 1.03 1.03
507 1000 Zinc Stainless Steel 1.06 1.05 1.04 1.04 1.04
760 1500 1.07 1.06 1.05 1.05 1.05
1013 2000 1.09 1.07 1.06 1.06 1.05
1267 2500 1.10 1.08 1.07 1.06 1.06
1520 3000 1.10 1.09 1.07 1.07 1.07
507 1000 Stainless Steel Stainless Steel 1.11 1.10 1.10 1.10 1.09
760 1500 1.12 1.11 1.10 1.10 1.10
1013 2000 1.12 1.12 1.11 1.11 1.10
1267 2500 1.13 1.12 1.12 1.11 1.11
1520 3000 1.13 1.12 1.12 1.12 1.12
Dielectric Liquid Fluid Resistances Ap = surface area of inner pipe wall per meter,
These values are determined for both laminar and tur- m2.
bulent flow conditions. H = constant, = 14 if pipe is cabled, 5.5 if pipe
is clear, dimensionless.
R so is the shield-to-fluid resistance
Wetted surface areas of cable, A c , and pipe, A p, are
2.75D defined in Equations 11-138 and 11-139:
h
Laminar: R sol = C°-m/W
kA c ( Gr i Pr )
0.25
11-134
A c = π Ds m 2 11-138
28 A p = π Dp m 2
Turbulent : R sot = 11-139
0.8 0.33
k ⋅ Re Pr 11-135 Where:
Ds = cable shield diameter, m.
R op is the fluid-to-pipe resistance Dp = pipe inside diameter, m.
3D Coating Resistance
h
Laminar: R opl = C°-m/W R d is the protective-coating thermal resistance,
kA p ( GrPr )
0.25
11-136 C°-m/W, calculated from Equation 11-49. The number

3H of conductors should be set to 3.
Turbulent: R opt = C°-m/W
k Re 0.8 Pr 0.33 11-137
Earth Resistance
Where: R earth is the earth thermal resistance, C°-m/W, and is
Dh = hydraulic diameter, m. calculated by Equation 11-60. The loss factor should
k = thermal conductivity. W/C°-m. not be introduced in this resistance (e.g., the calculation
Ac = surface area of cable per meter, m2. should be done at 100% loss factor as would be done for
Gr = Grashof number, dimensionless. dielectric losses). The number of conductors should be
Pr = Prandtl number, dimensionless. set to 3. The mutual heating thermal resistance compo-
Re = Reynolds number, dimensionless. nent, the mutual thermal resistance component is
11-60
included in Equation 11-60. Both the self-heating resis- act at node p and heat (q) is removed by fluid flow in the
axial direction at the mid-fluid node, o. The resistances
tance, ( R ), and the mutual term can be applied to
earth from mid-fluid to ambient earth contain a multiplying
liquid supply or return lines without cable. factor of 3 since the losses are based on a single cable.
Also, the heat removal term, q, contains a one-third fac-
11.4.5 Axial Thermal Calculations tor to yield the proper temperature rise in the axial direc-
The basic forced-cooled thermal formulas for a single tion. The resistances in the network are as follows:
pipe are developed from the following equivalent resis- R so = cable-surface-to-fluid film resistance for
tance network shown in Figure 11-32. one cable, C°-m/W.
R op = pipe-to-fluid film resistance including 3x
Where nodes are defined as follows: factor, C°-m/W.
s = the combined surface of the three cables. R d = protective-coating resistance including 3x
o = the mid-fluid point. factor, C°-m/W.
p = the steel pipe. R e = effective earth resistance including 3x fac-
d = the earth interface. tor, C°-m/W.
e = the earth at the horizontal centerline of the
pipe. The following differential equation can be written for
Wφ = heat injected at the mid-fluid node (Wc + the model given in Figure 11-32.
Wd + Ws), W/m.
dTo ( x )
Wp = pipe loss, W/m.
λ = To ( x ) -To ( u )
dx 11-140
Cable surface temperatures are assumed to be isother- Where:
mal, and therefore, conductor (Wc), dielectric (Wd), and λ = forced-cooling characteristic = –CQRoe/3,
shield (Ws) losses are lumped together to act as a single m/°C
source Wφ at node s. Eddy-current losses in the pipe (Wp) To (u) = ultimate temperature of the liquid without
forced cooling, °C.
To (x) = temperature of the liquid at any distance x,
°C.
Integrating Equation 11-140 explicitly and applying the

boundary condition at x = Φ, and To (x) = To (i) initial
oil temperature at the inlet, °C, yields the following use-
ful forms for a cooling section of length, L:
Relationships for the Single-Pipe Case
−L
Q= ft 3 / s
⎛ T − ΔTc' − To (u ) ⎞
C ⋅ R oe ln⎜⎜ c ⎟⎟
Figure 11-32 Equivalent resistance network for forced- ⎝ T o (i ) − To ( u ) ⎠
cooling calculations.
11-141
⎡⎛ Ri⎞ + Rso⎤ − Toi ⋅ exp⎛ L ⎞ − ( 3 ⋅ Wd ⋅ Roe100 + Te) ⋅ ⎛ 1 − exp⎛ L ⎞ ⎞

Tc − Wd ⋅ ⎢⎜ ⎟ ⎥ ⎜ ⎟ ⎜ ⎜ ⎟⎟
⎣⎝ 2 ⎠ ⎦ ⎝λ⎠ ⎝ ⎝ λ ⎠⎠
I :=
⎡ ⎛ ⎛ L ⎞ ⎞⎤
Rac ⋅ ⎢( Ri + Rso) + Roa ⋅ ⎜ 1 − exp⎜ ⎟ ⎟⎥
⎣ ⎝ ⎝ λ ⎠ ⎠⎦ 11-142
11-61
Case 1—Thermally Remote Pipes

⎛1⎞ ⎛ ⎛ L ⎞⎞
To (1) = To ( i ) exp ⎜ ⎟ + To ( u ) ⎜ 1 − exp ⎜ ⎟ ⎟ °C R mut < 0.1R e : Pipes are treated individually by setting
⎝λ⎠ ⎝ ⎝ λ ⎠⎠
the outlet of one pipe equal to the inlet of the other at
11-143
Tc (1) = To (1) + ΔTc' °C

the turnaround point: To ( )1 = To (i)2
11-144
Case 2—Parallel Identical Cable Pipe
The auxiliary equations for the above are: 0.1 ⋅ R e < R mut < 0.6 ⋅ R e : Set R e = R e + R mut
−CQR oe and set L=2L
λ= m / °C
3 11-145
(
⎡(1 + Yc' ) R + R so
ΔTc' = I 2 R dc ⎢ i )⎤⎥
⎢+Y R ⎥
⎣ s so ⎦
⎛R ⎞
+ Wd ⎜ i + R so ⎟ C° Case 3—Separate Supply Pipe
⎜ 2 ⎟
⎝ ⎠ 11-146
0.1 R e <R mut <0.6 R e : The supply pipe (#1) is
neglected, and the fluid inlet temperature to the cable
To ( u ) = I 2 R dc (1 + Yc ) R oa pipe is set to the fluid temperature into the supply pipe:
( )
+ Wd R op + R d + R e' + Te °C
11-147
To ( i )2 = To ( i )1
R oa' = qs R op + q e R pe TOM
11-148
( )
R op = R op + R d + R e . ( LFfc ) TOM
11-149
R oe100 = R op + R d + R e
11-150
( )
Case 4—Separate Return Pipe
R pe = R d + R 3 . ( LFfc ) TOM
11-151 0.1 R e <R mut <0.6 R e : The pipes are treated individ-
Ys ually, as in Case 1, but mutual resistances are added to
qs =1+ dimensionless
1+Yc R e1 .SetR e1 = R e1 + R mut and set R e2 = R e2 + R mut
11-152
Ys +Yp
q e =1+ dimensionless
1+Yc 11-153
R oe100 = The combined oil and earth thermal
resistance taken at 100% loss factor.
Each of these terms is defined in the Nomenclature at Case 5—Symmetrical Systems

the end of this chapter. Symmetrical systems can be accommodated by adding
the mutual resistances of the second system to the first:
Procedures for the Multipipe Case
A liquid supply or return pipe is required to establish a set R e2 = R e2 + R mut2 •2 '
circulation loop. This pipe could be a parallel cable pipe
or an uncabled pipe specifically incorporated to form
the cooling loop. If parallel pipes are in thermal proxim-
ity to one another, then a true thermal solution requires
the inclusion of thermal interference effects, which sig-
nificantly complicates the mathematical approach.
Under many conditions, though, the single-pipe solu-
tion yields fairly accurate results. The following guide-
lines may be used for applying the single-pipe equations
to multi-pipe configurations.
11-62
11.4.6 Pressure-Loss Calculations Fluid properties should be evaluated at the one-third

The pressure loss in the cable system can be calculated point:
from Equation 11-154.
2Tin + Tout
Teval = °C
ΔP = K f ⋅ f
(γ ⋅V 2
⋅ Δx ) 3 11-157 11-157
2 ⋅ Dh 11.4.7 Application of Equations

11-154
Where: Cooling Plant
Kf = correction factor, dimensionless. The cooling plant heat load, H, is calculated from
f = friction factor, dimensionless. Equation 11-158.
γ = density, kg/m3.
V = velocity m/s. CQ ( To ( ) -To ( i ) )
Δx = length, m. H = kW
Dh = hydraulic diameter, m.
1000 11-158
Where:
Pressure drop in psi may be obtained by multiplying val- C = specific thermal capacity, W-s/m3-C°.
ues in Pascals by 1.45 x 10-4. Q = flow rate, m3/s.
To ( ) =fluid temperature at outlet.
A correction factor, K f = 1.2, is recommended to To (i) = fluid temperature at inlet.
account for additional losses due to cable snaking and
accessories such as joints and diffusion chambers. The The fluid temperature can typically be lowered to within
friction factor, f, for cabled pipe can be determined from 5-6°C of the heat-sink temperature with a heat
Figure 11-33. exchanger. If additional temperature differential is
required, then mechanical refrigeration must be
For uncabled pipe, the friction factor, f, can be approxi- employed. The split between mechanical refrigeration,
mated from Equation 11-155 and 11-156:
64
f= Re < 1800
Re 11-155
f = 0.0356 Re > 1800 11-156
Figure 11-33 Friction factor versus Reynolds Number.
11-63
H mr, and the heat exchanger, H he, is calculated from are thoroughly discussed in EPRI Report EL-2233,
Equations 11-159 and 11-160. High Ampacity Terminations (G&W Electric Specialty
Co. 1982.). On the other hand, the porcelain is likely to
To ( ) -The +6 be the weakest pressure component in the system. This
H he = H kW
To ( ) − To ( i ) may limit the maximum allowable flows or dictate the
11-159 system flow configuration.
Hmr = H-Hhe kW 11-160
Hydraulic Components
The power requirements for the cooling plant, P c , Fluid exit and entrance chambers must be included to
including the pump, can be estimated from Equation complete the hydraulic circulation scheme. Entrance
11-161 for a pressure drop, ΔP. (diffusion) chambers are designed to prevent impinge-
ment of fluid normal to the cable surface while exhibit-
⎛ QΔP ⎞ ing fairly low pressure losses. The interior surface must
Pc = 1.1⎜ 0.05H he +0.25H mr + ⎟ kW also allow cable pulling. Although an exit chamber may
⎝ 1000 x Eff ⎠ be a simple “T” fitting, diffusion chambers are some-
11-161 times employed in case flow direction needs to be
reversed. Other hydraulic components such as valves
Pump efficiency (Eff) can be as low as 0.50, but usually and fittings should be selected with low pressure loss in
an efficiency of 0.80 can be obtained under the design mind.
working conditions.
Splices
The ultimate operating pressures under forced-cooling
The higher insulation thermal resistance ( R ) of should be reviewed so that hydraulic components meet
ij the required pressure capabilities.
splices, as compared to cable, should not be ignored at
the higher loadings afforded by forced cooling. An System Design
investigation by Engelhardt (Engelhardt 1972) suggests The degree of forced cooling is usually limited by the
the multiplication factors in Table 11-27 for various maximum allowable pressure drop around the cooling
splice types. loop. The high pressure is limited by the pressure capa-
Table 11-27 Splice Insulation Thermal Resistances bility of components, and the low pressure is limited by
Multiplication Factors either the minimum cable pressure or minimum pump-
suction pressure. The minimum cable pressure is usually
R ij 1000-1200 kPa (145-175 psi) and pump-suction pres-
Construction sures should be maintained above 310 kPa (45 psi) to
eliminate cavitation during startup or transient condi-
Crepe Paper 1.5 Ri tions. A pressure profile can be drawn along the cooling
loop, including pressure losses in the cable pipe, supply
Flat Paper 1.33 Ri or return line, heat exchanger, and other components.
The static profile due to elevation changes is then super-
Cast Epoxy 1.25 Ri imposed on the pressure profile due to flow. All system
components must be capable of withstanding the pres-
LPP 1.13 Ri sure associated with its location in the cooling loop.
Likewise, a fluid-temperature profile around the cooling
loop can be generated. Conductor-temperature rises,
The temperature rise of the splice conductor over fluid including those of the splices, can be superimposed on
can be estimated by substituting R for R in Equa- the fluid-temperature profile. If pressure or temperature
ij i
tion 11-146. limits are exceeded for a given cable loading and cable
Terminals design, the length of the cooling loop must be
Unlike splices, which are almost always thermally limit- decreased. In a multiple cooling-loop design, tempera-
ing in forced-cooled cable systems, terminals and risers tures and pressures must be balanced at the fluid turn-
are likely to be satisfactory without modification. Under around point. These concepts are treated more
typical conditions, terminals can carry 35% more cur- thoroughly in Chapter 9.
rent than the cable in pipe. When fluid is circulated
through the risers, even more capacity is realized due to 11.4.7 Alternative Calculation Approach
axial conduction of heat away from the terminal toward In addition to the methodology summarized above
the risers. Terminal ampacities and cooling procedures based on the EPRI handbook on forced cooling,
11-64
CIGRE (CIGRE 1987) developed a methodology for considerations for an ampacity audit are briefly dis-
evaluating forced-cooled ratings on pipe-type cables. cussed below.
The methodology is based on consideration for addi-
tional pressure drop (up to 20%) that may occur as a The basic concepts include performing a soil thermal
result of cable snaking. survey to determine soil characteristics and ambient
temperatures, evaluating load history, and then calculat-
The approach outlined in the reference requires the ing ampacity. If the AEIC guidelines are being followed
solution of simultaneous equations and up to an “nth” on a circuit that previously had used assumed soil
order polynomial, where there are “n” cable pipes being parameters (requiring a 10C° drop in maximum allow-
evaluated. The basic forms of the equations are as fol- able conductor operating temperatures), regaining the
lows (for a two pipe system): 10C° in conductor temperature through a thermal sur-
vey by itself generally allows a 10% or more increase in
ΔT1 ( x ) = K1 ⋅ a11 ⋅ e P1 ⋅ x + K 2 ⋅ a12 ⋅ e P2 ⋅ x ampacity assuming no other mitigating factors.
11-162
The ampacity audit is geared towards verifying the
ΔT2 ( x ) = K1 ⋅ a 21 ⋅ e P1 ⋅ x + K 2 ⋅ a 22 ⋅ e P2 ⋅ x ampacity by whatever means are available and assessing
11-163 which locations along the route limit the overall circuit
ampacity.
Where K1 and K2 reflect the uncooled temperatures of
Cable Construction Identification
Pipe 1 and Pipe 2, a11, a12, a21, and a22 are constants that
It is necessary to identify what cable constructions are
relate the mutual heating between the cable pipes, P 1
installed along a given circuit, particularly if the cable
and P 2 are constants that reflect the dielectric liquid
system is older and may have had sections replaced dur-
flow rate and change in temperature with respect to dis-
ing repairs. A careful review of owner records often
tance, x, along the cable pipes. Using the above equa-
reveals manufacturer’s data sheets that show the con-
tions, the change in temperature from the cooled to the
ductor size, insulation thickness, shield construction,
uncooled case can be determined with respect to dis-
etc. One-line diagrams are sometimes annotated with
tance along the pipe. The reader is encouraged to review
cable manufacturing data. If only conductor size and
the references at the end of this chapter for additional
material can be identified, most of the other parameters
information on this methodology.
for the cable can be developed based on the system
operating voltage and industry standards.
11.5 AMPACITY VERIFICATION AND
UPRATING Develop Installation Data
Cable system owners sometimes are faced with the real- As-built installation records are very helpful during an
world challenge of understanding a cable circuit for ampacity audit, since these details provide the burial
which complete information is initially unavailable. In depth, relative cable spacing, etc., needed for the ampac-
conjunction with obtaining a better understanding of ity calculations. Sometimes senior engineering staff may
the parameters affecting ampacity—called an “ampac- have first-hand knowledge of a particular circuit’s
ity audit”—utility engineers in particular are called installation data that can be helpful. Failing that, an
upon to get more capacity out of existing infrastructure above-ground or underground survey may help with
or perform “uprating.” identifying the installation conditions, including using
test pits to locate cable circuits.
This section is not intended to be a comprehensive dis-
cussion on these topics. Various publications, including One particular challenge to getting the needed installa-
EPRI’s Increased Power Flow Guidebook-Underground tion data is evaluating interference temperature effects
Cables (1001818, December 2003), have a detailed dis- that may be raising the temperature of the cable system
cussion of these topics. Following is a summary of the and, consequentially, robbing amperes from the circuit
issues to consider. or accelerating cable loss-of-life, because the circuit is
unknowingly being heated above its rated temperature.
11.5.1 Ampacity Verification Route Thermal Survey
An ampacity verification study is focused on gaining an A route thermal survey is primarily focused on identify-
increased understanding of the parameters needed to ing the soil thermal resistivities along the cable route
calculate ampacity. Often, particularly with older cable where the circuit of interest is installed. A qualified soil
system installations, the basis for the ratings used on the testing specialist would be called in to work with a local
line are not well documented or understood. Some basic driller. Soil thermal resistivity tests can be done in the
11-65
field—called in situ—to identify the soil conditions as hot spot would need to be developed using the owner’s
they naturally exist. The field measurement also pro- plan and profile information. Once the hot spots are
vides an ambient soil temperature at the depth and time identified, a mitigation plan could then be developed
of year when the measurements are taken. In addition to and implemented.
field measurements, samples would be obtained for lab-
oratory analyses to develop thermal dryout curves 11.5.2 Uprating
showing the impact of moisture content on thermal The following methods are mostly applicable to pipe-
resistivity. This information would then be used to type cables, although there are applications that could
update the cable ratings. extend to extruded or self-contained cables.
Distributed Fiber Optic Temperature Sensing Dielectric Fluid Circulation
Distributed fiber optic temperature sensing (DFOTS) Fluid circulation in HPFF cables is a common technique
uses a specialized optical time-domain reflectometer for mitigating relatively short thermally-limiting sections.
(OTDR) to measure the temperature along a multimode Circulating the dielectric liquid within the cable pipe
optical fiber; this type of measurement is also called dis- allows the heat generated in the hot section to be trans-
tributed temperature sensing (DTS). The process works ferred to other portions of the route, mitigating the hot
by taking advantage of temperature-dependent reflec- temperatures at that location. One requirement for this to
tions (called “backscatter” based on the Raman Effect) be implemented is to have a fluid return pipe or a parallel
in the fiber. The special OTDR instrument (see Figure cable pipe that permits a continuous circulation path.
11-34) records the magnitude of the reflection (propor- Flow rates are generally in the range of 20-160 l/m (5-40
tional to temperature) and the time for reflections to gpm) depending on the length of the hot spot. Fluid oscil-
return after sending an incident 1080 nm laser pulse into lation can sometimes be applied where no fluid return
the fiber, which, when combined with the fiber’s propa- pipe is available but is less efficient than circulation.
gation velocity, gives the distance to the measurement
location. By successively sending light pulses into the Fluid circulation can be extended to forced cooling, as
fiber, the special OTDR can scan the entire fiber and discussed earlier in this chapter, to actively increase the
obtain a temperature trace along the fiber with a spatial capacity of a pipe-type cable.
resolution of approximately 1 m and a temperature
accuracy of 1°C for multimode fibers (less accuracy is Heat Pipes
obtained with single-mode fibers). The obvious advan- Heat pipes are passive devices that take advantage of the
tage to DFOTS is that a continuous end-to-end temper- heat of vaporization of an internal fluid to remove heat
ature measurement is possible, allowing the ampacity from a location and move the heat a few meters away. A
study to reveal all of the hot spots along the cable route. heat pipe is constructed using an evaporative coolant
Later, these hot spots could be instrumented with ther- such as an alcohol-water or ammonia-water mixture in a
mocouples for extended temperature measurements at partially filled copper tube. A partial vacuum may be
key locations. drawn on the tube to adjust the vapor pressure to the
operating temperature range for the particular applica-
DFOTS offers the possibility that all the hot spots along tion. The heat pipe is then installed at an angle with the
a cable circuit could be identified. The reason for each low point installed near the heat source (cable, steam
main, etc.). Heat from the source is absorbed by the
coolant solution, causing a phase change to vapor, which
rises, carrying the heat away. The gaseous vapor then
condenses at the remote location back to a liquid, away
from the hot spot, and then drains back to the hot spot
location. This continuous process removes heat from the
hot spot location. The number and installation geometry
of the heat pipes are typically designed by a specialist.
Real-Time Dynamic Ratings
Comparison of Book and Real-Time Dynamic Ratings
Book or “precalculated” ratings of buried transmission
lines have historically been based on assumptions about
Figure 11-34 Two types of DTS equipment being used in the nature of the load cycle and the thermal parameters
the field for an ampacity audit (courtesy of PDC). of the surroundings. The thermal properties of cable
components and containment systems have been mea-
11-66
sured in laboratories and are well known. Although the The Dynamic Thermal Model is a computer-based
thermal resistivity and diffusivity of the soil surrounding mathematical thermal model based on the geometry
the cable systems can be determined in the laboratory, and design of the cable system that is able to dynami-
these properties are weather dependent, and to a certain cally respond to cable loading and external thermal con-
extent dependent on cable loading. They vary with time di t i o n s. I t c o n t i nu o u s ly o b t a i n s d at a f ro m t h e
and in space due to variation in the soil type, density, measurement system and updates its thermal state on a
and moisture content. The latter is a function of rainfall, frequent basis (typically 1 minute).
location of the water table, and loading. In addition,
ambient temperature varies with depth and with season. The parameter identification algorithm uses the com-
puted and measured temperatures and adjusts or com-
Normal ratings are strongly influenced by the earth por- putes the unknown thermal parameters, such as the
tion of the thermal circuit. The temperature rise in the earth thermal resistivity and diffusivity, so as to mini-
earth due to cable loading is normally greater than the mize the difference between measured and computed
temperature rise in the cable from the earth interface to temperatures. The result is a thermal model that can be
the conductor. Without knowledge of the thermal state used to compute accurate ratings.
of the cable system, precalculated or “book” ratings
must use conservative assumptions for earth properties The ratings algorithm uses the thermal model to com-
and ambient temperature. pute ratings based on the present thermal state of the
cable system. Normal and long-time emergency ratings
Steady-state or normal ratings are a function of the load may be based on the actual load cycle. All ratings reflect
cycle, since the thermal time constant of the earth is the actual ambient conditions rather than assumed con-
long compared to the length of the daily load cycle. stant ambient earth temperature and earth thermal
resistivity.
Transient or emergency ratings are a function of the
Location of Measurements
load shape and the starting load or temperature of the
Measurements can be made either at discrete “cross sec-
cable system. Without real-time measurement and a
tions” or continuously from end to end by the use of
modeling system, these assumptions for preloading
optical fiber. When discrete measurements are used,
must also be conservative
locations to be measured are selected taking into con-
sideration variation in depth, location of the water table
Knowledge of the ambient earth temperature, load his-
and crossing or parallel cables, and other heat sources.
tory, and the temperature response of the cable system
Locations at shallow depth tend to limit ratings in sum-
to this load enables real-time dynamic ratings to be
mer. Deeper sections tend to limit ratings at other times
computed based on actual rather than assumed condi-
of the year.
tions. The result is ratings that are frequently higher
than those based on assumptions about the parameters Typical Real-Time Data
enumerated above. In particular, dynamic emergency Data available from a real-time ratings system include
ratings can be significantly higher than ratings that the inputs—load currents, ambient earth temperatures
assume loading at 100% of the normal rating. Normal at various depths, and cable or containment surface
ratings can also be higher in the case where the designer temperatures—and outputs, such as normal and emer-
has assumed conservative thermal resistivity and ambi- gency ratings. Also available are quantities used to com-
ent temperatures. pute the ratings, including computed cable temperatures
and identified thermal resistivity, daily and weekly load
Components of a Real-Time Dynamic Rating System
and loss factors, computed currents at points along the
The principal components of a real–time dynamic rat-
route, power factor, and power delivered.
ings system are: (a) continuous data collection, (b) a
dynamic thermal model, (c) a parameter identification System Operational Considerations
algorithm, and (d) a ratings algorithm. Since some calculated real-time ratings vary consider-
ably from day to day or week to week, system operators
Data collected are: (1) continuous load current, (2) may require continuously updated minimum and/or
ambient earth temperature, and (3) temperature mea- average ratings computed over a period of time such as
sured on the cable system such as the surface of a pipe 30 days.
or cable, an empty duct, or a temperature inside the
Comparison of Real-time and Precalculated Ratings
cable jacket.
Normal dynamic ratings will be above precalculated val-
ues if the assumptions regarding soil properties were
11-67
conservative. It is possible that normal ratings will be ings depends on the level of loading at the time that the
below the precalculated values in cases where higher ratings are computed.
than assumed thermal resistivity exists. In any case, a
Quasi-Dynamic
real-time dynamic rating system maximizes the rating
While real-time ratings are very helpful to cable system
capability of the cable system and ensures that it is oper-
engineers and operators, sometimes a comprehensive
ating safely within its temperature limits. Typically
real-time rating system cannot be implemented. Instead,
emergency dynamic ratings are significantly above pre-
many of the techniques used for full real-time rating can
calculated values.
be applied on an intermittent basis to gain an under-
standing of the operating history of the circuit and its
Typical normal rating increases range from 3 to 7%.
overall capability. Also, this may help a cable system
Typical 300-hr emergency rating increases range from 4
owner get past a capacity shortfall until another means
to 14%. Short-time emergency rating increases are typi-
of increasing capacity can be achieved.
cally much greater. The increase in the emergency rat-
11-68
11.6 WORKED EXAMPLES

This portion of the chapter contains worked examples using the equations and concepts discussed earlier to demon-
strate for the reader how to perform ampacity calculations.
11.6.1 Example 1 – 230-kV Crosslinked Polyethylene Cable Normal Rating

Ampacity Example for Extruded Dielectric (XD) Cable
The ampacity calculation is for a 230-kV 1200 mm2 segmented copper conductor, with 23 mm XLPE insulation, cop-
per wire shield and copper foil laminate moisture barrier, two circuits installed in a vertical duct bank.
Cable Data:
A = 2368.2 Conductor area, kcmil
ρconductor = 1.7241 x 10-8 Conductor electrical resistivity at 20°C, ohm-m
τconductor = -234.5 Inferred temperature of zero resistance, °C
kS = 0.435 Conductor skin effect factor, dry segmental
kp = 0.6 Conductor proximitty effect factor, dry segmental
Dc = 1.774 Diameter of the segmental conductor, inches
Tc = 90 Maximum allowable normal conductor operating temperature, °C
tcs = 0.063 Thickness of conductor semiconducting shield, in.
ti = 0.906 Insulation wall thickness, in.
ρinsulation = 3.5 Thermal resistivity of insulation, C°-m/Watt
tis = 0.067 Thickness of insulation semiconducting shield, in.
num_wires = 46 Number of wires in stranded copper shield
11-69
wire_diameter = 0.095 Diameter of stranded shield wires, in.
ρwire = 1.7241 x 10-8 Electrical resistivity of wire shield (copper in this case) at 20°C, ohm-mm2/m
τwire = -234.5 Inferred temperature of zero resistance, °C
tsheath = 0.006 Thickness of foil moisture barrier/metallic sheath, in.
ρsheath = 1.7241 x 10-8 Electrical resistivity of metallic sheath (copper in this case) at 20°C, ohm-mm2/m
τsheath = -234.5 Inferred temperature of zero resistance, °C
tj = 0.122 Jacket thickness, in.
ρ jacket = 3.5 Thermal resistivity of jacket, C°-m/Watt
εr = 2.3 Dielectric constant of the insulation
tan δ = 0.001 Dissipation factor of the insulation, numeric
f = 60 Power frequency, Hz
n' = 1 Number of conductors per duct
Ta = 25 Ambient earth temperature, °C
LF = 0.62 24-hour daily loss factor, per unit
E = 230000 System line-to-line voltage, volts
circuits = 2 Number of circuits
N = circuits·3 Number of occupied ducts
Duct Data:
ODduct = 6.625 PVC Duct outside diameter, in.
tduct = 0.28 Thickness of duct wall, in.
IDduct = ODduct-2·tduct
IDduct = 6.065 Duct inside diameter, in.
ρ duct = 6.00 Thermal resistivity of duct/conduit, C°-m/Watt
11-70
Installation Data:
x_loc1 = -10
x_loc2 = -10
x_loc3 = -10
Horizontal locations of conduit centers, in.
x_loc4 = 10
x_loc5 = 10
x_loc6 = 10
burial1 = 65
burial2 = 75
burial3 = 85
Burial depths to conduit centers, in.
burial4 = 65
burial5 = 75
burial6 = 85
In this particular vertical configuration, the middle cable has the highest temperature, so that will be used as the basis
for ampacity calculations.
ρnative = 1.20 Thermal resistivity of the native earth, C°-m/Watt
ρbackfill = 0.60 Thermal resistivity of the duct concrete, C°-m/Watt
Widthcbf = 33 Width of backfill, in.
Heightcbf = 33 Height of backfill, in.
Center_Depthcbf = 75 Depth to center of backfill envelop, in.
Calculate Cable Geometry:
DCS = DC + 2·tCS Diameter over the conductor shield, in.

Dcs = 1.900
Dinsulation = Dcs + 2·ti Diameter over the insulation, in.

Dinsulation = 3.712
Dis = Dinsulation + 2·tis Diameter over the insulation shield, in.

Dis = 3.846
Dms = Dis + 2·wire_diameter Dia. over the metallic shield, in. If a taped shield, this is 2 times the tape
thickness.
Dms = 4.036
Dmss = Dms + 2·tms Dia. over the metallic sheath, in.

Dmss = 4.048
Djacket = Dmss + 2·tj Diameter over jacket, in.

Djacket = 4.292
11-71
Clearance = IDduct – Djacket Clearance in conduit, in.

Clearance = 1.773
Metric Conversion of Variables:

A
Area conductor = Areaconductor = 1200.0 mm2
1.9735
Dconductor = Dc·(25.4) Dconductor = 45.060 mm
Dcond_shield = Dcs·(25.4) Dcond_shield = 48.260 mm
Dinsulation = Dinsulation·(25.4) Dinsulation = 94.285 mm
Dinsl_shield = Dis·(25.4) Dinsl_shield = 97.688 mm
Dmet_shield = Dms·(25.4) Dmet_shield = 102.514 mm
tms = tms·(25.4) tms = 0.152 mm
lay_wires = Dmet_shield·6.7 Length of lay, mm (= 6.7 times diameter of layer)

lay_wires = 686.846 mm
wire_diameter wire_diameter = 2.413 mm

= wire_diameter·(25.4)
Dmet_sheath = Dmss·(25.4) Dmet_sheath = 102.819 mm
Djacket = Djacket·(25.4) Djacket = 109.017 mm
ODduct = ODduct·(25.4) ODduct = 168.275 mm
IDduct = IDduct·(25.4) IDduct = 154.051 mm
Dearth = ODduct earth diameter, mm (jacket OD for direct buried cables)
i = 1..N i = 1..6
⎛ - 254 ⎞ ⎛1651 ⎞
⎜ ⎟ ⎜ ⎟
x_loci = x_loci·(25.4) buriali = buriali·(25.4) ⎜ - 254 ⎟ ⎜1905 ⎟
Widthcbf = Widthcbf·(25.4) ⎜ - 254 ⎟ ⎜ 2159 ⎟
x_loci = ⎜ ⎟ buriali= ⎜ ⎟ mm
Heightcbf = Heightcbf·(25.4) ⎜ 254 ⎟ ⎜1651 ⎟
Center_Depthcbf = Center_Depthcbf·(25.4) ⎜ 254 ⎟ ⎜1905 ⎟
⎜ ⎟ ⎜ ⎟
⎜ 254 ⎟ ⎜ 2159 ⎟
⎝ ⎠ ⎝ ⎠
Calculate the dielectric losses:

2
⎛ E ⎞
2 ⋅π ⋅ f ⋅ε r ⋅ ⎜ ⎟ ⋅ tan δ ⋅10 −9
Wd = ⎝ 3⎠
⎛ D ⎞
18 ⋅1n⎜ insulation ⎟
⎜D ⎟
Wd = 1.268 W/m/phase
11-72
Calculate Cable Thermal Resistances:
ρinsulation ⎛ Dinsl _ shield ⎞

Ri = ·1n ⎜ ⎟
2⋅π ⎝ Dconductor ⎠
R i = 0.431 C°-m/W
ρ jacket ⎛ D jacket ⎞
Rj = ·1n ⎜ ⎟
2⋅π ⎜ D met _ sheath ⎟
⎝ ⎠
R j = 0.033 C°-m/W
U = 5.2 V = 0.91 Y = 0.01 Constants for fiber duct in concrete (Use earthenware duct constants for soap
stone or terracotta ducts)
Estimate of mean temperature in duct, °C (This is an iterative calculation based on rated ampacity. See the end of
this worked example for calculation method.)
Tmduct = 75.002 °C
U
R cable_to_d uct = R cable_to_duct = 0.2723 C°-m/W
1 + 0.1·(V + Y·Tm duct )·(D jacket )
ρ ⎛ OD duct ⎞
R duct = duct ⋅ 1n ⎜⎜ ⎟⎟
2⋅π ⎝ ID duct ⎠
R duct = 0.084 C°-m/W
Calculate External Thermal Resistances

Calculate geometric correction factor for backfill envelope:
x = Widthcbf x = 838.2 y = Heightcbf y = 838.2 y is always the longer dimension of the backfill
⎡ 1 ⎛ x ⎞ ⎛ 4 x ⎞ ⎛ y2 ⎞ ⎤
⎢⎛⎜ ⎞⎟⋅⎜ ⎟⋅⎜ − ⎟⋅1n ⎜1+ ⎟ +1n ( x ) ⎥ Db = 921.455 mm
⎜ ⎟ ⎜ ⎟
⎢⎝ 2 ⎠ ⎝ y ⎠ ⎝ π y ⎠ ⎜ x 2 ⎟ ⎥
Db = e ⎣ ⎝ ⎠ ⎦
Lb = Center_Depthcbf Lb = 1905 mm
⎛ 2·L + 4·L 2 − D 2 ⎞
G b = 1n⎜ b b b ⎟ Gb = 2.098 (dimensionless)
⎜ Db ⎟
⎝ ⎠
Calculate diameter Dx
6.71·10 4 mm 2
α soil' = Thermal diffusivity of native soil α soil ' = 1456.725
(ρ native ·100) 0.8 hour
D x = 1.02· α soil' ·24 Diameter beyond which 24 hour average Dx = 190.719

losses apply, mm
11-73
The self and mutual earth thermal resistances each consist of two terms. The first term assumes
that the concrete resistivity applies everywhere, and the second term corrects for the excess of the
native earth resistivity over the concrete resistivity. The calculation assumes that the middle cable in the vertical posi-
tion is the hottest cable.
⎡ ⎡ 2·burial + 4(burial ) 2 − D 2 ⎤⎤
ρ backfill ⎢ ⎛ D x ⎞ ⎥ ⎥ + LF· (ρ native − ρ backfill ) ⋅ n '⋅G
Rearth = ⋅ 1n ⎜⎜ ⎟⎟ + LF·1n ⎢ 2 2 x
b
2·π ⎢ ⎝ OD duct ⎠ ⎢ Dx ⎥⎥ 2·π
⎣⎢ ⎣ ⎦ ⎦⎥
Rearth = 0.3544
Earth thermal resistance with backfill correction
for AC losses (at load factor of losses)
⎡ ⎡ 2·burial + 4(burial ) 2 − D 2 ⎤ ⎤
ρ backfill ⎢ ⎛ D x ⎞ x ⎥ ⎥ (ρ native − ρ backfill )
Rearth' = ⋅ 1n ⎜⎜ ⎟⎟ + 1n ⎢ 2 2
+ ⋅ n '⋅G b
2·π ⎢ ⎝ OD duct ⎠ ⎢ Dx ⎥⎥ 2·π
⎣⎢ ⎣ ⎦ ⎦⎥
Rearth' = 0.5644
Earth thermal resistance with backfill correction
for dielectric losses (100% loss factor)
Mutual thermal resistance among cable phases, plus correction for backfill
Distances from Cable #2 to images of other cables, mm
d' 21 = (x _ loc 2 − x _ loc 1 ) 2 + (burial 2 + burial 1 ) 2 d'21 = 3556.0
Distances from Cable #2 to other cables, mm
d 21 = (x _ loc 2 − x _ loc 1 ) 2 + (burial 2 − burial 1 ) 2 d21 = 254.0
⎛ d' ⎞ ⎛ d' ⎞ ⎛ d' ⎞ ⎛ d ' 25 ⎞ ⎛ d ' 26 ⎞

Ffactor = ⎜⎜ 21 ⎟⎟ ⋅ ⎜⎜ 23 ⎟⎟ ⋅ ⎜⎜ 24 ⎟⎟ ⋅ ⎜⎜ ⎟⎟ ⋅ ⎜⎜ ⎟⎟ Ffactor = 77297.9
⎝ d 21 ⎠ ⎝ d 23 ⎠ ⎝ d 24 ⎠ ⎝ d 25 ⎠ ⎝ d 26 ⎠
11-74
ρ ⎛ρ − ρ backfill ⎞
R mutual = backfill ⋅ n '⋅LF ⋅ ln(Ffactor ) + LF⎜⎜ native ⎟ ⋅ n '⋅( N − 1) ⋅ G b
⎟
2⋅π ⎝ 2 ⋅ π ⎠
R mutual = 1.287 C° - m/Watt
ρ ⎛ρ − ρ backfill ⎞
R mutual' = backfill ⋅ n '·1n (Ffactor ) + ⎜⎜ native ⎟ ⋅ n '⋅( N − 1) ⋅ G b
⎟
2⋅π ⎝ 2 ⋅ π ⎠
R mutual' = 2.076 C° - m/Watt
Sp = d21 minimum phase spacing, mm Sp = 254 mm

Calculate Conductor Resistance:
ρ conductor
R dc 20 = Rdc20 = 14.368 x 10-6 ohm/meter at 20°C
Area conductor ⋅ 10 − 6
Assume 2.5% Stranding of Conductor
Rdc20 = 14.727 x 10-6 ohm/meter at 20°C
R dc20 = R dc20 ⋅ 1.025
⎛ Tc − τ conductor ⎞
R dcT = R dc 20 ·⎜⎜ ⎟⎟ RdcT = 18.777 x 10-6 ohm/meter at 90°C
⎝ 20 − τ conductor ⎠
Calculate conductor skin effect factor:
2
⎛ 8·π·f ·(k S )·10 − 7 ⎞
XS = ⎜
4 ⎟ X S 4 = 12.204
⎜ R dcT ⎟
⎝ ⎠
XS4
YCS = YCS = 0.060
4
192 + 0.8 ⋅ X S
Calculate conductor proximity effect factor:
2
⎛ 8·π·f ·(k P )·10 − 7 ⎞
XP4 = ⎜ ⎟ X P 4 = 23.218
⎜ R dcT ⎟
⎝ ⎠
⎡ ⎤
2 ⎢ 2 ⎥
⎛ 4 ⎞ ⎛ ⎞ ⎢ ⎛ ⎞ ⎥
YCP = ⎜
XP ⎟·⎜ conductor ⎟ ·⎢0.312·⎜ D conductor
D ⎟ +
1.18
⎥
⎜ 192 + 0.8 ⋅ X 4 ⎟ ⎜ Sp ⎟ ⎜ Sp ⎟ ⎛ 4 ⎞
⎝ P ⎠⎝ ⎠ ⎢ ⎝ ⎠ ⎜ X P ⎟ + 0.27 ⎥
⎢ ⎜ 192 + 0.8 ⋅ X ⎟
4 ⎥
⎢⎣ ⎝ P ⎠ ⎥⎦
YCP= 0.008
Rac = RdcT·(1 + YCS + YCP)
Rac = 20.116 x 10-6 ohm/meter
11-75
Calculate the shield losses:

The bonding scheme for this case is cross-bonded so there is no circulating current. Eddy current
losses for the foil shield are included in the calculations. For the purposes of calculations
(for this and all cases), the temperature of a metallic insulation shield and the metallic sheath
will be assumed equal.
Wire Shield
This particular case does not include a taped metallic shield (wire shield instead), so the shield does
not have an impact on the calculated eddy current loss increments. If a taped shield is present, the
taped shield and metallic sheath (if present) must be considered together when calculating eddy
current losses. If only a taped shield is present with no metallic sheath, then only the taped shield is
considered for eddy current losses. If only a wire shield is present with no metallic sheath, then
there are no eddy current losses.
num _ wires·π·(wire _ diameter) 2

Area sheild = Areashield = 210.360 mm2
4
2
⎡ (D met _ shield + D insl _ shield ) ⎤
ρ wire ⎢ π· ⎥
R shield 20 = · 1+ ⎢ 2 ⎥
Area shield ⋅ 10 −6 ⎢ lay _ wires ⎥ Rshield20 = 90.142x10-6 ohm/meter
⎢⎣ ⎥⎦
Continuous Metallic Sheath (assumes foil laminate is lapped 12.7 mm (0.5 in):
Area sheath = π·
[(D met_sheath )2 − (D met_shield )2 ]
+ t ms ·0.5 ⋅ 25.4 Areasheath = 51.090 mm2
4
ρ sheath
R sheath 20 = Rsheath20 = 337.462x10-6 ohm/meter
Area sheath ⋅ 10 − 6
Calculate Shield/Sheath Resistances, adjusted for temperature:
Ts=79.419°C
Ts - τ wire
R shieldT = R shield20 . RsheathT = 111.188x10-6 ohm/meter
20 - τ wire
Ts - τ sheath
R sheathT = R sheath20 . RsheathT = 416.251x10-6 ohm/meter
20 - τ sheath
1
Rs =
1 1 Rs = 87.749x10-6 ohm/meter
+
R shieldT R sheathT
Eddy current loss increment:
2·π·f ·10 −7 m = 0.091

m=
R sheathT
11-76
⎡ ⎡⎛ D ⎞⎤
2⎤
met _ sheath + D met _ shield
⎛ m ⎞ ⎢ ⎢⎜⎜
2 ⎟⎥
⎟
⎥
YSe0 = 6 ⋅ ⎜ ⎟. ⎢ 2 ⎥
⎜ 1 + m 2 ⎟ ⎢ ⎢⎢⎝ ⎠⎥ YSe0 = 1.994 x 10-3
⎥ ⎥
⎝ ⎠⎢ 2 ⋅ Sp ⎥
⎢ ⎥⎦
⎣⎣ ⎦
1.4 ⋅ m + 0.7
⎡⎛ D met _ sheath + D met _ shield ⎞⎤
⎢⎜ ⎟⎥
YSe1 = 0.86 ⋅ (m )3.08 ⎢⎜⎝ 2 ⎟
⎠⎥ YSe1 = 1.405 x 10-4
⎢ ⎥
⎢⎣ 2 ⋅ Sp ⎥⎦
8 ⋅ π 2 ·f
X ec = Xec = 165.764
⎛ ρ sheath ⎞ 7
⎜⎜ ⎟⎟ ⋅ 10
⎝ 10 6 ⎠
⎡ ⎡ ⎤
1.74 ⎤
⎢ ⎢ ⎥ ⎥
⎢ ⎢ t ms ⎥ ⎡ ⎛ D met_sheath + D met_shield ⎞ −3 ⎤⎥
g s = ⎢1 + ⎢ ⎥ ·⎢X ec ·⎜ ⎟·10 − 1.6⎥ ⎥ gs = 1.000
⎢ ⎢ ⎛ D met_sheath + D met_shield ⎜ ⎟
⎞⎥ ⎢⎣ ⎝ 2 ⎠ ⎥⎦ ⎥
⎢ ⎢ ⎜⎜ ⎟
⎟⎥ ⎥
⎢ ⎣⎝ 2 ⎠⎦ ⎥
⎣ ⎦
⎛R ⎞⎡ (X .t )4 ⎤
⎟⎟.⎢g s .YSe0 .(1 + YSe1 ) + ec ms ⎥
YEC = ⎜⎜ sheathT Yec = 0.041
⎝ R ac ⎠ ⎣⎢ 12 × 1012 ⎦⎥
Since this cable is cross-bonded, no net circulating current is assumed.
YCC = 0
Calculate AC Resistance Including Metallic Shield/Sheath Losses:
Q S = 1 + YEC + YCC QS =1.041
Temperature rise from dielectric heating:

⎛ Ri ⎞
ΔTd = Wd .⎜⎜ + R j + R cable _ to _ duct + R duct + R earth ' + R mutual ' ⎟⎟ ΔTd = 4.116 C°
⎝ 2 ⎠
Allowable conductor temperature rise:
ΔTc = Tc – Ta – Td ΔTc = 60.884 C°

Summation of electrical and thermal resistances:
( (
RacΣQ i R thermal i = R ac · R i + Q s · R j + R cable_to_duct + R duct + R earth + R mutual ))
RacΣQ i R thermal i = 51.212x10 -6 C°/ampere 2
11-77
ΔTC
I rated = Irated = 1090.3
RacΣQ i R thermali
Itotal = 2.Irated Itotal = 2180.7
Tc = 90.000
⎛ W ⎞
Ts = Tc − ⎜ I rated 2 .R ac + d ⎟.R i Ts = 79.419
⎝ 2 ⎠
( ⎛
) ⎞
R cable _ to _ duct Calculate temperature to middle of
Tm duct = Ts _ calc − I rated 2 ⋅R ac ⋅Q s + Wd .⎜ R j + ⎟ air space in duct, °C
⎜ 2 ⎟ Tmduct = 75.002
⎝ ⎠
( )(
Tearth _ int erface = Ts _ calc − I rated 2 . R ac ⋅ Q s + Wd . R j + R cable _ to _ duct + R duct ) Tearth_interface = 69.232
11.6.2 Example 2 – 230-kV Crosslinked Polyethylene Cable Emergency Rating

Using the data from Example 1, calculate a 100-hour emergency for the cable system.
Calculate Thermal Capacitances for Emergency Ratings
Specific heats of various cable materials used in this cable construction:
Cp conductor = 3.45·10 6 copper conductor, J/C°-m3
Cp insulation = 2.40·10 6 XLPE insulation, J/C°-m3
Cp shield_sheath = 3.45·10 6 copper wires and sheath, J/C°-m3
PE jacket, J/C°-m3
Cp jacket = 2.400·10 6
PVC conduit, J/C°-m3
Cp conduit = 1.70·10 6
Area conductor C conductor = 4140.0

C conductor = .Cp conductor J/C°-m
6
10
C insulation =
Cp insulation π
10 6 4
[( )
. . D insl _ shield 2 − (D conductor )2 ] C insulation = 14161.0 J/C°-m
2
π·(wire_diameter )2 Cρshield _ sheath ⎛ π·D met_shield ⎞
Cshield _ sheath = num _ wires· · · 1 + ⎜⎜ ⎟
⎟ C shield _ sheath = 971.1 J/C°-m
4 6
10 ⎝ lay _ wires ⎠
+
Cpshield _ sheath π
106
[( ) (
. . D met _ sheath 2 − D met _ shield 2
4
)]
C jacket =
Cp jacket π
10 6 4
[( )
. . D jacket 2 − (D met _ sheath ) 2 ] C jacket = 2474.7 J/C°-m
11-78
C conduit =
Cp conduit π
10 6 4
[
. . (OD duct )2 − (ID duct ) 2 ] C conduit = 6121.5 J/C°-m
Calculate thermal diffusivity of soil (convert from mm2/hour to m2/second)
6.71⋅104 mm 2
δsoil ' = Thermal diffusivity of native soil
( ρnative .100 )0.8 hour
2 mm 2
1 1 ⎛ 1 ⎞
δsoil = δ soil ' . . .⎜ ⎟ δ soil = 4.046 x10 −7 sec ond
60 60 ⎝ 1000 ⎠
Check duration of emergency to see if “long duration” or “short duration”
R check = R i + R j = 0.464
C check = C conductor + C insulation + C shield _ sheath + C jacket

= 21746.9
If emergency duration is greater than this time, then

R .C t check = 3360.9 seconds
t check = check check it’s considered “long duration”. Otherwise, it's
3 t check “short duration”.
= 56.015 minutes
60
Calculate Long Duration Emergency and Transient Temperature Response Constants
RA = Ri R A = 0.431
R B = Q s .(R j + R cable_to_duct + R duct ) R B = 0.405
1 1
p= − p = 0.381
⎛D ⎞ 2
2·ln⎜⎜ insulation ⎟⎟ ⎛⎜ D insulation ⎞⎟ − 1
⎝ Dconductor ⎠ ⎜⎝ D conductor ⎟⎠
CA = Cconductor + p·Cinsulation C A = 9538.1
1
Cshield_sheath + C jacket + ·Cconduit
C B = (1 − p) · C insulation + 2
C B = 15011.6
Qs
1
[ ( )
M = · C A · R A + R B + C B ·R B
2
] M = 7030.6 s
N = C A ⋅ R A ·C B ·R B N = 2.501 x107 s2
M + M2 − N
a= a = 4.786x10-4 1/s
N
11-79
M − M2 − N
b= b = 8.353x10-5 1/s
N
Ra =
1 ⎡ 1
·⎢
a − b ⎣ CA
( ⎤
− b· RA + RB ⎥ ) R a = 0.08 C°-m/W
⎦
R b = R A + R B − Ra R b = 0.748 C°-m/W
Define function to calculate exponential integrals: -Ei(x) = f(x), Ei(x) = -(f(x))
⎡
Ei _ max(x ) ≡ − ⎢− 0.5772156649 − In(− x ) − ∑
∞
(− 1)n ·(- x )n ⎤ Function to calculate exponential integrals.
⎥
⎣⎢ n =1
n·n! ⎦⎥
Calculate Long Duration Emergency (100 hour) for 105°C

t = 100·60·60 emergency duration (seconds) Current is normalized (1 ampere)
i = 0,1..N – 1 Calculate the temperature rise of each cable
105 − τ conductor
Rac emergency = (R ac )(Q s ) Racemergency = 21.914x10-6 Ohm/m
90 − τ conductor
I100 = 1232.5 amperes Irated = 1090.3 amperes
(assumed emergency rating value to give 105°C (from early “normal” rating calculation)
after 100 hours)
Wi = I1002·Racemergency - Irated2·Rac Qs Wi = 8.386 – Calculate additional heat added by emergency,
W/m
Wc = I1002·Racemergency - Irated2·Rac Wc = 9.373 W/m
Cable transient temperature rise:
[ ( ) (
Δ Tcable = Wc R a · 1 - e -a·t + R b · 1 - e − b ⋅ t )] ΔTcable = 7.839 C
Cable surface attainment factor:
ΔTcable
αt = αt = 1.000
(
Wc · R A + R B )
11-80
Calculate the earth transient temperature rise:

⎡ ⎡ ⎛D ⎡ ⎛ d ' ⎞2 ⎤ ⎤
⎞ ⎤ ⎡ ⎛ burial ⎞2 ⎤ ⎡ ⎛ d ⎞2 ⎤
2
⎢ ⎢ − ⎜ earth ⎟ ⎥ ⎢−⎜ 2
⎟ ⎥ ⎢ − ⎜ 21 ⎟ ⎥ ⎢ − ⎜ 21 ⎟ ⎥ ⎥
⎢ ⎢ ⎝ 1000 ⎠ ⎥ − − Ei ⎢ ⎝ 1000 ⎠ ⎥ + −Ei ⎢ ⎝ 1000 ⎠ ⎥ − −Ei ⎢ ⎝ 1000 ⎠ ⎥ ... ⎥
⎢ − Ei
⎢ 16·t·δsoil ⎥ ⎢ t·δsoil ⎥ ⎢ 4·t·δsoil ⎥ ⎢ 4·t·δsoil ⎥ ⎥
⎢ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎥
⎢ ⎢⎣ ⎥⎦ ⎣ ⎦ ⎣ ⎦ ⎣ ⎦ ⎥
⎢ ⎥
⎢ ⎡ ⎛ d ⎞ ⎤ 2 ⎡ ⎛ d' ⎞ ⎤ ⎡ 2 ⎡ ⎛ d ⎞ ⎤ 2 ⎡ ⎛ d ' ⎞ ⎤ ⎤⎥
2
⎢ ⎢−⎜ 23
⎥ ⎢−⎜ 23
⎥ ⎢ ⎢−⎜ 24
⎟ ⎥ ⎢−⎜ 24
⎟ ⎥ ⎥⎥
ρ 1000 ⎟⎠ ⎥ 1000 ⎟⎠ ⎥ ⎢ 1000 ⎠ ⎥ 1000 ⎠ ⎥ ⎥ ⎥
= native i ⎢ + − Ei ⎢ ⎝ − − Ei ⎢ ⎝ + − Ei ⎢ ⎝ − − Ei ⎢ ⎝
·W
ΔTearth ...
4·π ⎢ ⎢ 4·t·δsoil ⎥ ⎢ 4·t·δsoil ⎥ ⎢ ⎢ 4·t·δsoil ⎥ ⎢ 4·t·δsoil ⎥ ⎥ ⎥
⎢ ⎢ ⎥ ⎢ ⎥ ⎢ ⎢ ⎥ ⎢ ⎥ ⎥⎥
⎢ ⎢⎣ ⎥⎦ ⎢⎣ ⎥⎦ ⎣⎢ ⎣ ⎦ ⎣ ⎦ ⎦⎥ ⎥
⎢ ⎥
⎢ ⎡ ⎛ d ⎞ ⎤ 2 ⎡ ⎛ d' ⎞ ⎤ 2 ⎡ ⎡ ⎛ d ⎞ ⎤ 2 ⎡ ⎛ d ' ⎞ ⎤ ⎤⎥
2
⎢ ⎢ ⎜ − 25
⎟ ⎥ ⎢ ⎜ − 25
⎟ ⎥ ⎢ ⎢ ⎜− 26
⎟ ⎥ ⎢ ⎜− 26
⎟ ⎥ ⎥
1000 ⎠ ⎥ 1000 ⎠ ⎥ ⎢ 1000 ⎠ ⎥ 1000 ⎠ ⎥ ⎥ ⎥
⎢ + − Ei ⎢ ⎝ − − Ei ⎢ ⎝ + −Ei ⎢ ⎝ − −Ei ⎢ ⎝ ... ⎥
⎢ ⎢ 4·t·δsoil ⎥ ⎢ 4·t·δsoil ⎥ ⎢ ⎢ 4·t·δsoil ⎥ ⎢ 4·t·δsoil ⎥ ⎥ ⎥
⎢ ⎢ ⎥ ⎢ ⎥ ⎢ ⎢ ⎥ ⎢ ⎥ ⎥⎥
⎣ ⎢⎣ ⎥⎦ ⎢⎣ ⎥⎦ ⎣⎢ ⎢⎣ ⎥⎦ ⎢⎣ ⎥⎦ ⎦⎥ ⎦
ΔTtransient = ΔTcable + αt·ΔTearth ΔTtransient = 15.0 C°
Tcemergency = Tc+ΔTtransient Tcemergency = 105.0 °C I100 = 1235.7
This is the 100-hour emergency rating. Value is determined by

iterating I100 until the emergency conductor temperature is at
rated emergency temperature (in this case, 105°C).
11.6.3 Example 3 – Calculation of Shield/Sheath Fault Current Capability

Calculate the fault current capability for a 15-cycle fault at 60Hz for the copper wire shield
and copper foil laminate sheath
Several values are used from the ampacity calculation:
Shield/sheath temperature prior to the fault: Ts = 79.419 °C
Since the shield/sheath are in contact with cross-linked polyethylene (Tmax=250°C) and a medium density polyethylene
jacket (Tmax = 150°C), the fault current temperature will be limited to the lesser of these values (e.g., 150°C).
MDPE Jacket maximum allowable temperature: T2 = 150°C
Shield/sheath is Copper: k = 50237 T = τ sheath T = -234.5°C
timefault = 0.25 seconds
Area = Areashield + Areasheath
Calculate fault current:
k ⎛T −T⎞
I fault = Area· .In⎜ 2 ⎟ I fault
= 52.8 kA (thousands of amperes)
time fault ⎝ Ts − T ⎠ 1000
11-81
11.6.4 Example 4 – 345-kV Pipe-Type Cable Normal Rating
Pipe-Type Cable Ampacity Worked Example

Cables are 345-kV HPFF cables with 2500-kcmil segmental copper conductors, 600
mils laminated paper-polypropylene insulation, 0.1 x 0.2, 2 x 3 in. lay stainless steel
skid wires in a nominal 8-in. cable pipe, 2 circuits, 0.62 loss factor.
Cable Data:
A = 2500 Conductor area, kcmil
ρconductor =1.7241x10-8 Electrical resistivity, copper conductor, ohm-m at 20°C
τconductor = -234.5 Inferred temperature of zero resistance, °C
ks = 0.39 Conductor skin effect factor, in oil
kp = 0.46 Conductor proximitty effect factor, in oil, cradled
Dc = 1.824 Diameter of the segmental conductor, inches
Tc = 85 Maximum normal conductor operating temperature, °C
tcs = 0.005 Thickness of conductor semiconducting shield, inches
Values for LPP Insulation
ti = 0.600 Insulation wall thickness, inches
εr = 2.7 Dielectric constant of the insulation
tan δ = 0.0007 Dissipation factor of the insulation, numeric
ρinsulation = 6.00 Thermal resistivity of the insulation, C°-m/Watt
tis = 0.005 Thickness of insulation shield tapes, inches

ρshield = 70x10-8 Electrical resistivity of stainless steel shield tapes, ohm-m at 20°C
Skid Wire Size/Type 0.1in. x 0.2in., 2 x 3in. lay
ρskidwire = 70x10-8 Electrical resistivity of stainless steel skid wire, ohm-m at 20°C
f = 60 Hz
LF = 0.62 Daily (24-hour) loss factor, per unit (entered below)
n=3 Number of cables within pipe or conduit
N=2
11-82
Pipe Data:
Pipe is HPFF
ODpipe = 8.625 Pipe outside diameter, inches
IDpipe = ODpipe -2(0.25) IDpipe = 8.125 Pipe inner diameter, inches
tcoating = 0.07 Pipe coating (Somastic) thickness, inches
Thermal resistivity of the pipe coating, C°-m/W
ρcoating = 3.5
Installation Data:
x1 = –15 x2 = 15 Horizontal location of pipe center, inches
burial1 = 42 burial2 = 42 Burial depth to pipe center, inches
Ta = 25 Ambient earth temperature, °C
f = 60 Power frequency, Hz
E = 345000 x 1.05 System maximum line to line voltage, volts
Thermal resistivity of the native earth, C°-m/W
ρ native = 0.9
Thermal resistivity of the duct concrete,
ρ backfill = 0.5 C°-m/W
Widthcbf = 50.5 Heightcbf=26.5 Width and height of backfill, inches
Center_Depthcbf = 39 Value for center-line depth of cbf is arbitrary
Calculate Cable Geometry:
Dcs = Dc + 2tcs Diameter over the conductor shield, inches
Dinsulation = Dcs + 2ti Diameter over the insulation, inches
Dis = Dinsulation + 2tis Diameter over the insulation shield, inches
Dskidwire = Dis + 1.5 x 0.1 Diameter over 0.1" skid wires, inches
(1.5 factor assumes skid wires “nest”)
Dearth = ODpipe + 2tcoating
(ID − D skidwire ) ⎡ ⎛
Clearance in pipes, inches
0.5
ID pipe D skidwire ⎞⎤
⎢1 − ⎜⎜ ⎟⎥
pipe
Clearance = − 1.366 D skidwire +
⎟
2 2 ⎣⎢ ⎝ ID pipe − D skidwire ⎠⎦⎥
Dcs = 1.834
Dinsulation = 3.034
Dis = 3.044
Dskidwire = 3.194
Dearth = 8.765
Clearance = 1.163
11-83
Metric Conversion of Variables
A
Area conductor = Areaconductor = 1266.8 mm2
1.9735
Dconductor = Dc(25.4) Dconductor = 46.33 mm
Dcond_shield = Dcs(25.4) Dcond_shield = 46.584 mm
Dinsulation = Dinsulation(25.4) Dinsulation = 77.064 mm
Dinsl_shield = Dis(25.4) Dinsl_shield = 77.318 mm
Dskidwire = Dskidwire(25.4) Dskidwire = 81.128 mm
ODpipe = ODpipe(25.4) ODpipe = 219.075 mm
IDpipe = IDpipe(25.4) IDpipe = 206.375 mm
Dearth = Dearth(25.4) Dearth = 222.631 mm
burial1 = buriall(25.4) burial2 = burial2(25.4) burial1 = burial2 = 1066.8 mm
x1 = x1(25.4) x2 = x2(25.4) x1 = -381 mm; x2 = 381 mm
Widthcbf = Widthcbf(25.4) Widthcbf = 1282.7 mm
Heightcbf = Heightcbf(25.4) Heightcbf = 673.1 mm
Center_Depthcbf = Center_Depthcbf(25.4) Center_Depthcbf = 990.6 mm
Calculate the dielectric losses:

2
⎛ E ⎞
2πfε r ⎜⎜ ⎟⎟ tan δ10 −9
⎝ 3⎠
Wd =
⎛ D ⎞
18 ln⎜ insulation ⎟
⎜ D cond _ shield ⎟
⎝ ⎠ Wd = 3.44 W/m/phase
Calculate Conductor Resistance:
ρ conductor
R dc 20 =
Area conductor ⋅ 10 − 6
⎡ T − τ conductor ⎤
R dcT = R dc 20 ⎢ c ⎥ Rdc20 = 1.3610 x 10-5 ohms/meter
⎣ 20 − τ conductor ⎦
RdcT = RdcT1.025 Assume 2.5% Stranding of Conductor, Ohms/meter
2 2
⎛ 8πf (k P )10 − 7 ⎞ ⎛ 8πf (k P )10 − 7 ⎞
XS = ⎜
4 ⎟ XP =⎜
4 ⎟
⎜ R dcT ⎟ ⎜ R dcT ⎟
⎝ ⎠ ⎝ ⎠
X S 4 = 11.277 X P 4 = 15.688
Xs4
YCS = S = Dskidwire YCS = 0.056
192 + 0.8 ⋅ X s 4
11-84
⎡ ⎤
⎢ ⎥
⎛ 4 ⎞ 2 ⎢ 2 ⎥
YCP = ⎜
XP ⎟·⎛⎜ D conductor ⎞⎟ ·⎢0.312·⎛⎜ D conductor ⎞
⎟ +
1.18
⎥ YCP=0.088
⎜ 192 + 0.8 ⋅ X 4 ⎟ ⎝ S ⎠ ⎢ ⎝ S ⎠ ⎛ XP4 ⎞
⎝ P ⎠ ⎜ ⎟ + 0.27 ⎥
⎢ ⎜ 192 + 0.8 ⋅ X 4 ⎟ ⎥
⎢⎣ ⎝ P ⎠ ⎥⎦
Rac = RdcT[1+1.5(YCS+YCP)] Rac=21.29x10-6
Calculate the shield and skid wire loss increments:
Area of shield tape is width of tape times thickness. Resistance of shield tape is the area
times the helical length times the resistivity divided by the area. Shield tape thickness is
0.005 in. stainless steel, with 1/8 in. lapped, with typical tape width of 7/8 in. Assume there
are two tapes.
7 1
widthshield_tape = (25.4) thicknessshield_tape = 0.005x25.4 lapshield = (25.4)
8 8
(
area shield tape = thickness shield tape width shield tape)( ) layshield = widthshield_tape – lapshield
2
ρ shield ⎛ π ⋅ D insl _ shield ⎞
R shield = 1 + ⎜⎜ ⎟
⎟
area shield tape ⋅ 10 − 6 ⎝ lay shield ⎠
Rshield = 3.172 x 10-6 ohm/meter
Skid wire resistance:
Area of ellipse is pi * major_radius * minor_radius. Area of skid wire is half this. Skid
wire is 0.1 x 0.2, 3-inch lay, 2 wires. Material is stainless steel.
25.4 area sw = π (minor radius )(major radius )

minor_radius = 0.1x25.4 major_radius = 0.2 2
2
layskidwire = 3x25.4
2
⎛ π ⋅ D skidwire ⎞
Length of a helix is 1 + ⎜⎜ ⎟⎟
⎝ lay skidwire ⎠
2
ρ skidwire ⎛ πD skidwire ⎞
R skidwire = 1 + ⎜⎜ ⎟⎟
area sw ⋅ 10 − 6 ⎝ lay skidwire ⎠ Rskidwire = 0.241 x 10-6 ohm/meter
1
Rs =
1 1 1 1
+ + + Rs = 0.1121 x 10-6 ohm/meter
R skidwire R skidwire R shield R shield
Calculate Mutual Reactance (assume cradled configuration):
⎛ 2.3D insl _ shield ⎞

X m = 2(2 πf )10 −7 ln⎜⎜ ⎟⎟ Xm = 5.917 x 10-5 ohm
⎝ Dskidwire ⎠
11-85
Rs 1
YCC =
R ac 2
⎛R ⎞ YCC = 5.263 x 10-3
1 + ⎜⎜ s ⎟⎟
⎝ Xm ⎠
Q S = 1 + 1.5(YCC ) QS=1.008
Calculate the pipe loss increments to AC resistance:
YPL =
0.0438(D skidwire ) + 0.0226 ID pipe ( )
YPL = 0.469
R dcT 10 6
Q P = 1 + 1.5(YCC ) + YPL QP=1.477
Calculate Cable Thermal Resistances
ρ insulation ⎛ D insl _ shield ⎞

Ri = ln⎜⎜ ⎟⎟ R i = 0.489 C°-m/W
2π ⎝ D conductor ⎠
U = 0.26 V=0 Y = 0.0026 Constants for HPFF
Tmoil = 68.919 Estimate of mean temperature in duct, °C. (This is an iterative
calculation based on rated ampacity. See the end of this worked
example for calculation method.)
n⋅U
R oil = R oil = 0.196 C°-m/W
(
1 + 0.1(V + Y ⋅ Tm oil ) D insl _ shield ⋅ 2.15 )
ρcoating ⎛ D ⎞
R pipe _ coating = n ln⎜ earth ⎟
2π ⎜ OD ⎟ R pipe _ coating = 0.027 C°-m/W
⎝ pipe ⎠
Calculate External Thermal Resistances:

Calculate geometric correction factor for backfill envelope:
x = Heightcbf x = short dimension of backfill y = Widthcbf y = long dimension of backfill
⎡ 1 ⎛ x ⎞⎛ 4 x ⎞ ⎛ y 2 ⎞ ⎤
⎢⎛⎜ ⎞⎟⎜ ⎟⎜ − ⎟ ln ⎜1+ ⎟+ ln( x ) ⎥
⎢⎝ 2 ⎠⎜⎝ y ⎟⎠⎜⎝ π y ⎟⎠ ⎜ x 2 ⎟ ⎥ Db = 909.533 mm
Db = e⎣ ⎝ ⎠ ⎦
Lb = Center_Depthcbf Lb = 990.6 mm
⎛ 2L + 4 L 2 − D 2 ⎞
⎜ b b b ⎟
=
G b ln⎜ ⎟
Gb = 1.414 (dimensionless)
⎜ Db ⎟
⎝ ⎠
Calculate diameter Dx
6.71x10 4 mm 2
α soil = Thermal diffusivity of native soil αsoil = 1.834 x 103 mm2/hour
(ρ native 100)0.8 hour
11-86
D x = 1.02 α soil 24 Diameter beyond which 24 hour average losses apply

Dx = 213.978 mm
⎛ ⎛ 2 ⋅ burial + 4 ⋅ burial 2 − D 2 ⎞ ⎞ Earth thermal resistance for AC losses at loss
ρ ⎜ ⎛ Dx ⎞ ⎟ ⎟ factor
R earth = backfill n ⋅ ⎜ ln⎜⎜ ⎟⎟ + LF ln⎜⎜ 1 1 x
⎟⎟ ⎟
2π ⎜ ⎝ Dearth ⎠ ⎜ Dx ⎟
⎝ ⎝ ⎠ ⎠ R earth = 0.433 C° − m/W
⎛ ⎛ 2 ⋅ burial + 4 ⋅ burial 2 − D 2 ⎞⎞ Earth thermal resistance for dielectric losses at

ρ ⎜ ⎛ Dx ⎞ ⎟⎟
R earth ' = backfill n ⋅ ⎜ ln⎜⎜ ⎟⎟ + ln⎜ 1 l x
⎟⎟ ⎟ 100% loss factor
2π ⎜⎜
⎜ ⎝ Dearth ⎠ Dx ⎟
⎝ ⎝ ⎠⎠
R earth' = 0.704 C° - m/W
Mutual thermal resistance between two pipes (just for normal ampacities)
Distance between one cable pipe and image of
d'12 = [(x1-x2)2 + (burial1 + burial2)2]0.5 other cable pipe
d’12=2266 mm
Distance between one cable and the other
d12 = [(x1-x2)2 + (burial1 - burial2)2]0.5
d12=762 mm
d'
F12 = 12 F12=2.973
d12
ρ
R mutual = backfill n ⋅ LF ⋅ ln(F12 ) R mutual = 0.161 C°-m/W
2π
ρ
R mutual' = backfill n ⋅ ln(F12 ) R mutual' = 0.260 C°-m/W
2π
Thermal Resistance correction for native soil for
⎛ρ −ρ ⎞
R correction = LF⎜ native backfill ⎟n ⋅ N ⋅ G b AC losses
⎝ 2π ⎠ R correction = 0.335 C°-m/W
Thermal resistance correction for native soil for
(ρ native − ρ backfill ) dielectric losses
R correction ' = n ⋅ N⋅Gb
2π R correction' = 0.54 C°-m/W
Calculate Normal Ampacity on primary cable
( (
RacΣQ i Rthermal i = R ac R i + Q S R oil + Q P R pipe_coating + R earth + R mutual + R correction ))
RacΣQ i Rthermali = 4.469 x10 −5 C°/A2
⎛R ⎞
ΔTd = Wd ⎜ i + R oil + R pipe _ coating + R earth ' + R mutual' + R correction ' ⎟ ΔTd = 6.784 C°
⎜ 2 ⎟
⎝ ⎠
ΔT = Tc- ΔTd-Ta ΔT = 53.216 C°
ΔT
I rated = Irated = 1091.2 A Itotal = 2 · Irated Itotal = 2182.4 A
RacΣQ i Rthermali
[ ( (
Tc = I rated 2 R ac R i + Q S ⋅ R oil + Q P ⋅ R pipe_coating + R earth + R mutual + R correction ))]+ ΔTd + Ta
Tc= 85 °C
11-87
Ri
Tshield = Tc − I rated 2 R ac R i − Wd Tshield = 71.761 °C
2
(
Tmoil = Tshield − I rated 2 R ac ⋅ Qs + Wd ) R2oil
Tmoil = 68.919 °C
(
Tearth _ int erface = Toil − I rated 2 R acc ⋅ Qs + Wd ) R2 − ( I
oil
rated
2
)
R acc ⋅ Q p + Wd R pipe _ coating Tearth_interface = 64.977 °C
Temperature Rise Due to Charging Current:
E −9
2 ⋅π ⋅ f ⋅ ε r 10
I charging = 3 Icharging = 0.023 Amperes per meter
⎛ D ⎞
18 ln⎜ insulation ⎟
⎜D ⎟
For a 16km (10-mile) circuit, the total charging current would be:
Icharging(16000) = 375.9 Amperes
WIc_max = (Icharging16000)2RaccQp WIc_max = 4.444 Watts per meter
⎡ ⎛ ⎛ R pipe coating + R earth + ⎞ ⎞⎤

( )
ΔTIc max = I ch arg ing (16000 )⎢R ac ⎜ R i + Q S ⋅ R oil + Q P ⋅ ⎜
⎢ ⎜ ⎜R
⎟ ⎟⎥
⎟ ⎟⎥
ΔTIc_max = 6.316 C°
⎣ ⎝ ⎝ mutual + R correction ⎠ ⎠⎦
ΔT − ΔTIc max
I rated =
RacΣQ i Rthermali
IratedIc = 1024.4 A
Alternate Approach to Calculating Allowable Real Current
(
I real = I rated 2 − I ch arg ing 16000 2 ) Ireal = 1024.4 A
11-88
11.7 NOMENCLATURE
Symbol Description Units

a absorptivity —
a thermal diffusivity in.2/hr
a, b internal cable time constants based on the R-C network l/s
A difference between maximum and minimum air C°
temperatures
A free area within the tunnel m2
AF transient attainment factor dimensionless
Area cross-sectional area of metal mm2
C capacitance F/m
Ci thermal capacitance of material, i J/C°-m
CP air specific heat kJ/kg-C°
Cpn volumetric specific heat of material n J/C°-m3

D1 height of the skid wire mm
D2 width of the skid wire mm
Db circumscribing circle diameter for special backfill mm
d’c outer diameter of equivalent solid conductor of an mm
equivalent hollow-core, solid conductor
Dcable outer diameter of the cable when one cable is installed in mm
the conduit, or is equal to 2.15 x outer diameter of the
cable when three cores are installed inside a common
conduit or pipe
Dcond_shield diameter over the conductor shield mm
DConductor outside diameter of conductor mm
De diameter of cable, conduit, or pipe to environment mm
Dearth diameter over conduit, pipe, or cable jackets mm
di inner diameter of a hollow-core conductor mm
Dinsulation diameter over the insulation mm
d1N distance from cable 1 to cable n mm
d'1N distance from cable 1 to image of cable n mm
DInner inner diameter of thermal resistance layer mm
DLayer average diameter of the helical layer mm
dphase axial spacing between adjacent cables mm
DOuter outer diameter of thermal resistance layer mm
DS diameter over cable sheath mm
DSkid wire diameter over cable skid wires mm
DSM mean diameter of the metallic shield, sheath, and/or skid mm
wire
Dstrand diameter of a single-wire shield strand mm
Dx diameter beyond which the effect of average daily losses mm
applies
ΔT allowable conductor (or other) temperature rise C°
ΔTC conductor temperature rise due to current-dependent C°
losses
11-89

ΔTd conductor temperature rise due to dielectric losses C°
ΔTIc temperature rise due to charging current losses C°
ΔTint temperature rise due to extraneous heat source or C°

interference temperature rise
ΔTsurface − to − ambient cable, conduit, or pipe surface temperature rise above C°
ambient air
Δz incremental distance over which mutual heating effect is m
assessed; “N” such increments are evaluated from 1 to
“v”
E constant from Table 11-16 —
“Ei” notation for exponential integral —
f power frequency Hz
F geometric factor representing mutual heating between —
multiple cables
Fint mutual heating factor between the external heat source dimensionless
and cable
g constant from Table 11-16 —
Gb geometric correction factor for special backfill dimensionless
gS calculation coefficient —
H Total heat quantity to raise temperatures of both metals kg-J/C°-m3
1 C°
h1 computational parameter mm
I normal ampacity amperes
IC charging current amperes/m
Ic max maximum charging current amperes
IDPipe inner diameter of the cable pipe mm
Ii hourly load values amperes
Iint conductor current in other interfering heat sources amperes
Imax maximum hourly load value during the cycle amperes
Isc maximum short-circuit current amperes
It transient ampacity amperes
k constant for conductor or shield/sheath material dimensionless
kP conductor proximity effect factor —
kS conductor skin effect factor —
L depth from surface to center of cable, conduit, or pipe mm
Lb depth from surface to center of special backfill envelop mm
LayLayer distance along the cable for one turn of the tape or wire mm
LF 24-hour loss factor for the load on the cables —
LFfc forced cooled loss factor dimensionless
LFint loss factor of other heat source —
Lh burial depth of crossing cable m
m calculation coefficient —
• air mass flow rate kg/s
m
11-90

n number of cables within the conduit, pipe or earth —
diameter
N number of shield wires, shield tapes, or skid wires; also, —
number of cables, conduit, or pipes in a group
nint number of cables at the location of the interfering heat —
source
q heat removed from forced circulation W
Qi the ratio of losses in the “ith” layer plus the ratio of —
losses in all inner layers (including the conductor) to
that of the losses of only the conductor
QP ratio of losses in conductor, shield, and pipe to those —
just in the conductor
QS ratio of losses in conductor and shield to those just in
the conductor
Qsun solar heat input W/m2
R any thermal resistivity C°-m/W
R1 , R2 ac resistances of metals 1 and 2 ohm/m

thermal resistances from the equivalent R-C network C°-m/W
Ra , Rb
during the time period being evaluated
Rac electrical resistance ohm/m
Rac ac conductor resistance, including skin and proximity ohm/m
effects
Rac conductor ac resistance at Tmax ohm/m
Rac-int ac resistance of other heat source ohm/m
R cable − to − duct thermal resistance between cable surface (jacket) and C°-m/W
conduit (duct)
R cable − to − pipe thermal resistance between cable surface and pipe C°-m/W
R correction correction factor for controlled backfill or concrete– C°-m/W

encased duct
RdcT resistance at temperature, T ohm/m
R duct duct thermal resistance C°-m/W
R earth Thermal resistance of earth plus mutual heating (may C°-m/W

include surface transfer)
R earth earth thermal resistance C°-m/W
R insulation insulation thermal resistance C°-m/W
R int thermal resistance between an external heat source and C°-m/W

a cable
R Jacket thermal resistance of jacket C°-m/W
R mutual thermal resistance to account for multiple cables within C°-m/W

the trench
11-91
R pipe coating pipe-covering thermal resistance C°-m/W
RS total shield/sheath/skid wire resistance, or resistance of ohm/m

layer for eddy current losses only
R source, earth thermal resistance to ambient earth for external heat C°-m/W
source
R Thermal thermal resistance C°-m/W
R total total thermal resistance from conductor to ambient for C°-m/W

constant load
S axial spacing between nearest phases mm
Sc circuit spacing mm
Sh horizontal displacement of other cables grouped with m
the crossing cables
SP axial spacing between cable, conduit or pipe centers mm
t time since maximum air temperature occurred hr
t time of short circuit seconds
t time (duration) of the response seconds
tShield thickness of the shield or sheath layer for eddy current mm.
loss calculations
T temperature of the conductor °C
T temperature °C
t0 length of period hr
T1 pre-fault temperature of metal °C
T2 allowable short-circuit temperature of the insulation °C
material
Ta ambient earth temperature °C
Tambient ambient soil temperature “remote” from any heat °C
sources
TAverage average annual air temperature °C
tan δ insulation dissipation factor
TC allowable conductor temperature °C
Tmax final conductor temperature °C
Tmean mean temperature of the duct air, or nitrogen gas or °C
dielectric liquid in the pipe
Tφ initial conductor temperature °C
TOUT , TOUT air temperature leaving and entering the tunnel °C

tShield thickness of the shield layer mm
Tsource surface temperature of external heat source °C
T(t-1) shield/sheath temperature from last time step °C
U numerical constant from Table 11-12 —
v incremental counter for evaluating the temperatures —
V numerical constant from Table 11-12 —
Wconductor ac losses generated in the conductor W/m
Wd insulation dielectric losses W/m
11-92

Wdint dielectric loss in other heat source W/m
Wf final watt loss W/m
Wh heat output from cable, conduit, or pipe W/m
Wi losses developed in the conductor W/m
Wi heat output of a particular cable, pipe, or conduit W/m
including shield and sheath losses
WIc max heat output from charging current W/m
Wo initial watt loss W/m
WPipe pipe losses W/m
Wshield ac losses generated in the shield/sheath W/m
Wsun heat absorbed by cable from solar radiation W/m
x depth below the earth's surface in.
x short dimension of the backfill envelope mm
Xm mutual reactance ohm/m
y long dimension of the backfill envelope mm
Y numerical constant from Table 11-12 —
YCC increment in ac losses for shield/sheath circulating —
currents
YEC increment in ac losses for shield/sheath eddy currents —
YPL increment in ac losses for pipe heating effects —
Yse0 calculation coefficient —
Yse1 calculation coefficient —
Z constant from Table 11-16 —
11-93
Greek Symbols
α soil thermal diffusivity of soil mm2/hour
β crossing angle of heat source °

β heat transfer coefficient at surface C°-m2/W
β volumetric coefficient of thermal expansion, dimensionless
δ thermal diffusivity m2/s
ε emissivity of the cable, conduit, or pipe surface —
εo = permittivity of free space = 8.854 x 10-12 F/m
εr = relative permittivity of the insulation (specific inductive dimensionsless
capacitance, dielectric constant)
γ axial thermal resistances of conductor C°-m/W
ρ air density kg/m3
ρ electrical resistivity of metal ohm-m
ρ thermal resistivity C°-m/W
ρ backfill backfill envelope thermal resistivity C°-m/W
ρ native native soil thermal resistivity C°-m/W
ρ shield electrical resistivity of shield at operating temperature ohm/m
τ inferred temperature of zero resistance °C

υ air flow velocity within the tunnel m/s
11-94
REFERENCES CIGRE. 1986. Working Group 08 of CIGRE Study

Committee. “Forced Cooled Cables—Calculation of
AIEE. 1953. “Symposium on Temperature Rise of Thermal Transient and Cyclic Loads.” ELECTRA. No.
Cables, AIEE Committee Report.” AIEE Transactions 104.
on Power Apparatus and Systems. Vol. 72. June. pp. 530–
561. (This symposium contained four papers on cable CIGRE. 1987. Working Group 08 of CIGRE Study
ampacity.) Committee 21. “The Calculation of Continuous Rating
for Forced-Cooled High-Pressure Oil-Filled Pipe-Type
Anders, G. J. 1997. Rating of Electric Power Cables. Cables.” ELECTRA. No. 113. July. pp. 97–121.
McGraw-Hill. New York.
CIGRE. 2003. Technical Brochure 218. Gas Insulated
ANSI/IEEE. 1986. Std. 738-1986. “IEEE Standard for Lines (GIL). February.
Calculation of Bare Overhead Conductor Temperature
and Ampacity Under Steady-State Conditions.” CIGRE. 2005. ELECTRA. No. 220. Technical Bro-
chure 272. “Large Cross Sections and Composite
ASTM. 2000. Standard Test Methods for Laboratory Screens.” CIGRE Working Group B1.03. June.
Compaction Characteristics of Soil Using Standard
Effort (12,400 ft-lbf/ft3 (600 kN-m/m3). ASTM Stan- El-Kady, M. A., and D. J. Horrocks. 1985. “Extended
dard D 698. Values for Geometric Factor of External Thermal
Resistance of Cables in Duct Banks.” IEEE Transac-
Buller, F. H., and J. H. Neher. 1953. “The Thermal tions on Power Apparatus and Systems. Vol. PAS-104,
Resistance Between Cables and a Surrounding Pipe or No. 8.
Duct.” AIEE Transactions on Power Apparatus and Sys-
tems. Vol. 72. pp. 1260–1275. Engelhardt, J. 1972. “Ampacity Considerations of
HPOF Joints.” Proceedings of the 1972 IEEE Under-
Burghardt, R. H. Matthews, et al. 1983. “Full Scale ground Transmission Conference. Pittsburgh, PA.
Forced Cooling Tests at Waltz Mill.” IEEE Paper, 83
WM, 009-8. EPRI. 1977. Backfill Materials for Underground Power
Cables. Phase I. EPRI Report no. EL-0506.
Chato, J. C. 1977. Free and Forced Convection Cooling
of Pipe-Type Electric Cable Systems. August . EPRI EPRI. 1982. Determination of AC Conductor and Pipe
EL-147. Loss in Pipe-Type Cable Systems. EPRI Report no. EL-
2256.
Chato, J. C., J. M. Crowley, et al. 1980. Free and Forced
Convective Cooling of Pipe-Type Electric Cables. EPRI EPRI. 1984. Designer’s Handbook for Forced-Cooled
Report no. EL-1872. Vol. 1. High-Pressure Oil-Filled Pipe-Type Cable Systems. July.
EPRI Report no. EL-3624.
CIGRE. 1983. Working Group 02 of CIGRE Study
Committee 21. “Computer Method for the Response of EPRI. 2003. Increased Power Flow Guidebook-Under-
Single-core Cables to a Step Function Thermal Tran- ground Cables. Report No. 1001818. December.
sient.” ELECTRA. No. 87.
Fink, L. H. 1972. “The Role and Practice of Cable Rat-
CIGRE. 1985a. Working Group 02 of CIGRE Study ings.” IEEE Proc. Underground Trans. Conf. Supple-
Committee 21. “The Calculation of the Effective Exter- ment. pp. 63-67.
nal Thermal Resistance of Cables Laid in Materials
Having Different Thermal Resistivities.” No. 98. Janu- Furukawa Electric Co. 1969. Sumitomo Electric Indus-
ary. tries Ltd. “Increase of Thermal Resistivity of the Insula-
tion Layer of Forced Cooled 275 kV LOF, HPOF
CIGRE. 1985b. “Calculation of the Continuous Rating Cables.” Technical Report. September.
of Single Core, Rigid Type, Compressed Gas Insulated
Cables in Still Air with No Solar Radiation.” ELEC- G&W Electric Specialty Co. 1982. High Ampacity Ter-
TRA. No. 100. pp. 66-75. minations. EPRI Report no. EL-2233.
11-95
Glicksman, L. R., J. V. Sanders, et al. 1978. “Heat Con- Martin, Jr., M. A., R.A. Bush, et al. 1981. “Practical
duction in the Cable Insulation Forced-Cooled Under- Aspects of Applying Soil Thermal Stability Measure-
ground Transmission Lines.” IEEE Transactions on ments to the Rating of Underground, Power Cables.”
Power Apparatus and Systems. Vol. 97. No. 1. January– IEEE Transactions on Power Apparatus and Systems.
February. pp. 134–139. Vol. 100. No. 9. September. pp. 4236–4249.
Hartley, J. G., and W. Z. Black. 1979. “Predicting Ther- Mattews, L., and P. Malburg. 1977. “Calculation of
mal Stability and Transient Response of Soils Adjacent Transient Temperatures and Pressures in Oil Filled
to Underground Power Cables.” 7th IEEE/PES Trans- Cables.” IEEE Summer Power Meeting. July.
mission & Distribution Conference and Exposition.
April 1–6. Morris, M., and R.W. Burrell. 1954. “Current-Carrying
Capacity of Pipe-Type Systems under Steady-State and
IEC. 2002. International Electrotechnical Commission. Transient Cyclical Loading Conditions.” AIEE Trans-
“Electric Cables—Calculation of the Current Rating – actions on Power Apparatus and Systems. June. pp. 650–
Part 1-3: Current Rating Equations (100% Load Fac- 660.
tor) and Calculation of Losses – Current Sharing
between Parallel Single-core Cables and Calculation of Neher, J. H. 1949. “The Temperature Rise of Buried
Circulating Current Losses.” Publication 60287. Parts Cables and Pipes.” AIEE Transactions on Power Appa-
1-3. ratus and Systems. Vol. 68. pp. 9–21.
IEC. 1989. International Electrotechnical Commission. Neher, J. H., and M. H. McGrath. 1957. “The Calcula-
“Calculation of the Cyclic and Emergency Current Rat- tion of Temperature Rise and Load Capability of Cable
ing of Cables, Part 2, Cyclic Rating of Cables Greater Systems.” AIEE Transactions on Power Apparatus and
Than 18/30 (36) kV and Emergency Ratings for Cables Systems. Vol. 76. October.
of All Voltages.” Publication 60853. Part 1 (1989). Parts
2-3 (2002). Nelson, R. J., T. F. Brennan, and J. S. Engelhardt 1988.
“The Application of Real-Time Monitoring and Rating
IEEE. 1990. “Loss Evaluation for Underground Trans- to HPOF Pipe Cable Systems.” IEEE Proc. 1988 Sum-
mission and Distribution Cable Systems.” IEEE Trans- mer Meeting.
actions on Power Delivery. Vol. 5. No. 4. November. pp.
1652–1659. Numerical Recipes. 1986. Section 5.5. Cambridge: W.H.
Press. Cambridge University Press.
IPCEA. 1976. “Ampacities Including Effect of Shield
Losses for Single-Conductor Solid-Dielectric Power Radhakrishna, H. S., and J. E. Steinmanis. 1981. “Ther-
Cable, 15-kV through 69-kV.” IPCEA Publication P-53- mal Resistivity Survey of Lake Erie Sediments for the
426. May. Ontario Hydro-GPU Interconnection.” Symposium on
Underground Cable Thermal Backfill. September 17 &
Katz, C., et al. 1978. “Progress in the Determination of 18. pp. 96–109.
AC/DC Resistance Ratios of Pipe-Type Cable Systems.”
IEEE Transactions on Power Apparatus and Systems. Saleeby, K. E., W. Z. Black, and J. G. Hartley. 1979.
Vol. PAS-97. No. 6. “Effective Thermal Resistivity for Power Cables Buried
in Thermal Backfill.” IEEE Transactions on Power
Koci, P. F., L. R. Glicksman, et al. 1977. “Prediction of Apparatus and Systems. Vol. 98. No. 6. Novem-
the Friction Factor in Laminar Flow Through Three- ber/December.
Phase Pipe-Type Power, Transmission Line.” IEEE
Transactions on Power Apparatus and Systems. Vol. 96. Sandiford, P. 1981. “Cable Backfill Materials— State-
No. 4. July–August. pp. 1305–1310. of-the-Art.” Symposium on Underground Cable Ther-
mal Backfill. September 17 & 18. pp. 3-9.
Lee, G., R. Patton, J. W. McCourt, and J. Engelhardt
1981. “Progress in the Field Validation of the Cable Shannon, W. L., and W. A. Wells. 1947. “Tests for Ther-
Monitoring and Rating System.” IEEE Trans. P.A.S. mal Diffusivity of Granular Materials.” ASTM Proce-
Vol. 100. No.3. March. dures. Vol. 47. Pp. 1044–1055.
11-96
Slutz, R., L. R. Glicksman, et al. 1975. “Measurement Wantanabe, T., N. Yoshida, et al. 1975. “Forced-Cool-
of Fluid Flow Resistance for Forced Cooled Under- ing Tests 500-kV Pipe-Type Oil-Filled Cable.” Fujikura
ground Transmission Lines.” IEEE Transactions on Technical Review. No. 7.
Power Apparatus and Systems. Vol. 94. No. 5. Septem-
ber–October. pp. 1831–1834. Weedy, B. M. 1988. “Thermal Design of Underground
Systems.” Chapter 8. Chichester, England and New
Soulsby, D. R., and A. J. Donovan. 1981. “The Effect of York, NY: John Wiley & Sons, Ltd..
Backfill of the Temperature Distribution in a Buried
Circuit.” Symposium on Underground Cable Thermal Williams, Jr., J. L., A. Zanona, and R. J. Nease. 1971.
Backfill. September 17 & 18. pp. 167-180. “Forced-Cooling Tests on a 138 kV Pipe Cable Circuit.”
IEEE Transactions on Power Apparatus and Systems
Steinmanis, J. E. 1981. “Thermal Property Measure- Vol. 90. No.3. May—June. pp. 1246–55.
ments Using a Thermal Probe.” Symposium on Under-
ground Cable Thermal Backfill. September 17 & 18. Williams, J. L. et al. 1976a. “Comprehensive Forced-
Cooling Tests on Pipe-Type Cables at Waltz Mill.”
Thomann, G. C., T. Aabo, E. C. Bascom, III, R. Ghafu- CIGRE 21-07.
rian, and T. M. McKernan. 1991. “A Fourier Transform
Technique for Calculating Cable and Pipe Temperatures Williams, J. A., E. D. Eich, and T. Aabo. 1976b. “Forced
for Periodic and Transient Conditions.” IEEE PES Cooling Tests on 230 kV and 345 kV HPOF Cable Sys-
Winter Meeting. New York, NY. Paper 91 WM 248-5. tems.” IEEE Power Engineering Society Winter Meet-
ing, New York, NY. January 25-30. Published by IEEE
VanWormer, F. C. 1955. “An Improved Approximate (Cat. no. 76, CH1075-1).
Technique for Calculating Cable Temperature Tran-
sients.” AIEE Transactions on Power Apparatus and Sys- Williams, G. P., and L.W. Gold. 2003. “Ground Tem-
tems. April. peratures.” Canadian Building Digest. CBD-180. Octo-
ber.
Walldorf, S., A. Ernst, and S. Pancholi 1999. “A Cost
Effective Solution for Increasing Ratings on Under- Wollaston, F. O. 1949. “Transient Temperature Phe-
ground Transmission Cables - A Case Study.” IEEE. nomena of 3-Conductor Cables.” AIEE Transactions on
Proc. 1999 T&D Conference. Power Apparatus and Systems. Vol. 68.
Wantanabe, T. 1974. “Several Fundamental Problems

on Forced-Cooled Techniques.” Technical Brochure.
Tokyo Electric Power Company, Inc.
11-97
CHAPTER 12 Installation Design

Authors: Dennis E. Johnson, POWER Engineers, Inc.
Deepak Parmar, Geotherm, Inc.
Reviewer: John S. Rector, Black & Veatch
Design engineering for an underground transmission cable system has been called a busi-
ness of details. No single activity is especially complex, but long-term, reliable cable oper-
ation requires attention to a number of details at every stage of system design.
This chapter addresses many of the common design considerations for designing an
underground cable circuit: cable type selection, route considerations, surveying require-
ments, soil thermal properties, mechanical considerations, installation modes, and design
calculations.
Related information may be found in other chapters. Chapter 2 describes cable system
types and their common applications. Chapter 11 covers ampacity principles and calcula-
tion procedures. Chapter 13 provides detailed descriptions of installation procedures.
Chapter 7 describes special application cables, such as submarine, gas-insulated transmis-
sion line, and superconductors. Chapter 10 discusses grounding of underground cable
systems and cathodic protection.
Dennis Johnson is a Senior Project Engineer with POWER Engineers,

Inc. (POWER). He received a BSEE degree from Brigham Young Uni-
versity (1985). He joined Arizona Public Service in 1986, where he ini-
tially was responsible for distribution system design. He later worked
for Black & Veatch from 1988 to 2001 as a design and project engineer
on numerous substation, overhead, and underground transmission
projects at voltages ranging from 69 to 345 kV. Mr. Johnson has worked
as a Senior Project Engineer in POWER’s underground transmission
design group since 2001. Mr. Johnson is a Member of the IEEE Power Engineering Soci-
ety and a Voting Member of the IEEE Insulated Conductors Committee (ICC). He is a
member of various ICC subcommittees that are developing guides and standards for
high-voltage underground cable systems. He is a registered professional engineer in the
states of California, Connecticut, Iowa, Kansas, Kentucky, Nevada, New York, Pennsyl-
vania, Texas, Vermont, and Virginia.
Deepak Parmar is President of Geotherm Inc. He holds a B.Sc. in civil

engineering from Woolwich Polytechnic (1966, London, U.K.) and a
Diploma in management studies (1972, Slough Polytechnic, U.K.).
During undergraduate studies, Mr. Parmar worked at Soil Mechanics
Ltd. in London. From 1973 to 1978, he worked with two civil engineering
consulting firms in Canada, specializing in geotechnical and material test-
ing, and instrumentation of tunnels and shafts. In 1978, Mr. Parmar
founded Geotherm Inc., specializing in underground cable-related soil
and backfill testing. Under contract with Ontario Hydro Research for two years, Mr. Par-
mar worked on EPRI-funded projects for the design and development of the Thermal
12-1
Chapter 12: Installation Design EPRI Underground Transmission Systems Reference Book
Property Analyzer and thermal aspects of soil and conducted numerous seminars and courses on under-
backfills. Since 1980, Mr. Parmar and Geotherm have ground T&D cable ampacity. Mr. Parmar is a member
undertaken numerous research contracts for Canadian of the IEEE Power Engineering Society and Insulated
government agencies as well as for electric utilities and Conductors Committee, Canadian Society for Civil
cable manufacturers in Canada, North America, and Engineers, Canadian Geotechnical Society, Tunnelling
overseas. Mr. Parmar and Geotherm have performed Association of Canada, Canadian Electrical Associa-
more than 400 route thermal survey projects for under- tion, and CIGRÉ. He has authored and coauthored
ground and submarine cable crossings. Mr. Parmar was numerous papers on the application of thermal parame-
a member of the team of experts retained to investigate ters of soils and backfills, and was the principal author
the cable failures in Auckland, New Zealand, and he has for the IEEE Guide for Soil Thermal Resistivity Testing.
12-2
EPRI Underground Transmission Systems Reference Book Chapter 12: Installation Design
12.1 INTRODUCTION one time in city streets with trench openings of only
Design of an underground electric transmission cable 300–600 ft (90–180 m). The cable is installed in a sepa-
system installation involves anticipating a host of issues rate operation, with minimal traffic disruption.
related to the system’s planning, construction, and oper- Suburban
ation. While every installation will inevitably confront In suburban locations, extruded-dielectric, pipe-type,
site-specific issues, there are, nevertheless, a number of and SCFF cable can be used, depending upon the spe-
common factors that can be expected with most system cific application. Careful route-specific technical analy-
installation designs. sis and costing are required to determine the best cable
type for the application.
This chapter describes the common challenges of instal-
lation design. It covers three primary areas. Section 12.2 Rural
reviews typical design considerations—including selec- Rural installations are, by their nature, well-suited for
tion of the cable type, choosing the best route, surveying most cable systems. Low traffic volume and long trench
practices, transitions to overhead systems, soil thermal openings allow flexibility for the designer to consider all
properties, mechanical considerations during cable the different types of cable systems.
installation, and monitoring systems.
Special Cases
Long underwater crossings often dictate use of SCFF or
Section 12.3 describes the main modes of installation,
extruded-dielectric cables, primarily because it is diffi-
which are trenching (directly buried or in conduit),
cult to make pipe-type splices underwater. SCFF or
trenchless installations, bridges, and underwater instal-
extruded-dielectric cables are generally preferred for
lations.
bridge crossings, because the weight and expansion
characteristics of pipe-type cable systems require resolu-
Section 12.4 briefly explains the series of calculations
tion of the bending forces, which tends to complicate the
that must be performed, principally related to cable
design. Short sections of cable requiring high power
pulling tension and sidewall pressure.
transfer, such as substation ties, are often best served
with compressed-gas-insulated cable systems. For short
12.2 COMMON DESIGN CONSIDERATIONS underground dips, extruded-dielectric cables are better
suited. Generally, pipe-type cable systems are not eco-
12.2.1 Cable-Type Selection nomical in short lengths.
Often the primary procedure in installation design is to
decide on which type of transmission cable system to 12.2.2 Route Considerations
use. Pipe-type cables have been the most commonly A utility should investigate several alternative routes for
used cable system at higher voltages in North America an underground cable. The optimal route can save sub-
until recently. Now, with the advances in purity of stantial time and project cost, and in many U.S. states,
extruded-dielectric cables and the general reluctance to the siting commission requires that at least two routes
install fluid-filled systems in some parts of the United be investigated. Overhead routes are typically not the
States, extruded-dielectric cable systems are becoming best route for an underground cable.
much more common at increasingly higher voltage lev-
els and longer circuit lengths. XLPE-insulated cables are In general, the shortest route is typically the least expen-
being designed and installed at voltages up to 230 kV sive. In urban and suburban environments, the allowable
and 345 kV in long lengths. Self-contained fluid-filled routes are usually limited to existing thoroughfares, and
(SCFF) cables are generally only used in specialized existing rights-of-way. Frequently, the cable route may be
extra-high-voltage applications. Conditions on individ- the same as the route determined suitable for an overhead
ual utility systems vary, which results in different cable line. Detailed route considerations are discussed below.
types being a better choice on one system than another.
Additional discussion on cable system selection is pro- Environmental
vided in Chapter 2. A primary routing consideration is to avoid activity,
where possible, in or near the following types of environ-
Urban mentally sensitive areas. If an environmentally sensitive
Extruded-dielectric cables installed in duct banks and area cannot be avoided, consideration should be made
pipe-type cables are frequently used in urban environ- to limit or mitigate the impact.
ments in the United States because of their ruggedness
and the ability to install short lengths of pipe or duct at
12-3
Wetlands Magnetic field readings directly over a cable are gener-

Wetlands provide habitat for many plant and wildlife ally higher than directly under an overhead transmis-
species in addition to providing a method for replenish- sion line, depending on the type of cable system and the
ing the earth’s reserve of fresh water. Three characteris- overhead conductor arrangement, generally because the
tics define a wetland: (1) standing water, (2) aquatic cables are closer to the measuring device. However, the
vegetation, and (3) hydric (or hydraulic) soils. If any of magnetic field strength falls off faster as you move away
these three characteristics are present, a high probability from the centerline of the cables, as compared to an
exists that this location will be designated a wetland overhead line. Similar to overhead transmission lines,
area. In North America, these types of areas are feder- magnetic fields produced by cables systems can be
ally protected and require a special permit for cable sys- reduced by various mitigation methods. Chapter 16 dis-
tem crossings. cusses various mitigation methods.
Wetlands should be avoided if at all possible by choosing Land Use

an alternate route. Depending on the type of wetland, When routing an underground line, it is important to
open trenching may be used, but more likely other meth- know the type of area that the line will be traversing. The
ods of installation such as horizontal directional drilling type of land use often determines the amount of ease-
may need to be considered when crossing these areas. ment or right-of-way available for the underground line.
Aesthetics Underground lines are frequently located in existing

One of the main reasons for selecting an underground roadway rights-of-way. Typically, no easement is
cable system is the aesthetic benefit. The line is placed required to install the underground line in public rights-
underground, so the public cannot see the transmission of-way; however, the owner of the road usually reserves
line until it exits the ground at the substation or at an the right to have the utility relocate the underground
overhead-to-underground transition structure. Aesthet- line if a future conflict occurs. For this reason, some
ics should be considered when locating and designing utilities prefer to install underground circuits within a
transition structures. dedicated easement adjacent and parallel to the public
Archeological/Historical right-of-way, and to accept the added cost.
Construction activity in and around known archeologi- Urban
cal sites is typically regulated in the United States by the Urban areas are becoming more and more congested
state through a State Historical Preservation Office with traffic and underground utilities. This makes the
(SHPO). If the SHPO knows or suspects that an archeo- installation of a new underground transmission line dif-
logical site exists at or near the site of the planned facil- ficult. When choosing routes in urban areas for new cir-
ity, further investigation may be required. An cuits, extreme care is required to locate the existing
archeologist may need to be engaged to deal with the underground facilities. The typical location for a new
process of working in or around these sites. underground circuit in an urban area is within the road
right-of-way. There is usually very little undeveloped
If an area is archeologically sensitive, the best option is land available that could be used for installing an under-
to reroute the line. There are no construction methods ground line. Major thoroughfares should be avoided
that will limit the impact to an archeologically sensi- because of the large amount of traffic that would have
tive area. to be controlled. The designer should be aware that a
Electromagnetic Fields significant cost of installing circuits in urban locations is
Electromagnetic Fields (EMF) may be a significant fac- traffic control.
tor in the regulatory and environmental process for site Suburban
selection. Even though there is no federal standard Suburban areas, like urban areas, are becoming con-
defining allowable EMF levels, local jurisdictions may gested with traffic and construction activities. Schools,
have restrictions on EMF levels related to the siting of churches, and homes will likely be located along the
power lines near schools, daycare centers, children’s route selected through suburban areas, requiring addi-
playgrounds, and residential areas. tional safety considerations during construction. These
areas should be avoided, if possible. During construc-
As with overhead lines, underground lines generate tion, the entire road may need to be closed to provide
magnetic fields. However, because the cable shielding sufficient working space for the installation of the
contains the electric field within the cable, no external underground cable system.
electric field is generated.
12-4
Rural underground cable deeper underground, which results

Rural areas are generally easier locations in which to in reduced ampacity.
construct underground lines because they usually have
fewer existing underground utilities; however, they also Other Utilities
lend themselves to overhead transmission lines more In many areas, especially well-established cities, there
easily than suburban and urban areas. are usually extensive underground utilities already in
place. Avoiding these utilities can be costly and time-
Construction and Maintenance Considerations consuming. In addition, if the cable must frequently dip
Regardless of whether the cable is installed in urban, beneath other utilities, pulling lengths between man-
suburban, or rural locations, the cable route must be holes may need to be decreased, and ampacity may
patrolled routinely, and access maintained for quick decrease because of the greater burial depth. Although
repairs in the event of a failure. An open space of it is generally not possible to avoid these utilities com-
15–20 ft (4.5–6 m) is required for vehicle access along pletely, some routes may have fewer subsurface obstruc-
the route (Blau 1975). During construction, an additions than others. Where this cannot be avoided, it may
tional temporary easement of approximately 30–50 ft be desirable to install the cable system shallower, above
(9–15 m) is needed to provide sufficient space for the the existing utilities.
construction equipment along the route. Figures 12-1
and 12-2 show typical access requirements. Heat Sources
Routes having other heat sources, such as multiple dis-
Recently Paved Streets tribution circuits or steam mains, should be avoided.
Metropolitan areas frequently have moratoriums on These are usually found in downtown urban locations.
new construction within newly paved or resurfaced The designer should try to maintain at least 12 ft (3.6 m)
roadways. Sometimes it is possible, through long-range clearance from these other heat sources to avoid
planning and coordination with the local officials, to decreasing the cable’s ampacity or installing a larger
install cable pipes or ducts in advance of when the cable conductor. Thermal barriers and heat pipes have been
is actually needed, prior to repaving activities. Where used in special cases, although at a significant added
this is not possible, selection of another route is often cost (Iwata et al. 1991). Each situation should be evalu-
the best option. ated individually.
Obstacles Traffic Control
Obstacles such as rivers, major highways, and railroads Traffic control is becoming a significant consideration
should be avoided where possible. Crossing these obsta- for street construction in all parts of the country. Most
cles adds significantly to permitting requirements, con- jurisdictions require that a complete traffic control plan,
struction cost, and installation time. Additionally, on a block-by-block basis, be provided during the per-
crossing these obstacles usually requires installing the mitting process. Busy thoroughfares may require elabo-
rate traffic control measures, including severe
restrictions on the hours during which construction may
take place. Smaller streets usually have fewer restrictions
and may even allow closure for short periods of time.
The selected route must provide access for construction
material and equipment both during the installation and
after, should there be a cable failure.
Figure 12-1 Typical and minimum access requirements for Permitting

urban and suburban areas. Longer underground cable routes frequently pass
through several different jurisdictions, adding to the
complexity of obtaining permits. There are some routes,
even in metropolitan areas, where construction is dis-
couraged or prohibited because of traffic congestion,
tourist attractions, etc. Zoning classifications affect the
allowable construction hours and noise levels, and dif-
ferent routes may have varying degrees of public opposi-
tion. Areas with potential archeological and historical
Figure 12-2 Typical and minimum access requirements for significance should be avoided wherever possible.
unpaved areas.
12-5
Soil Types • All sewer, water, communication, electric, and gas

Trenching costs vary significantly as a function of soil manholes and valve covers.
type, amount of pavement, amount of rock, amount of • Elevations to water/gas valve nuts or top of pipe, size
fill material, passage through old landfill areas, and
of pipes.
amount of contaminated soil (e.g., near old gasoline
stations with leaky underground tanks). Some soils • Culverts and invert elevations on open pipes.
require more shoring than others, and extensive dewa-
tering may cause subsidence in some soil types. In addition, as part of the planimetric and topographi-
Although there may not be many alternatives, the cal survey, the following steps should be taken:
designer should consider avoiding difficult trenching • Ground elevations should be shot across the width of
areas. Detailed geotechnical evaluation is required in the surveyed area to allow a 3-D Triangular Irregular
the early stages of project design. Network (TIN) model (file) to be generated. This will
allow the route to be relocated and a new profile to be
Soil Thermal Resistivity
easily developed.
Soil thermal resistivity can also vary widely along differ-
ent routes. Initial soil thermal studies should be made at • The local underground locating service should be
the same time as the geotechnical evaluation. The cost notified to have all existing underground facilities
of corrective backfill must be considered when evaluat- located prior to the survey commencing, so this infor-
ing different routes. A more detailed discussion of this mation can also be documented by the surveyor.
subject is included in Section 12.2.6. • Efforts should be made to locate other unmapped,
abandoned, and private underground facilities and
12.2.3 Surveying structures (fuel tanks, private lines [gas, electric, com-
An accurate route survey, showing both above-grade munication, and water], old foundations, etc.).
and below-grade structures and features, is a very • Contact all the local underground utilities and obtain
important design element. The survey can be either their facility drawings. This is a crucial step in verify-
land- or aerial-based; however, land-based surveys gen- ing the accuracy of the survey, and identifying data
erally provide better detail of the surface features and missing from the survey.
the below-grade obstacles that the designer will need to
know. Aerial-based photography and surveys provide a Underground Locating
good overall picture, but lack the detail and accuracy In addition to the field markings, in highly congested
considered necessary for an underground design. areas, it may be necessary to field-locate existing under-
ground facilities to determine actual location and
Level of Survey depth. The American Society of Civil Engineers
For an underground line survey, it is important to get the (ASCE) has developed a guide for collecting existing
most detailed planimetric and topographical survey pos- subsurface utility data (ASCE 2003). The ASCE guide
sible. Even though a route centerline may have already has established four levels of data to determine the
been determined, it is important to make certain that accuracy of the underground facility locations. These
sufficient area on either side of the centerline is surveyed levels are as follows:
to identify open areas in case the cable system must be
rerouted during the detail design and construction stage. • Utility Quality Level A—Precise horizontal and
Typically, a distance of about 50 ft (15 m) on either side vertical location of utilities by the actual exposure
of the route centerline is sufficient for this purpose. and subsequent measurement of the existing subsur-
face facility.
Pre-Design Survey
• Utility Quality Level B—Information obtained
The planimetric and topographical survey should
through the application of appropriate surface geo-
include the following information:
physical methods to determine the existence of sub-
• All property lines and street/utility rights-of-way. surface utilities. Methods such as electromagnetic,
• Curb lines, street lanes, manholes, streetlights, fire magnetic, elastic wave, and other high-cost special-
hydrants, traffic lights, driveways, and all pertinent ized methods.
physical information. • Utility Quality Level C—Information obtained by
• Location of storm sewer inlets including widths, surveying and plotting visible above-ground utility
depths, and elevations. features.
• Utility Quality Level D—Information derived from
existing records or oral recollections.
12-6
At a minimum, the designer should develop drawings Extruded-dielectric Cables

using information up to Level C for all underground Direct-buried extruded-dielectric cables require that a
projects. If the project is in an urban area, or there is splice pit be opened at each end of the relocated area. A
evidence of a significant amount of existing under- trench is excavated between the two pits, and the new
ground facilities, it is recommended that the documen- cable is placed in the trench. The existing cable is cut,
tation be to Level B or A. If exact location is needed, and the new cable is spliced to the existing cable. Some-
Level A is required. times manholes are used in place of a splice pit, but are
not required.
The ASCE guide indicates that the total savings on a
typical underground project using Quality Level B and If the cable is installed in a duct bank, the relocation
A data may range from 10 to 15% compared with costs would follow the process described above. Special care
from a project using Quality Level C and D data should be taken to protect the existing cable while the
(ASCE 2003). duct bank concrete is being removed.
Post-Design Survey Pipe-type Cables
Once the design is completed, it is important to have A gas-filled pipe-type cable system would follow a simi-
the route resurveyed to determine if any conflicts exist. lar procedure to the one described above. The main dif-
This step is usually performed as part of the construc- ference is the need to reduce the gas pressure prior to
tion survey. cutting into the cable pipe and replacing the gas after
the cable has been spliced.
12.2.4 Rerouting Existing Lines
Occasionally, an existing underground circuit must be For a fluid-filled pipe-type cable system, the main con-
relocated or rerouted to accommodate new infrastruc- cern is controlling the flow of the dielectric fluid. If the
ture or the addition of a new substation. line length is relatively short, it may be cost-effective to
remove the dielectric fluid from the pipe. If removing the
General fluid is not practical, the most common method of con-
Regardless of the type of cable system installed, rerout- trolling the fluid is by freezing the fluid, thus solidifying
ing or relocating an existing underground circuit is the fluid. This is commonly known as a “pipe freeze.”
costly, time-consuming, and usually requires an Chapter 13 discusses the method of installing a “pipe
extended outage. If only small adjustments of a few freeze.” The “pipe freeze” is placed on either side of the
inches are needed, it may be possible to expose the new splice points. Once the freeze is in place, the fluid
underground line by excavation and raise or lower the between the freezes can be removed, and the pipe and
pipe or ducts, as required. If more than a few inches are cable can be cut. After the new installation is complete,
needed, a new length of cable must be installed. Because and prior to releasing the “pipe freeze,” the new pipe is
most utilities do not have large quantities of transmis- filled with fluid.
sion cable in stock, they are unlikely to be able to per-
form the relocation quickly. The time required for the SCFF Cables
repair is normally based on how quickly a utility can As with pipe-type cable, the dielectric fluid in an SCFF
obtain the required skilled labor and materials. Depend- cable must be controlled. This is commonly done by the
ing on the cable type and its location, repairs can take freezing method similar to the pipe-type cable system.
from 1 month, if the materials are on hand, to as much Otherwise, the relocation process is similar to the pro-
as 6 months if the cable has to be manufactured. cess associated with extruded dielectric cable.
In general, the relocation or rerouting process consists 12.2.5 Overhead-to-Underground Transition

of the following activities. The connection of an underground circuit directly to an
• Installation of new manholes around the existing overhead line at a transition structure is becoming more
cable. common. The designer should give considerable atten-
tion to the design of the transition structure.
• Installation of new pipe or duct between the new
manholes. General
• Installation of the new cable. The two most common ways of making an overhead-to-
underground transition is to use a single-shaft structure,
• Splicing of the new cable to the existing cable. or a small transition site. When planning for an under-
• Removal of old pipe or duct. ground-to-overhead transmission line transition, the
designer must consider several issues relating to the sit-
12-7
ing of the structure or station. Some of the issues to be As with overhead line structures, various types of struc-
considered are: tures have been used for the transition pole, including
• Environmental/community impacts, such as wetland, the following: self-supporting steel, guyed direct-
embedded steel, guyed wood, and guyed laminated. The
archeological, land contamination, public lands,
most common type of transition pole is self-supporting
noise, aesthetics, and EMF.
steel. A self-supporting steel structure eliminates the
• Physical site considerations, such as topography, geo- potential clearance issue that would exist if the structure
logical, access, and overall land use. needed to be guyed. Since steel typically lasts a long
• System requirements, such as switching, outage dura- time, it eliminates the potential problem of needing to
tion, and maintenance. replace a wood pole after a few years due to normal
• Equipment requirements, such as load-break discon- deterioration. Additionally, wood structures dry out and
shrink, resulting in loose hardware fittings and guy wires.
nect switches, breakers, and/or reactors.
Preferably, the cable should be installed on the outside
In general, a complete system study should be per-
of the structure. There are special instances where the
formed to determine the effect that the new under-
cable may be required to be installed on the inside of the
ground line will have on the overall transmission system.
structure by local ordinances. Special design and detail-
Chapter 16 discusses the effect that an underground
ing are required for a structure in which the cables pass
cable system can have on a utility’s electrical system. In
through the base plate and are installed internal to the
addition, an analysis should be made on the equipment
structure.
to determine if the switches or breakers are capable of
switching out the cables given the anticipated charging Transition Site Design
currents, and withstanding the anticipated switching The layout of a transition site is determined by the
surges and overvoltages. amount of equipment that is needed, such as disconnect
switches, reactors, breakers, control house, etc. The
Single-structure Design
design of this site is similar to a small switching station.
Design of the transition structure must provide the
The utility should use their normal substation design
proper electrical clearances defined in the latest version
specifications and standards.
of the National Electric Safety Code (NESC). These
clearances address the electrical clearance from the con- Extruded-dielectric Cables
ductors, jumper loops, equipment, and other energized Transition structures for voltages up to 230 kV are usu-
parts to the surfaces of the supporting structures and to ally a single-pole structure, an H-frame structure, or an
the spaces that will be designated for climbing and A-frame structure. The single-pole structure is common
working on the structures. Consideration should be on rights-of-way, whereas the other structures are com-
made to provide adequate climbing and working spaces monly found within substations. At voltages higher than
for safely performing maintenance on energized over- 230 kV, extruded-dielectric cables generally transition to
head conductors using hot-line tools. The need to pro- overhead within a transition station.
vide hot-line climbing and working areas should be
reviewed on a project-by-project basis, and the require- Extruded-dielectric cables can be installed on the inside
ment waived where line configuration and installed or outside of the transition structure, and terminations
equipment on the structure would not permit hot-line are placed on support arms, as shown in Figure 12-3.
work or where the structure will not be climbed or Most manufacturers recommend clamping the cable on
maintained with the line energized. the vertical face of the structure at 5- to 6-ft (1.5-1.8 m)
intervals. A support channel from the pole to the base
Structures must be designed for the wind, ice, equip- of the terminator is sometimes installed to provide addi-
ment, seismic, and code loads. Structure loads should tional support to the cable. Most terminations require
include all typical attachments for the structure with two cable clamps immediately beneath the terminator
conductors, static wires, and insulators. base plate. Basket-weave grips could also be used to
support the section of cable from the pole to the sup-
Support arms must be designed for personnel loading port arm. Basket-weave grips may also be used at inter-
and all construction/installation loads. The attachment mediate points to support long vertical runs of single-
plate for the terminations should have an open side to conductor cable.
prevent circulating currents from flowing in the plate.
Vertical steel members may be required below support
arms to provide support to the cable.
12-8
Most cable manufacturers and cable installers prefer 12.2.6 Soil Thermal Properties and Special
having the cable installed on the outside of the structure Backfill
because of the ease of the installation.
Importance
Pipe-type Cables The earth portion of the thermal circuit is responsible
Pipe-type cable systems must have a pressurization plant for the greatest percentage of thermal resistance for bur-
at one end of the cable system. If the new underground ied cables, often accounting for more than half of the
line is in the middle of the overhead line, a transition site total resistance. The geometric component of resistance,
is required at one end of the underground line to accom- which is depth of burial, can be accurately determined
modate the pressurization plant. Pipe-type cable termi- during route surveys. The intrinsic material component,
nations can be placed on single-pole structures, but they soil thermal resistivity (commonly called Thermal Rho
would normally incorporate a small fenced area around or TR), is the most variable of the components, chang-
the facility. The reason for this is to provide protection ing with both time and distance along the route. Accu-
to the public in the unlikely event of a riser pipe failure. rately determining soil thermal resistivity permits much
more accurate ampacity calculations, and special back-
Another design consideration when installing a pipe- fill allows higher ampacity of the circuit. Thermal diffu-
type cable on a single-pole structure is to limit the sivity is a measure of the ability of soil to undergo
mounting height of the terminations. The terminations temperature change. Thermal diffusivity is important
should be kept as low as possible for hydraulic reasons. for transient calculations and for calculating the effects
of daily load cycles as described in other sections below.
SCFF Cables
Most SCFF cable systems require a transition site Factors Affecting Thermal Resistivity and Diffusivity
because of their pressurization system requirements. An Several factors affect soil thermal resistivity and ther-
SCFF cable could be placed on a single-pole structure, mal diffusivity. Figure 12-4 shows the effects of two of
but would require the pressurization reservoirs to be the major determinants, soil composition (soil type) and
mounted on the structure. moisture content, for the most common soil types and
for some corrective thermal backfills. Similarly,
450
Soft Organic Clay

400 Clay
Silt
Silty Sand with Gravel
350 Uniform Sand
Stone Screenings
FTB
Thermal Resistivity ( C-cm/W)
300
o
250
200
150
100
50
0
0 5 10 15 20 25 30 35
Moisture Content (% by dry weight)
Figure 12-3 Extruded-dielectric cable installed on a steel Figure 12-4 Soil thermal resistivity as a function of soil type
transition structure (courtesy POWER Engineers, Inc.). and moisture content (courtesy Geotherm Inc.).
12-9
Figure 12-5 shows the effect of moisture content on Soil Composition and Texture
thermal diffusivity for some common soil types. Soil is a composite material consisting of solid particles,
water, and air. Heat flows through soil primarily by con-
Moisture Content
duction from particle to particle. In some cases, it is by
Even soils that have poor composition can have accept- conduction and/or convection through the moisture or
able thermal resistivity if the moisture content is kept air that occupies the pore space between solid particles.
high. For most soils, thermal resistivity increases only Even highly compacted soil can have up to 30% free vol-
marginally when moisture decreases from a wet to a ume that is occupied by moisture and/or air. In a totally
moist condition. If the moisture content drops below dry condition, this volume is filled with air, which has a
the critical moisture content level, the thermal resistivity very high thermal resistivity (about 4000 C°.cm/watt
increases substantially. Moisture moves away from (40 K.m/W)), and thus the resulting soil resistivity is
higher temperatures and tends to migrate from the high. A good thermal backfill is made up of well-graded
cable/soil interface to ambient earth. If the moisture solid particles that provide more points of inter-particle
content for material other than clay is above the critical contact for conduction of heat. Two benefits result:
value (the knee of the curves in Figure 12-4), moisture resistivity is low even at low moisture contents, and
moves back toward the cable by capillary action quickly moisture is less likely to migrate away from the cable/soil
enough to prevent drying. Below the critical moisture interface. These benefits are attributed to the tight pack-
content, moisture does not return as rapidly, and the ing of the particles (high density, low porosity, and low
soil will continue to dry. If this condition persists, a ther- hydraulic conductivity). Figure 12-6 shows the mecha-
mal runaway condition can occur. Thermal runaway nism of heat transfer through wet and dry soil.
happens when lower moisture content causes an
increase in thermal resistivity, which causes a higher Specifications for controlled thermal backfill often
cable temperature rise, which in turn drives more mois- include a sieve analysis to ensure that the material falls
ture away, causing further increases in thermal resistiv- within specified limits of gradation characteristics. Fig-
ity, etc. until there is a cable failure. Ambient moisture ure 12-7 shows an envelope with upper and lower limits
content can be very low for sands in dry areas, giving for a good thermal backfill. This type of granular mate-
high thermal resistivities. Moisture content under paved rial is made up of sound (nonporous) fine aggregate;
surfaces generally is higher than under unpaved sur- when installed at its optimum moisture content and at
faces. Clays fall in a special category, since field mois- 95% standard Proctor density, it will give low and stable
ture content is very high, typically 20–30%, and clays thermal resistivity. The widespread use of the Thermal
dry and rewet at a much slower rate than other soils.
Moisture content is expressed as a percentage of the
weight of water to the dry weight of soil solids, as deter-
mined by oven drying at 105°C.
0.014
0.012
0.01
Thermal Diffusivity (cm /s)
2
0.008
0.006
0.004
stone screenings
sandy silt
0.002 sand
clay
0
0 2 4 6 8 10 12 14 16
Moisture Content (% by dry weight)
Figure 12-5 Soil thermal diffusivity as a function of soil Figure 12-6 Heat flow through soil—the effect of moisture
type and moisture content (courtesy Geotherm Inc.). content on thermal resistivity (courtesy Geotherm Inc.).
12-10
Property Analyzer (TPA) (EPRI 1981) has reduced the Soil Compaction (Density)
reliance only on gradation characteristic (sieve analysis), Once installed, the density of a backfill does not change;
because users can readily make direct, accurate measure- it is only the moisture content that may change due to
ments of thermal resistivity rather than inferring values various factors. Backfill material must be carefully com-
based only on sieve analysis. Soil texture refers to soil pacted. Compaction decreases thermal resistivity, espe-
grain size, shape, and particle size gradation. In engi- cially at low moisture contents, since the interparticle
neering applications, a soil is often qualitatively catego- heat path is improved. An increase in density of one
rized by a visual description using accepted adjectives to pound per cubic foot (16 kg/m3), which is approximately
indicate the fractional amount of each component (i.e., 1% additional compaction, can give a 2–3% decrease in
gravel, sand, silt, clay, etc.). resistivity (EPRI 1977). This improvement is especially
important at low moisture levels. Compaction is best
The 5% moisture content is a typical dry-season level carried out at moisture contents of 8–12% for granular,
for soils in many parts of North America. The first four noncohesive soils, and much higher for fine-grained
materials listed are very stable and will not reach zero (silty-clayey) soils. Great care should be taken to obtain
moisture content unless heat flux is very high. Ambient good compaction along the entire cable route to prevent
moisture contents for clays are often greater than 15%, hot spots.
but if subjected to high heat flux, they can become ther-
mally unstable. The numbers in Table 12-1 are only rep- Table 12-1 Representative Thermal Resistivity Values in
resentative. Any backfill material considered for use on C°.cm/watt (K.m/W)
a cable system should be tested thoroughly. Soil particle 5% Moisture 0% Moisture
soundness (porosity), mineral type (limestone, granite, Fluidized thermal backfill 40 (0.4) 80–100 (0.8–1.0)
basalt, mica, etc.), and organic content also affect the Concrete (no air) 30 (0.3) 70 (0.7)
thermal resistivity to some degree. Porosity and organic Stone screenings 40 (0.4) 100 (1.0)
content higher than 4% can increase the dry thermal Thermal sand 50 (0.5) 100 (1.0)
resistivity by as much as 15%. The user should evaluate Uniform sand 70 (0.7) 200 (2.0)
the cost of changing routes to avoid poor soils such as cin- Clay 100 (1.0) 250 (2.5)
ders or organic material, versus installing special backfill Lake bottom (organic) 100a (1.0) >300 (>3.0)
in the trench. Highly organic soil >300 (>3.0) >600 (>6.0)
a. 50% moisture.
GRAIN SIZE DISTRIBUTION

unified soil classification system
GRAVEL SAND FINES
COARSE FINE COARSE MEDIUM FINE SILT CLAY
3" 1" 3/4" 1/2" #4 #8 #16 #30 #50 #100 #200 U.S. standard sieve
100 100
90 90
80 80
70 70
60 60
Percent Passing
50 50
40 40
30 30
20 20
10 10
0 0
100 10 1 0.1 0.01 0.001
Grain Diameter (mm)
GRADING LIMITS FOR GRANULAR TYPE

THERMAL BACKFILL
Figure 12-7 Gradation limits for granular-type thermal backfill (courtesy of Geotherm, Inc.).
12-11
Compaction is commonly specified in percentage of Soil Interface Temperature

Proctor Density (ASTM 1978). A Proctor Density of For many soils, if the temperature at the interface
100% gives the highest density possible for a given mate- between the heat source and the soil is kept below
rial at the optimum moisture content. Proctor densities 50–60°C, a balance will be maintained between mois-
of 90 to 95% are commonly specified for backfill in ture migration away from the heat source and capillary
cable trenches. In addition to improving thermal proper- action, which returns the moisture. The soil will, there-
ties, high compaction reduces the chances of the pave- fore, remain stable. Published ampacity tables for distri-
ment settling. Many municipalities specify high Proctor bution cables list three values of allowable ampacity:
Densities to prevent the settling of road-base material. that governed by conductor temperature, and those
(generally lower) governed by 50° and 60°C soil inter-
Compaction (density) and moisture content of the face temperatures (IPCEA 1976). Selection of the higher
material affect both thermal resistivity and diffusivity. interface temperature may be done only after testing the
Care should be taken to maintain the field (in situ) con- soil; otherwise, the lower value should be used.
ditions for the laboratory tests. A utility evaluating con-
Heat Flux
trolled backfills may have laboratory tests performed at
Moisture migration actually occurs as a function of heat
several compaction levels to permit specification of field
flux at the surface of the heat source. Heat flux is deter-
compaction (density). Figure 12-8 demonstrates the
mined by heat input in watts/foot (W/m) and cable
relationship between soil density and thermal resistivity.
diameter in inches (mm). One approach to evaluating
Factors Affecting Thermal Stability soil stability is to energize a thermal probe using the
Unlike thermal resistivity and diffusivity, soil thermal same heat input per unit length as the cable system (e.g.,
stability is a system-driven parameter. In other words, 30 W/ft [98 W/m]) and measure thermal resistivity as a
soils can have widely differing thermal stabilities function of time (Martin et al. 1981; Hartley and Black
depending on a number of operating and installed con- 1979). The time at which the slope changes (increase in
ditions. Two criteria to reduce chances of dryout and to thermal resistivity) gives an indication of time to dryout.
evaluate thermal stability are discussed—soil interface The corresponding dryout time for the cable is obtained
temperature and heat flux. from Equation 12-1:
2
⎛D ⎞
tc = t p ⎜ c ⎟ (min) 12-1
⎜D ⎟
⎝ p ⎠
Where:
tc = time for soil near cable to dry, min.
tp = time for soil near probe to dry, min.
Dc = diameter of cable/earth interface, in.
Dp = diameter of probe, in.
Figure 12-9 gives typical test results for sandy silt. A

dryout time of 20 minutes for a 157.5-mil (4-mm) probe
corresponds to a dryout time of about 112,500 minutes
or 78 days for an 11.81-in. (300-mm) cable pipe. Calcu-
lated dryout times for the cable system should be several
weeks to ensure stability over the summer season. This
approach can be conservative since cyclic loading and
rewetting due to rainfall serve to extend dryout time.
Soil thermal stability may also be evaluated during field

thermal surveys. A thermal probe may be energized at a
nominal heat input (typically 10 W/ft [33 W/m], which is
30–50% of typical cable heat input). After the slope of
temperature vs. log time has stabilized, the heat input is
increased tenfold. If a new, constant slope is achieved,
Figure 12-8 Effect of soil grain size and distribution on the soil is stable. If the slope continues to increase, the
density and thermal resistivity (courtesy Geotherm Inc.).
12-12
soil is not considered stable. Figure 12-10 shows a typi- Laboratory Thermal-Resistivity Measurement
cal curve for this test. Undisturbed soil samples retrieved in thin-wall Shelby
tubes or bulk soil samples should be tested for thermal
Concrete duct bank encasement and controlled backfill and geotechnical parameters by a qualified and experi-
both have excellent thermal stability. Their use transfers enced laboratory. The tests should include natural mois-
the interface temperature and heat flux concern to the ture content, density, sieve analysis, compaction, thermal
interface with the native soil. Temperatures and heat resistivity, thermal stability, and thermal resistivity vs.
flux are low at that interface, and stability of the native moisture content (dryout tests). Nominal 3-in. (76-mm)
soil is not generally a problem. diameter Shelby tubes should be used to retrieve undis-
turbed samples. Care should be taken to maintain the
natural moisture and density of the samples during
retrieving and transportation. Figure 12-11 shows vari-
ous types of samples being tested in the laboratory for
thermal resistivity using a Thermal Property Analyzer.
A laboratory-type thermal needle, 0.125-in. (3.2-mm)

diameter, is inserted into the test sample and connected
to a Thermal Property Analyzer (TPA). A series of ther-
mal-resistivity measurements is made for various mois-
ture contents by stage-drying the sample. Enough time
should be allowed to ensure moisture redistribution
throughout the sample at the start of each new stage.
When testing disturbs bulk samples or granular soils, a
test mold similar to a Shelby tube can be used to repack
the soil at a specified density and moisture content. The
curves shown in Figure 12-4 were developed from these
laboratory tests.
A similar setup can be used to determine the thermal

diffusivity of the soil. The test sample, either taken as a
section from a Shelby tube or reconstituted in a mold, is
Figure 12-9 Soil thermal stability for sandy silt. allowed to equilibrate at a constant elevated tempera-
ture overnight in an oven. This temperature should only
be about 8–10°C above the ambient room temperature.
The warm sample is immediately transferred to a circu-
lating water bath at a constant room temperature, and
Figure 12-10 Typical thermal stability test. Figure 12-11 Various types of samples being tested in
the laboratory (courtesy Geotherm Inc.).
12-13
the temperature of the thermal needle is monitored with It is important to perform the tests at the expected cable
time. The resultant plot of temperature vs. log time is depth. Soil thermal resistivity is usually lower at greater
analyzed by the theory of Shannon and Wells (Shannon depths because of higher soil moisture content.
and Wells 1947) to calculate the thermal diffusivity. Fig-
ure 12-12 shows this setup. The TPA (EPRI 1981) was developed in the late 1970s
to automate field thermal-resistivity tests for greater
Field Thermal-Resistivity Measurement accuracy, simplicity, and reproducibility of test results.
In situ tests provide rapid and accurate values of soil When a thermal probe is used with a TPA, a laptop com-
resistivity along a cable route. The thermal probe has puter controls the test input parameters. Test results—
been used for many years for this purpose. It consists of including probe power, time, temperature, and thermal
a stainless-steel tube that contains a precisely wound resistivity—are all displayed and stored during a test run.
heater and a temperature sensor of high resolution and The TPA also facilitates the thermal-stability measure-
accuracy. The annular space is filled with a thermal ments by extending the standard test over longer time
epoxy, and all components are electrically insulated. periods or by ramping the probe power to a higher value.
Commonly used field thermal probes are 0.25 in. IEEE Standard 422-1981, “IEEE Guide for Soil Ther-
(6 mm) in diameter and 12 in. (300 mm) long. Probes of mal Resistivity Measurements,” (IEEE 1981) should be
larger diameter and lengths up to 6.5 ft (2 m) with multi- followed for soil thermal-resistivity measurements.
ple temperature sensors have also been used successfully
in special conditions. Prior to conducting a field thermal survey, any available
soil data along the proposed cable route should be
The probe head is designed for easy adaptation to con- reviewed. Data may be available from government agen-
ventional soil-drilling equipment. The probe is simply cies, other utilities, or borehole logs for existing over-
inserted into the undisturbed native soil by pushing or head power line tower foundations. Soil thermal-
driving it to the desired depth at a test location. Power is resistivity measurements should be made at least twice
applied to the probe heater, and temperature rise is in each soil formation, and at least every 0.5 mile
monitored with time. The slope of temperature rise vs. (800 m) depending on the length of the cable route. At
log time determines the value of soil thermal resistivity. least two samples of each soil type should be tested in
Figure 12-13 shows a drill rig setup for in situ thermal
resistivity testing and soil sampling. A TPA powered by
a portable generator or a power inverter is shown in use.
Figure 12-12 Shannon and Wells thermal diffusivity test Figure 12-13 In situ thermal resistivity testing
applied to a Shelby tube sample. and soil sampling (courtesy Geotherm Inc.).
12-14
the laboratory for thermal stability and resistivity vs. The addition of small quantities of cement to granular
moisture content. Since thermal-resistivity values are materials, such as well-graded sands and screenings, can
extremely moisture-dependent, it is best to make field provide excellent thermal resistivity and stability (Sandi-
measurements during the driest times of the year. Esti- ford 1981). The cement creates interparticle bonding
mates of moisture content at other times of the year can and lowers thermal resistivity by reducing contact resis-
be used to extrapolate thermal resistivities based on tance and filling fine air voids. This type of cement-
curves similar to those shown in Figure 12-4. based mixture must be moisture-conditioned to its opti-
mum value and should be placed by a compaction
Often a soil-testing firm breaks pavement and augers to method similar to that for granular-type backfill.
the desired depths where thermal-resistivity measure-
ments are to be taken, and the probe is inserted into Fluidized Thermal Backfill™ (FTB™) is a concrete-like
undisturbed soil below the augured hole. Soil samples material that has been used extensively over the past
for laboratory testing are normally retrieved from the 20+ years. It is a mixture of natural mineral aggregate,
same borehole, at depths in-between the test depths. At sand, cement, water, and a fluidizer that is formulated
locations where a soil drill-rig cannot be used, in situ to meet specific thermal and strength requirements. Fly-
thermal testing and soil sampling can be conducted in ash is a commonly used fluidizer. However, if flyash is
test pits dug by a backhoe. not readily available, a water-soluble resin or slag can be
used. FTB can be mixed in a regular concrete truck, and
Controlling Thermal Resistivity
most ready-mix concrete suppliers can handle this
It is not generally feasible or practical to modify the
material without any problem. FTB is delivered to the
thermal resistivity of native soil surrounding the cable
site in a fluid state and can be poured or pumped into
trench. Selected materials having good thermal resistiv-
trenches very easily using conventional concrete place-
ity and thermal stability can be placed in the trench,
ment equipment. FTB is often an ideal choice for cable
however, to reduce overall thermal resistivity. The entire
trench backfill application because of the advantages
trench width is filled with special backfill to a height
that it offers over granular-type backfill. Figure 12-14
typically about 1 ft (300 mm) above the cable. In some
shows the application of FTB on a typical duct bank
instances, the trench must be widened or entirely filled
project. The process is the same for pipe-type or direct
with special backfill to offset the effects of poor
buried applications.
native soil.
Preliminary work has been done to develop special
Backfill used for this purpose, commonly termed con-
additives that may be injected into native soils to lower
trolled backfill, is made of well-graded crushed stone
thermal resistivity and increase thermal stability (EPRI
screenings or well-graded sands with some fines that can
1977). Soil-interface temperature and heat-flux values
achieve high densities at relatively low moisture con-
are seldom problems if special backfill is used. A typical
tents. There are several sources of controlled backfill:
backfill thermal resistivity requirement is the range of
• Suitable natural sands having the required gradation 50–70 C°-cm/W at 3–4% moisture content.
characteristics can be found in many parts of the
country. The material must be well compacted in the
trench, in layers of 6–12 in. (150–300 mm) or so. A
1-ft (300 mm) layer compacts to about 8 in.
(200 mm). Standard civil engineering (construction)
practice must be followed where applicable.
• In other areas, well-graded crushed limestone screen-
ings are used to give good thermal resistivities.
The screenings must also be well compacted as
described above.
Granular backfills of this type should be tested in the

laboratory to evaluate maximum dry density and opti-
mum moisture content. In situ density of the backfill
when placed in the trench should be about 95% of this
maximum dry density and at the optimum moisture
content. A quality control/assurance program should be Figure 12-14 Fluidized Thermal Backfill™ (FTB™) being
implemented for all field applications. applied to a duct bank (courtesy Geotherm Inc.).
12-15
It is important that the backfill in the immediate vicinity in a few years, with the thermal capability of the cable
of the cables is installed at the best possible density in suffering accordingly. Problems associated with cables
order to achieve a low thermal resistivity. Figure 12-15 sinking in soft sediments should also be addressed.
shows that the temperature drop is inversely propor-
tional to the radial distance from the cable surface. Special attention must be paid to soil thermal resistivity
and stability measurements for most underwater cross-
Calculating Effective Thermal Resistivity ings, and specialized instrumentation approaches are
Placing special backfill in the cable trench lowers required. Reference (Radhakrishna and Stienmanis
the effective thermal resistivity compared to native soil. 1981) provides a good description of test equipment and
The heterogeneous earth portion is conveniently han- procedures. Figure 12-16 shows a drill rig on a barge
dled by using Equations 11-58 to 11-75, as detailed in used for submarine thermal surveys.
Chapter 11. This treatment is considered sufficiently
accurate for ratios of controlled backfill with long Figures 12-17 and 12-18 show a vibra core sampler and
dimensions to short dimensions up to 3/1. its adaptation to thermal probe for in situ thermal resis-
tivity measurements.
More sophisticated approaches, such as finite element
analysis (Soulsby and Donovan 1981), provide greater Thermal Resistivities for Cables in HDD (Horizontal
accuracy at the expense of more complicated calculation Directionally Drilled) Installation
procedures. Installation of cables or casing containing power cables
in directionally drilled bores is quite common when
Thermal Resistivities for Submarine Cables
Sub-bottom marine sediments are usually fully satu-
rated, so in situ thermal resistivities are generally low if
the material is nonorganic. However, depending upon
the composition of the material (very soft clay or mud),
the thermal stability may be poor. Heat generated dur-
ing normal power cable operation can cause the mois-
ture to migrate from the cable vicinity, creating areas of
very high thermal resistivity. The dryout problem is
especially severe when the bottom sediments are highly
organic. Fully saturated organic sediments with high
moisture contents are known to have thermal resistivi-
ties greater than 150 C°.cm/watt (1.5 K.m/W). It is very
difficult—and in most cases, impractical or expensive—
to install controlled thermal backfill in submarine appli-
cations. Thus, the cable design must accommodate the
bottom conditions with respect to thermal resistivity. Figure 12-16 Drill rig on a barge for submarine thermal
Care must be taken in areas where silting takes place. surveys (courtesy Geotherm Inc.).
One meter of cover today might be four meters of cover
Figure 12-15 Thermal gradients in a radiant flow field are Figure 12-17 Vibra core sampler (courtesy Geotherm Inc.).
inversely proportional to the distance from the heat source
(courtesy Geotherm Inc.).
12-16
crossing railroads, highways, and rivers. For such instal- 12.2.7 Mechanical Considerations
lations, the geotechnical as well as thermal, properties The principal mechanical considerations for under-
of the soil at the cable/casing elevation must be investi- ground transmission cables consist of pulling tensions
gated. This is conducted in the same manner as for shal- and sidewall pressures during cable installation,
low borings using a conventional soil drill rig. The as described in Section 12.3. Other considerations
major difference is that, for river crossings, the cable ele- include Thermomechanical Bending (TMB), proper
vation may be below the water table, and bedrock may manhole racking, supports for vertical cable sections,
be encountered at other locations. and vibration.
Below the water level, soil is saturated with water, and Thermomechanical Bending
the thermal resistivity is not of concern unless organic Cables expand and contract with load cycling. This
material or peat is encountered. This is quite common in motion is usually accommodated by cable snaking
the upper soil layers below the bottom and near the river within the pipe of a pipe-type cable, or within the
banks. It is impossible to install corrective backfill at duct for single-core duct systems. However, for large
deep elevations, and therefore, measurement of ambient conductor and cable sizes, it is critical to make an in-
temperature and thermal resistivity is very important. depth investigation into this issue. For directly buried
single-core cables, the earth constrains most motion,
Most bedrock has fairly low thermal resistivity ranging except for the small amount that may occur in man-
between 40 and 65 C°.cm/W (0.40–0.65 K.m/W). The holes, if present.
annular space within the casing and around the cables Extruded-dielectric Cables
or cable conduits must be filled with a pumpable ther- There have been a number of splice failures caused by
mal slurry of low thermal resistivity. excessive unrestrained TMB forces within the manholes.
Ambient Earth Temperature There has been significant discussion on the best way of
Figure 12-19 depicts the variation in the earth ambient handling the TMB for large single-core cables. The com-
temperature at depths of up to 10 ft (3 m), as a function mon practice for determining the required duct size for
of the time of year. Air temperature near the surface a particular installation is to size the inside diameter of
changes significantly and quickly as a function of time, the conduit 1.5 times larger than the outside diameter of
whereas the temperature at depths is slow to react to the cable. For example, if the cable has an outside diam-
this change. Although this data was collected for eter of 3.5 in. (89 mm), then the minimum-sized conduit
Toronto, Canada at latitude 40°, similar trends hold ID would need to be 5.25 in. (133 mm), resulting in a
true for other locations. In southern states, the varia- 6-in. (166-mm) duct requirement. Historically, no major
tions in the upper and lower limits are smaller, and the circuit failures have been caused by thermomechanical
difference between the peaks at different depths (lag forces when this practice has been followed. However, if
time) is shorter. this practice is not followed, a detailed TMB analysis
should be made.
25
20
15
Temperature ( C)
o
10
5 Air temperature
1.5 m depth temp.
0 3.0 m depth temp.
Average ground temp.
-5 Freezing temperature
Min. & max. temp.
-10
0 30 60 90 120 150 180 210 240 270 300 330 360
Day
Figure 12-18 Vibra core sampler with thermal probe for Figure 12-19 Variations in earth temperature at
testing (courtesy Geotherm Inc.). various depths as a function of time of year (courtesy
Geotherm Inc.).
12-17
The following variables affect the overall TMB design, after many years of load cycling. Proper reinforcement
and each should be considered when determining the of the sheath/sleeve transition and the use of pressed
best and most cost-effective design. connectors have solved these two problems.
• Size of conduit Manhole Racking and Sizing
• Size of cable Manhole racking is only required for SCFF and
extruded-dielectric cable. For pipe-type cables, no rack-
• Size of the manhole
ing is typically needed, because the steel cable pipe/joint
• Strength and number of cable clamps sleeve usually supports itself across the manhole. If the
• Route geometry manhole is too long, the designer may want to provide
some type of support.
• Strength and design of the cable splice
• Cable restraint and support brackets Extruded-dielectric Cables
As discussed previously, there are two common methods
There are two common methods for handling the TMB of installing cable in a manhole: the straight-through
forces resulting from the cable load cycling: the first is method and the S-bend method. The main design con-
the straight-through approach using sufficient clamping cern with manhole racking is to protect the splice from
to force the cable expansion back into the duct; the sec- experiencing excessive thermomechanical forces.
ond is to install an S-bend within the manhole and allow Regardless of the installation method, this is accom-
the S-bend to flex and absorb the expansion. With the plished by placing cable clamps on either side of the
straight-through approach, the size of the conduit plays splice to prevent the cable from moving near the splice.
an important part in the overall thermomechanical
design. When the cable is fully restrained and no move- When designing the cable supports for the straight-
ment of the cable is permitted in the manhole, the through method, it is important to calculate the
thermal expansion or contraction must be accommo- expected TMB forces and design the cable clamp and
dated in the duct. This could result in high internal com- support hardware accordingly. This method allows the
pressive and tensile forces. This design allows the manholes to be smaller. Sizing of the manhole is based
manholes to be shorter. on the splice design. There must be enough space within
the manhole to allow the splice sleeve to slide out of the
With the S-bend approach, compressive and tensile way during splicing procedures. The typical inside
forces are not developed within the cable. However, dimensions for a 138-kV manhole are 6 ft wide by 20 ft
because the cable is allowed to flex, special consider- long and 7 ft high (1.8 x 6.1 x 2.1 m). Figure 12-20
ation must be made to prevent the cable from exceeding shows a typical manhole clamping arrangement.
its bending limits. This results in the manhole needing to
be wider and longer. For the S-bend design, the manholes must have enough
room to allow the cable to move smoothly to accommo-
Another factor that affects the overall TMB forces is date thermal expansion and contraction. This can gen-
the number of vertical and horizontal bends that the
cable will go through between the manholes. If the route
has many bends, then the resulting forces will be smaller.
If the route is flat, the resulting forces will be signifi-
cantly higher.
Pipe-type Cables
Early 345-kV pipe-type cables have experienced TMB in
the splice area, where the pipe diameter increases in the
joint casing (Aabo and Moran, Jr. 1988; Moran et al.
1984). Early designs had long unsupported cable lengths
in this area, and the joints could flex too easily. These
early joints were retrofitted with additional joint sup-
ports, and new designs include the proper supports.
SCFF Cables
SCFF joints have experienced mechanical fatigue in Figure 12-20 Typical cable clamping arrangement for
manholes. Sheaths have cracked at the transition to the cables passing straight through a manhole (courtesy
joint sleeve, and soldered connectors have pulled out POWER Engineers, Inc.).
12-18
erally be accomplished by having the cable ducts enter If an extruded cable is installed along a long steep slope,
the center of the manhole and then train the cables to an intermediate clamping manhole may be needed to
their proper position. The cables form two reverse prevent the cable from ratcheting down the slope.
bends, as shown in Figure 12-21, as the joints are racked
Pipe-type Cables
on the manhole sidewall. The typical inside dimensions
For vertical and steep slope designs, a riser cable design
for a 138-kV manhole with this arrangement are 8 ft
is used for pipe-type cables. Long-lay stainless-steel
wide by 25 ft long and 7 ft high (2.4 x 7.6 x 2.1 m). Some
tapes are placed under the skid wires. These ribbons
utilities make a full 360° loop in the manhole to ensure
support most of the cable weight. Anchor joints are
that movement is not concentrated in one location,
placed at the top of the slope, and skid joints are placed
which would require an even larger manhole. A straight
at the bottom. Chapter 8 describes these types of joints.
distance of at least two cable diameters should be left at
each end of the reverse bend. Enough straight distance SCFF Cables
must be provided near the joint to allow the joint sleeve The method for installing SCFF cables would be similar
to slide out of the way during splicing procedures. to the methods described above for extruded-dielectric
cables. However, special consideration should be
Pipe-type Cables
given to the hydraulic issues associated with large eleva-
Since there is typically no support requirements, the siz-
tion differences.
ing of the manhole for a pipe-type cable system is based
on the joint-casing design. A sufficient straight distance
12.2.8 Monitoring
from the face of the reducer to the manhole end wall
must be provided to allow the joint sleeve to slide out of Monitoring systems can help to ensure long, trouble-
the way during splicing procedures. The joint casings free cable life, and rapidly determine and locate troubles
can be designed in multiple telescopic sections. This on the cable system. Monitoring systems that are part of
design allows the manhole length to decrease. However, cable system design are described below.
additional welding is required with multiple sleeves Temperature Monitoring
while the cable is in the pipe. The recommendation is to For many years utilities have installed thermocouples on
limit the number of welds, if possible. cable pipes several hundred feet (50–100 m) on either
Supports for Vertical Cables side of manholes. The thermocouples are terminated in
Special consideration of conductor stresses and boxes in the manholes, and can be read as desired. In
methods to support cable weight is required for cable the 1970s, elaborate temperature monitoring was
systems designed to be installed on steep slopes or verti- applied as part of early dynamic rating systems (Patton
cal applications. et al. 1979). Temperature monitoring is part of more
sophisticated dynamic rating systems designed in the
Extruded-dielectric Cables early 1990s (Engelhardt and Purnhagen 1991). Distrib-
Extruded-dielectric cables can be installed on the inside uted fiber-optic temperature monitoring offers the pos-
or outside of single-pole deadend structures. The cables sibility of monitoring temperatures along the entire
are typically clamped to the pole, as shown in
Figure 12-22. Cable clamps or basket-weave grips can
be used at intermediate points to support long vertical
runs of single-conductor cable. Thermal expansion and
contraction can be accommodated by cable snaking
between the clamps.
Figure 12-21 Training cables in a manhole (courtesy Figure 12-22 Cables clamped to a transition pole (courtesy
POWER Engineers, Inc.). POWER Engineers, Inc.).
12-19
length of a cable. These systems are suitable for all cable the power cable screen, laid close to the screen or situ-
system types. ated in the interstices of a number of cables (e.g., a trefoil
arrangement). The optical fibers may be installed inside
Distributed Temperature Monitoring
a plastic or metal tube. A disadvantage of this method is
A distributed temperature monitoring system (DTS)
that it requires splicing the fiber at each manhole.
consists of installing a fiber optic cable, which acts as a
temperature sensor, along the entire length of a cable
The most common method in the majority of existing
and some type of monitoring. The monitoring system
and new cable installations is to install the temperature-
reads the temperature of the fiber. The monitoring sys-
sensing fibers nearby the cables. This can be accom-
tem can be permanent or portable. Whether the moni-
plished by either attaching the optical fibers to the cable
toring system is permanent or temporary should be
using cable ties or installing the optical fibers in an adja-
based on the needs of the utility. If real-time monitoring
cent conduit. If the cable surface is accessible along the
is desired, a permanent installation is necessary. If only
entire length, then the sensor can be fixed to the outside
historical or benchmark information is desired, a porta-
surface of the power cable. This would be the preferred
ble unit would be sufficient. With a portable system,
method for direct-buried cables. There will, however, be
multiple systems could be periodically monitored.
uncertainty in knowing what is being measured: the sur-
face of the cable or the layer of air or protection around
For the most accurate temperature reading, the best
the fiber in its vicinity. When a temperature-sensing
location to install the temperature-sensing fiber is inside
fiber is attached to a power cable, it may be installed in a
the conductor itself. This method eliminates outside
plastic or metal tube, or it may consist of an optical
interference and any misinterpretation of results. How-
fiber cable with additional strength members which may
ever, there is currently no manufacturer that can accom-
include metallic parts. If metallic parts are included,
plish this type of installation.
then consideration needs to be given to the possibility of
induced circulating currents generating additional heat-
The second best choice is to place the temperature-sens-
ing effects and causing temperature measurement errors.
ing fiber, during cable manufacturing, under the outside
In addition, the bonding and grounding of metallic
jacket. The calculation of the conductor temperature
parts needs to be considered. If the cable with the tem-
based on the measured temperature is straightforward in
perature-sensing fiber attached is to be pulled into a
this case, since the thermal characteristics of the cable
duct, special consideration needs to be given to the
materials are well known. When a temperature-sensing
forces that the fiber will encounter during the pulling of
fiber is integrated into the power cable, it may be part of
the power cable.
Table 12-2 Advantages and Disadvantages of Fiber Optic Cable Location for Temperature Sensing
Advantages Disadvantages
Integrated into the power cable
Provides a better indication of the conductor core temperature Manufacture of power cable is more complex
More responsive to current loading Optical splicing at each cable joint is required and is more complicated
Fiber is protected by the power cable More fiber splices are needed (determined by power cable drum length),
resulting in higher dB losses.
Well-suited where power cable pulled into ducts May require more monitoring loops.
Fiber needs to be replaced and respliced in the event of a cable failure.
Externally attached to the outside of the power cable
Fiber cable can be run in long lengths without splicing, resulting Less responsive to load changes than integrated design
in lower dB losses
If installed in a tube, fibers can be blown out and new ones Less representative of conductor temperature
blown in, should a fiber failure occur
Can be easily attached to a conventional cable in an open Increases installation work
trench (before backfilling)
Can be easily attached to a power cable in a tunnel, or to the Not suitable for attaching to power cables that will be pulled through
roof or floor of a tunnel long ducts
In a separate duct or close to the power cable
Relatively easy to install or may already exist Least responsive method to load changes
Fiber cable can be retrofitted Least representative of conductor temperature— remote from conductor
Can be run in long lengths without splicing, resulting in lower
dB losses
Can use spare fibers in an already installed telecom cable
12-20
In a duct bank, any empty duct can be used to house the Leak location cables were developed in the late 1980s
temperature-sensing fiber. However, the farther the tem- (Ghafurian et al. 1989). These cables are sensitive to
perature-sensing fiber is from the hottest conductor, the dielectric fluids, and may be placed in the trench adja-
more uncertainty is introduced in the correlation cent to a cable pipe or an SCFF cable. They indicate
between the measured temperature and the actual tem- that a leak has occurred and locate a leak to within
perature of the conductor, and thus the rating. When a about 6 ft (2 m).
sensor fiber is separately installed and positioned a
short distance from the power cable, the cable construc- 12.2.9 Grounding/Cathodic Protection
tion can be of various types. The cable may be part of a Another important design consideration is the type of
telecommunication system and one or two fibers made grounding needed to ensure the proper operation of an
available for temperature sensing. However, typically a underground cable system. A detailed discussion of the
separate fiber cable is installed. grounding and cathodic protection for underground
cable systems is included in Chapter 10.
The advantages and disadvantages of using optical fiber
sensors in different configurations are given in Table 12-2.
12.3 INSTALLATION MODES
Leak Detection
Rapid detection of a fluid leak reduces the amount of 12.3.1 Splices: Directly Buried vs. in Manholes
fluid loss. Leak detection is especially important for Cable splices are required to join sections of cable. The
high-pressure fluid-filled (HPFF) cables, because of the distance between splices is governed by pulling tensions
large quantities of fluid involved. Fluid loss is also or the length of cable that can fit on a reel, and varies
important in SCFF cable due to the significantly lower from less than 1000 ft (300 m) to more than 7000 ft
reserves of dielectric fluid and the possibility of mois- (2100 m). Splices are described in detail in Chapter 8.
ture ingress and ultimate failure of the cable. As Splices must be made in a clean, dry environment.
described in Chapter 9, pressurizing plants contain sim- Humidity control must be provided for higher-voltage
ple leak detection in the form of fluid-level alarms on cables, especially paper-insulated cables. Splicing is
the storage tanks or frequent-operation alarms on generally done in a permanent manhole, although splic-
pumps. Cumulative flow meters add further sophistica- ing requirements can be met using a temporary man-
tion and reduce the amount of fluid loss before detec- hole, which is then backfilled, thus creating a directly
tion. More elaborate systems, combining software that buried splice.
evaluates expected fluid expansion/contraction as a
function of cable loading, provide even further sensitiv- Most pipe-type cable splices are installed in permanent
ity (Engelhardt and Purnhagen 1991). Incorporating manholes. Directly buried splices are often used with
these devices in the cable design and installations stages directly buried SCFF or extruded-dielectric cables.
is more effective and less costly than adding them after Some users provide permanent manholes for easy access
the line is in service. to splices even if the cables themselves are directly bur-
ied. Permanent manholes are almost always used for
Leak Location
cables in duct. Because of the increasing importance of
Once a leak is detected, rapid location minimizes fluid
dissolved gas analysis, more utilities are installing sam-
loss. An EPRI project in the 1980s investigated many
pling valves on splice casings for HPFF and SCFF
leak location methods (Williams et al. 1983). Chapter 15
cables (or are using the evacuating/filling valves that are
addresses the operating procedures for leak location and
already in place), so that permanent manholes are nec-
repair. The engineer can take several steps at the design
essary for access to the valves.
stage to speed future leak location efforts.
A temporary manhole can consist of a precast or
Probes have been developed that can detect the very
poured concrete pit, with plywood walls and tarpaulins,
slow fluid flow in the direction of a leak (Ghafurian et
or a temporary chamber placed over the pit to protect
al. 1989). These probes must be inserted into the cable
the cables during splicing. Typical dimensions are
pipe. Valves can be added at any time. However, if they
6–10 ft wide x 15–20 ft long (1.8–3.0 x 4.6–6.1 m). After
are installed at the top of joint casings or perhaps on the
the splice is complete and all tests are done, the pit is
line pipe in cable manholes before cable installation, the
filled with clean sand. Pavement is restored if necessary,
cost of installing them is lower, and there is no chance of
and a marker is placed so the splice can be located
damage to the cable. These valves are also useful for
quickly if repairs are needed.
fluid sampling for dissolved gas analysis.
12-21
Permanent manholes are typically 8 ft wide x 18 ft long person determines that the soil conditions do not
x 7 ft high (2.4 x 5.5 x 2.1 m), although sizes can vary require shoring. Along roadways, shoring may be
depending upon the number of circuits, cable type, and needed to maintain the trench sides and prevent damage
voltage. Manholes can be precast, which are often less to the rest of the roadway.
costly if facilities exist nearby for casting and if many
manholes are ordered. Alternatively, field-poured man- During the construction of the underground transmis-
holes may be installed. Field-poured manholes are espe- sion line, obstacles, such as major roadways, large utili-
cially useful in areas where other utility lines are nearby, ties, and railroads may be encountered that require
preventing the installing of a precast manhole. alternatives to open trenching. In cases where open
trenching is not an option, jacking and boring or hori-
Almost all transmission cable manholes have two zontal directional drilling (HDD) may be used. Discus-
accesses (chimneys) to facilitate cable installation. Man- sion of these installation methods are provided later in
holes include significant steel reinforcement and the cor- this chapter.
rect grade of concrete to ensure that there is no problem
with heavy truck traffic, and reinforcements are pro- Beginning in the mid-1980s, utilities began installing
vided to permit installing cable-pulling accessories such fiber-optic cables or spare ducts for future fiber-optic
as sheaves. Figure 12-23 shows a precast manhole being cables. The fiber-optic cables are used for cable system
lowered into place. communication or other utility requirements. In some
cases, the utility may make the fiber-optic cables avail-
12.3.2 Trenching able for outside communication company use. Recently,
additional ducts were added for temperature-monitor-
General ing purposes.
The most common method for constructing an under-
ground cable system is by open-cut trenching. This con- Extruded-dielectric Cables
sists of using a backhoe to remove the concrete, asphalt Extruded-dielectric cables can be installed either directly
road surface, topsoil, and subgrade material to the buried, as illustrated in Figure 12-25, or in duct, as illus-
desired depth, as shown in Figure 12-24. The material trated in Figure 12-26. General practice in North Amer-
can be stockpiled near the trench, if acceptable by the ica is to install the cables in ducts to facilitate removal
local authorities, or removed to an appropriate off-site and replacement in event of cable failure. However,
location for disposal or used for fill as appropriate. The directly buried cables have 10–15% higher ampacities
pipe/duct is installed in the trench and the trench back- than duct cables because there is no dead air space to
filled. Fill materials can be thermal sand and/or a con- impede heat transfer to the earth. Reference (Tarpey et al.
crete mix. Depending on the depth, shoring of the 1991) describes an extruded-dielectric cable installation.
trench may be required. The U.S. Occupational Safety
and Health Administration (OSHA) requires any trench The advantages of direct-buried cable systems are that
deeper than 5 ft (1.5m) to be shored unless a competent they are significantly less expensive to install than duct
Figure 12-23 Precast manhole being lowered into place Figure 12-24 Backhoe excavating a trench (courtesy
(courtesy POWER Engineers, Inc.). POWER Engineers, Inc.).
12-22
bank cable systems. Direct-buried cable systems can Duct banks designed for the utility industry are fre-
require that significant lengths of trench be left open quently located within or near public rights-of-way such
until the cable is installed. Repairs to direct-buried cable as along roads and highways, beneath sidewalks, or in
systems require the trench to be opened and the cables green spaces within the rights-of-way that are assessable
exposed, which in itself may further damage the cable or to the general public. It is common practice in these
even the adjacent good cables. When the trench is cases to encase the duct system in high-strength con-
located within commercial or residential property, land- crete of at least 3000 psi for protection from mechanical
owners can expect significant disruption to their busi- damage. The concrete encasement also provides a mea-
ness or property due to the excavation required to sure of safety for the personnel who may be installing
expose the cable system. Any type of cable can be direct another facility in close proximity to the duct system in
buried, provided that it is designed appropriately for the future. The latest edition of the National Electric
direct-buried applications. Safety Code should be consulted to ensure that all
applicable code requirements for minimum burial depth
Directly buried cables should be protected from above are met.
to reduce the chances of dig-in. Precast or free-poured
concrete slabs are typically placed 1 ft (30 cm) or so In some cases, the municipal authorities may allow the
above the cables. Trench widths typically vary from 36 installation of a duct bank system backfilled with a spe-
to 100 in. (91 to 254 cm), depending upon the number of cial soil or thermal sand instead of high-strength con-
circuits and spacing among cables, and depths vary crete. In this instance, the utility should install and
from 46 to 52 in. (117 to 132 cm). Trenches for directly maintain a “High-Voltage Buried Cable” warning sys-
buried cables often have controlled backfill installed to tem of signs indicating the hazard potential, and the
improve ampacity. Spacing between adjacent phases is utility should regularly patrol the cable route, especially
usually 9–12 in. (23–30 cm), and typical spacing in areas where excavation is most likely. A good practice
between circuits is 24 in. (60 cm) or greater. Fiber-optic to protect the ducts from mechanical damage is to place
cables can also be installed. In addition, ground conti- 4-in.-thick (10-cm) preformed concrete slabs on top of
nuity conductors can be installed in the trench to carry the thermal sand backfill directly over the ducts, or to
fault currents if required by the bonding scheme. simply pour a 3- to 4-in. (7- to 10-cm) layer of concrete
directly from the concrete truck about 1 ft (30 cm) above
It is possible to install single-conductor cables in duct the cable. Ducts installed in this fashion without con-
using the flat configuration shown in Figure 12-25. crete encasement should be at least Schedule 40 in thick-
Ampacity will be reduced because of the dead air space ness. Thinner walled ducts may become oval in shape
in the ducts. when encased in concrete if proper pouring practices are
not followed.
Figure 12-25 Typical directly buried Figure 12-26 Typical duct bank trench detail
trench detail (courtesy POWER Engineers, Inc.). (courtesy POWER Engineers, Inc.).
12-23
The designer should consider including one additional this is acceptable for lightweight distribution cables, it is
duct per circuit as a spare conduit to be used in the not satisfactory for transmission voltage cables.
event that a new replacement cable needs to be pulled in.
Should a cable experience a fault, the heat and duration Although more expensive that PVC, fiberglass duct has
of the fault may bond the cable to the duct material, the advantage of being less brittle and generally stiffer
prohibiting the removal of the cable. If there is a possi- than PVC, making it suitable for nonburied duct runs
bility of a future circuit being installed along the same that are supported at regular intervals between hanger-
alignment, consideration needs to be given to the type supports without significant bowing between sup-
cost/benefit of installing additional ducts during the ini- ports. Fiberglass is also a good alternative for installa-
tial installation. The incremental cost of including additions that will be above grade and exposed to the
tional ducts during the design stage of the project is elements. Fiberglass is highly resistant to weathering
small compared to the initial cost of excavating the with good UV radiation stability (sunlight resistance).
trench and repaving or relandscaping. If future circuits Fiberglass also has a lower coefficient of friction, allow-
are planned but not going to be installed at this time, the ing longer cable pulls with less reliance on lubricants.
ampacity requirements for the ultimate circuit build-out Another advantage of using fiberglass conduits is that
needs to be incorporated into the duct bank design. fiberglass has a lower coefficient of thermal expansion,
which makes it an excellent option for bridge attach-
A t y p i c a l d u c t b a n k c o n fi g u rat i o n i s s h ow n i n ments and other exposed installations.
Figure 12-26. The ducts are installed in a vertical
stacked arrangement. Spacers hold the ducts in position PVC conduits are available in several ratings: DB or
until a concrete envelope is poured around the ducts. Schedule 20, Schedule 40, and Schedule 80. Schedule 40
The concrete envelope provides mechanical protection is the most commonly used PVC rating, although many
and aids heat dissipation. Since concrete has a low ther- utilities will allow DB-rated PVC if it is encased in con-
mal resistivity, controlled backfill is not generally crete. DB is thinner walled and is more easily broken
needed outside the envelope; clean fill is usually ade- during construction activities, especially by workers
quate. The spare ducts may be used for fiber-optic stepping on the conduit and by concrete dropping onto
cables, ground continuity conductor, or replacement it during encasement. Schedule 80 is considered heavy
power cable. wall and is often used where above-grade protection of
the cable is important, such as cable risers on termina-
The duct materials primarily used today are polyvinyl tion structures. Be aware that, without special formula-
chloride (PVC), polyethylene (PE), and reinforced ther- tion for UV resistance, PVC will grow brittle with
mosetting resin (fiberglass). Because transmission volt- extended exposure to UV light.
age cables are larger and generally heavier than
distribution cables, and because it is economically justi- Further discussion on conduit installation is provided in
fiable to design longer pulls rather than shorter pulls, Chapter 13, Section 13.4.2.
the coefficient of friction of the duct material becomes
an important consideration in the selection of duct Trench widths can vary from 12 to 54 in. (46 to 137 cm),
material. The designer must consider the type of cable depending upon the configuration of the cable or con-
jacket that will be used on the cable to avoid selection of duit. Occasionally, wider trenches are used to keep the
two materials (cable jacket and duct material) that pro- depth of the trench as shallow as possible. Most utilities
duce a higher coefficient of friction, thus limiting the require the top of the concrete to be a minimum of 36 in.
length of the cable pull and potentially requiring an below grade. Typically, the minimum spacing between
additional manhole. For instance, a PE jacket and a PE the edges of the ducts is 3 in. (76 mm). There are spacer
duct material would produce a higher coefficient of fric- designs with this spacing. If a greater distance is needed,
tion than a PE jacket in a PVC duct. the spacer would have to be special ordered. The mini-
mum distance from the duct to the edge of trench
Most concrete-encased duct bank systems and direct should be 3 in. (76 mm). A distance greater than 3 in.
buried conduit systems use PVC duct. This material (76 mm) from the duct to the edge of trench may be nec-
comes in straight sections and, when joined together, essary, due to the trench depth, to facilitate the installa-
makes for an extremely straight duct system without tion of the duct into the trench, or to provide sufficient
undulations. On the other hand, PE duct can generally space for reinforcing steel. Reinforcing steel may be
be delivered to the site on reels. This material has suffi- installed if additional duct bank strength is needed, but
cient elastic memory to cause the individual duct runs to should never metallically encircle the duct bank.
undulate, which increases pulling tensions. Although
12-24
Pipe-type Cables than 5 ft (1.5 m) deep. In many cases, however, the depth
The cables for high-pressure fluid-filled (HPFF) and must be greater to avoid existing utilities, or to maintain
high-pressure gas-filled (HPGF) pipe-type cable systems sufficiently large bending radii, etc. Intersections require
are installed in steel pipes, which withstand pressure, greater depths because of the need to pass under exist-
provide protection, and facilitate replacing failed cables. ing utilities in the cross streets.
The pipe can be installed directly buried or encased in a
low-strength thermal concrete, as shown in Figure 12-27. The trench is commonly filled with a special low-ther-
Many utilities install two cable pipes per trench, with mal resistivity controlled backfill to aid heat dissipation
one line able to carry the full load for the duration of a and improve ampacity. The amount of backfill varies
repair on a failed companion circuit. Reference (Hatcher with native soil conditions. Many utilities standardize
et al. 1966) describes pipe-type cable installations. on 12 in. (30 cm) above the cables. See Section 12.2.6 for
a discussion of controlled backfill. The material above
Trench widths can vary from 18 to 54 in. (46 to 137 cm), the controlled backfill may be native soil if it has suit-
depending upon pipe size, whether there are one or two able thermal and mechanical properties. If the line is
pipes per trench, and whether welding is done alongside under pavement, roadbed sub-base is then installed, fol-
or in the trench. Occasionally, wider trenches are used lowed by the base pavement and final pavement.
for three-pipe installations, or for circuits containing
SCFF Cables
fluid-return pipes for forced cooling. Typically, mini-
SCFF cables are installed in a similar manner as
mum 24-in. (61-cm) spacing is maintained between
extruded cables. Reference (Kozak et al. 1965) describes
cable pipes to permit working on the pipes during instal-
an extensive SCFF cable installation.
lation and during possible future repairs. Greater spac-
ing also reduces mutual heating effects, thereby
12.3.3 Trenchless Installations
improving ampacity.
During the process of planning a route layout for an
Trench depth is selected to give a cover over the pipe of underground cable system, extensive research must be
36–42 in. (91–407 cm) if possible, since a shallower performed to locate any encumbrances that would pro-
depth provides greater ampacity. At these depths, shor- hibit open trench excavation and force the utility into
ing may not be required, since the trench will be less using various trenchless technologies: horizontal direc-
tional drilling, jack and bore, and tunneling/microtunnel-
ing. These encumbrances could be due to the following:
• Numerous subsurface services that require a protec-
tive barrier between them and the cable system to
be installed.
• Major street intersections that cannot be excavated
by the open trench method during certain times.
Some city governments do not allow any open trench
excavation or work during normal business or com-
muter hours in particular areas of their city. All work
must then be scheduled for the night hours.
• Interstate or state highway systems including major
city parkways and boulevards where open trenching
is not permitted.
• Railroad systems where there is no available existing
underpass that could be utilized.
• Water or wetlands crossings.
Trenchless installation costs are generally greater than

costs for open trenching. However, it is quite possible
to reduce some installation costs by comprehensive
planning. When roads, highways, parks, airports, rail-
roads, etc., are built or rebuilt, utility planners should
take the opportunity to install bores, tunnels, man-
Figure 12-27 Typical pipe-type cable trench cross section holes, or casings while open trenching is permitted,
(courtesy POWER Engineers, Inc.).
12-25
thus reducing future installation costs. Being aware of General

current and planned civil construction projects and The HDD method consists of three processes. First, a
comparing them to anticipated cable routings can result small-diameter pilot hole is drilled from entry to exit.
in significant savings, depending on the present worth Next, the pilot hole is enlarged by reaming. Finally, the
of future construction. product pipe is pulled into the enlarged hole. Each indi-
vidual process is described in detail.
Descriptions of the various trenchless technologies are
given in the following subsections. There are several factors to be considered in the design
process of an HDD crossing. Among these are depth of
Horizontal Directional Drilling (HDD) cover requirements, underlying soil conditions, existing
Horizontal boring operations have become quite popu- utilities, entry and exit angles, radius of curvature, and
lar with utilities. This began with small directional bor- compound curves and available drilling and pull back
ing operations for utility distribution services (electric, space. Each individual design factor is discussed in
gas, and telephone). Large directional boring machines detail below. Equations used in the design process are
were then developed for long lengths of gas and oil also provided.
transmission pipe systems. They were previously only
used for river or water crossings where the cost of exca- The first to be discussed is the depth of cover. A plan
vation would exceed that of a directional bore, or where and profile view of a typical HDD crossing is presented
excavation of the river bottom would present environ- in Figure 12-28.
mental problems. This type of boring is now used by
electric utilities for underground transmission projects The underlying soil strata of an HDD should be investi-
where conventional trenching is not an alternative. The gated to determine the suitability of the HDD method.
North American Society for Trenchless Technology Soil samples should be taken at each end of the crossing
(NASTT) has prepared a guide entitled Horizontal and at the center point, and/or the deepest point, along
Directional Drilling Consortium HDD Good Practices the crossing. This information will allow the designer to
Guidelines (NASTT 2004). determine if there are any unusual or unexpected mate-
rials that the drill string will encounter. This informa-
HDD can be a preferable construction method because tion will enable the designer to plan for the appropriate
of its minimal impact to the surrounding environment. equipment and drill head, select the casing material, and
One of the drawbacks to using a HDD is the impact understand the risks of the operation before mobilizing
on the circuit ampacity. A typical HDD installation to the site.
results in the cable being installed at least 20 ft (6.1 m)
below grade. Soil conditions such as silts, sands, and clays tend to be
well suited to directional drilling. Soils containing gravel
HDD installations are used for both pipe and duct and cobble make all phases of the HDD process much
installations. Typically, the longer installations (greater more difficult. Even though facilities have been installed
than 1.5 miles, or 2.4 km) use pipe-type cable because of in numerous types of rock formations, HDD in rock is
two main factors: (1) the pulling forces are distributed oftentimes much more costly than a comparative instal-
over three cables, instead of just one, and (2) pipe-type lation in soil. This is due to the need for specialized
cables frequently have a smaller diameter, thus, allowing rock-drilling tools. These tools are discussed in detail in
more cable to be carried on one reel; therefore longer a later section. During the design phase, a geotechnical
distances can be installed. exploration of the subsurface should be investigated to
determine the suitability of the HDD method. The
depth of the installation is determined primarily by the
Figure 12-28 Plan and profile view of a horizontal directional drill (courtesy POWER Engineers, Inc.).
12-26
type of soils encountered. One of the primary design Horizontal Distance of Curve = R (sin α ) 12-2
concerns is the potential of “frac out,” the loss of drill- Where:
ing fluid, during the drilling process. Loss of drilling R = Radius of Curvature
fluid can cause significant environmental problems. The α = entry or exit angle (degrees)
deeper the installation, the less likely a “frac out” will
occur, but the greater the impact on circuit ampacity. Vertical Distance of Curve = R (1 − cos α ) 12-3
Where:
Existing utilities to be crossed and other underground R = Radius of Curvature
facilities and structures in the proximity of a planned α = entry or exit angle (degrees)
HDD crossing must be identified, just as with any Rπα
underground digging activity. Length of Curve = 12-4
180
Where:
The entry angle of the directional drill needs to be
R = Radius of Curvature
understood when planning the route and the set-up
α = entry or exit angle (degrees)
area. The entry angle is defined as the angle that the bit
enters the ground relative to horizontal. Most direc-
Compound curves should be avoided in the design
tional drilling rigs are capable of an entry angle of
of HDD crossings whenever possible. A compound
between 8 and 16°, although many drilling companies
curve occurs when there is a change in direction in both
prefer an entry angle of 11 to 12°. A steeper entry angle
the horizontal and vertical planes simultaneously. In
allows the drill path to reach its maximum depth rela-
certain cases, it is an unavoidable situation. A method
tive to the entry point in a shorter horizontal distance.
for estimating the compound radius is given in the fol-
The exit angle must also be considered. Exit angles typi-
lowing equation:
cally range from 6 to 20° or more relative to horizontal.
Shallower exit angles are generally reserved for larger-
( Rv ) 2 * ( R h ) 2
diameter steel pipeline installations. Shallow exit angles Rc = 12-5
(6 to 10°) reduce the amount of pipe-handling require- ( Rv ) 2 + ( R h ) 2
ments (cranes, side booms, etc.) on the exit side. The
downside is that a greater horizontal distance is Where:
required when small exit angles are used, which Rc = Compound Radius of Curvature
increases the length of the crossing. Steeper exit angles, Rv = Vertical Radius of Curvature
defined as in excess of 15° are used where distance con- Rh = Horizontal Radius of Curvature
straints are a prime consideration in the design process.
Steep exit angles also increase the need for additional Site Layout
pipe-handling equipment. In addition to the necessary space required to set up the
drilling machine, sufficient space is needed for preparing
One of the most important factors in the design of an the duct/pipe for pull back. This pullback area can be
HDD crossing is the radius of curvature. This is of narrow, but can be extremely long. Optimally, the length
greater importance when steel pipe is used. A general of the pull back area should equal the length of the drill;
rule-of-thumb in the design process is 100 ft (12 m) of however, sometimes that may not be possible. Another
radius per inch (cm) of outside diameter (OD) of prod- thing to consider is the weight of the pipe/conduit that is
uct pipe. For example, the radius for a nominal 10-in. being installed.
OD (25.4 cm) steel pipe would be 1000 ft (304.8 m).
Larger radii are desirable if sufficient room is available The size of the drilling area is dependent on the type of
in the length of the crossing. For high-density polyethyl- drill rig required. HDD rigs are available in many sizes,
ene (HDPE) pipe, much tighter radii can be used. As a ranging from 7000 lbs (31,150 N) of pullback force to
rule-of-thumb, the minimum radius of curvature for 1,000,000+ lbs (4450 kN). Also, the rigs have rotary
HDPE pipe is given as 50 times the outside diameter. torque ra nging from approximately 4000 ft-lbs
Radii this small are generally not used, because the steel (5420 N-m) to in excess of 80,000 ft-lbs (108,465 N-m).
drill pipe used to drill the pilot hole is incapable of toler- The ancillary equipment that is used with the rig is a
ating relatively small radii. Equations used in the design mud pump, mud-cleaning system, drill pipe, a tempo-
of HDD drill paths are presented below. rary site office, a control unit for the rig, power unit for
the rig, water supply, and a tool/spares container. A typ-
12-27
ical HDD site layout is shown in Figure 12-29. A typical common casing material is HDPE. Steel casings are also
footprint would be about 150 x 250 ft (45.7 x 76.2 m). used, but use of steel casings may result in a significant
A photo of an HDD rig is presented in Figure 12-30. reduction in ampacity due to additional losses from hys-
teresis and eddy current losses in the steel pipe. Also an
Extruded-dielectric Cables
HDPE casing is less likely to affect the tracking of a
Extruded cables are always installed in inner ducts
parallel bore as compared to a steel casing. The next
within a bore, both with and without casings. Though
step is to determine the size and thickness of the casing.
technically feasible, a cable system has never been
The diameter of the casing is determined by the number
installed directly into a bore: most owners consider this
and size of the ducts that will be installed in the casing.
risky since the cable could be damaged during installa-
The casing thickness is determined by the calculated
tion, and there would be no opportunity to pull out
stresses associated with the HDD design, as discussed
damaged cable in the event of a failure. Historically, the
previously. A plastic casing offers some advantages over
method of installing single-core cable with HDD tech-
steel in that the material is softer and more flexible, so
nology was to install an outer casing with the individual
bending radii may be smaller than a comparable steel
conduits installed inside.
casing. However, a steel casing has a higher tensile load
limit for pullback.
Installing a cable system without a casing has some
advantages, including avoiding the cost of the casing
When pulling the duct bundle into a casing, special
and possibly smaller trenchless bending radii, perhaps
spacers with rollers are usually used to maintain the
meaning the length of the bore is shorter since the con-
conduit spacing and lower the forces required to pull in
duit system can get down to depth more quickly without
the duct package. The most common material used for
as much setback. When no casing is used, the conduits
ducts within casings is HDPE, although PVC (polyvinyl
or cable pipes are installed directly during pullback, and
chloride) and fiberglass can also be used. Once the cas-
there is no opportunity to place an engineered grout
ing and ducts have been installed, it is recommended
backfill around the conduits or cable pipes. If the
that the casing be filled with a thermally approved grout
ampacity goals can be achieved, this is typically the low-
to eliminate the air pocket inside the casing and improve
est-cost alternative. In addition, there is one less pull-
the ampacity of the cable circuit. Depending on the
back step since the inner ducts or cable pipes do not
length and depth of the HDD installation, this can be
need to be pulled back into the casing pipe after the cas-
very problematic. The main design concern is the grout
ing is installed.
pressure, associated with filling the casing with grout.
Casing Design Considerations The designer needs to be sure that the duct material can
A number of design considerations are associated with withstand the grout pressure. If grouting is not used, the
installing a casing using the HDD method. The first designer must consider the derating effect of the air
consideration is to determine the casing material. A pocket on the circuit ampacity.
Figure 12-29 Typical HDD site layout (courtesy POWER Engineers, Inc.).
12-28
Bundle Design Considerations stresses. This may cause some problems when the
The main design consideration for installing the ducts thicker HDD cable pipe is connected to the normal
as a bundle is to adequately size the ducts to withstand cable pipe. It is recommended that the HDD cable pipe
the installation stresses. This normally requires a larger be extended to the manholes. Otherwise, a special cou-
wall thickness for the duct than what is needed for a pling is needed.
normal trench installation. Because the larger wall
thickness affects the inside diameter, a larger duct may The recommended coating for an HDD cable pipe is a
be needed. This will cause some problems with the inter- fusion-bonded epoxy with an external POWER-
face of the HDD duct to the normal trench duct. Either CRETE™. This provides a stronger coating to with-
the HDD duct is taken directly into a manhole, or a spe- stand the abrasions that could occur during the
cial coupling is needed to make the transition. The con- pullback. Another option is to increase the normal
cern with the transition coupling is to make sure the coating thickness.
interface is smooth to prevent damage to the cable dur-
SCFF Cables
ing installation. The primary duct material used for this
SCFF cables are installed similar to the method
type of HDD installation is HDPE due to the strength
described above for the extruded-dielectric cable.
of the joint. The HDPE is usually butt-fused together,
Because of the associated depth, it is important to con-
resulting in a very strong joint. Depending on the length
sider any pressure-related issues.
of the drill and the type of geography, the designer may
consider pressure-grouting the hole to improve the ther- Jack and Bore
mal environment. If grouting is not used, the designer Another trenchless alternative is boring, which involves
must consider the thermal properties of the drilling mud augering a holes and jacking a casing into the hole
to determine the impact to the circuit ampacity. simultaneously. The boring method is generally used to
provide an opening for cable systems under roadway
Pipe-type Cables
crossings, street intersections, and railroad crossings,
Generally, an outer casing is not needed for pipe-type
where no bends are necessary for the cable route. Bores
cables, unless specifically required by local regulations.
are normally less costly to install than other trenchless
For all cable systems, but particularly for pipe-type
methods, because less manual labor is required and they
cables, the casing offers protection and a degree of sec-
can be installed faster.
ondary containment in the event of a dielectric fluid
leak. If no casing is required, the cable pipe is pulled General
directly into the hole similar to a casing. The thickness One of the main design considerations with a jack-and-
and coating of the pipe are the main design concerns. bore operation is to ensure that there is adequate space
Depending on the length of the drill, the pipe thickness to perform the bore. To initiate a casing installation, a
may need to be increased to withstand the installation pit having a minimum size of 40-ft (12.2-m) long by 10-
to 15-ft (3- to 4.6-m) wide is required by the boring
equipment and for placing and welding 20-ft (6.1-m)
sections of casing pipe. Also, prior to starting the boring
process, an exit pit approximately 10 ft (3 m) in length
should be excavated. Since a bore is basically at a
greater depth than an open trench, the entrance and exit
pits require shoring (and possibly tight sheeting) in
accordance with OSHA regulations. Secondly, if the
boring is to be done in poorly consolidated soil (soil that
begins to slough or flow if unsupported after a few min-
utes), solid sheeting is required to enclose the entire
entrance pit, allowing only an opening for inserting and
installing the casing and auger.
Casing sizes can vary for cable systems from 14 to 84 in.

(36 to 213 cm), depending upon the type of cable sys-
tem, number of circuits being installed, and length of
the bore. It is suggested that the design engineer select a
casing size with sufficient space or clearance to permit
an easy and thorough backfilling operation after the
Figure 12-30 Horizontal directional drilling rig (courtesy
cable system has been installed. Generally, casings
POWER Engineers, Inc.).
12-29
should be sized 8 in. (20 cm) larger than the OD of the with rollers are recommended to allow for easier instal-
cable or cable pipe. Borings should be limited to lengths lation of the conduit bundle. Detailed ampacity calcula-
less than about 400 ft (122 m). The longer the bore, the tions should be made to determine the optimal duct
more difficult it is to control the direction of the bore spacing for each installation. Typically, a minimum sep-
and to fill the casing after the pipe/duct installation. aration of 3 in. is recommended.
Pipe-type Cables
Selection of the casing material is also very important.
Normally, only one cable pipe is installed in a single cas-
Historically, most bores have been installed with steel
ing. Polymer spacers are needed to protect the pipe coat-
casings, but there has been a trend to nonmetallic cas-
ing by keeping it off the casing. The spacer runners can
ings such as concrete, HDPE, fiberglass-reinforced con-
be from 0.75 to 12 in. (19 to 305 mm) in length. With
crete, and composites. The designer must consider the
longer runners, other ducts can be attached to the pipe
derating effect on cable ampacity when using a steel cas-
and pulled in with the cable pipe. These ducts are typi-
ing or a concrete casing containing steel reinforcing.
cally located on the outside of the spacer between the
Using a steel casing may reduce the circuit ampacity by
runners and tied to the cable pipe using nylon ties.
5–10%, for non-pipe-type cable systems. Each of these
materials has its own advantages and disadvantages and SCFF Cables
may not be applicable to a particular application. Addi- SCFF cables are installed similar to the method
tionally, the owners of the facilities granting the crossing described above for extruded-dielectric cables.
permit will frequently have specific requirements for the
casing materials, thus, requiring the designer to use a Tunneling/Microtunneling
specific material. Contractors specializing in jack-and- Other trenchless alternatives are tunneling or microtun-
bore installations are the best source of information neling, which involve digging and building the tunnel
regarding the application of the various casing materials simultaneously. These methods may be selected if the
and their limitations in particular soil types. size of the opening is too large to be accommodated by
either HDD or jack and bore.
Casing diameters for typical cable systems will usually
be between 14 and 84 in. (36 and 210 cm), depending on Conventional tunneling is similar to pipe jacking. The
the type of cable system, number and size of inner con- tunnel is excavated manually or mechanically, and the
duits, number of circuits being installed, and length of spoil is removed using an auger system, carts, or con-
the bore. Additionally, the wall thickness of the casings veyor. The difference is in how the pipe is installed. In
will vary considerably, depending on the casing material the case of tunneling, the pipe or tunnel liner is added at
selected. For instance, a polyethylene casing will typi- the face of the tunnel. A temporary casing precedes the
cally have a wall thickness of 2 to 3 in. (5 to 7 cm), liner plates, which are assembled at the face of the tun-
whereas a typical steel casing may have a wall thickness nel and added on. The temporary casing is then jacked
of 0.5 in. (1.25 cm). Therefore, the resulting inner diam- forward against the liner, as opposed to the entire length
eter of the casing may be smaller for a nonsteel casing being jacked forward. Additional pieces are added, and
than for a steel casing. It is suggested that the design the procedure continues. The diameters of the casings
engineer select a casing size with sufficient space or are usually greater than 3.5 ft (1.2 m), with essentially
clearance to permit an easy and thorough backfilling no upper limit. The driving distance also is unlimited.
operation after the cable system has been installed. Gen- This method may also be selected if cooling of the cir-
erally, casings should be sized 8 in. (20 cm) larger than cuits is required to achieve the desired loading require-
the outside diameter (OD) of the cable or cable pipe. ments. This method is very costly and should only be
used when the other methods cannot be used effectively.
Once the casing is installed, the pipe/duct is installed
using specially designed spacers. The pipe should then Microtunneling is a horizontal boring technique that is
be filled with a thermally approved grout. Since the used in instances where the diameter of the final casing
grout needs to be pumped into the casing, consideration is greater than what is customarily used in traditional
must be made to ensure that the grout is pumpable and jack-and-bore operations or where rock is encountered.
still retains good thermal properties. Microtunneling uses highly sophisticated, laser-guided,
remote-controlled equipment. This enables the user to
Extruded-dielectric Cables monitor its precise location and to ensure accurate line
Since the number and size of ducts vary between and grade. This system typically consists of a steerable
projects, the duct spacers are typically designed on a tunneling machine that is jacked from a shaft ahead of
project-to-project basis. The spacers are needed to the permanent pipes or a temporary casing. The spoil is
ensure the proper separation between the ducts. Spacers removed by using either the auger method or the slurry
12-30
method. The auger method uses a continuous flight of contract, but not damage the coating. Analysis should
augers to remove the spoil. The auger is surrounded by be made on the cable pipe to determine the design limi-
a casing placed inside the permanent pipes or the tem- tations. The pipe of pipe-type cables can be restrained,
porary casing. The slurry method uses a slurry mixture and expansion can be accommodated, but restraining
to pump the spoil back to the surface. the cables themselves is difficult in applications where
there are large elevation changes. Riser-type cable,
Any of the cable types may be used in tunnels. Com- which has long-lay stainless-steel outer tapes to hold
pressed-gas-insulated cable can be an attractive most of the cable weight, may be required. As described
approach for tunnel installations, especially those with in Chapter 8, special anchor and skid joints are required
large elevation changes, since there is no fire danger and with the riser cable.
no problem with hydrostatic pressures. Ventilation is
SCFF Cables
important should a gas leak occur. Thermal expansion
SCFF cables are installed similar to the method
is an important consideration since the aluminum enclo-
described above for extruded-dielectric cables.
sure has a high coefficient of thermal expansion. It may
be necessary to anchor every three to five 40–60 ft
12.3.4 Bridges
(12–18 m) sections, and to provide expansion relief in
the joints and enclosure between anchors. Further dis- Existing or new bridges can be an attractive alternative
cussion on tunneling and microtunneling is provided in to submarine cables for some water or railroad cross-
Chapter 13, Section 13.3.1. ings. When considering installing cable on a bridge, it is
very important to contact the owner of the bridge and
General obtain all available design drawings. Some existing
It is important to note that, because of the size of the bridges are not designed to accommodate additional
tunnels, it is not practical to fill the tunnels, so special attachments. So, it is very important to have an existing
consideration needs to be taken to calculate the ampaci- bridge analyzed to determine whether it can handle the
ties of the circuits based on the cables being installed in additional load. Another concern is the possibility that
air. Forced ventilation may be needed to achieve the the bridge may need to be rebuilt in the near future. This
desired ampacities. may cause additional cost to the utility, because the util-
Leak detection and fire protection are very important

for tunnel installations of HPFF and SCFF cables. Leak
detection wires may be installed alongside the cables,
and fluid sensors can be installed in joint bays. Smoke
and fire monitoring can be provided using conventional
detectors or newly developed, fiber-optic temperature-
monitoring cables. For extensive tunnel installations in
Japan, the cables are typically placed in sand-filled
troughs, which provide cable restraint, contain possible
leaks, and provide fire protection (Nakagawa et al.
1973). Concrete protection has been provided in other
tunnels (Chaaban et al. 1991). Figure 12-31 shows an
SCFF installation in a tunnel.
Extruded-dielectric Cables
A design concern for installing a single-core cable in a
tunnel is to determine how the cable is to be supported.
Extruded-dielectric cables are normally installed in tun-
nels on support brackets spaced at 5 to 10 ft (1.5 to 3 m)
intervals, attached to the sidewall or to floor-mounted
stanchions. By allowing the cables to hang in a catenary
between supports, the TMB forces are relieved.
Pipe-type Cables
As with the single core, a major design concern is the
method used for supporting the cable pipe within the
tunnel. Because of the cable coating, it is important to Figure 12-31 Cable installation in a tunnel (Ray et al. 1974)
design a system that allows the cable pipe to expand and (courtesy IEEE).
12-31
ity may be required to relocate the cable. An additional Pipe-type Cables

alternative to attaching to an existing or new bridge is to Many short, pipe-type cables have been installed on
install a utility bridge. Some design considerations for a bridges in North America (Sheirer and Winistsky 1958;
separate utility bridge are the following: Gillette 1959). Pipe expansion is accommodated by 90°
• Protection from sun. Solar radiation will derate or S-bends, or less commonly by expansion joints. Cable
weight, pipe expansion, bridge expansion joints, and
the cable.
corrosion protection from road salts are design consider-
• Physical protection. Consideration should be made ations. Additionally, where the pipe is allowed to move
to protect the cable system from potential through its expansion range, Teflon-coated skid plates or
external damage. similar material can be used to protect the pipe coating.
• Cable/Duct radii. In many cases, sharp radii may be
SCFF Cables
required to attach to a separate utility bridge. This, of
There are significant numbers of SCFF installations on
course, will affect the pulling tensions and sidewall
bridges outside North America (Marquez et al. 1981),
pressures on the cable.
including a 14-mile (22-km), 500-kV line installed in the
General late 1980s (Minemura and Maekawa 1989). During
Cable/pipe clamps, cable weight, vibration, and methods winter months, SCFF cables can experience excessive
to account for bridge expansion joints are the major pressures when load is restored after the cables are
design issues. Fairly elaborate schemes have been de-energized, because of high fluid viscosity at low
employed to account for the large motion that occurs at ambient air temperatures on the bridge. Lower viscosity
bridge expansion joints (Minemura and Maekawa 1989). fluids and larger fluid areas are used to reduce the pres-
sure problems.
Compressed-gas-insulated cables are not generally
considered for bridge use because of their large SCFF cables are installed similar to the method
space requirements. described above for extruded-dielectric cable.
Extruded-dielectric Cables 12.3.5 Underwater Installations

Single-conductor cables are more easily installed on
Underwater power transmission cables have been success-
bridges than pipe-type cable, since bending can be
fully installed for over 65 years. This experience has
accomplished by cable snaking, and elaborate expan-
included all of the transmission cable types (with the
sion joints are not required. If the bridge is short, ducts
exception of compressed-gas-insulated and superconduct-
may be used. The advantage of ducts is that they pro-
ing cables) in a wide range of installation environments.
vide added mechanical protection and make installing
the cable easier. The main concern for installing ducts is
While the longest submarine cables are dc, primarily
to have a duct that will not degrade over time due to UV
because of charging-current limitations associated with
rays. Fiberglass duct is the preferred duct material for
long ac cables, the majority of submarine cable installa-
bridge applications. Figure 12-32 shows a fiberglass duct
tions are ac. Chapter 7 provides a review of worldwide
bridge crossing.
submarine cable experience. Submarine cable installa-
tions involve a variety of potential hazards during
installation, operation, and even during retrieval for
repair. They are far more expensive to install than land-
based cables and extremely expensive to repair. Thus,
careful planning and thorough engineering of underwa-
ter cable installations are essential.
Installing cables underwater involves many technical

and economic design considerations beyond those in a
conventional underground cable installation. Selecting a
cable type and designing it specifically for underwater
installation involve more than defining the power trans-
mission capability and the desired transmission voltage.
Requirements for installation, operation, effects of
marine life and environment, fault location, cable
Figure 12-32 Bridge crossing with fiberglass duct (courtesy retrieval, and cable repair, as well as the effects of the
POWER Engineers, Inc.). installation on the environment, are all factors to be
12-32
studied and evaluated in the planning stage of an under- surizing plant tank or reservoir can provide a warning of
water installation. This is necessary to select the correct potential sheath damage. However, the loss of dielectric
cable, cable configuration, route, handling equipment, fluid is an environmental concern. These cable types are
and installation method. only feasible for moderate water depths and short-to-
moderate lengths. Flow rate and pressure drop con-
Typical Design Aspects of Submarine Cables straints related to transferring fluid in and out of the
Cables for underwater applications are similar in their cable due to load changes preclude using these cables for
basic electrical design to conventional underground really long lengths, generally more than 30 miles
cables, but they differ in their exterior mechanical (48 km). As the installation depth increases, the cable
design, sheath requirements, and electrical bonding must be maintained at a higher and higher pressure to
methods. The most significant difference involves the prevent the ingress of water should a leak occur, since
addition of armor, both to protect the cable from exter- the specific gravity of the dielectric fluid is somewhat
nal mechanical damage and to provide mechanical lower than water. The above-surface portion of the cable
strength to allow for cable installation and retrieval. The and its accessories must be designed to withstand the
cable jacket and sheath design also have a critical role in fluid pressures necessary to provide slightly positive
ensuring the integrity of the cable. A more detailed dis- internal pressure at the deepest point in the submarine
cussion on submarine cable system design is provided in installation. For example, an 80 psi (550 kPa) pressure is
Chapter 7. The following are key design aspects. needed for the 1312-ft (400-m) depth of the Vancouver
• Spacing Island cables.
• Mechanical strength and protection High-Pressure Fluid-Filled Pipe-Type Cables
• Sheath/jacket requirements Pipe-type cable for submarine application has the same
construction as for land-based cables. Three impreg-
• Corrosion protection
nated-paper-insulated conductors are installed in a pipe,
• Sheath bonding which is filled with dielectric fluid and pressurized.
• Cable losses Design differences for submarine application may
involve size and strength of skid wires, pipe wall thick-
Types of Submarine Cable ness, and concrete pipe coating. Heavier-walled pipes
The cable/utility industry has had successful submarine are often used (0.5-in. or 0.375-in. vs. 0.25-in, wall
experience with a variety of cable types, including [12.7-mm or 9.5-mm vs. 6.4-mm wall]), and provide a
high-pressure fluid-filled pipe-type cables (HPFF), self- number of benefits, including increased mechanical
contained fluid-filled cables (SCFF), conventional strength, greater weight to reduce buoyancy, better abil-
solid-type mass-impregnated paper-insulated cables, ity to withstand a corrosive or electrolysis condition in
paper-impregnated gas-pressure-assisted cables, and the event of a coating failure, and reduced probability of
extruded-dielectric insulated cable. Selection of cable burn-through during a fault. A concrete coating is often
type depends on the desired transmission operating added to the exterior of the pipe for mechanical and
voltage and power transfer level, and the depth, length, corrosion protection and for negative buoyancy.
and profile of the circuit. The alternative cable types are
briefly described below. The advantage of the pipe-type design is that the pipe
Fluid-Filled, Fluid-Pressurized Cables provides mechanical protection for the cables and allows
The highest-voltage submarine cables are those that are the three phases to be installed in close proximity. This
impregnated and pressurized with dielectric liquid (pipe- minimizes the installation trench size, spoils removal,
type and self-contained fluid-filled). Pipe-type and and construction time, and reduces environmental
SCFF cables are considered suitable for voltages up to impacts during construction. Pipe-type cable is not nor-
765 kV. These cables possess three advantages: their mally considered for long underwater crossings, because
high power transfer; their ability to match overhead it is limited by the mechanical forces required to pull the
transmission voltages, which means transformers and cables into the pipe. The installed length cannot contain
other substation equipment could be eliminated at the field splices; splices do not fit in the pipe, nor do they
land /submarine transitions; and the fact that they are have the strength to be pulled into the pipe. In the event
self-protecting and provide warning in case of damage of an electrical failure, the cables can be pulled out and
to the sheath or pipe. In case of limited sheath damage, replaced from the land ends, with no need to deploy sur-
the pressurized systems will pump dielectric fluid out at face repair ships or perform at-sea splices. The longest
the site of the damage, minimizing water and other con- section of submarine pipe-type cable, installed across
taminants from entering and traveling up into the cable. Long Island Sound near New York City, is 6900-ft
Fluid pressure drop or reduced fluid levels in the pres- (2100-m) long (Bazzi 1982). One of the more ambitious
12-33
projects was a 230-kV, 2-mile (3.2-km) crossing of the This is only true if gas pressure is lost; the cable must be
Baltimore Harbor with five cable circuits and two designed for this contingency. The maximum voltage for
splices per circuit, which were made on platforms in the this cable is approximately 275 kVac and 250 kVdc
harbor (Ruekert and Bien, Jr. 1979). (Bazzi 1982).
SCFF Cables Extruded-dielectric Cables
Single-conductor cables of this type utilize a conductor This cable type has been used commercially for land
with a hollow core for feeding dielectric fluid to the installations at 230, 345, and 400 kV, and short sections
cable, which is insulated with paper tapes impregnated of cable are presently operating under test at 500 kV. It
with a low-viscosity cable fluid. Three-conductor consists of a conductor insulated with extruded solid
designs, which are not normally utilized above 138 kV, dielectric, such as cross-linked polyethylene or ethylene-
provide ducts in the spaces between conductors for flow propylene rubber (up to 138 kV). The advantage of this
of dielectric fluids. The fluid can be pressurized to over construction is that the dielectric is truly self-contained.
200 psi (1380 kPa) during operation. The insulated con- Without the need for internal pressurization, a heavy
ductor is covered with a lead-alloy sheath, insulating wall sheath is not required, further reducing the cable
jacket, bedding, and armor. This design can be installed size and weight. Length is also not limited by fluid-
to depths of 3280 ft (1000 m). The length of the cable feeding limitations. However, the dielectric performance
depends on dielectric fluid-feeding limitations. With of cross-linked polyethylene is significantly weakened in
very low-viscosity fluids, large duct size, and high fluid- the presence of even microscopic amounts of water.
feed pressures, this cable could be installed in lengths of Thus, although there is no internal pressurization, a
over 20 miles (32 km). watertight metallic sheath is still critical for cross-linked
polyethylene cables. The major disadvantage is the basic
Impregnated-Paper, Nonpressurized, Solid Cables
performance of the insulation. Unlike laminar impreg-
This cable type consists of a conductor insulated with
nated-paper dielectrics, extruded dielectrics are not for-
paper tapes impregnated with a very viscous dielectric
giving of defects in the insulation. One contaminant or
fluid. The insulated conductor is covered with a lead-
void can cause an eventual breakdown of the cable.
alloy sheath to prevent water ingress. An insulating
Thus, if extruded insulation is to be used for a subma-
jacket, bedding, and armor wire are applied outside the
rine transmission cable, great care should be taken to
lead sheath. This cable is considered suitable for opera-
ensure that manufacturing quality and factory testing
tion up to 45 kVac and 250 kVdc. Without changes to
are performed using state-of-the-art materials, equip-
the conductor and armoring design, this cable is limited
ment, and techniques.
to installation in water depths of approximately 1500 ft
(500 m) (Bazzi 1982). Its greatest advantage is for use on Splicing Submarine Cables
long submarine cables. It can be installed in very long The designs of splices for submarine cables differ from
lengths, because it is nonpressurized, and thus limita- those for land-based cables due to the mechanical
tions regarding fluid-feeding pressures and fluid chan- requirements associated with installation and retrieval,
nel size do not pertain. The disadvantage of the as well as operation under high water pressures. The
nonpressurized insulation is that the maximum operat- required depth of installation and the method to be used
ing voltage is limited, due to ionization limitations, par- during installation are critical. For example, installa-
ticularly with alternating current. Also, its maximum tions where the splice must pass over a capstan ten-
operating temperature must not exceed approximately sioner, through a linear cable tensioner, or over a
60°C or the impregnant in the cable will drain, causing moderately sized overboarding sheave, often place
dielectric failure. severe design constraints on the maximum diameter and
length of a splice. Therefore, SCFF and extruded-dielec-
Gas-Filled, Pressurized Cables
tric submarine cables are usually supplied in a single
This cable design utilizes a hollow-core, copper conduc-
length, with smaller “factory splices” being used to join
tor wrapped with paper insulation that is impregnated
factory lengths of conductor. In three-conductor sub-
with a high-viscosity (jellylike) liquid. The cable is pres-
marine cable, these factory splices are usually staggered
surized with gas. The advantages of this design, com-
over a length of 25–50 ft (7.6–15.2 m) to keep overall
pared to liquid-pressurized cables, are that the duct size
dimensions to a minimum. Shallow-to-moderate water
can be very small, thus reducing overall cable size and
depths, where the splice can be lowered off the side of
weight, and extremely long lengths of cable are feasible,
the cable installation vessel, would not necessarily
since no dielectric liquid-feeding limitations exist.
restrict the splice size in the same way, nor would the
Installation depth is limited to approximately 1300 ft
mechanical stresses be the same.
(400 m), based on gas pressure limitations. Below that
depth, the cable could be crushed due to water pressure.
12-34
The electrical design of a splice for submarine cable is There are three different types of splices to be consid-
no different from that of a land-based cable in terms of ered: flexible factory splices, field splices (at sea), and
maximum allowable radial and tangential electrical repair splices. Flexible factory splices are provided to
stresses. Requirements of the cable-handling equipment allow the maximum length of cable to be installed with-
tend to be primary factors in the configuration of the out the need for splicing at sea. Factory splices are supe-
splice. For example, limitations of the cable-tensioning rior to field splices in that their armor can still be applied
equipment make it desirable for the submarine splice to continuously, splice free, and they are made in highly
be the same or close to the diameter of the cable. controlled conditions with trained factory technicians.
Whereas splices for land application typically have a
m a x i m u m i n s u l at i o n o u t e r d i a m e t e r ( O D ) o f At-sea splices present greater risks, in that the splicing
1.5–2.0 times the cable outer diameter, splices for deep- must be accomplished on the vessel in varying weather
water application are often designed as a “reconstruc- and sea conditions, and at-sea splices involve the recon-
tion” of the cable insulation, matching the cable insula- struction of conductor, insulation, shields, jackets, and
tion outer diameter. Thus, the length of splice tapers, armor. Extra cable is needed to allow the cable to be
the dielectric tapes used, and the method of application brought up to the vessel, resulting in a “U-bend” in the
are defined in large part by the desired splice OD. cable. While splicing is performed in a temperature- and
humidity-controlled enclosure, the cable-tensioning
The mechanical design of the splice is also determined equipment and vessel-positioning systems must be capa-
primarily by how it is to be lowered to the sea bottom. ble of holding the cables steady during the splicing oper-
Splices that must pass through or over a cable-tension- ations. Heavy seas and severe deck motions can
ing device are designed to be “on size” with the cable. obviously affect the quality of the splicing or, at least,
These cable splices often use armoring similar to that of extend splicing time. In very heavy seas, it is sometimes
the cable. The armor wires from one side of the splice necessary to cut the cable free to protect the cable vessel
are hand-applied to the splice area and individually and its crew. At-sea splices should be completed
welded to the cable armor to effectively reconstruct as quickly as possible to minimize exposure to bad
the cable. weather and sea conditions; if possible, they should be
avoided entirely.
Cable splices to be installed by lowering over the side are
often installed in a submarine splice casing. Making the Even if a submarine cable is designed and installed with-
splice is similar to that for a land-based cable, including out splices, a repair splice design must be available in
installing a lead sleeve over the complete splice. The cas- case of damage to a cable in mid-run, unless replace-
ing has clamps on both sides, both to hold the cable ment of the total length of submarine cable, shore to
armor wires and to take the mechanical load from the shore, is economically and logistically feasible. There
armor wires. After the splice is made and installed inside have been cases where the submarine cable has been
the casing (lead sleeve and all), the casing is filled with damaged during the initial installation, in one case by
semi-hard insulating compound to provide further the anchor of the installation vessel. Thus, the repair
assurance that the splice is mechanically protected and splice design should be completed before the installation
water ingress is prevented. begins, and a suitable number of repair kits should be
available at the time of installation.
Generally, it is desirable to minimize or eliminate alto-
Underwater Installation Environment
gether the use of splices in submarine cables, because a
Designing an underwater crossing requires consider-
splice adds a site of potential weakness in the electrical
ation of a number of factors related to the installation
insulation and mechanical structure. Further, if splices
environment. These factors affect the selection of the
are to be made at sea, exposure time to possible storms
type of cable, detailed cable construction, overall cable
and shipboard failures is obviously increased. Installa-
transmission system configuration (i.e., how many cir-
tion and handling are made more complex depending
cuits, how many spare phases, etc.), and the route of the
on the splice design. Submarine cable circuits are more
cable crossing. Obviously, length of the cable and maxi-
reliable if they do not contain splices, at least not in the
mum water depth are key technical issues, but addi-
lead sheath, the reinforcement, or the armor. Most
tional factors must also be evaluated.
major underwater cables have been produced splice-free
in long lengths stretching from shore to shore. Where Underwater Profile and Bottom Topography
necessary, splices have been laid successfully in water of The underwater profile and bottom topography can be
moderate depth such as the English Channel (Gazzana- significant factors in selecting the type of cable, the
Priaroggia and Mascio 1973). route, and the method of installation. A mild, uniform
12-35
elevation change for a cable crossing is generally desir- pended. Fast subsurface currents can actually move the
able. Securing cables on steep embankments usually pre- whole cable downstream. In severe conditions, anchor
sents some difficulties, and once installed, they could be systems have been installed to hold the cable in position
subject to problems such as impregnant migration with (Edison Electric Institute 1957). These currents can also
solid paper cables, or creepage of the insulation. Cables cause movement of thermal backfill and silt, burying a
on steep embankments are also more prone to damage cable beneath piles of soil.
from submarine landslides. Many methods have been
Water Temperatures and Bottom Soil Thermal
used to secure cables on steep embankments—including
Characteristics
concreting or grouting in place, anchoring, snaking,
Water temperature and soil thermal characteristics must
etc., all of which appear to have worked, but require
be considered very early in the design stages of subma-
periodic inspection. The submarine cable design engi-
rine cables. Just as thermal characteristics of the soil
neer should select the most appropriate method based
have a significant effect on the ampacity of land-based
on site conditions.
cables, the same is true with submarine cables. Poor
thermal conductivity of river and lake beds is not
A rough, erratic bottom makes cable installation diffi-
uncommon. Further, silt can dry out if it is in contact
cult and time consuming, and therefore expensive. If the
with a hot cable, and as it dries, it can form a thermal
cable is to be laid without being embedded, it is impor-
insulation that is capable of initiating thermal runaway
tant for the cable to lay flat on the bottom. If the bot-
and breakdown of the cable insulation. The 1969 failure
tom is such that lengths of the cable are unsupported
of a 115-kV fluid-filled cable crossing Lake Champlain
and underwater currents are present, these sections of
was attributed to thermal breakdown due to high ther-
the cable could be subject to damage due to cable move-
mal resistivity of the bottom soil conditions (Electrical
ment. Damage due to vibration from underwater cur-
World 1971). Additional discussion of soil thermal char-
rents (called strumming) can include abrasion, chafing,
acteristics is provided in Section 12.2.6.
and sheath fatigue. If the cables are to be embedded, a
rough bottom makes the job of trenching and placing
Water temperatures in deep waters might be quite cool,
the cables far more difficult. Some waterways have sig-
but improved cable ampacity over the total circuit due
nificant numbers of man-made obstacles such as sunken
to the cold-water environment cannot automatically be
barges, anchor lines, etc.
assumed. Water temperatures at the shore ends are usu-
ally higher. Further, the positive effect of cold waters
Thus, bottom soundings along potential routes should
can be eliminated by the bottom sediments that cover
be taken to obtain accurate profile information. If the
the cable.
bottom conditions are not uniform along the route,
more detailed bottom information should be obtained.
If poor thermal bottom characteristics or warmer
Serious problems with bottom conditions can warrant
waters present an ampacity constraint for only a portion
the consideration of alternate routes and, at the least,
of an underwater crossing, such as at either end
dramatically increase the cost for installation.
approaching landfall, an alternative is to use a larger
Surface and Subsurface Currents conductor size in only those areas to eliminate the ther-
Surface currents must be evaluated in determining the mal restriction.
method and equipment to be used to install the subma-
Man-made Facilities
rine cable or retrieve it. Fast surface currents might
Man-made facilities that may represent hazards to the
require special anchoring or active propulsion systems
cable system should also be evaluated when determining
for positioning a cable ship or barge, or might preclude
the suitability of a route. For example, submarine cable
the use of surface-ship cable installation methods alto-
installations near sanitary sewers or chemical discharges
gether. Subsurface currents can also cause a variety of
might increase corrosion problems with the armor. Pil-
problems. If the cable is laid on the bottom without
ings, bridge abutments, and other structures built in
embedment, and it becomes suspended (for example,
waterways often cause turbulence nearby, particularly
due to the sediment below the cable being transported
during high-water conditions. Turbulence can cause
downstream), the currents can cause the cable to
cable vibration or chafing. Ice jams at bridges, piers, and
vibrate. As noted above, this in turn causes chafing dam-
other man-made facilities can also cause scouring of a
age to the cable armor and jacket and fatigue to the lead
river bottom with related damage to cables. Thus, cable
sheath. In rivers and streams, it is common practice to
routes should avoid these structures if possible. Finally,
lay cable in an arc with the bow upstream. If the cable
areas involving construction of other facilities should be
does move due to the currents eroding the bottom, slack
avoided simply to reduce the possibility of damage to
is provided, so that the cable does not become sus-
the cable by dig-ins, etc.
12-36
Ship Traffic and Commercial Fishing Activities usually scheduled during neap tides, when the difference
The greatest threat to the reliability of submarine cable between high and low tide is very small and tidal cur-
systems are people. Externally caused damage from rents are at a minimum.
trawling, dredging, or ship anchors is the most common
Regulatory Restrictions
problem for cables. Thus, commercially active water-
ways should be avoided if possible. If the cable route Regulatory restrictions from both federal and local
must cross an active waterway, the risks to the cables authorities should be clearly understood. Blocking of
must be minimized. Embedding the cable to protect it is navigable waters is often an issue that directly affects
one option to minimize risk of cable damage. Widely construction and repair. For example, a cable system
separating the cable phases (with single-phase cables) to floated across a river, blocking shipping traffic for a
minimize the possibility of damaging multiple cables period of time, might be acceptable if the proper author-
during a single incident should be considered to improve ities are notified in advance, and if local weather and
circuit availability. Installing additional spare cables or recreation situations allow. Also, there are often con-
cable circuits is another alternative. cerns and restrictions associated with trenching, bottom
disruption, and spoil removal, as well as maximum pos-
Underwater or Shore Environmental Sensitivity sible fluid-loss levels. In most cases where navigable
Although submarine power cables have not been found waterways are concerned, the U.S. Army Corps of Engi-
to present significant environmental problems in the neers has the ultimate jurisdiction, and can be contacted
past, environmental sensitivity, both underwater and on to assist in obtaining the proper permits and under-
shore, must be considered carefully in selecting the standing the limiting factors.
route, the installation method, and the cable system
itself. For example, installation can be scheduled during Environmental Issues
seasons that are environmentally safer. Installation tech- A wide range of potential environment interactions are
niques that minimize bottom soil disruption should be evaluated for submarine cable installations. Typical top-
evaluated. If cable fluid loss is a concern, for example, ics considered in an Environmental Impact Statement
flow limiters, special jacketing, or solid cable designs (EIS) submittal are listed below.
could be evaluated. Environmentally sensitive areas • Terrestrial ecology
should simply be avoided if at all possible. • Aquatic ecology
Tides, Weather, and Seasonal Constraints – Turbidity and sedimentation
Tides, weather, and seasonal constraints can limit the – Spoil disposal
time periods during which the cable can be installed and – Chemical quality
when access to the cable for retrieval and repair is possi- – Biological quality
ble. Limited access to the cable for retrieval and repair is
• Land traffic
the more serious of the two issues. Constraints with
regard to seasonal periods when a cable system can be • Water traffic
installed can be accommodated with good construction • Noise
planning in most cases. If the cable system is inaccessible
• Socioeconomics
for repair for a few months each year (due to freezing,
severe sea conditions, ice, bitterly cold weather, etc.), the • Aesthetics
system must be designed accordingly. For example, spare • Recreation
cables or additional cable circuits could be installed, so
that power transmission can be maintained until the Terrestrial ecology, land traffic, noise, socioeconomics,
damaged cable can be repaired when the weather will and recreation are primarily associated with impacts at
allow repair procedures to begin. Underwater cable cir- the land connections and terminations. Aquatic ecology
cuits should be rated to allow for the loss of a phase or has major relevance to submarine cable operation.
the total circuit for an extended period of time. Equip-
ment might also be installed to minimize the damage if Many of the impact areas (land traffic, water traffic,
the cable must remain unrepaired on the bottom for an noise, socioeconomics, and recreation) are associated
extended period of time, such as flow-limiting devices on with construction operations and are short term in
a fluid-filled cable system or double-lead sheaths to pro- nature. This is not to say that the short-term construc-
vide additional protection for the cable core. tion impacts are not important. In fact, these typically
have more significance in obtaining construction per-
Tides represent a significant factor affecting the installa- mits, Corps of Engineers’ approvals, and public accep-
tion of submarine cables. Installation and repair are tance than long-term operational aspects of the cable.
12-37
Underwater Cable System Impacts on Aquatic Ecology moderate-sized bodies of water have found the possibil-
Turbidity, sedimentation, and spoil disposal receive sig- ity of a cable fluid spill not to be of significant environ-
nificant attention in environmental evaluations. They mental concern (Chern 1981; Dames and Moore 1981).
are generally related to construction rather than opera- Environmental impacts in deeper open waters are even
tion, short term in duration, and moderate to insignifi- lower, since the open waters should see higher fluid dilu-
cant in magnitude unless previously deposited toxic tion rates due to winds and tidal mixing, and the dis-
substances are disturbed. Impacts related to these fac- solved oxygen levels of these waters are probably better
tors are likely to be higher in smaller, restricted bodies than the restricted waters.
of water (i.e., small river and lake crossings), where
marine life would be subjected to prolonged exposure to 12.4 CABLE SYSTEM DESIGN CALCULATIONS
high concentrations of suspended sediment.
Underground transmission cable design requires a series
of calculations that must be done by the utility, cable
The cable system operational impacts that are associ-
manufacturer, or a consultant. Initial system calcula-
ated with the chemical and/or biological quality of the
tions such as impedances, load sharing with other trans-
water and would typically be evaluated for a site-specific
mission lines, insulation coordination, etc. are described
EIS include:
in Chapter 16. Detailed accessory calculations, such as
• Electric and magnetic field effects electrical stresses in splices, are performed by the manu-
• Cable heat dissipation facturers. Utilities may specify allowable levels, but they
seldom perform the calculations. Cable installation
• Cable dielectric-fluid spills.
design calculations typically done by utilities are sum-
marized below.
Although there is little data on the effects of electric and
magnetic fields on the aquatic environment and com- • Extruded-dielectric cables
passes, due to the rather low fields and the localized – Steady-state and transient/emergency ampacity
impact area, these effects have been considered insignifi- – Cable-pulling tension and sidewall pressure
cant in magnitude (Dames and Moore 1981). – Sheath voltages
• Pipe-type Cables
Heat dissipation from the power cables is low, as are
cable/seabed interface temperatures. The possibility of – Steady-state and transient/emergency ampacity
cable heat dissipation increasing the overall temperature – Cable-pulling tension and sidewall pressure
of the water body is of interest because this could – Hydraulic requirements
decrease the dissolved oxygen level in the water. How- – Forced-cooling requirements
ever, when cable designs have been evaluated in • Self-contained fluid-filled cables
restricted bodies of water, calculations have shown that – Steady-state and transient/emergency ampacity
water temperature would not be elevated due to the – Cable-pulling tension and sidewall pressure
water’s high heat capacity, motion, and latent heat of – Hydraulic requirements
evaporation (Dames and Moore 1981). Any effects due – Sheath voltages
to cable heat dissipation are localized directly around
the cable. In some cases, the utility performs trench design
calculations for controlled backfill. The Underground
Fluid-filled cables, either pipe-type or self-contained, Transmission Workstation, developed by EPRI in the
also require consideration of environmental impacts due early 1990s, to be upgraded in the near future, provides
to possible release of the dielectric fluid. Although no applications programs for each of these calculations.
fluid would be released during normal operations, envi- The calculation procedures are summarized on the fol-
ronmental concerns relate to the possible release of lowing pages.
dielectric fluid during a fault or accident, a highly
unlikely event. Typically used cable impregnants include 12.4.1 Steady-State, Transient, and Emergency
highly refined napthenic mineral oils, synthetic poly- Ampacity
butenes, and alkybenzenes. These fluids have been
found to be relatively nontoxic (Chern 1981; Dames and Steady-state ampacity refers to the maximum current
Moore 1981; U.S. Army 1975; Leiboritz 1972), and, as that a cable can carry for long times, for a defined load
evaporation occurs rapidly and biodegradation occurs shape. Transient ampacity refers to allowable currents
within 30 to 60 days, long-term effects are not an issue. for very short periods where all heat is stored in the
Previous EIS documents prepared for restricted and cable system, and emergency ampacity refers to allow-
able currents for longer times when heat storage of the
12-38
earth enters into the calculations. These very detailed 1953) recommends a minimum 0.25-in. (6.4-mm) clear-
calculations are at the heart of a cable system design. ance, but most utilities today use the larger number. The
Chapter 11 is devoted to ampacity calculations. minimum clearance for pipe-type cables occurs when
the cables are in a triangular configuration. This clear-
12.4.2 Cable-pulling Tension and Sidewall ance, C, is calculated from Equation 12-6.
Pressure 2
Although the maximum length between splices is some- D 1 ⎡ d ⎤
C = − 1.366d + ( D − d ) 1 − ⎢ ⎥ (in. [cm])
times limited by the length of cable that can fit on a reel, 2 2 ⎣ (D − d ) ⎦
for most routes it is governed by allowable tension on 12-6
the cable conductor as the cable is pulled into the duct Where:
or pipe, or allowable sidewall pressure as the cable goes D = nominal pipe inside diameter, in. (cm)
around a bend. Allowable clearances between cable and d = nominal cable insulation shield outside
duct or pipe and jam ratio are also checked during pull- diameter + 1.5 times skid wire thickness,
ing-tension calculations. Route information necessary in. (cm)
to calculate pulling tensions and sidewall pressures is
summarized below: The clearance for three 2500-kcmil (1267-mm2), 345-kV
cables with a 3.654-in. (9.3-cm) shield diameter and a
• Section length 0.10 in. (2.54 mm) skidwire in a 10.25 in. (26 cm) inside
• Radius of bends and dips diameter pipe is 2.53 in. (6.4 cm).
• Arc length or angle of bends and dips Jam Ratio
• Position of bends and dips If the ratio of pipe inside diameter to cable diameter is
in the range of 2.9 to 3.1 (some designers use a smaller
Required cable system information includes: range of 2.95 to 3.05, although the 1957 Underground
Systems Reference Book [Edison Electric Institute 1957]
• Cable conductor material and size uses 2.8), there is a chance that the cables can jam in the
• Cable outside diameter pipe during pulling. This is a special problem as cables
• Cable weight are pulled through a bend in the pipe. Since the pipe
tends to flatten slightly when the bend is made, the
• Cable skid wire material, if present cables can cross, and then they can wedge as they enter
• Pipe or duct material, size, and internal coating, the circular cross section beyond the bend. If the jam
if any ratio is below about 2.9, the cables will not cross. If the
• Pulling rope data, including diameter and weight ratio is above about 3.1, they are unlikely to wedge
because of the loose fit. The jam ratio calculation is
• COF (Coefficient of Friction) between cable and pipe given in Equation 12-7.
or duct
⎛D⎞
• COF between pulling rope and pipe or duct Jam Ratio = 1.05⎜ ⎟ (dimensionless) 12-7
⎝d⎠
Most of this discussion addresses three conductors in a
common pipe and follows the procedures published by The jam ratio is 2.82 for three 2500-kcmil (1267-mm2)
Rifenburg in 1953 (Rifenburg 1953). Single-conductor 345-kV cables in a 10.25-in. (26-cm) ID pipe.
pipe-type cables for riser sections and single-conductor
SCFF or extruded-dielectri c cables in duct are Pulling Tensions
addressed as well. Three SCFF or extruded-dielectric Although the term “tension” is commonly used by cable
cables in a common duct are seldom used at transmis- engineers, and is the term used in Reference (Rifenburg
sion voltage. If calculations need to be done for this 1953), we actually calculate a force in pounds force, lbf
arrangement, the procedures for pipe-type cables may (kgf). This is the force required to pull the cables into the
be followed, and the user can refer to Reference (EPRI pipe or duct. In keeping with common usage, this book
1984) for additional information. refers to pulling tensions.
Clearances The force to pull the cable is applied to the conductor,

A minimum clearance of about 0.5 in. (12.7 mm) should and allowable pulling levels are based on the conductor
be maintained between the top of a single-conductor material and site conditions. Acceptable levels are
cable and the pipe or duct, or between the three cables 10 lbf/kcmil (8.9 kg f/mm 2) for copper conductors and
and the pipe for a pipe-type cable. Reference (Rifenburg
12-39
6 lb f /kcmil (5.4 kg f /mm 2 ) for aluminum conductor values at varying installation temperatures, resulting
cables (EPRI 1983; Aabo et al. 1979). from EPRI-sponsored research (EPRI 1984) on distribu-
tion cables in the 1980s, is given in Tables 12-4 and 12-5.
Therefore, a 2500-kcmil (1267-mm2) copper conductor These figures assume a soap-and-water-based lubricant.
can accept a 25,000-lbf (11,400-kgf) tension. Experimen-
tal evidence from EPRI experiments has shown that a
conservative assumption is that two of the three cables Table 12-4 Basic Dynamic Coefficients of Friction for
in a pipe-type system carry the full tension (EPRI 1983). Straight Pulls and Bearing Pressures Less than 150 lbf/ft
The total allowable force on three 2500-kcmil conduc- (224 kgf/m) (Soap-and-Water-Based Lubricants)
tors is therefore 50,000 lbf (22,000 kgf).
COF
Duct Material Cable Sheath/Jacket 75°Fa 25°Fa
The Rifenburg paper addresses copper conductors only,
XLPE 0.40 0.40
and proposes a maximum stress of 10,000 lb f /sq.in. PE 0.40 0.35
(7850 lb f /kcmil; 7.0 kg f /mm 2 ) (Rifenburg 1953). All PVC 0.50 0.25
PVC
three conductors are assumed to share the tension, but Neoprene 0.90 0.55
CN 0.40 0.40
the 10,000 lb f /sq.in. is reduced by 15% to account Lead 0.25 0.25
for unbalance in tension. Working through these num- XLPE 0.45 0.35
bers gives the same 50,000-lbf (22,000-kgf) tension for PE 0.25 0.20
2500-kcmil cables as the EPRI results. PVC 0.30 0.20
PE
Neoprene 0.65 0.45
CN 0.20 0.20
The tension, T, for a single cable pulled into a straight Lead 0.20 0.25
pipe or duct is calculated from XLPE 0.30 0.20
PE 0.25 0.35
T = K0 wL lbf (kgf) 12-8 PVC 0.40 0.20
Fiber
Neoprene 0.40 0.30
Where: CN 0.40 0.35
Lead - -
K0 = basic coefficient of friction, dimensionless.
XLPE 0.30 -
w = total cable weight, lb/ft (kg/m). PE 0.35 -
L = length, ft (m). PVC 0.55 -
Concrete
Neoprene 0.50 -
Coefficient of Friction CN - -
The coefficient of friction (COF) between cable and pipe Lead 0.55 -
or duct is a very important parameter for pulling-ten- XLPE 0.70 -
PE 0.70 0.35
sion calculations. The addition of a lubricant can reduce PVC 0.70 0.35
the COF significantly for SCFF or extruded-dielectric Transite
Neoprene 1.00 0.35
cables in duct. For pipe-type cables, the lubricant is typi- CN - -
Lead - -
cally a small amount of the filling fluid that will be used
XLPE 0.60 0.45
in the pipe. Many commercial lubricants are available PE 0.50 0.50
for distribution-voltage extruded-dielectric cables, and PVC 0.65 0.40
Steel
most can be used for transmission cables. The user Neoprene 1.05 0.70
CN 0.50 0.50
should check with the cable manufacturer before using Lead - -
any lubricant. For pipe-type cables, Table 12-3 gives
COF values for different skid wire materials in steel line a. Installation temperatures.
pipe having an internal epoxy coating. A range of COF
Table 12-5 Basic Dynamic Coefficient of Friction at Bends
Table 12-3 Basic Dynamic Coefficient of Friction for Where Sidewall Pressure is 150 lbf/ft (224 kgf/m) or Greater
Different Skid Wire Materials in Coated Steel Pipe (Aabo et (All Lubricants)
al. 1979) Cable Outer Covering Duct Material COF
Skid Wire Material COF XLPE, PE, N PVC, PE, Concrete 0.15
Brass 0.17 PVC PVC, PE, Concrete 0.30
Stainless Steel, Bright 0.13 XLPE, PE, N Steel 0.25
Stainless Steel, Matte 0.11 PVC Steel 0.30
Zinc 0.15 Lead Steel 0.20
12-40
Better lubricants have recently become available. weight correction factor is a function of cable and pipe
Table 12-6 shows the anticipated COF for Polywater ® diameters. A triangular configuration is assumed for
“J” lubricants. This lubricant has been the mostly com- ratios of pipe inside diameter to cable outside diameter
monly used. Other lubricants with better or equal COF of 2.5 or lower, and many utilities use the triangular
values are also available. The lubricant manufacturers configuration for all calculations. Some users average
should be consulted for updated COF values when the the numbers for cradled and triangular configuration
utility is designing a specific installation. No COF val- when the ratio is greater than 2.5. WC for cables in trian-
ues are available for fiberglass duct. When performing gular configuration is given in Equation 12-9.
pulling calculation for fiberglass duct, the designer
1
should consult with the lubricant manufacturers. WC = (dimensionless) 12-9
2
Weight Correction Factor ⎛ d ⎞
1- ⎜ ⎟
The weight correction factor, WC, accounts for the ⎝ D-d ⎠
wedging and resulting increase in effective COF that
takes place when three cables are pulled into the pipe. The weight correction factor for cables in cradled con-
(Three cables are rarely installed in the same duct for figuration is given in Equation 12-10.
SCFF or extruded-dielectric transmission cables.) The
2
4⎛ d ⎞
Table 12-6 Basic Lubricated and Unlubricated Coefficient of Wc = 1 + ⎜ ⎟ (dimensionless) 12-10
Friction for Polywater® “J”
3⎝D−d ⎠
Lubricated Unlubricated For a 2500-kcmil (1067-mm 2 ), 345-kV cable with a
Conduit Jacket COF COF
3.804-in. (9.67-cm) diameter in a 10.25-in (26-cm) ID
CPE 0.12 0.58
HDPE 0.07 0.14 pipe, WC is 1.24 for triangular configuration and 1.46
Hypalon 0.06 0.75 for cradled configuration.
LLDPE 0.11 0.35
Aluminum
Nylon 0.13 0.38
PVC 0.08 0.48 The effective COF, K, equals WCKO, and the allowable
Polypropylene 0.08 0.42 pulling tension is shown in Equation 12-11.
XLPE 0.08 0.38.
CPE 0.42 0.65 T = WC K O wL or T = KwL lbf (kg f ) 12-11
HDPE 0.19 0.42
Hypalon 0.38 1.10
Where:
LLDPE 0.24 0.43 KO = basic coefficient of friction (dimension-
EMT
Nylon 0.28 0.33 less).
PVC 0.43 0.80
Polypropylene 0.24 0.45
w = total cable weight (lb/ft [kg/m]).
XLPE 0.25 0.42 L = length (ft[m]).
CPE 0.09 0.80
Hypalon 0.07 0.90 Dips and Bends
LLDPE 0.06 0.30 Equation 12-11 gives the pulling tension for straight
IMC Nylon 0.12 0.28 pipe sections. Longer equivalent lengths are calculated
PVC 0.07 0.60
Polypropylene 0.07 0.38 to account for dips and bends. Figure 12-33 shows a
XLPE 0.09 0.58 typical pull. The increase in length, L1, for major hori-
CPE 0.35 0.85
HDPE 0.09 0.15
Hypalon 0.15 0.58
LLDPE 0.17 0.30
PVC
Nylon 0.24 0.38
PVC 0.10 0.60
Polypropylene 0.09 0.33
XLPE 0.14 0.23
CPE 0.30 0.85
HDPE 0.21 0.35
Hypalon 0.18 0.90
RIGID LLDPE 0.33 0.68
STEEL Nylon 0.40 0.50
PVC 0.15 0.55
Polypropylene 0.13 0.50 Figure 12-33 Typical section for calculating equivalent length
XLPE 0.29 0.50 of horizontal bends (Rifenburg 1953) (courtesy AIEE).
12-41
zontal bends such as Bend 1 in the figure, results in a SWP is affected by triangular or cradled cable configura-
new length, L2, as calculated in Equation 12-12. tion and the ratio of cable to pipe diameter. SWP for tri-
angular configuration is divided equally between the two
2
⎡R⎤ bottom conductors, and is shown in Equation 12-16,
L 2 = L1 coshKθ + L21 + ⎢ ⎥ sinhKθ ft (m) where WC is calculated from Equation 12-9.
⎣K ⎦
12-12 ⎛ W ⎞ T lbf ⎛ kg f ⎞
Where: SWP = ⎜ 3 C ⎟ ⎜ ⎟ 12-16
L1 = equivalent length at entrance to bend, ⎝ 2 ⎠ 3R ft ⎝ m ⎠
ft (m).
The SWP calculation for cradled configuration is given
L2 = equivalent length at exit from bend, ft (m).
in Equation 12-17, where W c is calculated from Equa-
K = effective COF, dimensionless.
tion 12-10.
R = radius of bend, ft (m).
θ = angle of bend or dip, in degrees. T lbf ⎛ kg f ⎞
SWP = ( 3 WC − 2 ) ⎜ ⎟ 12-17
3R ft ⎝ m ⎠
No correction is needed if the radius of a dip is less than
the coefficient of friction times the length of the dip. The
The cable manufacturer should be consulted before
increase in length, L2, for more severe vertical dips, as
sidewall pressure calculations are finalized. A range of
shown in Figure 12-34, is given in Equation 12-13.
typical allowable values is given in Table 12-7.
R 4Kθ
L 4 = L1e 4Kθ + ⎡⎣e - 2e3Kθ + 2e Kθ -1⎤⎦ ft (m) As noted in the EPRI research (EPRI 1983), these side-
K
12-13
wall pressures assume a smooth pipe. Ripples on the
inside of a field bend can cause cable damage at a lower
The total equivalent length is the straight-line length, SWP than the values listed in Table 12-7.
plus increases due to major dips and bends. Also, an
equivalent length of about 50 ft (13 m) is added to When the cable tensions and sidewall pressures are cal-
account for the force needed to pull the cables off the culated, the effect of the pulling rope should also be
reel; see Equation 12-14. evaluated. The tension measured during a pulling oper-
ation is the pulling-rope tension, so the rope tension
L tot =L1 +L 2 +L3 +...L n + 50 ft (13m) ft 12-14 also needs to be estimated to use calculated values as a
criterion for evaluating the actual cable-pulling tension.
Appendixes I–III of Reference (Rifenburg 1953) provide For pipe-type cable installation, a steel rope COF of
a series of sketches and equations for calculating cable- 0.25–0.5 is often used. The effect of the rope on the pipe
pulling tensions in many typical dips and bends. or duct is discussed in Chapter 13.
Sidewall Pressure
Hydraulic Calculations
Sidewall pressure (SWP) is the crushing action that
Fluid volume and pressure calculations are important
occurs when cables are pulled around bends. SWP is
for HPFF and SCFF cable design and operation.
defined as the tension, T, in lbf (kg), divided by the bend
Proper fluid pressures must be maintained over the full
radius, R, in feet (m) as shown in Equation 12-15.
length of the circuit for a wide range of loading condi-
T lbf ⎛ kg f ⎞ tions and rates of load changes. The fluid expelled from
SWP = ⎜ ⎟ 12-15 the cable system during increasing-load periods must be
R ft ⎝ m ⎠ stored in a reservoir tank, and the tank capacity must be
sufficient to provide fluid during cable contraction as
Table 12-7 Range of Allowable Sidewall Pressures for

Different Cable Types
lbf/ft Kgf/m
Impregnated Paper 1000 1490
Laminated Paper-Polypropylene 1000 1490
Crosslinked Polythylenea,b 2000 2980
Ethylene-Proplylene Rubbera,b 2000 2980
a. With overall jacket.

Figure 12-34 Typical section for calculating equivalent b. Reduce to 1500 lbf/ft (2235 kgf/m) if jacket with a wire or
length of vertical dips (Rifenburg 1953) (courtesy AIEE). tape shield is not applied tightly.
12-42
well as to supply a possible leak. Refer to Chapter 9 for the fluid expansion. The temperature differences are
typical reservoir tank sizing as a function of fluid vol- usually taken from the coldest ambient earth tempera-
ume changes. ture to the maximum operating temperature for each
component. Calculation procedures in Chapter 11 may
Fluid demand requirements, fluid flows, and pressure be used to calculate conductor, insulation, fluid, and
changes are complex calculations, which require the use pipe temperatures for a given load condition. The safest
of empirical data, usually in the form of curves. Detailed approach is to use the highest emergency operating tem-
calculation procedures for pipe-type cables are given in perature, which is a 105°C conductor temperature.
a 1955 AIEE committee report (AIEE 1955), and those Dielectric losses do not enter into normal daily volume
for SCFF cable are given in a 1956 paper (Buller et al. changes since dielectric losses are present whenever the
1956). This book reviews the general calculation proce- line is energized. However, dielectric losses are taken
dures, but the two referenced papers, and their back- into account when sizing pumps, reservoirs, and pipe
ground references, should be used for comprehensive sizes for pressure drop considerations, since the conser-
hydraulic calculations. vative assumptions are either de-energizing a fully
loaded cable, or energizing a cold cable to full load.
High-Pressure Fluid-Filled Cables
The fluid requirement for an HPFF cable pipe is calcu-
Volume changes calculated from Equation 12-19 are
lated from Equation 12-18. Volume in liters may be
given in cubic inches per foot of cable length. For U.S.
obtained by multiplying gallons by 3.785.
units, total volume change is obtained from multiplying
π 12 by the length in feet. Number of gallons is obtained by
V=L (D 2 − 3d s 2 ) gal 12-18
dividing cubic inches by 231, and number of liters is
4 231
obtained by multiplying gallons by 3.785.
Where:
L = line length, ft.
Fluid demand, a, determines the required pump capac-
D = pipe inside diameter, in.
ity to ensure that fluid can be supplied during rapid
dS = cable shield diameter, in.
cool-down. Fluid demand also determines the required
free fluid area in the pipe to ensure that pressure
This volume is typically increased by 5–10% to account
increases are not excessive during rapid heating. Fluid
for extra fluid in joint casings and piping, fluid absorp-
demand is determined by differentiating Equation 12-19
tion by the cable as pressure is increased, and tempera-
with respect to time, as shown in Equation 12-20. Vol-
ture changes as the fluid leaves tank trucks and enters
ume change in liters/meter-sec may be obtained by mul-
the cable pipe. The amount to fill the reservoir tank to
tiplying Equation 12-20 results by 0.054.
the desired level, plus a buffer in the tank for possible
fluid leaks, must be added to this volume. d
a= ΔV in.3 /ft-s 12-20
dt
Equation 12-19 gives the volume change, ΔV, with tem-
perature changes for a pipe-type cable: Solving Equation 12-20 requires knowledge of tempera-
ΔV = Cdc2ΔTc + 0.012(ds2−dc2) ΔTi ture changes with time for each cable system compo-
+ 0.0070(D2−3ds2) ΔTo nent. Reference (AIEE 1955) provides curves for several
− 0.0003D2ΔTp in3/ft 12-19 cable and pipe sizes. Rigorous temperature change cal-
culations may be performed using procedures described
Where: in Reference (Buller et al. 1956). Some commercial
C = constant, 0.0060 for concentric stranded ampacity programs are able to calculate fluid demand,
conductors, or 0.0036 for compact seg- and the EPRI Underground Transmission Workstation
mental conductors. includes hydraulic calculations including fluid demand.
dc = conductor diameter, in.
ds = cable shield diameter, in. Maximum fluid demand values given in Reference
D = pipe inside diameter, in. (AIEE 1955) range from 30 to 45 x 10-6 in.3/watt-sec per
ΔTc = conductor temperature change, °C. foot (1.6–2.4 μl/watt-sec per meter) for HPFF cables up
ΔTi = insulation temperature change, °C. to 230 kV. Fluid demand per mile is, therefore,
ΔTo = fluid temperature change, °C. 0.33 gal/min for a typical ohmic loss of 8 watts per con-
ΔTp = pipe temperature change, °C. ductor foot (demand per km is 0.75 l/min for 26 W/m).
Pump capacity should be greater than the calculated
The last term in Equation 12-19 is negative, since pipe fluid demand. Some utilities standardize on
expansion with increasing temperatures offsets some of 10–12 gal/min (38–45 l/min) pumps, since there is little
12-43
disadvantage to oversizing pumps, and larger pumps Self-Contained Fluid-Filled Cables

provide the opportunity for slow circulation to reduce Calculation principles and procedures for SCFF cables
hot spots along the route. are quite similar to those for HPFF cables, and the pro-
cedures of Reference (Buller et al. 1956) closely parallel
The design engineer should also calculate fluid pres- those of (AIEE 1955). For single-conductor SCFF
sures along the route for all loading conditions to ensure cables, fluid expansion and contraction occur princi-
that pressures are not too low during cooling transients, pally in the conductor core. The fluid requirement for
or too high during heating transients. Most utilities one phase of a SCFF cable is calculated from
maintain a nominal 200 psi (1380 kPa) pressure at the Equation 12-23. Volume in liters may be obtained by
high point of the circuit. Pressure at any other point multiplying gallons by 3.785.
must also include the hydrostatic head, Ph, due to eleva-
⎛ π ⎞ 12
tion changes as calculated from Equation 12-21. V = L ⎜ d o2 ⎟ gal 12-23
Ph = 0.434 γ Δh psi 12-21 ⎝ 4 ⎠ 231
Where:
Where: L = line length, ft.
γ = specific gravity, dimensionless. dO = cable core diameter, in.
Δh = elevation difference, ft.
This calculated volume is typically increased by about
Pressures in kPa may be obtained by multiplying psi 25% to account for fluid in joints, plus fluid in the insu-
by 6.895. lation and under the sheath. The amount required to fill
the reservoirs to the desired level and to provide a buffer
Additional pressures arise from dielectric fluid motion in the reservoir for possible fluid leaks must be added to
during fluid expansion and contraction. This friction the calculated volume. The sum of these volumes is the
pressure, P f is a function of fluid viscosity, distance, total fluid volume for one conductor in an operating
cable/pipe size, and cable position in the pipe, as shown cable system. The total volume for three phases is three
in Equation 12-22. times greater than that calculated for a single cable.
However, each conductor typically has its own oil reser-
⎛ K ⎞
Pf = 35.5 x 2 aμ ⎜ 4 ⎟ x 10−6 psi 12-22 voir. In addition to this amount, several times as much
⎝D ⎠ fluid may be used during cable installation to provide
Where: fluid flow during splicing and terminating. The cable
x = distance from closed end, ft. manufacturer should be consulted when determining
μ = fluid viscosity, centipoise. the fluid required during installation.
a = fluid demand, from Equation 12-20,
in.3/ft-s.
K = constant, depending upon cable/pipe
diameter ratio.
D = pipe inside diameter, in.
Values for the constant, K, can be obtained from

Figure 12-35. Fluid viscosity should be taken from Fig-
ure 3-35, for the coldest temperature anticipated during
cable system operation. Note that frictional pressure
drop is much less of a concern with the low-viscosity
pipe-filling fluids used today, compared to the much
higher viscosity fluids used in the 1950s.
Pressure drop due to rapid dielectric fluid circulation

will be in addition to the values calculated above. It is
possible for a differential pressure of 200 psi (1380 kPa)
to be required to force a 300 gal/mm (1135 l/min) flow in
a cable pipe. Refer to Chapter 9 for these calculations.
Figure 12-35 Factor K as a function of cable/pipe diameter
(From Figure 3, Reference [AIEE 1955]) (courtesy AIEE).
12-44
w
Fluid demands and pressures are calculated using pro- Effective ρ = ρf + (ρo − ρf) ε- h 12-24
cedures similar to those described previously for HPFF
cables. Refer to Reference (AIEE 1955) for detailed Where:
equations and empirical curves. w = width of trench.
h = depth to the bottom of the structure.
Trench Design for Corrective Backfill
ε = natural log.
Because all soils do not have the same thermal resistiv-
ρf = of fill.
ity, an underground transmission line designer must
ρo = of original.
evaluate the effect that the native thermal resistivity will
have on the circuit ampacity. For example, if the native
Figure 12-36 shows the effect of changing the width of
soil has a thermal resistivity of 150 C°.cm/watt
the controlled backfill.
(1.5 K.m/W), the resulting ampacity for a given circuit
will be lower compared to an identical circuit placed in a
Additional calculation procedures for the design of a
native soil of 90 C°.cm/watt (0.9 K.m/W). The main way
trench with corrective backfill are given in References
to improve the overall effective thermal resistivity seen
(Brookes and McGrath 1960; Williams et al. 1993). An
by a cable system is to install a thermal backfill material
effective overall soil thermal resistivity is assumed, and
around the duct/pipe. Equation 12-24 can be used to
trench size is determined along the different sections of
calculate the effective thermal resistivity for various
the route having varying thermal resistivities of the
widths and depths of backfill (Edison Electric Institute
native soil.
1957, pp. 10–15). This equation is good for a trench
width-to-depth ratio of about 3 to 1.
Figure 12-36 Effect of changing the width of a trench (courtesy POWER Engineers, Inc.).
12-45
REFERENCES Chamberlin, D. M. 1980. “Operating Experience with

Aabo, T., J. A. Moran, Jr., and J. R Shimshock. 1979. The Norwalk, CT—Northport, NY 138 kV Submarine
“Increased Pipe Cable Section Lengths.” Proceedings of Cable.” Presented at Transmission & Distribution Com-
the 1979 IEEE T&D Conference. April 1–6. mittee Meeting No. 57. Engineering Operations Divi-
sion. Electric Council of New England. May 22.
Aabo, T., and J. A. Moran, Jr. 1988. Thermomechanical
Bending Effects on Extra-High Voltage Pipe-Type Chern, C. 1981. “Ocean Thermal Energy Conversion
Cables. July. EPRI Report no. EL-5880. (OTEC) Project, Bottom Cable Protection Study, Envi-
ronmental Characteristics and Hazards Analysis.”
AIEE. 1955. “Oil Flow and Pressure Drop Calculations Washington, D.C. Ocean Engineering and Construction
for Pipe-Type Cable Systems.” AIEE Committee Project Office. Chesapeake Division. Naval Facilities
Report. AIEE Transactions. April. pp. 251–259. Engineering Command. February 19.
ASCE 2003 “Standard Guideline for the Collection and CIGRE. 2002. Working Group 19 of CIGRE Study
Depiction of Existing Subsurface Utility Data” Committee 21. “General Guidelines for the Integration
CI/ASCE Standard 38-02. of a New Underground Cable System in the Network.”
March.
ASTM. 1978“Moisture-Density Relations of Soils,
Using 5.5-lb (2.5 kg) Rammer and 12-in. (304.8 mm) Dames and Moore. 1981. Draft-Study Documentation
Drop.” ASTM Standard D 698. Report Ashe-State Environmental Impact Statement—
Crow Butte Crossing. Prepared for Bonneville Power
Barnes, C. C. 1977. Submarine Telecommunication and Administration. March.
Power Cables. Peter Peregrinus Ltd.
Edison Electric Institute. 1957. Underground Systems
Bazzi, G. 1982. “Advanced Technologies in HVAC and Reference Book.
HVDC Submarine Power Cable Transmission.” Pre-
sented at Fourth Conference on Electric Power Supply Eich, E. D. 1976. Installed Cost Comparison for Self-
Industry (CEPSI). Bangkok. November 22–26. Contained and Pipe-Type Cable. November. EPRI
Report no. EL-935.
Blau, D. 1975. Study of Environmental Impact of
Underground Transmission Systems. May. EPRI Report Electrical World. 1971. “Underwater Cable Fault Due to
no. 7826. Thermal Runaway.” May 1. pp. 52–53.
Brookes, A. S., and M. H. McGrath. 1960. “Soil Ther- Engelhardt, J. S., and D.W. Purnhagen. 1991. Dynamic
mal Characteristics in Relation to Underground Power Rating and Underground Monitoring System. July. EPRI
Cables—Part V. Practical Applications—Trench Design Report no. EL-7341.
and Construction.” AIEE Transactions. No. 51. Decem-
ber. pp. 792–856. EPRI. 1977. Backfill Materials for Underground Power
Cables. Phase I, EPRI Report no. EL-0506.
Buller, F. H., J. H. Neher, and F. O. Wollaston. 1956.
“Oil Flow and Pressure Calculations for Self-Contained EPRI. 1981. Thermal Property Analyzer. EPRI Report
Oil-Filled Cable Systems.” AIEE Transactions on Power no. EL-2128.
Apparatus and Systems. No. 23. April. pp. 180–194.
EPRI. 1983. Increasing Pipe Cable Section Lengths.
Chaaban, M. et al. 1991. “Evaluation of HVDC Cables March. EPRI Report no. EL-2847.
for the St. Lawrence River Crossing of Hydro Quebec’s
500 kV DC Line—Part III: Thermal Behavior.” Pro- EPRI. 1984. Maximum Safe Pulling Lengths for Solid
ceedings of the 1991 IEEE T&D Conference. Dallas, TX. Dielectric Insulated Cables. February. EPRI Report no.
September 22–27. pp. 117–121. EL-3333.
EPRI. 1997. Soil Manual for Underground Transmission.

EPRI Report no. TR-108919.
12-46
Foxall, R. G., K. Bjorlow-Larsen, and G. Bazzi. 1984. Kozak, S., J. T. Corbett, and F. J. Bender. 1965. “Fea-
“Design, Manufacture and Installation of a 525-kV tures of the New 138-kV Self-Contained Oil-Filled
Alternating Current Submarine Cable Link from Cable Systems.” IEEE Transactions on Power Apparatus
Mainland, Canada to Vancouver Island.” CIGRE and Systems. Vol. 94. No. 3. May.
21-04. 10 p.
Luoni, C., and P. Anelli. 1976. “Armor Corrosion in
Gazzana-Priaroggia, P., J. H. Piscioneri, and S. W. Mar- Single Core Submarine AC Cables.” IEEE Paper No. A
golin. 1971. “Long Island Sound Submarine Cable 76 190-9. Presented at IEEE Power Engineering Society
Interconnection.” IEEE Spectrum. Vol. 8. No. 10. Octo- Winter Meeting & Tesla Symposium. New York. Janu-
ber. pp. 63–71. ary 25–30.
Gazzana-Priaroggia, P., and C. Mascio. 1973. “Contin- Marquez, R. D., et al. 1981. “230 kV Self-Contained
uous Long Length AC and DC Submarine HV Power Oil-Filled Cable Line Installed Underneath a Bridge
Cables, The Present State of The Art.” IEEE Paper T- Located in Maracaibo, Venezuela.” IEEE Transactions
73-127-8. Presented at IEEE Power Engineering Society on Power Apparatus and Systems. Vol. 100. No. 7. July.
Winter Meeting. January 28–February 2. pp. 3153–3165.
Ghafurian, A. R., H. Chu, et al. 1989. “Detection and Martin, Jr., M. A., R. A. Bush, et al. 1981. “Practical
Location of Dielectric Fluid Leaks on Pipe-Type Cables Aspects of Applying Soil Thermal Stability Measure-
Systems.” IEEE Transactions on Power Delivery. July. ments to the Rating of Underground, Power Cables.”
pp. 1499–1503. IEEE Transactions on Power Apparatus and Systems.
Vol. 100. No. 9. September. pp. 4236–4249.
Gillette, R. W. 1959. “Installation of 69-kV High-pres-
sure Gas-filled Pipe-type Feeders on Queensboro Minemura, S., and Y. Maekawa. 1989. “500-kV Oil
Bridge, New York City.” IEEE Insulated Conductors filled Cable Installed on Bridges.” Proceedings of the
Committee Minutes. Spring. Appendix F. 1989 IEEE T&D Conference. New Orleans. Paper 89TD
354-2-PWRD.
Hatcher, C. T., R.W. Gillette, and R.W. Burrell. 1966.
“345-kV Underground Transmission on the Consoli- Moran, Jr., J. A., T. Aabo, and J. F. Shimshock. 1984.
dated Edison Company of New York System.” IEEE “Thermo-Mechanical Bending Effects in EHV Pipe-
Transactions on Power Apparatus and Systems. Vol. 85. Type Cables.” Proceedings of the 1984 IEEE T&D Con-
No.4. April. pp. 353–360. ference. Kansas City. April 29–May 4.
IEEE. 1981. IEEE Standard 422-1981. “IEEE Guide Morgareidge, K. 1965. “Feeding Studies with Poly-
for Soil Thermal Resistivity Measurements.” The Insti- vis-SH in Rats and Dogs.” Food and Drug Research
tute of Electrical and Electronics Engineers, New York. Laboratories, Inc. March 12.
IEEE. 1987. IEEE Standard 575-1988. “IEEE Guide Nakagawa, H. et al. 1973. “Installation of 275-kV
for the Application of Sheath-Bonding Methods for XLPE Cables in the Long and Steep Slope Tunnel.”
Single-Conductor Cables and the Calculation of IEEE Transactions on Power Apparatus and Systems.
Induced Voltages and Currents in Cable Sheaths.” Vol. 102. No. 12. December.
The Institute of Electrical and Electronics Engineers.
New York. NASTT. 2004. “Horizontal Directional Drilling Con-
sortium HDD Good Practices Guidelines.” American
IPCEA. 1976. Insulated Power Cables Engineers Society for Trenchless Technology.
Association. “Ampacities Including Effect of Shield
Losses for Single-Conductor Solid-Dielectric Power Patton, R. N., S. K. Kim, and R. Podmore. 1979.
Cable, 15-kV through 69-kV.” IPCEA Publication “Monitoring and Rating of Underground Power
P-53-426. May. Cables.” IEEE Transactions on Power Apparatus and
Systems. Vol. 98. No. 6. November/December.
Iwata, Z. et al. 1991. “Heat Pipe Local Cooling System pp. 2285–2293.
Applied for 145 kV Transmission Line in Copenhagen.”
Proceedings of the 1991 T&D Conference. Dallas, TX.
September 22–27. pp. 52–60.
12-47
Radhakrishna, H. S., and J. E. Steinmanis. 1981. “Ther- Soulsby, D. R., and A. J. Donovan. 1981. “The Effect of
mal Resistivity Survey of Lake Erie Sediments for the Backfill of the Temperature Distribution in a Buried
Ontario Hydro-GPU Interconnection.” Symposium on Circuit.” Symposium on Underground Cable Thermal
Underground Cable Thermal Backfill Backfill. September 17 and 18. pp. 167–180.
Ray, J. J., C. Arkell, and H. W. Flack. 1974. “525-kV Steinmanis, J. E. 1981. “Thermal Property Measure-
Self-Contained Oil-Filled Cable Systems for Grand ments Using a Thermal Probe.” Symposium on Under-
Coulee Third Powerplant Design and Development.” ground Cable Thermal Backfill. September 17–18.
IEEE Transactions on Power Apparatus and Systems.
March/April. pp. 630–639. Tarpey, J. W. et al. 1991. “Installation of a Solid Dielec-
tric 138-kV Underground Transmission System—From
Rifenburg, R. C. 1953. “Pipe-line Design for Pipe-type Concept to Completion.” Proceedings of the 1991 IEEE
Feeders.” AIEE Transactions. Vol. 9. December. T&D Conference. Dallas, Texas. September 22–27.
Rueckert, J. J., and J. I. Bien, Jr. 1979. “230-kV Cable U. S. Army. 1975. “Final Environmental Statement,”
Crossing of Baltimore Harbor.” IEEE Power Engineer- “Applicant’s Environmental Report,” and “Responses
ing Society Transmission and Distribution Conference of Commenting Agencies.” Documents submitted to
and Expo. Atlanta, GA. April 1–6. the U.S. Army Engineer District, New York. For an
environmental statement related to the “Con Edison–
Saleeby, K. E., W. Z. Black, and J.G. Hartley. 1979. Long Island Lighting Company’s Power Transmission
“Effective Thermal Resistivity for Power Cables Buried Cable Crossing of Long Island Sound,” Hempstead
in Thermal Backfill.” IEEE Transactions on Power Harbor, New York.
Apparatus and Systems. Vol. 98. No. 6. Novem-
ber/December. Weedy, B. M. 1988. Thermal Design of Underground
Systems. Chapter 8. Chichester, England and New
Sandiford, P. 1981. “Cable Backfill Materials—State-of- York, NY: John Wiley & Sons, Ltd.
the-Art.” Symposium on Underground Cable Thermal
Backfill. September 17 and 18. pp. 3–9. Williams, J. A., S. Kozak, and T. J. Rodenbaugh. 1983.
“Leak Location Methods for HV Underground
Shannon, W. L., and W. A. Wells. 1947. “Tests for Ther- Cables.” IEEE Transactions on Power Apparatus and
mal Diffusivity of Granular Materials.” ASTM Proce- Systems. Vol. 102. July. pp. 2029–2037.
dures. Vol. 47. Pp. 1044–1055.
Williams, J. A., D. Parmar, and M.W. Conroy. 1993.
Sheirer, E. W., and L. Winitsky. “Cable Installations “Controlled Backfill Optimization to Achieve High
on Bridges.” AIEE Transactions. Vol. 77. Part III. Ampacities on Transmission Cables.” IEEE 1993 Win-
pp. 39–42. ter Power Meeting. Columbus, Ohio.
12-48
CHAPTER 13 Cable System

Construction
Authors: John S. Rector, Black & Veatch
Frank Musick, WA Chester
Reviewer: Jay A. Williams, Power Delivery Consultants Inc. (PDC)
This chapter provides information on issues related to the construction of underground

transmission cable systems, including both the specifications and standards commonly
practiced in the industry. The engineering design principles and supporting calculations
are provided in Chapter 12.
The investment in underground cable systems is growing ever larger, requiring each
installation to be thoroughly understood, carefully planned, and executed by experienced
contractors. Construction remains the single most expensive cost item involved with the
development of an underground transmission cable system. As material prices rise, and
as cable technology continues to press the limits of voltage stress, the cost of the installa-
tion continues to drive the overall cost of these projects. Although materials and pro-
cesses are significantly improved from just a few years ago, the construction techniques
employed in this industry, with the exception of the advances in trenchless technologies,
have generally remained unchanged. The most technically advanced cable is still subject
to failure if the installation is faulty.
The information contained in this chapter is a collection of experience, standard prac-

tices, proven techniques, and manufacturer’s recommendations. Additional information
can be found through the IEEE, vendors and contractors, cable manufacturers, and a
multitude of websites.
John S. Rector is a Senior Project Manager with Black & Veatch Inc.
(B&V). He received a BSEE degree from Kansas State University
(1976). Mr. Rector joined B&V after graduating from KSU in 1977. He
began his career with B&V designing overhead transmission lines and
substations and later became the Section Leader for the overhead and
underground transmission line design groups. He then served as the
manager of the underground transmission line design activities for B&V.
Presently Mr. Rector is a Senior Project Manager in the power delivery
business line. He is a Senior Member of the IEEE Power Engineering Society, and a Vot-
ing Member of the IEEE Insulated Conductors Committee (ICC). He is a registered pro-
fessional engineer in the states of Kansas, Michigan, Texas, Maryland, and Vermont. Mr.
Rector has published several papers on the subject of underground transmission.
13-1
Chapter 13: Cable System Construction EPRI Underground Transmission Systems Reference Book
Frank Musick is a Senior Vice Presi- where he held various positions in underground and
dent with W. A. Chester, L. L. C. overhead engineering and project management. Mr.
(WAC). He received a BSEE degree Musick joined WAC in 1996 as Manager of Operations.
from Bluefield State College in 1985 His responsibilities have included being a Project Man-
and a Masters degree in engineering ager for underground transmission projects throughout
c o n s t r u c t i o n m a n ag e m e n t f ro m the United States and estimating underground transmis-
George Washington University in sion projects. Mr. Musick is a Member of IEEE Power
1990. Mr. Musick began his career Engineering Society.
with the Potomac Electric Power Company in 1985,
13-2
EPRI Underground Transmission Systems Reference Book Chapter 13: Cable System Construction
13.1 INTRODUCTION 13.1.1 General

This chapter describes the installation of underground The installation of underground transmission cable sys-
transmission cable systems. The cable types to be tems requires not only specialized equipment but experi-
discussed include crosslinked polyethylene (XLPE), enced personnel trained in the handling of underground
ethylene propylene rubber (EPR), high-pressure fluid- transmission cables. Underground transmission cables
filled (HPFF) and high-pressure gas filled (HPGF) require precautions beyond those of underground distri-
pipe-type cables, and self-contained fluid-filled (SCFF) bution cables to ensure that the bending radius is not
cables. Included in the discussions are equipment exceeded. The installation of these cables requires the
types, materials, installation methods, and necessary ability to monitor and record pulling tensions as well as
personnel qualifications. distance and installation speed. The increased size of the
reels necessary to accommodate the increase in the
The chapter is generally organized in order of the steps bending radius of these cables necessitates the use of
followed in installations. The chapter also has separate larger reel-handling equipment and higher-capacity
sections on practices that are unique to specific kinds winches (a 100,000-pound winch would be required for
of equipment. The first two sections cover two areas of pipe-type transmission cables).
general practices; Section 13.2 describes trenching
and manholes, and Section 13.3 reviews tunneling, jack- 13.1.2 Responsibility
and-bore, and directional drilling. Sections 13.4 and Areas of responsibility need to be defined and agreed
13.6 discuss pipe and conduit installations for HPFF upon by the project owner, design engineer, and installa-
and extruded-dielectric cables. Section 13.5 describes tion contractor prior to beginning a project. Procedures
direct-buried installations, which are not typically used for identifying and reporting problems that could poten-
in the United States but are more common in Europe. tially affect the design or completion date of the project
S e c t i o n 1 3 . 7 c ov e r s b a c k fi l l i n g m e t h o d s, a n d need to be established at this time.
Section 13.8 reviews cable installation techniques for
extruded-dielectric, pipe-type, and SCFF cables. Sec- 13.1.3 Project Planning
tions 13.9 and 13.10 outline procedures for splicing and
A project schedule needs to be developed prior to the
terminations. Sections 13.11, 13.12, and 13.13 briefly
start of any project. The project owner, design engineer,
describe installation of dielectric fluid, nitrogen gas, and
and installation contractor should review the extent of
pressurization plants. Section 13.14 reviews after instal-
the project and define an expected timeline for the com-
lation tests. The last three sections, Sections 13.15
pletion of the project. Areas of difficulty need to be
through 13.17, cover unique practices for installation of
identified, and appropriate time and resources allocated
pipe freezes, fiber optic cables, and submarine cable.
for these in the schedule. Possible difficulties that should
be taken into consideration include zoning issues and
The majority of underground transmission system
work hours, community resistance, material lead time,
installations are performed by contractors. Very few
areas with restricted work space, heavily traveled areas,
utilities have their own installation crews or equipment
and the necessity for a deep trench area.
required to perform underground installations.
13.1.4 Safety
Additional information can be found in the Association
of Edison Illuminating Companies (AEIC) “Guide for Project safety is the primary concern in any project. On
Installation of Pipe Type Cable Systems” (AEIC 2005) large underground transmission projects, it is common
and “Guide for Installation of Extruded Dielectric Insu- practice to require a dedicated project safety manager. It
lated Power Cable Systems Rated 69 kV Through would be the safety manager’s job to promote a culture
138 kV” (AEIC 1997), the 1992 Underground Transmis- of safety for the project and to ensure that all estab-
sion Systems Reference Book (EPRI 1992), and the 1957 lished safety rules and regulations are being adhered to
Underground Systems Reference Book (EEI 1957).Chap- by the personnel.
ter 12 of the present book addresses installation design.
Also, additional information can be found in the IEEE It also is common practice for the owner to request
P1120, “Draft Guide for the Planning, Design, Installa- the contractor to have a safety manual, perform daily
tion, and Repair of Submarine Power Cable Systems” tailgate sessions that relate to the project at hand,
(IEEE 2005). and request a copy of the contractor’s experience modi-
fication rate (EMR) and Occupational Safety and
Health Administration (OSHA) recordables for the last
three years.
13-3
13.1.5 Quality Control/Inspection of Work The contractor and/or the owner need to also address
To ensure a finished product that will provide many the security on a project. This includes protection of the
years of trouble-free service, it is essential that the con- public and personnel and securing the material on a
duits or pipes, cables, and accessories be installed in project. The project jobsite must have the proper sig-
accordance with the owner’s, the engineer’s, and the nage and barriers that alert the public to the jobsite and
manufacturer’s installation specification and guidelines. potential dangers.
The owner and the contractor should develop a quality
control checklist prior to starting any project. The Many projects are located in areas that may have prob-
checklist can be provided to an inspector as an indica- lems with theft. Because of these potential dangers,
tion that all critical areas have been completed within many contractors and/or owners hire full-time security
defined guidelines. personnel. The contractor and/or owner needs to evalu-
ate each project individually.
Many contractors who specialize in underground trans-
mission systems have a quality control manual. These 13.1.8 Environment
manuals are often used with the owner’s specification. Environmental requirements imposed on the construc-
tion activities are to minimize the impact of these activi-
13.1.6 Contractor Licensing ties on the existing natural resource base in the vicinity
Licensing requirements can vary from state to state, and of the project site. This includes minimizing distur-
most states require that a contractor have a license in bances to streams, wetlands, and natural drainage pat-
the state where the work is to be performed. Frequently, terns; vegetation; wildlife; archeological or historical
a state will require that a business and technical exami- relics; and all current land uses.
nation be completed by the contractor. Depending on
the state’s requirements, the licensing process can take 13.2 TRENCHING AND MANHOLES
several months. Prior to starting any underground Excavating for a manhole or trench will often be the
installation project, the contractor needs to verify com- most visible part of the construction project and,
pliance with state and local licensing requirements. because of its inherent danger, it will be of the most con-
cern to both the permitting authorities and the public. It
13.1.7 Permitting/Security is important that the contractor confirm that the project
Permitting has become a major factor in the scheduling is being constructed from drawings that have been certi-
and costs of a project. Concerns about traffic flow, fied as “released for construction” and that the drawings
safety and health hazards, the environment, or the per- are the most current design.
ceived lack of necessity for a project may be expressed
by the surrounding neighbors. Consequently, permitting Prior to the actual start of trenching, the owner or con-
agencies require assurance that the project under contractor must perform the following preliminary tasks:
struction will have the least possible negative impact on • Selecting a qualified civil excavation contractor. To
the public and environment. The contractor must be
ensure a smooth project, it is critical to select a con-
made aware of all permitting conditions and require-
tractor who has experience in excavation in city
ments prior to the start of the project. Generally, the
streets, who understands shoring requirements, who
owner/engineer will coordinate and apply for the per-
is aware of soil conditions in the area, and who is
mits, but it is the contractor who must deal with the
capable of providing additional crews as necessary to
public and permitting agencies on a day-to-day basis.
meet the project schedule. Experience in excavating
Therefore, it is essential that the contractor be informed
for duct or pipe installation for the relevant type of
of all permitting conditions and requirements prior to
cable system is important, and the civil contractor is
bidding a project. This will not only help the project
expected to work closely with the duct or pipe instal-
flow smoother, but it also will help minimize project
lation contractor.
costs and demonstrate goodwill to the public.
• Permissions. Determine and acquire any private per-
Most local authorities have noise restrictions placed on mission or easements necessary.
worksites. The normal restrictions are typically mea- • Approvals. The engineer and/or owner should iden-
sured at the property line. Usually the noise limit varies tify all local, state, and/or federal governmental
based on the time of day. It is often very difficult to approvals and easement for public right-of-way con-
restrict the noise generated by the equipment. struction in the planning stages of the project. The
lead time for permits can often range from weeks to
13-4
years. Projects have often been rerouted due to per- • Drawings. All drawings necessary to locate the final
mits. Manhole locations have often changed due to placement of the cable system should be provided to
approval issues, affecting the cable cut lengths. the construction manager. The complete drawing
• Permit Requirements. Acquire all environmental, package typically includes plan and profile drawings
building, and/or excavating permits and exhibit in the of the route, manhole placement drawings, typical
construction manager’s office. cross sections, and borings.
• Police and Fire Department Notifications. If traffic is The construction manager should always verify that
to be affected, both the police and fire departments they are constructing the project from the latest set
need to be provided with names and contact numbers of revised drawings. On large projects, the
for the construction manager and they need to be owner/engineer often has a person responsible for
kept current with construction progress so that traffic document control.
patterns can be adjusted and emergency equipment • Project Review. The contractor and owner should
routed around the construction areas. review the drawings prior to the start of construction
• Public Relations. Notify all property owners/tenants to identify sensitive utility crossings and the difficult
along the cable route of the project and the proposed and/or easy areas so work schedules can be planned
start date. The notifications are normally handled by accordingly. They also should review drawings
the utility by either using door hangers or placing a to determine if any special interaction or advance
notice in the monthly utility invoice, however some- notification will be required with the public
times the contractor will perform this activity on and/or environment.
behalf of the utility. • Traffic Control Plans. Most large underground trans-
All property owners must be provided with pedes- mission projects require an approved traffic control
trian and vehicular access to their property. Work plan. These plans may be included as part of the per-
schedules should be planned to minimize inconve- mit or design; however, the contractor may be
niences. Do not encroach on private property during required to provide the traffic control plan for
the installation process without written permission approval prior to starting a project.
and do not place material or equipment on private The traffic control plans typically show the require-
property without either written permission or a wit- ments for the required signage and barricades, noting
nessed verbal approval. changes in pedestrian walkways, speed limits for
Many utilities require that the entire alignment be vid- vehicular traffic, and lane changes. At night, illumi-
eotaped prior to the start of construction, during con- nated flashing signs and light plants must be used to
struction, and after construction is complete to guide vehicular traffic and identify the work area.
provide information on existing conditions and poten- The workers should wear reflective vests when work-
tial damage if there are claims by property owners. ing in public space.
• Call for Utility Locate Service. Prior to any excava- While construction crews are working, vehicles not in
tion, the contractor must request the local utility ser- constant use should be parked off site to minimize
vice to locate all utilities in the impacted area. These the impact to traffic.
agencies have different names throughout the United
States. Utilities normally require 48 hours advanced 13.2.1 Survey of Route
notice to perform the location service, and the The contractor should be responsible for contacting the
marked-out locations will only be valid for a specific governing authorities for the locations of existing utili-
period of time. Should a requested “locate” be on ties and for obtaining a survey of the route. The con-
private property, the locating service may not tractor should also be responsible for marking the
respond to the request. In these instances, the con- centerline stationing and offset stationing along the pro-
tractor should notify the other utilities directly (for posed cable route with either paint on streets and side-
example, gas, water, electric distribution, telephone, walks or with stakes when in open field environments.
sewer, transit systems, central heating, and fuel) of Offset stationing marked every 10–20 ft (3–6 m) is nec-
the pending excavation and request that their services essary so that the trench location can be verified once
be located and marked. If at all possible, the contrac- the centerline of the trench has been excavated.
tor should obtain historical data from the other
underground utility agencies about soil and obstruc- The contractor and/or owner will often perform a visual
tions encountered but not included on the drawings survey of the route prior to the start of construction.
for construction. The visual survey is usually videotaped and dated so
13-5
there is a record of the condition of the site before and an attempt to provide uninterrupted installation of the
after construction. piping. Street crossing excavations often require addi-
tional man-hours due to the high probability of encoun-
13.2.2 Excavating the Trench tering multiple underground services and the necessity
Once the previously mentioned tasks have been com- to hand excavate to avoid damaging these services.
pleted, the excavation and associated activities can
begin. At this time, the contractor should compare the Figure 13-1 shows a duct bank installation within a
marked route with the most current certified drawing city street.
and select the most logical location to permit the work-
ers to become familiar with the equipment and installa- The soil excavated from the trench should be inspected
tion techniques while maintaining an acceptable to determine whether it is suitable to be used as backfill
production rate. material. The soil should be inspected for thermal char-
acteristics, compaction, and contamination.
Manhole placement is generally more difficult than
trench placement; therefore, it is common to spot and If the soil is to be used as trench backfill, the material
install the manholes (or at least dig test pits to verify will likely need to be hauled and stored off site. The soil
acceptability of manhole location) prior to trenching in should be covered to prevent it from being saturated by
the vicinity of the manhole locations. potential rain.
After determining the starting point, the contractor will 13.2.3 Water Removal/Well Pointing
sawcut pavement several hundred feet in advance of the When water is encountered during a trenching opera-
actual excavation to minimize disturbances to the road tion, a trash or ditch pump can be used to handle the
surface. Depending on the resurfacing requirements removal of the water and permit the work to continue.
established by the local authorities, cut lines may need However, in locations where the soil is sandy or has a
to exceed the proposed trench width. coral base or the presence of water is either constant or
cyclical (e.g., due to tidal action), the trench can be
It is advantageous at this point in construction to have maintained in at a relatively dry condition by well-
an out-front or exploratory crew assigned to opening pointing. Well-pointing is the vertical insertion of 2-in.
street crossings and excavating manholes. This crew can (5-cm) diameter perforated steel or plastic pipes into the
function as a “practical” survey crew, documenting on earth to an average depth of 8 ft (2.4 m) and about 8 ft
the engineering drawings the locations of missing or (2.4 m) on centers. These well pipes are then connected
incorrectly located services that would require modifica- to a 6-in. (15-cm) or 8-in. (20-cm) header (manifold)
tions to the location of the circuit or its route. This activ- pipe that, in turn, is connected to a large, vacuum-
ity is commonly referred to as “potholing.” This crew assisted centrifugal pump for water removal. The aver-
excavates crossings ahead of the main trenching crew in age assembly for a single pump is 1000 ft (300 m) in
length with sufficient assemblies installed to permit
uninterrupted trenching, pipe installation, or duct work
and backfilling operations. Once the well-point assem-
bly is installed, the pumping operation continues with-
out stopping until the trench backfilling operation is
approximately 60% complete. Prior to starting the exca-
vation, the well-point system should be operated for
about 24 hours or longer. As the trenching and backfill-
ing operation is completed, the well-point system is
advanced as required. Normally, the 1000-ft (300-m)
assembly is arranged in two 500-ft (150-m) sections for
advancing to prepare the next trench subsurface path
section, allowing continued excavation in a safe, dry,
and economical mode.
Not all sections of the trenching route (especially street

crossings and areas with numerous underground
Figure 13-1 Illustration of a duct bank installation in a city services) can be completely dewatered by using well-
street environment (with permission of ITCTransmission). pointing if the installed system would be a hindrance to
13-6
traffic or could possibly cause subsidence and subse- ings or protective system recommendations of a regis-
quent damage to existing underground services or even tered professional soil engineer in place of shoring and
building foundations. Pipes, ducts, and/or cables can be sheeting; but, regardless of the material being excavated,
installed under marginally dewatered conditions by if the trench is to be deeper than 5 ft (1.5 m), shoring
using large or numerous trash (ditch) pumps placed in and sheeting must be implemented if the trench cannot
the trench as it is being excavated. However, in some be benched as described below.
costal areas where strong tidal action is present, dewa-
tering of a trench may not be possible. In such cases, the The terms “shoring” and “sheeting” are defined as
construction schedule should be planned around the follows:
tidal influx. • Shoring (shoring system) refers to a structure such as
a hydraulic, mechanical, or timber shoring system
Unless test hole data are provided in the bid package,
that supports the sides of an excavation and is
well-pointing will not normally be included in the con-
designed to prevent cave-ins. In short, shoring is a
tractor's pricing for excavating, and the owner will need
simple framework of horizontal braces across the
to state in the request for proposal (RFP) that any well-
width of the trench holding either vertical (uprights)
pointing is to be included in the trenching proposal.
or horizontal (stringers) members against the wall of
the trench.
In some states, a discharge permit may be required prior
to pumping water from a trench, and the owner/con- • Sheeting is that part of a shoring system that retains
tractor will have to make a formal submittal to the the earth in position. Sheeting is a continuous wall of
appropriate agencies. The information required will wood or metal held in place by the horizontal braces
include but not be limited to (1) a sketch of the job site of the shoring system.
location showing what area is to be dewatered, (2) the
location where the water is to be pumped, (3) the catch Pedestrian and vehicular traffic require protection from
basin, and (4) where the catch basin discharge will be. a possible trench hazard. One method in use is to posi-
Contaminated water generally must be collected in tion vertical shoring pieces or sheeting so that their top
holding tanks for proper disposal off-site. elevation becomes a guardrail on which an applicable
safety sign can be prominently displayed. In the past, a
When catch basins in tidal areas are drained into rivers soil bank from the excavation would be used as a natu-
or canals that are tidal, the water may be forced back ral vehicle barrier. The extension of shoring pieces or
into the trench during high tides and then spill over sheeting above the street surface has become more
onto the streets, becoming an unplanned traffic hazard. widely accepted because most municipalities require
that excavated spoil be removed from the construction
Another problem that could occur during the dewater- area and stored at an off-street location. In suburban
ing of a trench while using the well-point method is the and rural areas, a trench may be “benched,” (i.e.,
extraction of fine grains of sand from the ground adja- stepped back) to prevent collapse. Because benching
cent to the trench. The contractor may begin to notice widens the trench at the ground line, it is seldom used in
sinkholes in locations adjacent to the well-point which urban areas.
may cause damage to paving and structural foundations.
13.2.5 Traffic Control/Protective Barriers and
13.2.4 Shoring/Sheeting Signs
“Shoring” and “sheeting” are terms used to describe the As with all construction work, safety is of paramount
type of protective system used to prevent the cave-in or concern. As part of permitting conditions, most large
collapse of trenches. In general, the depth of a trench projects are required to submit a traffic control plan
should be equal to or greater than the width. The 2004 along with the permit application. Traffic control plans
OSHA standards state that trenching need not require range from using “typicals” (i.e., typical plans) provided
shoring or sheeting if the excavation is less than 5 ft by the highway department to having drawings prepared
(1.5 m) deep and shows no indication of a cave-in. How- by a certified traffic control consultant. A traffic control
ever, excavation contractors will often shore a trench consultant will normally be required on large projects
that is less than 5 ft (1.5 m) deep if the trench is in a that involve either heavy traffic volume or when traffic
public space that has exposure to vehicular traffic. will be traveling at a high rate of speed.
The need for shoring and sheeting can be debated when The public needs to be protected from construction site
excavating rock. OSHA will sometimes accept the find- hazards at all times. In general, construction sites
13-7
require obvious barricades and numerous signs that In areas where a clearance is indicated as 1 ft (0.3 m),
identify changes in pedestrian walkways or speed limits additional reinforcing such as placing or pouring a bar-
for vehicular traffic. Without impeding traffic, construc- rier wall of concrete should be incorporated to protect all
tion vehicles that are not in constant use should be services involved. As a provision for future maintenance,
parked at the trailing end of an open trench and along utilities will generally try to maintain a 3-ft (0.9-m) work-
the vehicular side of the trench. Unless required by space clearance when cables parallel other services.
municipal regulations to be parked “off street,” digging
equipment should be strategically placed at the end of In many instances, utilities will state in the bid specifica-
the workday so as to ward off vehicles approaching too tions that the excavating contractor must have a small
close to an open trench or pit area. Steel plates are gen- exploratory excavating crew dig at certain locations
erally required to cover exposed trenches, and illumi- along the proposed cable route to determine question-
nated flashing signs should be used to guide vehicular able subsurface conditions. As mentioned previously,
traffic flow. At night, battery-operated flashing lamps contractors normally have an out-front exploratory
attached to standing barriers should be placed along trenching crew at work. When undesirable subsurface or
the trench. poor soil conditions are encountered, the responsible
utility engineer should be notified daily, or as required,
On roads with high traffic volumes during commuter and a decision requested for a change in trench width or
periods, some jurisdictions require that the trench be depth and/or route. All changes must be incorporated
temporarily restored (with steel plates or temporary into “as-built” drawings for the utility and municipal
pavement) during the high traffic periods. This leads to records. At a later date, these records can protect the
a slower and more expensive installation. installed cable system from dig-in failures by providing
the exact cable route information to any utilities or exca-
13.2.6 Bed of Trench vating contractors planning to dig in the area. Many
The trenching depth and width dimensions should be municipalities have established an information center
included on the route layout drawings created by the that can be contacted when excavation is planned any-
design engineers and given to the contractor. The bases where within their jurisdiction.
for these dimensions (e.g., cable ampacity requirements
and number of circuits to be installed) can be found in After having determined the location, plan, and profile
the previous chapters. of the trench, the excavator should complete the trench-
ing so that the bottom is smooth, uniform, and free of
When excavating a trench, the depth and width may loose rocks, dirt, stones, or other foreign matter. Also,
need to be modified because actual locations of objects the trench walls should not have any protruding rocks
vary from the locations indicated on the layout drawings or foreign matter. Nothing should be left in the bottom
or because previously unknown subsurface obstacles or walls of the trench that might damage the pipe coat-
and/or soil with poor thermal conductivity are discov- ing, pipe, conduit, or cable during installation.
ered. Poor thermal conductivity can be due to soil that
contains such things as ashes, cinders, rubble, or other 13.2.7 Backfilling the Trench
materials containing large voids. When encountering The engineer normally specifies the backfilling require-
poor soil conditions, removal of soil for a distance of ments around the pipe or conduit. The specification
2 ft (0.6 m) in all directions from the final position of usually states the required soil thermal resistivity (com-
the pipe, cable, or ducts is common practice. Providing monly called Thermal Rho or TR) value of the material
that the design drawings do not indicate otherwise, the
minimum clearances (as shown in Table 13-1) should be Table 13-1 Minimum Design Clearances for Trenching
used to determine the trench plan and profile when sub-
Minimum Design Clearance
surface obstacles are encountered. Measurements
Parallel Crossing
should be from the installed pipe, cable, or duct-bank
Obstacle ft m ft m
surface. Effects on ampacity should be investigated for
Storm Sewer 1 0.3 1 0.3
any heat source within 10–12 ft (3–4 m) of the cable. The
excavator or inspector also should identify soils with Water Line 1.5 0.45 1.5 0.45
potentially high thermal resistivity (e.g. organic materi- Gas Line 1 0.3 1 0.3
als) and notify the owner/engineer to determine if the Electric Duct Bank 10 3 2 0.8
trench design should be modified. Telephone Duct Bank 1 0.3 1 0.3
Steam Line 10 3 4 1.2
13-8
and whether the backfill material is to be a select ther- vated spoil earlier in the day is now used to place the
mal sand, concrete, or Fluidized Thermal Backfill™, or controlled backfill around the pipes and the balance of
FTB™ (similar to a weak concrete mix; see Chapter 12). the trench filled with good-quality excavated spoil. If
The envelope around the pipe is usually no more than concrete or FTB is used for the controlled backfill, then
2 ft (0.6 m) over the top of the pipes, and the remaining cement trucks deliver it, and the material is slumped into
fill may be the soil that was removed from the trench if it the trench. When placing FTB around a duct bank, the
is suitable. As described later in this chapter, most duct contractor needs to use a vibrator to prevent air pockets.
banks have a high-strength concrete envelope for The contractor should also be aware that curing con-
mechanical protection and to aid in heat dissipation. crete has, upon occasion, softened PVC conduits.
The thermal resistivity of the backfill material defines On pipe-type cable systems, it is good practice for the
how well the soil dissipates heat away from the contractor to test the pipe coating for damage (com-
pipes/conduits. Typical backfill thermal resistivity monly called jeeping the pipe) just prior to backfilling to
requirements are discussed in more detail in Chapter 12, ensure there are no coating defects. See Section 13.6.3
Section 12.2.6. for a discussion of electrical tests on pipe coatings.
When placing the backfill around the pipes, care should 13.2.8 Street and Highway Crossings
be used not to damage the pipe coating. When using Almost all underground transmission cable system
FTB, it is preferred not to have the concrete trucks routes encounter streets having numerous underground
dump directly onto the pipes. Using a sheet of plywood utilities. These streets and the underground utilities
to deflect the FTB is preferable. The plywood is must be thoroughly researched to determine the best
removed from the trench once the pipes are covered. way to cross them. Some street crossings have few or no
Before placing controlled backfill around cable pipe, the obstacles involved in the installation of a cable system,
pipe should be supported at least 6 in. (15 cm) off the but the majority, especially in urban areas, have numer-
trench bottom. These supports can be sandbags. If more ous obstacles, especially utility services, to avoid.
than one pipe is being installed, the pipes should be Figure 13-2 shows an example of subsurface obstruc-
spaced in accordance with the design drawings so that tions that can be encountered in city streets.
there will be sufficient controlled backfill between the
pipes and the trench wall. The design engineer, in planning an installation through
these areas, must specify one or a combination of several
When other subsurface services are encountered and of the following techniques: open cutting, tunneling,
cannot be avoided, protective barriers should be used, boring, or horizontal directional drilling. These tech-
such as synthetic barrier boards or steel plates, to niques are briefly described in the following paragraphs.
mechanically and thermally protect the cable system.
Where steam lines can not be avoided, utilities have on • Open Cut. Open cut excavation with a backhoe can
occasion wrapped the cable pipes with copper tubing be used when all known services are below the trench
connected to a water source, creating a buried heat-
transfer mechanism to assure the proper pipe-surface
temperature. Also, utilities have built concrete boxes
around the pipes open to the street surface but covered
with latticed grating, allowing air ventilation in this
area. Table 13-1 lists minimum clearances between
transmission cable pipes and other utility services.
Another mechanical protection that has been used is

placing steel plates along the entire trench above the
backfill. This protective method has been used when the
pipes were installed near the street surface because
installation at deeper depths was impossible due to
obstructions such as a railroad tunnel.
The backfilling procedure for a pipe-type cable system is

a function generally performed during the last hours of
the workday. The equipment used in removing the exca- Figure 13-2 Subsurface obstructions (with permission of
Northeast Utilities).
13-9
bottom or below a manually excavated pathway are the allowable pulling tension on the conductor(s),
beneath the known, and at times unknown, services sidewall pressure on the cable(s) in traversing a bend,
to permit the installation of the pipe, cable, or duct maximum cable footage on a reel that is sized to be
bank. An open cut trench requires a very strong, legally transported (height, width, and weight) over fed-
solid shoring and sheeting system and, likewise, a eral, state, and local roads, and access to the manhole at
continuous bridging of the trench with 1-in. (2.5-cm) the present and in the future. The typical manhole is
steel plates to permit unimpeded vehicular and sized and designed for the type of cable system to be
pedestrian traffic. The steel plates must handle the installed and the equipment and the tools needed to pull
weight of heavy trucks and be anchored to prevent and splice the cable(s).
any sliding or bouncing.
• Tunneling. When an open cut trench is not feasible The contractor normally has a choice of either a precast
because of such factors as depth, number of subsur- or cast-in-place manhole. In the earlier days of installa-
face services exposed, and municipal regulation, a tion of underground transmission systems, cast-in-place
tunnel or bore must be installed. Tunnel installation manholes were typically used. However, today nearly all
is basically a mining operation where manual labor is manholes are precast. The reasons for the change are
used to dig and construct a tunnel that is lined with price, and the time required to install a cast-in-place man-
either continuous oak or steel roofing and shoring to hole is usually much greater than for a precast manhole.
prevent cave-ins. Tunnels can be installed to negotiate
a 90° turn in the cable routing. The steel tunnel men- Precast manholes are typically shipped directly from the
tioned is commonly known as a liner-plate tunnel factory to the installation site, and the contractor has
and is used more frequently than an oak tunnel. The the hole excavated and shored prior to delivery.
cross-section view of an oak tunnel is basically a Figure 13-3 shows a precast manhole section being off-
square, while that of the steel liner-plate tunnel is cir- loaded at the location of the installation. The contractor
cular. Newer installation technologies use automated needs to ensure that there is an adequate-sized crane
tunneling devices in addition to the previously men- available to set the manhole. Precast manholes typically
tioned hand-dug procedures. come in two sections, and, based on their size, each sec-
tion can weigh up to 60,000 pounds (27,200 kg).
• Bores. Bores are an alternative to tunneling for the
installation of cable, pipe, or duct systems. Boring is The contractor needs to inspect the site to ensure that
an auguring operation that, at the same time, pushes there are sufficient overhead clearances to install the-
(jacks) a casing (generally steel, HDPE, fiberglass, manhole. If overhead transmission/distribution circuits
or reinforced concrete pipe [RCP]) through the soft are present, the contractor may need to schedule an out-
ground beneath the street or highway crossing. The age with the local utility and, in some cases, even have
boring/casing method is the most popular design the lines temporarily moved for the manhole installation.
when installing pipe/conduits when no bends are
required in the route layout, and the total length does
not exceed 300 to 400 ft (90 to 120 m).
• Horizontal Directional Drill. As price and time
increases with additional distance of a crossing,
another alternative that is often more attractive is
horizontal directional drills (HDD). Over the last
several years, the advancements in HDD technology
have increased the accuracy and decreased the pricing
of this technology.
HDD, tunneling, and boring are discussed in further

detail in Section 13.3.
13.2.9 Manhole Installation

For underground transmission cable systems, manholes
are installed at the locations where it is necessary to pull
and splice cable. The factors contributing to the neces- Figure 13-3 Precast manhole installation (with permission
sity and placement of manholes along the circuit route of ITCTransmission).
13-10
The contractor needs to ensure that adequate traffic • Crossings under waterways such as open drains or
control plans are in use. The setting of manholes often across public spaces such as parks are best accom-
requires the closure of additional lanes of traffic. plished by either jack and bore casings or by using
directional drilling procedures to minimize impacts
It is a good idea to install the manholes at the beginning to the surface and to the environment.
of a project. This helps ensure that unknown utilities
will not necessitate moving manhole(s). Relocating 13.3.1 Tunneling and Microtunneling
manholes will affect the cable cut lengths that have been Conventional tunneling is similar to pipe jacking. The
previously determined. If manholes are installed as the tunnel is excavated manually or mechanically, and the
pipe or conduit installation progresses, then sufficient spoil is removed using an auger system, carts, or con-
lead time for manhole installation should be planned to veyor. The difference is in how the pipe is installed. In
allow continued installation of pipe through the man- the case of tunneling, the pipe or tunnel liner is added at
hole without disruption of the conduit installation. the face of the tunnel. A temporary casing precedes the
liner plates, which are assembled at the face of the tun-
In flood plains or tidal action areas, the manhole place- nel and added on. The temporary casing is then jacked
ment area should be constantly dewatered. Dewatering forward against the liner, as opposed to the entire length
in this type of area should continue until the cable pull- being jacked forward. Additional pieces are added, and
ing and splicing has been completed, since manholes the procedure continues. The diameters of the casings
can float at times. After all work has been completed are usually greater than 3.5 ft (1.2 m), with essentially
within the manhole, the space can be backfilled with no upper limit. The driving distance also is unlimited.
sand or suitable backfill.
Historically, there have been two main types of tunnels
13.3 TUNNELING, JACK-AND-BORE, AND used for underground cable systems: box tunneling that
DIRECTIONAL DRILLING uses oak wood and circular steel liner-plate tunneling. In
When new conduit systems are to be installed in estab- both cases, the excavation was done manually because
lished neighborhoods or along existing streets and high- the size of the required earth cavity for cable systems
ways, it becomes increasingly difficult to find adequate does not permit digging machines to be used. The mini-
space within the public utility right-of-way to accommo- mum dimension of an oak tunnel is a 3-ft (0.9-m) square
date the new system. Often it is necessary to install cas- opening to facilitate digging, and the minimum for a
ings beneath existing facilities, either by tunneling, steel liner-plate tunnel has a diameter of 4 ft (1.2 m).
microtunneling, directional drilling, or jack and bore Basically, a box tunnel is a square or rectangular hole
methods. This section discusses these methods in gen- that is dug approximately horizontally into the earth
eral and provides some guidance for the designer who and lined with oak slats that are 2 x 8 in. (5 x 20 cm) in
may need to consider these alternatives. cross section that will support the roof and prevent the
sides from collapsing. A steel liner-plate tunnel is a cir-
The most common cases where a casing or tunnel might cular hole that is dug horizontally into the earth and
be required on a typical underground construction supported from collapse by arc sections 16-in. wide x
project are listed below: 38-in. long (0.4 x 1 m) of steel plates bolted together.
• Major streets and highways cannot be open cut due The tunnel was excavated only enough to permit the
to (1) high volume of traffic, (2) the disruption to immediate installation of the oak lumber supports or
traffic of open cutting and the ensuing traffic control, the liner plates. If the earth was excavated too far in
(3) the potential safety hazard, and (4) a moratorium front of the supported tunnel, the overhead earth could
on cutting newly installed pavement. begin to collapse. This is referred to as “exceeding the
• Municipality does not allow major intersections to be stand-up time.” Normally, this type of tunnel construc-
open cut, thereby requiring casings to be installed. tion is selected only for cohesive clay or clay/sand soils,
• Railroads require casings to be bored or drilled not a “running-ground” soil that flows into the exca-
beneath their right-of-way. vated area while the tunnel is being dug. The minimum-
size work pit to begin the tunnel construction was 5-ft
• Intersection is so crowded with underground utilities, wide x 12-ft long (1.5 m x 3.7 m). Because street traffic
open cutting and installation of conduits from vibrations can reduce the “stand-up time” of the tunnel
the road surface is not practical or possible. In overhead excavation, a suggested depth from street level
these cases, boring or drilling would be the only fea- to top of tunnel roof of 6 ft (1.8 m) was desirable. This
sible recourse. dimension varied based on the soil surveys taken and
13-11
analyzed prior to any excavation, as well as the presence casing material. This is accomplished by the use of a
of subsurface obstructions. trackhoe working from ground level reaching into the
pit to extract sections of casing. These two pits are usu-
Although these types of tunnels are still used today, ally deep enough to require sheeting and shoring per
newer technology and work practices have largely OSHA rules. Soil samples should be taken at each bore
replaced these processes. location to determine the sheeting/shoring needs and to
determine if dewatering will be expected.
Microtunneling is a horizontal boring technique that is
used in instances where the diameter of the final casing The boring operation is performed by a machine that
is greater than what is customarily used in traditional jacks (pushes) the casing forward into an excavation
jack and bore operations. Microtunneling uses highly that is simultaneously being bored (augured) by a rotat-
sophisticated, laser-guided, remote-controlled equip- ing auger within the casing. Figure 13-4 shows a typical
ment. This enables the user to monitor its precise loca- boring machine within the jacking pit. Directional accu-
tion and to ensure accurate line and grade. This system racy is obtained by (1) careful alignment of the
typically consists of a steerable tunneling machine that machine’s rail-type guidance system, (2) carefully plac-
is jacked from a shaft ahead of the permanent pipes or a ing and welding or jointing each casing section, and (3)
temporary casing. The spoil is removed by using either applying the appropriate directional force on the casing
the auger method or the slurry method. The auger while being pushed forward into the augured hole. Nor-
method uses a continuous flight of augers to remove the mal boring operations have a variety of boring tips
spoil. The auger is surrounded by a casing placed inside available for use on the augers, dependant on the
the permanent pipes or the temporary casing. The soil/rock encountered. The rotating auger is located at
slurry method uses a slurry mixture to pump the spoil the casing/earth interface or slightly in front of the cas-
back to the surface. ing. When boring under existing aboveground facilities,
such as railroads and highways, it is necessary to keep
13.3.2 Jack-and-Bore Installations the boring tips inside the casing lip to prevent any earth
The jack-and-bore method is normally used to install a settlement that might adversely affect the integrity of
casing for a cable system under roadway crossings, street the railroad track system or highway pavement.
intersections, railroad crossings, and small waterways
where the cable system crosses in a straight and horizon-
tal line. Casings installed this way are usually limited to
between 300 and 400 ft (91 and 122 m) in length,
although occasionally longer lengths can be possible,
depending on soil conditions. Installations longer than
this distance are more difficult to steer accurately. Also,
the jacking forces applied to the casing on longer runs
may become too great for the casing material. Bores are
usually the least expensive of the three casing installa-
tion methods; however, each method has its own unique
applications. The boring operation usually takes less
time than tunneling or drilling and requires the simplest
and least amount of equipment.
The jack-and-bore operation is begun by excavating two

pits; one for boring (auguring) and jacking and one for
receiving the casing. The jacking pit is approximately
40-ft (12.2-m) long by 10- to 15-ft (3- to 4.6-m) wide and
is set to the required depth of the bottom of the casing
to be installed. The receiving pit is usually smaller than
the jacking pit because no equipment needs to be set
into this pit. The only activity at the receiving pit is the
recovery of the casing pipe and, in some cases, removal Figure 13-4 Jack-and-bore operation (with
of a temporary steel casing that preceded the permanent permission of ITCTransmission).
13-12
Figure 13-5 shows the final casing pipe being jacked amount of ground area on each side of the crossing. In a
into the bored hole, pushing the temporary steel casing normal HDD operation, the drilling rig and approxi-
out into the receiving pit. mately six to eight equipment trailers are located in an
area at the insertion side of the drill. Figure 13-7 shows
There are a variety of casing materials being used today the HDD drilling rig, control trailer, and equipment
besides conventional steel. These materials are discussed trailers at the insertion side. The trailers contain the
in Chapter 12, Section 12.3.3. control room, auxiliary power units, the drilling mud
control and cleaning system, a construction office, drill
Figure 13-6 shows the PVC conduit package being pipes, and spare parts and tools. The other side of the
installed into the final casing. Notice the conduit spac- crossing must have a suitably large space for the casing
ers holding the conduits in the required configuration. or conduit string that is to be pulled back into the hole
to be assembled, an area to contain drilling mud, and an
13.3.3 Horizontal Directional Drill Installations assortment of tools and equipment to assemble the cas-
HDD technology has advanced significantly in the past ing/conduits. Usually this can be accommodated in a
few years, and is becoming widely accepted. A detailed long, although relatively narrow, space.
discussion of HDD technology is provided in
Chapter 12, Section 12. 3.3. HDD requires a significant As with jack-and-bore installations, casings—if used—
may be made of steel, plastic, polyethylene, or compos-
ite materials. It is possible to install the conduit or pipe
directly into the drilled hole without a casing, depend-
ing on the nature of the soils and the thermal perfor-
mance required from the cable system. When casings
are installed, the casing sections are welded together in
long sections on the pullback side of the crossing to
reduce the amount of welding during the pullback oper-
ation. Similarly, polyethylene casings are welded using
butt fusion, electro fusion, socket fusion, or compres-
sion couplings. Special care must be taken when choos-
ing the casing material and the wall thickness to ensure
that the casing is capable of sustaining the forces
encountered during the pullback operation. Consult-
ants or contractors who specialize in HDD installations
will recommend the appropriate casing material and size
required to accommodate the number of conduits
Figure 13-5 Jacking final casing into place (with permission required in the crossing.
of ITCTransmission).
Figure 13-6 Installing the conduit package in casing (with Figure 13-7 HDD drill rig (with permission of PG&E).
permission of ITCTransmission).
13-13
The HDD process begins by inserting the drill head/drill backfill is pumped into the filling tubes, the tubes are
pipe assembly into the ground at an entry angle of slowly withdrawn from the casing, allowing the annular
approximately 11 to 12º, relative to horizontal. A drill- space between the conduits and the casing to be com-
ing fluid conditioning system is set up to lubricate the pletely filled with backfill.
drill head, remove cuttings from the hole, and recycle
and clean the drilling fluid for reuse. The pipe string will Once the annular space between casings is completely
contain a control cable connecting multiple magnetome- filled, the ends of the casing are sealed to prohibit the
ters and accelerometers in the drill head to the control thermal fluid from escaping and to prevent the backfill
trailer. This is used for direction control and tracking of from drying out.
the drill head while the pilot hole is being drilled. The
drill begins from the point of entry moving toward the 13.4 CONDUIT INSTALLATION FOR SCFF AND
exit point on the other side of the crossing. As the drill is EXTRUDED-DIELECTRIC CABLES
advanced, more sections of drill pipe and more sections
Conduit installation for a transmission cable system is
of control cable are added to the drill string. Once the
much more critical than for a distribution cable system.
drill head reaches the intended exit point on the other
Transmission cables are heavier, have generally longer
side of the crossing, any drilling mud that escapes from
distances between manholes, are much more expensive,
the hole is contained in a pit and is recycled and condi-
have larger bending radii, and operate at higher voltage
tioned as necessary. The pilot hole is then enlarged with
stresses when compared to distribution cables. Addi-
a series of back reaming pulls until the diameter of the
tionally, the duct bank system for transmission cables
hole is of the required size to accept the casing (or the
needs to be stronger to resist the installation loads
conduit bundle).
required of longer pulls and heavier cable. It is impor-
tant that the duct alignment, including the joints, be as
The casing, if used, will be assembled on the other side
straight as possible to keep pulling tensions as low as
of the crossing and pulled from the exit point toward the
possible. Transmission cable systems have failed because
entry point once the final hole diameter has been
of cable damage resulting from improper conduit instal-
reached. As the pullback operation proceeds, sections of
lation. Refer to Chapter 12, Section 12.3.2, for a discus-
casing that were preassembled will be welded together
sion of designing conduit systems.
so that the pullback operation is not stopped, except for
brief periods. Because HDDs are typically long, the cas-
The installation procedure for a duct bank system has
ing that is being pulled back will be quite long. It is
been well documented by utilities, manufacturer’s asso-
advantageous to have as much of the casing preassem-
ciations, and the manufacturers of the materials and
bled prior to the start of pullback as possible. Therefore,
accessories used to construct a duct bank. A duct bank
a long continuous temporary right-of-way is preferred
is defined as a group of conduits that are installed in a
on the pull back side of the crossings.
common trench with spacers to arrange the ducts in a
defined alignment. Duct banks can be encased in struc-
The next operation is to install the pre-assembled con-
tural concrete, thermal concrete, or thermal sand.
duit string, which is held in proper alignment by cus-
Unless the conduits runs are restrained, they will have a
tom-made spacers at approximately every 5 ft (1.5 m),
tendency to float when covered with concrete. A fre-
although there have been many installations made with-
quently used procedure to address this concern is to
out spacers. Occasionally, if a casing is used, the conduit
restrain them by tying them off with wire to short stubs
string is inserted into the casing before the casing is
of rebar every 5 to 10 feet along the trench length. Fig-
pulled back into the hole. Overall drill length, number
ure 13-8 shows a PVC duct bank under construction in
and size of conduits, and overall weight usually dictate if
a paved highway. Refer to Chapter 12, Section 12.3.2,
this can be performed in a single operation or if they
for a discussion of conduit system design.
must be installed separately.
13.4.1 Joining Sections of Conduit
HDD casings, like jack-and-bore casings, sometimes
need to be backfilled with a suitable thermal fluid to Conduits normally come in lengths of 10 and 20 ft (3 m
allow the cable system to achieve the desired design and 6 m) and, depending on how the sections will be
ampacity. To facilitate the backfilling operation, it is connected together, can be purchased in two different
customary in long HDDs to install one or more fill end configurations. The two configurations are (1) bell
tubes in addition to the conduit string. As the thermal and spigot, and (2) straight.
13-14
Bell and spigot conduits have two different end configu- The other type of conduit is simply referred to as
rations per conduit; the spigot end is straight and the straight conduit. This type has two straight ends and is
other end is belled. Because of this difference in the joined together by the use of straight couplings. The
ends, it is imperative that the conduit be installed in the coupling has an inside diameter that is larger than the
proper direction as shown on the construction drawings. outside diameter of the conduits being joined, and a
Failure to install the conduits in the appropriate direc- stop seam at the inside center of the coupling. Conduit
tion will result in the cable being pulled against the edge is inserted into the coupling until the end of the conduit
of the spigot end, possibly damaging the cable jacket. comes to rest at the stop seam in the center of the cou-
The straight end of one conduit is inserted into the bell pling. This lets the installer know that the conduits have
end of another conduit, and the sections are glued been inserted fully into the coupling. The sections are
together using normal PVC glue techniques. Usually, glued together with the couplings using normal PVC
this type of conduit is readily available in sizes up to a glue techniques.
6-in. (15.2-cm) nominal diameter. As stated before, for
transmission cable systems it is imperative that the con- Similar to the bell ends found on bell and spigot con-
duit runs are extremely straight, and all joints are duits, couplings come in two lengths, regular and long.
straight and in alignment. When bell and spigot conduit The long line coupling is approximately 6 in. (15.2 cm)
is installed around a radius, care must be taken to long on each side of the center, and the regular coupling
ensure that the spigot end does not partially remove is approximately 3 in. (7.6 cm) long on each side. The
itself from the bell end and create an elevated and advantage of the long coupling is that it provides a
exposed edge that could scrape the cable jacket during greater gluing surface area, ensures the alignment of the
pulling operations. One safeguard is to specify the use of joint when installed around radii, and provides addi-
bells that are approximately 6-in. (15.2-cm) long (or tional protection against backfill leakage into the joints
long bells) instead of normal length bells, which are only at bends.
3 in. (7.6 cm) long. The longer bell provides a greater
gluing surface, will hold the alignment of the conduits 13.4.2 Field Bends
around radii better, reduces the likelihood of a protrud- Conduit will usually require some field bending to fit the
ing edge developing, and provides greater protection alignment of the trench. When this is required, it is
against leakage of backfill material, such as concrete or important to ensure that the conduit is not flattened
flowable backfill, into the joint. Another good practice during bending. Flattening usually happens when con-
is to bevel the inside edge of the straight end of each duit is either not preheated, an improper bending tool is
conduit piece to 45º. In case a spigot end does slip out used, or the conduit is bent in a radius that is too small
of alignment with the bell end, the protruding edge will for its diameter. These problems can be avoided by fol-
be beveled off and should not protrude above the bot- lowing the manufacturer’s recommended procedures for
tom of the conduit. field bending of conduit. It is imperative that the con-
duit remain circular in cross-section so that it will pass a
mandrel of predetermined length and diameter during
proof testing. Gradual bends can often be made with
long lengths of conduit by laying the conduit string into
the trench and bending it using the trench wall as the
fulcrum. Short radii bends that are near or at the cable
manufacturer’s minimum guidelines, should not be field
bent, but should be purchased from the factory as “fac-
tory bends.” It is always a good idea to mandrel each
factory bend before placing it into the duct bank.
13.4.3 Proof Testing

Proof testing is an important procedure carried out at
the completion of a duct run between manholes. It is a
repetitive process where a string of cleaning swabs,
disks, and mandrels are pulled through each conduit
run between manholes. The goal of proof testing is
three-fold; (1) clean construction debris such as dirt,
Figure 13-8 Duct bank installation (with permission glue, and concrete from the duct system; (2) ensure the
of PG&E). duct system is continuous without breaks or blockages;
13-15
and (3) ensure the roundness of the conduits so that the not anticipated, in or along areas that are restricted to
cable won’t become jammed in a bend that is out of the general public. Examples of this might be along and
round. The designer should specify the size and length within a fenced railroad right-of-way, within a utility’s
of the mandrel based on the inside diameter of the con- fenced substation, along an overhead transmission line
duit and the radius of the smallest bend. For example, it right-of-way, or on commercial or industrial private
is customary to allow one-half inch clearance between property. Direct-buried installations are not practical in
the outside of the mandrel and the inside of the conduit. locations such as in existing road rights-of-way where
A mandrel 5.5 in. (14 cm) in diameter and 16.5 in. numerous utilities or local service laterals are being
(42 cm) long—with a 4-in. (10-cm) taper on each end, crossed or under major roads or intersections where
having a center straight section of 8.5 in. (21.6 cm) in open cutting is not allowed. Recently, it has been
length—would be suitable for a 6-in. (15.2-cm) PVC reported by European manufacturers and utilities that
Schedule 40 conduit system with no bend smaller in there is a trend of moving away from the direct-buried
radius than 8 ft (2.4 m). Refer to AEIC G3-2005, Guide installation in favor of the concrete-encased duct bank.
for Installation of Pipe Type Cable Systems, for addi-
tional information regarding the proper sizing of man- 13.5.1 Trench Preparation
drels. The designer should work with the contractor to Figure 13-9 shows a cable trench suitable for direct bur-
make sure the mandrel dimensions are appropriate for ied cable. Notice the trench’s sloped sides which meet
the specific project. OSHA requirements. The trench is excavated to design
depth, the bottom is cleaned of debris and smoothed,
13.5 DIRECT-BURIED CABLE INSTALLATION and the trench side walls are smoothed and stabilized
A direct-buried cable installation is where the cable is sufficiently to keep loose soil from falling back into the
laid or pulled directly into a suitably prepared open trench. A 4- to 6-in. (10.2- to 15.2-cm) bed of thermal
trench rather than being pulled into a pipe or duct sys- sand is then placed on the bottom of the trench, com-
tem installed in the trench. The trench is backfilled pacted, and moistened to design specifications. The
directly over the cables with native soils, granular ther- sand bed is leveled so that the cable rollers, if used, will
mal backfill, or a flowable thermal backfill material. Pre- sit solidly on the trench bottom.
formed concrete slabs or free-poured concrete caps are
placed over the cables to provide mechanical protection. 13.5.2 Cable Installation
Cables can be installed in a trench by either placing
Although it is not commonly used for transmission them on rollers and then pulling the cable within the
cables in the United States, a direct-buried cable instal- trench, or by transporting the cable along the top of the
lation has been the common installation method trench while still on the reel, and placing the cable
throughout Europe. The most appropriate locations for directly into position in the trench. Normally, cables
a direct-buried cable system are generally in open rural installed in trench without conduit are placed on rollers
areas where road improvement or other construction is and then pulled into position. In this case, the cable roll-
ers are placed in the trench at intervals that average
between 5 and 10 ft (1.5 and 3.0 m) for the length of the
cable pull. Only heavy-duty rollers designed specifically
for large diameter and heavy transmission cables are to
be used. Rollers also must be positioned on all sides of
the trench wherever the cable could come in contact
with the side walls during pulling, especially around
bends. Where the trench crosses under existing facilities,
care must be taken to protect the existing facility and
the cable from possible damage during pulling.
Once the rollers are in place, a pulling line is pulled

across the rollers to begin the cable installation process.
The cable is pulled into the trench across the rollers until
the pulling end is completely through the trench. Inspec-
Figure 13-9 Trench for direct-buried cable (with permission
tors are stationed at intervals along the pull to maintain
of SMECO).
13-16
watch on the cable as it moves through the trench. Fig- rate small vault is typically used to provide access to the
ure 13-10 shows the cable on rollers in the trench. When link box. Direct-buried cable systems in the United
the cable pull has been completed, the pulling line is dis- States are more common on 69-kV through 138-kV
connected from the pulling eye on the cable, and the cable systems, although any voltage can be direct bur-
rollers are repositioned to the next cable-phase location ied. It is common practice in the United States to install
in the trench. This process is repeated for the third 230-kV and 345-kV splices in manholes.
phase. Finally, the cable is raised off of the rollers using
either nylon slings or by hand, then gently placed into 13.6 HIGH-PRESSURE FLUID-FILLED (HPFF)
their final position. The rollers are removed from the PIPE INSTALLATION
trench, and the cables are repositioned on the bottom of
The pipe installation is the first construction phase of
the trench, maintaining the specified separation between
building an underground transmission system. It also is
phases. Backfilling is then begun using either granular
the phase of the project that will be the most visible to
thermal material or native spoils that have been cleaned
the public and cause the most inconvenience. Therefore,
of debris. If granular materials are used as the backfill
to help ensure that the project is completed successfully
material, the material is placed in 6- to 8-in. (15.2- to
with the least amount of impact, it is imperative that all
20.3-cm) lifts, moistened to the specified level, and com-
of the permit conditions are followed, the job site is kept
pacted by mechanical and manual means until the
clean and orderly, and a safe work ethic is maintained.
trench is brought to the appropriate level.
It is imperative that the pipes are installed with the
13.5.3 Direct-Buried Cable Joints
utmost quality control. If the pipe is not properly
Cable joints on direct-buried cable systems can be installed, then the cable also will be impacted during the
located in either splicing vaults (manholes) or directly cable installation.
buried. Direct-buried splices, like the cable itself, should
be located only where access to the general public is 13.6.1 Pipe Installation
restricted, and some means of protection from mechani-
This section covers the transporting, storage, handling,
cal protection can be provided, such as precast concrete
joining (welding), installation practices, and testing of
caps placed above the backfill over the splices. If direct-
the pipe in preparation for the cable installation.
buried splices are used where cross-bonding or open cir-
cuit sheath connections are required, an adjacent sepa-
Although pipe is sometimes shipped by railroad from
the coating suppliers, the majority of pipe is shipped by
trailer for expediency and to minimize storage yard
rental and excessive handling costs. The trailers used
can be either the flatbed or pipe-carrier type. Regardless
of the type used, all pipe should be protected with bed-
ding and spacers to prevent shipping damage to the
coating or pipe flares. The pipe, in double random
lengths (43 ft [13 m] average) should be supported
approximately every 8 ft (2.4 m) on centers with padded
boards at least 2 x 6 in. (5 x 15 cm) to preclude any dam-
age to the pipe coating. The same size boards are used
to space the pipe horizontally to prevent any displace-
ment or shifting which could damage the pipe flares.
The stacking of pipe on the trailer should be limited to
five tiers or less, depending upon the pipe diameter and
weight, to protect the pipe coating.
Coating suppliers have recently begun to use a more

substantial type of truck bolster, in the form of rubber
rings placed on each length of pipe approximately 8 ft
(2.4 m) apart. In this manner the pipe is better pro-
tected, and it is easier to load for shipping and handling
in the field. This type of spacer is well suited for trans-
Figure 13-10 Direct-buried cable installation porting pipe with either a flatbed or pipe-carrier trailer.
(with permission of SMECO).
13-17
Both the external corrosion coating and internal coating 300 ft (30–91 m) per day. The production will vary based
of the pipe must be protected. When the coating sup- on the number of pipes being installed, number of utili-
plier finishes the internal coating procedure, an end cap ties encountered, the depth of the trench, and the traffic
or plug is placed on both ends of the pipe to prevent and soil conditions.
dust and debris from entering the pipe during shipment.
These caps or plugs should always be in place on the Trench being installed in streets and sidewalks is nor-
pipe ends during storage or handling in the trench to mally secured each night by steel plating over the open
maintain a clean and dry pipe. trench. In rare circumstances, the trench can be barri-
caded with either snow fence or jersey barricades. This is
To store the pipe for any length of time at the construc- usually only allowed where traffic is not a problem or
tion yard, the suggested method for protection is to use the trench is extremely wide or deep.
a bedding of sand or sand bags, with no more than five
tiers. If the pipe is stored in cold weather, the sand or The pipe should be inspected prior to installation. The
sand bags could freeze to the pipe coating. If this occurs, inspection should include checking the coating for holi-
a torch must be used to heat the area and free the pipe days, temporarily removing the protective caps, swab-
before lifting. bing the pipe with clean dry rags, and checking for split
ends or damaged pipe ends (flares).
Upon receiving the pipe at the storage yard, a thorough
inspection should be made for any damaged pipe ends The pipes should be carefully offloaded from the trailer,
(flares) or pipe coating and missing caps or plugs. As the using either two nylon slings per section, or using a
pipe is being removed from the transport trailer, the spreader bar with nylon slings. When placing sections of
pipe corrosion coating should be tested for holidays pipe into the trench, the contractor must be careful to
with a detector, and any coating damage should be ensure the pipe coating is not damaged. Often a
marked for immediate repair. One of the first activities spreader bar with nylon slings is used to lower lengths of
that the contractor should do before installing pipe on pipe into the trench.
HPFF systems is to submit a weld procedure to the
owner for approval. Once the weld procedure has been There are different methods for installing the pipe. Con-
approved, the contractor needs to have welders tested by tractors often weld the pipe on top of the trench and
performing a weld and having it tested by an indepen- feed the pipe into the trench by means of a winch line.
dent testing agency. The weld needs to conform to API This method allows the pipe to be fed underneath exist-
Standard 1104. ing utilities and shoring. In preparing to pull the pipe,
the length needs to be determined from the point which
The installation of pipe begins with the trench excava- the pipe will be pulled to where there is easy access for
tion. Normally a contractor would like to have in the the continuation of pipe installation from the street
range of 200–400 ft (61–122 m) of trench open prior to level. Once the pipe length is known, the crew can
starting the installation of the pipe. Most permitting assemble and weld as many double random pipe lengths
authorities limit the length of open trench (often as needed, test and cover all welds, and install a proper
defined as a trench that has not been backfilled to pipe pulling head. The assembly can take place above or
within 6–8 in. [15–20 cm] from the original surface) in the trench. If above, the pipe is laid on sand bags sup-
allowed daily to minimize the impact on vehicular traf- ported by timbers over the trench. If in the trench, the
fic and protect pedestrians from harm. Of course, this pipe can be assembled and laid on the sand backfill
length limit is influenced by population and business material over the previously installed pipe. Or if space
density and can vary along the cable route. However, in permits and the previously installed pipes have not yet
most cases, and especially in urban areas, the rate of been covered with the thermal backfill, the assembly can
installation of pipe or duct will be determined by the be made alongside the installed pipe and laid on sand-
rate of trench excavation. bags in preparation for the pull.
The pipe installation work area is typically a 600- to The proper method to pull the pipe assembly is to use a
800-ft (182- to 244-m) moving work zone along the pulling head, designed to grasp the pipe so that the pull-
route. The first 200 to 300 ft (61 to 91 m) is the excava- ing forces will be evenly distributed to prevent ovality
tion of the trench, the following 200 to 300 ft is the and damage to the flared end. Also, this pulling head
installation of the pipe, and the final 200 to 300 ft (61 to should act as a seal to prevent earth spoil or moisture
91 m) is the backfilling of the trench. This work area from entering the pipe. It is not advisable to pull a pipe
moves along the route at a rate of approximately 100– assembly with any type of nylon sling wrapped around
13-18
the pipe end unless the pipe is completely sealed to pre- When preparing the pipe end, the chill ring is inserted in
vent dirt and moisture entry. Quality control is still vital the pipe end having the smallest diameter flare. This can
at this point. Once the pulling head is applied and a be judged by observing the smallest slot-opening dimen-
suitable and properly sized winch is in place down- sion when the chill ring is inserted into each pipe end.
stream from the pipe assembly, the pull can proceed. As The chill ring should be inserted into the pipe(s) with
the pipe assembly is pulled, constant inspection is neces- the slot opening on the horizontal and toward the out-
sary to watch for any unforeseen obstacle that could side of a pipe bend so that during a cable pull no dam-
damage either the pipe coating or the pipe. Before start- age is incurred. Once positioned in the open end, the
ing the pipe assembly pull, the trench should be thor- chill ring should be tack- or spot-welded to prevent slip-
oughly sanded with final backfill material and/or page. The final move butts both pipes together over the
sandbagged, especially along the trench walls to guide chill rings’ raised triangular ridge spacer, which is used
the pipe around potentially damaging obstacles. as a template and positions each end properly for the
butt weld.
The other method used in areas with less congestion is
Bell and Spigot
to install the pipe end to end within the trench. When
This type of pipe weld connection requires that only one
this method is used, the contractor may need to widen
end of each pipe length be flared. The flared (bell) end is
the trench at the locations of pipe bell ends where the
actually a swaging (expansion) of the pipe for a distance
sections of pipe are welded together. The widened
of approximately 3 in. (8 cm) so that the inside diameter
trench at these locations is necessary to give the welder
of the flare (bell) is approximately 0.125 in. (0.3 cm)
adequate clearance during the welding process. These
larger than the outside pipe diameter. This pipe connec-
widened locations in the trench are called “bell holes.”
tion is made simply by inserting the unflared (spigot)
pipe end into the bell end and making a fillet weld.
There are three types of weld connections that can be
Although weld beads do not occur with this type of con-
made for joining the pipes together. Two require the
nection, the installation still warrants skilled and tested
flaring of one or both pipe ends. The connections are
welders. While cable pulls can be made in either direc-
as follows.
tion with a bell and spigot pipe connection, provided
Chill or Backer Ring that the inside pipe edges are properly prepared, most
Both pipe ends are expanded on a 5º taper to an inside utilities prefer to pull so the cables exit the spigot end.
diameter matching the pipe outside diameter. This facil-
Sleeves
itates the placement of a chill ring between the two
A sleeve connection is used when the pipe ends are not
lengths of pipe being joined. The chill ring is a spacer to
flared. The sleeve has an inside diameter approximately
provide the welder with an alignment template and to
0.125-in. (0.3-cm) larger than the outside diameter of
help prevent weld beads on the inside of the butt weld
the pipe and at least a 0.375-in. (0.95-cm) wall thickness.
that could damage the cable during a pulling operation.
The length is equal to or greater than the pipe diameter.
The chill ring, however, does not positively prevent weld
With this type of connection, the inside pipe edges are
beads from occurring if inexperienced welders are used.
beveled and rounded smoothly to allow the cable to pass
Qualified welders are the key to a quality pipe installa-
through freely and safely without damage during the
tion; therefore, all welders should be tested as previously
pulling operation. After centering the sleeve over the
mentioned prior to performing any work.
pipe ends, fillet welds are then completed. This type of
connection is normally used in tunnel and bridge instal-
A chill ring weld is used in the majority of installations
lations where the pipe could be pre-stressed and
since this type of pipe connection allows cable pulls
anchored. It requires more welding than the chill ring or
from either direction.
bell and spigot approaches.
Prior to welding the chill ring, the pipe ends are cleaned
Welders should be qualified or certified by test in accor-
and brushed. The welder will normally use power tools
dance with API 1104 regardless of their previous experi-
to wire brush or grind the inside and outside of the pipe
ence. Likewise, all pipe, reducer, split-joint, riser,
end. In some cases, the welder might have to remove
and trifurcate welds should conform and be tested to
excessive dents to the pipe end flares with a heavy-
the same standard with the exception of the following
toothed hand file. As the welder prepares the pipe end,
two paragraphs:
ovality is checked, and, if minor, the pipe can be turned
to match as closely as possible the pipe end it is being 1. Radiographic inspection of all welds should be per-
joined to. If the flared end is damaged beyond repair, formed by an experienced testing organization
then the flare can be replaced as described later. approved by the utility. The testers will review and
13-19
make recommendations as to approval or rejection of clean to remove any dirt or sand from the damaged
a weld, but the qualified utility field engineers make area, then as much of the original pipe coating laid back
the final decision. in place as possible. Having done this, the polymeric
2. If a weld is rejected for any defect exhibited during corrosion protection tape should be wrapped around
welding, it should be repaired by grinding or chip- the pipe (if there is sufficient space to allow taping), then
ping the weld material down to the first pass or fur- a “special” split heat-shrink sleeve, made especially for
ther, then re-welding. Most utilities require the entire repairing damaged pipe coating, should be applied.
pipe weld (including chill ring) to be replaced, and, if
rejections continue, the welder should be requested to To ensure that the pipe is adequately heated per the
submit another test weld for continued qualification sleeve manufacturer’s recommendations, it is recom-
and welding on the project. mended that a “thermal gun” be used to verify the pre-
heat temperature.
All pipe welds should be tested by radiographic inspec-
tion as soon as the weld has cooled enough to allow for 13.6.3 Electrical Tests on Pipe Coatings
the placement of the films so weld integrity can be All pipe coatings, whether factory- or field-applied,
approved as soon as possible to permit production to should be tested with a holiday detector for damage or
continue without undue interruption. imperfections during pipe installation and just prior to
backfilling the trench. The test equipment suppliers sug-
13.6.2 Corrosion Coatings and Repairs gest voltage settings ranging from 8–15 kV depending
After the welds have been radiographically tested and on the wall thickness for extruded or heat-shrink coat-
approved, the weld area must be protected (as the pipe ing materials. The testing should be in accordance with
is) from corrosion. Earlier pipe coatings such as Somas- the NACE Standard RP 02 74—latest edition (NACE
tic or enamel wraps used a weld coating material of hot- 2004). Any holidays found must be repaired and
applied enamel. Since these coatings are rarely used any- retested, and all pipe coatings should then test positive
more, the weld coatings used with today’s extruded before being covered with backfill material. A discus-
polyethylene or polypropylene pipe coatings are dis- sion of cathodic protection testing is provided in Chap-
cussed, as follows. ter 10, Section 10.3.8.
A most popular and excellent pipe-weld covering, when After the backfilling, trench coating-resistance tests
installed properly, is a polymeric heat-shrink sleeve. In should be done on the pipe from manhole to manhole.
using this covering, a sleeve or possibly several sleeves The measured resistance should not be less than the val-
are slipped onto the pipe just prior to the joining and ues shown in Table 13-2.
welding process. After the weld has been completed and
radiographically tested, the weld and adjacent area The values listed are for polyethylene and polypropylene
should be thoroughly cleaned with a wire brush to be pipe coatings and are based on a unit resistance of
sure no short radius protrusions (such as weld beads) 10 megohms per square foot. The required resistance (in
could penetrate the heat-shrink sleeve when installing. ohms) for any length of pipe installed is determined by
The sleeve is pulled into place and centered. Heat is then dividing the minimum coating resistance listed in ohms
applied to shrink the sleeve onto the weld area. Heat per 1000 ft of pipe length by the total length of pipe
should be applied to the center section first and slowly under test in thousands of feet.
worked outward toward both ends. Some users want
additional coverage and prefer that several sleeves be Table 13-2 Minimum Electrical Resistance of Pipe Coatings
installed, one overlapping the other. Another method of Minimum Coating
providing additional coverage is to apply a polymeric Surface Area Sq. Resistance in Ohms Per
tape over the weld area and overlap it onto the pipe Pipe O.D. (in.) Ft. Per 1000-Ft Pipe 1000-Ft Pipe Length
coating. This is the same tape used to cover the splice 2.625–4.5 - 8500
sleeves in manholes. 5.5625 1460 7000
6.625 1740 5750
Should the pipe coating be cut, sliced, abraded, or oth- 7.0 (tubing) 1840 5500
erwise damaged after being placed and installed in the 8.625 2260 4500
trench, heat-shrink sleeves and/or tape can be used to 10.75 2820 3600
repair the damage. The damaged area must be wiped 12.75 3000 3000
13-20
Should the resistance values of any section of pipe radii possible to permit this method of pipe installation.
tested be less than as specified above, an over-the- Using this method gives a continuous radius without
ground survey should be performed to locate any coat- local bending and a smooth surface for cable pulling.
ing defects. This survey is easier and the pipe coating This method of pipe bending requires bend radii of at
repairs are less costly if accomplished before the finished least 50 ft (15.2 m) for 8.625-in. (21.9 cm) pipe. The
paving is put down. However, most local governments larger the pipe diameter, the larger the radius required
require the finished paving be done as close to the trench to avoid kinking and ovality. For example, a 10.75-in.
backfilling operation as possible. Therefore, the pipe (27.3-cm) pipe diameter requires a bend radius of at
coating holiday detection testing should be quite thor- least 100 ft (30 m).
ough—quality control at its utmost—to prevent costly
re-excavations of the pipe line. If the bend radius is less than the acceptable radius for
pulling pipe around the bend, then suitable mechanical
13.6.4 Typical Work Area pipe-bending machines can be used to pre-bend the pipe
The best way to describe a typical work area is to refer to lengths in the field. The minimum recommended radius
Figure 13-11. This sketch shows the typical excavation of curvature for field bending by mechanical equipment
crew setup, the pipe setup, and the backfilling operation. is 40 times the outside diameter of the coated pipe.
However, even at this minimum radius, an experienced
13.6.5 Field Bends pipe bender is still required to produce a bend without
kinking, flattening, or excessive ovality. The pipe-bend-
Pipe can be field-bent by either using a portable bending
ing equipment was developed and used by the gas and
machine or by using air bags and trench jacks to con-
oil transmission companies and has proved quite suc-
form the pipe to the trench. In many situations, the lay-
cessful. If a utility does not own such equipment, it is
out of the trench and the number of utility crossings
often possible to rent it from pipeline contractors.
prevent the use of factory bent bends.
Figure 13-12 illustrates a field bend made with a bend-
One of the biggest problems with field bending a pipe is
ing machine at a job site. This particular bend was a
the “egg” shaping of the pipe. Anytime a field bend is
120°, 50-ft radius bend.
made, the contractor should verify the roundness of the
pipe by mandrelling the bend prior to tying the pipe in.
Regardless of the method used to install pipe around
bends, it is imperative that an appropriate mandrel be
The majority of horizontal pipe bends can be made by
pulled through the bent pipe to check for ovality prior
pulling the pipe around the bend, using the entire trench
to welding into the existing installed pipe system. If the
wall as the bending die. Most design engineers lay out
mandrel cannot pass freely, then the bent section should
the cable route bends with the largest allowable practical
be rejected and not welded into the existing system. The
Figure 13-11 Typical work area. Figure 13-12 Field bend on a 8- 5⁄8 O.D. steel pipe.
13-21
suggested mandrel to be used for various sized pipes is the factory flares have been damaged beyond repair.
given in Table 13-3. Remember, all welds must be radiographically tested
before acceptance.
For vertical grade changes, the pipe bend radii should
be limited to 500 ft (150 m) where possible. This is the 13.6.6 Quality Control
guide to be used during the excavation phase to avoid The purpose of this section is to list some of the critical
obstacles when installing the pipe. For convenience in quality control checkpoints for the installation of pipe.
making such bends, Table 13-4 is provided. However, the contractor and owner should always
ensure that the pipe is being installed in accordance with
After placing and mandrelling the pipe bends, a tie-in the owners/engineers and industry standards/specifica-
weld has to be made into the previously installed pipe. tion as a minimum.
Placing the pipe ends in the exact position to have the
proper spacing for chill-ring welding or any other type Some of the key quality control checkpoints that the
of flare welding is practically impossible. Normal prac- owner/contractor should ensure are being adhered to
tice has the two ends overlapping, so one pipe end are the following:
(flare) will be cut a distance of approximately 6 in.
(15 cm) from the end of the other pipe. The pipe cut • Keep moisture out of the pipe. Pipe plugs should
should be done with a mechanical pipe cutter if space is always be installed and at least 5 psi of dry nitrogen
available. Otherwise, the pipe can be cut with a burning kept on the pipe.
torch, but only after extensive precautions, such as • The bends are smooth and gradual. There should not
inserting a pipe plug beyond the cutting location to pre- be any kinks or ovaling of the pipe.
vent metal dust and slag from entering the pipe interior.
Once the cut has been made, the cut end is cleaned and • The pipe coating is not damaged during the installa-
beveled so a field flare or “pipe pup” can be butt-welded tion. A jeeping test should be performed routinely to
to it. Once welded, the interior of the pipe weld is ensure that the coating integrity will be maintained.
ground smooth to prevent damage to the cable during Many utilities and contractors also elect to perform a
pulling. After flares are installed on both pipe ends, the coating resistance test at the end of each day to verify
pipes can be joined with a chill-ring weld. Field flares the coating integrity.
are provided with the pipe shipment by the coating sup- • All welds are checked by an independent X-ray, prop-
plier upon request. Field flares also are installed when erly coated, and the coating properly tested.
Table 13-3 Mandrel Dimensions for Typical Pipe Sizes for After the pipe installation crew has placed and welded
0.250-in. (6.35-mm) Wall Pipe the pipe in the trench, protecting the pipe against dirt,
Mandrel debris, and water becomes even more important. No
Length
Mandrel Mandrel without
longer will the temporary plastic pipe caps (or plugs)
Pipe O.D. Pipe l.D. O.D. Flat Length Eyes suffice to protect the pipe interior. A commercially avail-
6.625 in. 6.125 in. 5.78 in. 8.5 in. 16.5 in. able sealing plug is used to assure that no moisture or
(16.8 cm) (15.6 cm) (14.7 cm) (21.6 cm) (41.9 cm) free water enters the pipe. Personnel should not walk on
8.625 in. 8.125 in. 7.60 in. 10.5 in. 20.5 in. the coated pipe at any time.
(21.9 cm) (20.6 cm) (19.3 cm) (26.7 cm) (52.1 cm)
10.750 in. 10.250 in. 9.80 in. 12.5 in. 25.0 in.
(27.3 cm) (26.0 cm) (24.9 cm) (31.8 cm) (63.5 cm)
At the end of the workday, nitrogen gas should be used
to purge the pipe slowly for 5 minutes. The pipe ends
should be sealed and the nitrogen pressure increased to a
Table 13-4 Length of Pipe Required to Achieve minimum of 5 psi (34 kPa) above atmosphere. In loca-
Required Offset tions where the trench is deeper than the normal 5–6 ft
Offset Length (1.5–2 m), the pressure should be increased to offset
ft m ft m potentially higher hydrostatic heads. The use of nitrogen
1 0.3 45 13.7 must include a safety sign denoting “Caution—Danger.”
2 0.6 63 19.2 If nitrogen is being vented in a confined area, a person
3 0.9 77 23.5 could be asphyxiated from lack of oxygen. If the nitro-
4 1.2 90 27.4 gen gas is being vented in an open trench, the workers
5 1.5 110 33.5 should still remain a safe distance from the pipe being
10 3.0 140 42.7 vented. Venting, in this case, is defined as discharging a
15 4.6 170 51.8 large volume of the gas from the pipe. If working in a
13-22
manhole, hoses should be used and trained to discharge so the work area within the tunnel becomes rather con-
the nitrogen above ground on the downwind side of the stricted, especially with smaller diameter tunnels.
manhole chimney. A number of utilities, because of this
possible danger, use dry air during the pipe installation Installing pipe or ducts on a bridge is somewhat similar
and then revert to nitrogen during the cabling phase of to the procedure for installing pipe in a tunnel. Support
the project. Still, caution should always prevail. arms with sliding clamps or hangers with rollers are
used to hold the pipe or ducts crossing the bridge.
Physical protection for the pipe not yet buried is rather Should the bridge be a short span, such as a span cross-
difficult in some city areas because of vandalism, so the ing a six-lane express highway having thick heavy abut-
installation crew might be required to repair and retest the ments on each side, the pipe could be supported and
pipe coating each morning before backfilling continues. anchored at the abutments after tensioning the pipe. In
some cases, the pipe is anchored at the center of the
13.6.7 Installing Pipe in Tunnels or Casings and bridge and large radius offsets are placed in the pipe or
on Bridges duct at either end to accommodate thermal expansion
The installation of pipe in tunnels or casings (bores) and and contraction. When installing a duct system across
on bridges is basically the same as given in the previous any length of bridge, expansion couplings must always
sections. The following paragraphs present a few high- be used. Regardless of whether a bridge crossing instal-
lights for these special types of installations. lation is short or long or done with pipe or ducts, the
installation must be in accordance with the design draw-
If a cable system is being installed in a tunnel—a con- ings and as approved by the government agency owning
crete-box passageway construction—the pipe installa- and maintaining the bridge.
tion should begin at all bends and progress outward into
the straight sections. Normally all the pipe bends will be 13.6.8 Installing Pipe in Manholes
provided pre-bent from the supplier, and sometimes the When there are two cable pipes in the trench, the trench
pipe will be without end flares since sleeves are used in should be widened starting about 50 ft (15 m) from the
some cases to connect each pipe length. Also, to con- manhole, and the pipe spacing should be increased to
tend with pipe expansion when the cable is fully loaded, allow working room in the manhole for splicing.
the pipe should be pre-stressed under tension, at a pre-
determined value based on ambient temperature, and This trenching and pipe spacing are to be handled the
welded to anchors embedded in the wall or floor. same way on the egress side of the manhole. Before the
pipe crew begins installing the pipe through the man-
When pipe is being installed in a casing, it is pulled into hole, the manhole should be tested for possible oxygen
the casing by the same method as pulling pipe in a deficiency and combustible gases. After the gas tests, an
trench (i.e., with a pulling head attached to the pipe and air blower should be used to ventilate the manhole
a winch handling the steel pulling rope). However, when whenever workers are inside the manhole.
puffing the pipe into the casing, pipe casing spacers are
used to protect the pipe coating. A casing spacer is a When installing the pipe through the manhole, no welds
device molded in halves that is strapped to the pipe to will be allowed within the manhole (other than the
prevent the pipe from touching the casing wall. Spacers splice joint reducers that are applied later) nor within
can be supplied for any diameter pipe being installed, 12 ft (3.7 m) of the manhole exterior wall. It is suggested
and, if multiple pipes are to be installed within the same that the longest single length of pipe (48–50 ft [about
casing, special spacers are available. The spacers are 15 m]) available be installed through the manhole and
normally installed at intervals of 6–8 ft (1.8–2.4 m). welded to the already installed pipe. Another method is
When the casing (bore) is installed, a standpipe should to join two pipe lengths and pull the pipe into the man-
be welded to the end (or just inside the end) of the cas- hole, centering the pipe weld in the area that will be
ing pipe for backfilling. removed later for splicing. After placing the pipe, center
the pipe in the sleeves in the manhole end walls and seal
If pipe is being installed in a liner-plate tunnel, which is the area between the pipe and the casing. Most utilities
a circular tunnel comprised of flanged steel plates are now using rubber sectionalized spacers designed to
bolted together with a minimum diameter of 4 ft support and seal the pipe. Also, in a very wet soil condi-
(1.2 m), the pipe should be pulled after a bedding of tion, utilities are augmenting the spacer seal with cylin-
thermal backfill (sand) has been placed to protect the drical neoprene boots that are strapped to the manhole
coating while pulling the pipe. Normally, the sand bed- sleeve and cable pipe. With the pipe centered and in as
ding will place the pipe at about the center of the tunnel, true a horizontal and vertical position as possible to
13-23
avoid bending and stressing it, the trench should be in place prior to the prefitting of the spreaderhead, pot-
backfilled on both sides of the manhole. When the pipe head base plates, and riser pipes. Anchoring can be
is in a firm position, the section of pipe where the splice accomplished by either enclosing the termination bend
is to be located can be cut using a mechanical pipe cut- in concrete or securing the bend to the pothead struc-
ter. The pipe end should be cleaned and beveled and the ture foundation with steel beams.
pipe coating should be removed for approximately 10 in.
(25 cm) to allow the installation and welding of the Quality control must be practiced during the installa-
splice joint reducers. After the reducers are in place, the tion of either type of termination assembly. All parts
bare metal should be primed to prevent rusting and test must be cleaned and kept free of debris and water. The
caps rated for high pressure (500 psi [3500 kPa]) should stainless-steel riser pipes, as well as the 90° carbon-steel
be installed. These test caps are designed to facilitate the termination pipe bend, must be mandrelled before final
vacuum and pressure testing of each pipe section assembly and welding.
between manholes.
13.6.10 Cleaning and Testing the Installed Pipe
13.6.9 Installing Termination Assemblies System
At each end of a pipe cable system, the three single-con- To prove the integrity of the installed pipe prior to cable
ductor cables within the pipe must be trifurcated into pulling, all sections (manhole to manhole or termina-
single-phase, nonmagnetic pipes, each to be terminated tion) should be swabbed, mandrelled, and pressure- and
with a termination mounted on a steel structure. This vacuum-tested. Swabbing cleans the pipe interior. Man-
separating of the three phases of cable is defined as a drelling checks for ovality and dented pipe to assure a
termination assembly, which consists of a spreaderhead proper passageway for cable pulling. Pressure testing
assembly, prefabricated stainless-steel riser pipes, and proves the structural integrity of the pipe and welds.
termination base and mounting plates. A termination Vacuum testing dries the pipe and provides a negative
assembly can be either an above- or below-grade design. pressure test to assure a hermetically sound pressure
Although both assemblies serve the same function (to system when the cable installation is complete and ready
separate the three phases), the cable industry uses the to be energized.
term “spreaderhead” for an aboveground design and
the term “trifurcator” for a belowground design. Normally, swabbing and mandrelling the pipe can be
done in the same operation. To start this operation, a
Trifurcating joints placed in manholes away from the mouse or pig (a series of tight-fitting urethane cups
terminations can eliminate the need for placing large resembling pistons) is blown by sufficient nitrogen pres-
cable reels in the substation. Regardless of the design, sure from one manhole to the next manhole or trifurca-
the termination assembly in accordance with the layout tor with a trailing 0.375-in. (0.95-cm) wire rope. At the
drawings should be prefitted, adjusted, and assembled exiting manhole, a swab and mandrel are attached in
in a final position before the final welding is done. series to the wire rope and pulled back through the pipe.
The swab is a cylindrical wrapping of lint-free white rags
The installation of the stainless-steel riser pipes requires placed around a short (1- to 3-ft [0.3- to 0.9-m]) wire
that after all welding has been completed and radio- rope sling and sized to fit tightly within the pipe to per-
graphically tested, a layer of corrosion-protective tape mit a wiping action. The mandrel is sized and shaped
be applied half-lapped over all below-grade sections. It per the recommended AEIC pipe installation specifica-
is suggested that a second layer be applied at all weld tion described previously.
joints. This also should be done over the trifurcator
assembly after the cable has been installed and all welds If the swab shows moisture after being pulled through
have been completed. the pipe, another swab or a series of swabs must be
pulled through the pipe until the swabs emerge dry.
The aboveground spreaderhead assembly is much sim- Also, if the swab shows a ripping action rather than a
pler than the trifurcator since no temporary or perma- wear action, then mandrelling is necessary again to
nent manhole is needed. Accurate placement, in recheck, through close inspection of the mandrel sur-
accordance with the design drawings of the pre-bent 90° face, for possible pipe or weld imperfections. Any imper-
termination pipe bend, is very important. This bend, fections must be located and repaired to avoid cable
having a minimum radius of 6 ft (1.8 m), a pipe flare on damage during installation.
the lower end welded onto the line pipe, and a straight
length of approximately 4 ft (1.2 m) on the upper end for Pressure testing the pipe is done on a section-by-section
final elevation adjustment, should be securely anchored basis after the splice reducers have been welded into
13-24
place and a temporary sealing pressure plate has been vacuum reading at the far end as a function of time.
bolted onto the reducer. Most utilities use nitrogen for This pressure rise should not exceed 100 μm in one
the pressure test, but some prefer air to eliminate the hour. Most utilities accept a one-hour test, but some
asphyxiation hazard of nitrogen. Today, most utilities prefer a longer test period of 4–24 hours where a pres-
have and require the use of independent life-support sure rise of 10 μm per hour is allowed for each hour
breathing apparatus for any personnel entering poten- beyond the first hour.
tially hazardous manholes. Because of the high quality
and low price of nitrogen and air supplied today, air McLeod vacuum gauges are still in use by the installers
compressors with drying apparatus attached are seldom that have and maintain them. However, this gauge is
used. The dew point of the gas provided by the manu- in disfavor because of the mercury it contains. The
facturers is -65°F (-54°C) or lower and is certified on the present substitute for the ranges of vacuum readings
required test reports. For circuits of 1 mile (1.6 km) or experienced in pipe-type cable systems is a thermocou-
more in length, the most economical delivery of nitro- ple vacuum gauge, which is readily available through
gen for pressure testing is by a liquid nitrogen tank several suppliers.
trailer with heater and compressor or tube trailers
(45,000 ft 3 @ 2200 psi [1275 m 3 @ 15,000 kPa]). All During the pumping operation, vacuum leaks may
hoses, valves, pipe fittings, pressure gauges, and other occur at any threaded or flanged and bolted pipe and
equipment used for the test program must be rated at valve connection. To assure a reasonable vacuum pull-
600-psi (4140-kPa) working pressure—the utility should down time, the work personnel should seal or cover
purchase and accept nothing else. The test pressure can these connection surfaces with clay, duxseal, petroleum
be from 300–500 psi (2070–3450 kPa) with a minimum jelly, petrolatum, or other appropriate material.
one-hour time period after pressurizing to observe any
loss of pressure indicating a weld or a pipe structural Once the vacuum test has been accepted by the respon-
leak. Most utilities prefer a longer test period of sible utility engineer, the pipe vacuum should be broken
72 hours, observing any pressure change every 12 hours with dry nitrogen and a positive pressure of 5–10 psi
while recording and correcting for temperature changes. (34.5–69 kPa) maintained until the cable-pulling opera-
The quality of pipe today and the radiographic inspection begins.
tion of welds have reduced the probability of a leak to
very low levels. 13.7 BACKFILLING
The selection and installation of backfill material are
The vacuum test essentially accomplishes the drying of
critical to achieving the required, or designed ampacity
the pipe system. The procedure is to work section by
of the cable, and the expected life of the cable system.
section (manhole to manhole). The equipment used is
Backfill material will be a part of any cable system that
normally a 150 ft3/min (4.3 m3/min) or larger vacuum
is installed in the ground, regardless of whether it is a
pump that is connected to the pipe end pressure plate
pipe cable system, a duct bank cable system, or a direct
with a 2- or 3-in. (5.1- or 7.6-cm) vacuum hose. The
buried cable system. The effect of different backfill
larger the hose, the faster the vacuum draw-down time.
materials on cable ampacity has previously been dis-
However, when starting the vacuum procedure, the flow
cussed in Chapter 12.
of nitrogen from the pipe into the vacuum pump should
be controlled and throttled with valving. Although the
13.7.1 Placing of Backfill Material
nitrogen has a low dew point, the pipe could contain a
slight amount of moisture. If the vacuum drops too rap- Backfill material that is specially selected or designed
idly, the moisture could freeze as hoarfrost, and test for its thermal characteristics is commonly referred to as
equipment would not indicate its presence in the pipe. “thermal backfill.” Thermal backfill can be divided into
two main types: granular and flowable. Granular ther-
When evacuating the pipe section, the vacuum pump mal backfill generally consists of a specifically graded
should be placed at one manhole, and the vacuum gauge material such as limestone screenings, bank-run sand,
should be placed at the pipe end in the adjacent man- or other similar material that has been tested to assure
hole. The suggested procedure for vacuum drying and that it contains the appropriate amount of moisture and
testing is to evacuate the pipe, and when a vacuum read- has the desired compaction properties. Providing that
ing of 150 μm is attained, to continue for an additional larger stones, rock, and foreign debris have been
4 hours of pumping. At this time, the pipe should be removed, and that it has the required thermal proper-
valved off and the pump operation stopped. Upon shut- ties, native spoils from the trench excavation will some-
down, a vacuum-rise test is performed by recording the time be used as thermal backfill. Fluidized thermal
backfill is usually either a specially designed mixture of
13-25
locally available ingredients, including sand, cement, ferred to allow easier recovery of the pipe if needed
aggregate, and fly ash, or a low-strength (weak mix) for repairs.
concrete having the necessary thermal properties. What-
ever type of backfill material is selected, the material 13.7.2 Surface Restoration
must be installed properly for it to perform as expected. Surface restoration or paving is done in accordance with
This means it must have the proper compaction, mois- local, state, or federal specifications, or if on private
ture content, and density. Figure 13-13 shows fluidized easements, by mutual agreement with the property
thermal backfill being installed over the structural duct owner. Usually, paving of urban and suburban streets
bank concrete. requires a layer of steel-reinforced concrete of between
6 and 9 in. (15 to 23 cm) in thickness, serving as either
Should the backfill material be a type of sand or soil, the final surface or as a sub base to an asphalt overlay of
the backfill is placed in layers and compacted thor- generally 2 to 4 in. (5 to 10 cm) in thickness.
oughly using a vibratory compactor. Particular care
should be applied in compacting the sand or special soil Temporary surface restoration or temporary patching is
in and around the cables, pipes, or ducts to eliminate usually required by the municipalities as a means to
voids and attain maximum density while not exposing allow traffic along the affected streets as quickly as pos-
the cable to damage. For maximum compaction and sible. The temporary patch can be either a thin quick-
density, the best backfilling procedure is to place the setting concrete, (such as a 9 bag mix) or an asphalt
controlled backfill material with the balance of the patch, depending on the local requirements or on the
trench backfill material up to subsurface grade in 6-in. contractor’s choice. Depending on the amount of per-
(15-cm) layers, compacting each layer as added. The manent resurfacing required, the concrete patch may be
earth backfill to subgrade level can be compacted with quicker and less costly, including the removal costs, than
large pneumatic or mechanical tampers. Compaction the asphalt patch.
density should be 95% as measured by the Proctor den-
sity test. Generally, municipalities will require seam-to-seam
pavement replacement as a minimum restoration
If weak-mix concrete is used around the pipe for the requirement. The designer and the construction con-
controlled backfill material (a duct bank system would tractor should take into consideration the location of
normally be encased in concrete), an electrically driven the trench within the pavement seams when laying out
cylindrical vibrator (which is much like a pipe attached the cable route and minimize the number of pavement
to a hose) may be used to remove any air entrainment sections that are affected by the cable route. Pavement
and assure the density of the concrete. The strength of restoration is a significant cost element for projects of
the weak-mix concrete usually varies from 100 to 300 psi several miles or more in length that are installed in
(0.7 to 2.1 MPa). When encasing a duct bank, 3000 psi urban and suburban streets. Additionally, since this is
(21 MPa) high-strength concrete should be used, but usually the last activity performed by the installation
when covering the pipes for HPFF cable systems, contractor, it can affect the project schedule since it is
a weaker strength of 1000 psi (6.9 MPa) or lower is pre- somewhat weather dependant.
13.8 CABLE INSTALLATION

Once the pipes/conduits have been certified, the cable
installation phase can commence. Prior to starting the
cable installation phase of the project, a delivery sched-
ule of the cable lengths should be agreed upon with the
manufacturer. This enables the construction manager to
lay out the most advantageous and least costly cable
installation schedule.
If a direct buried cable is the type being installed, then

the optimum schedule would be for the cable to be deliv-
ered just when the trench preparation is completed. In a
pipe-type cable or a duct bank cable installation, select-
ing the sections to be pulled from manhole to manhole
Figure 13-13 Trench backfilling with flowable fill (with
will depend on the cable lengths that have been delivered
permission of ITCTransmission). or are in storage, or possibly on other priorities such as
13-26
municipal restrictions or community schedules. Also, All open manholes should have a manhole guard, a
the pulling schedule might consider starting with the highly visible folding steel fence, placed around the
easiest setup area for the entire project to provide a opening to indicate a dangerous area that is to be
learning curve for the installation crew. A manhole-to- approached only with caution. Some manholes have
manhole pull should be accomplished for training pur- permanent steel ladders attached to the chimney and
poses before a termination cable pull (which is consid- wall, but most utilities prefer to use a portable wooden
ered the most difficult) is attempted. or fiberglass extension ladder because the steel ladder
could be corroded and, therefore, dangerous to person-
Prior to the cable-pulling phase of the construction nel. Also, extension ladders can be set to extend above
project, all traffic conditions that influence cable trans- the street level 3 to 4 ft (0.9–1.2 m) to provide a safer
portation from the warehouse or storage area to the entrance and exit.
pulling location should be thoroughly investigated to
eliminate any delays and cost overruns. This investiga- While personnel are in the manhole, there should always
tion must include time restrictions on moving the be a person on the surface at the manhole opening to
cable reels through city streets and possible weight and maintain contact with personnel in the manhole. Also,
height limitations. all personnel working in the manhole should be
equipped with a constantly operating gas detector unit
In the cable equipment setup process at both manholes, and wear a harness with a lifting tripod available at the
an area should be cordoned off by police or flagmen to top of the manhole.
facilitate placing the equipment on arrival at the job site
and to warn pedestrian and vehicular traffic of an 13.8.1 Storage of Cable
impending obstruction or detour. On the arrival of the A portion of the cable is normally required to be stored
rolling stock equipment (e.g., cable-pulling winch, onsite to ensure that adequate cable is on hand during
mousing winch, and cable trailer), the work area should the cable installation. On rare occasions, the cable can be
be defined and protected by the placement of this equip- shipped from the factory to the site the day of the pull.
ment. In addition, large flashing signs indicating proper
vehicle traffic flow must be strategically placed at each When cable is required to be stored onsite, a secure yard
end of the work island, along with sufficient luminous is required. Based on the weather conditions existing
traffic cones defining the work area. As noted earlier in during the scheduled pull, a heated indoor storage facil-
this chapter, many municipalities require formal traffic ity may be required. The owner or contractor should
control plans as part of the permitting process. verify with the cable manufacturer the minimum tem-
perature the cable can be installed.
Safety should be the foremost concern of the utilities.
When working on an underground cable system, 13.8.2 Extruded-Dielectric Cable Installation
whether existing or new, all personnel must be taught
Installing extruded-dielectric and self-contained fluid-
about the hazards of gases and liquids seeping into a
filled (SCFF) transmission cables in a duct system pri-
manhole. When preparing to remove a manhole cover,
marily follows the same procedure as used when install-
the cable crew should first check for explosive and toxic
ing a distribution cable system except that the
gases if the cover has a small opening hole. Gas detec-
equipment is larger and quality control procedures are
tion equipment is quite portable and highly reliable and
much more stringent. All safety procedures practiced
is a standard tool without which a manhole cover should
for other types of cable systems are applicable in this
not be lifted for further work. On removing the manhole
case and must also be practiced. This includes using air
cover and storing it in a safe location, the manhole must
blowers to ventilate the manholes being worked in, even
be ventilated with a blower that is placed on the down-
though nitrogen gas is not used in installing these cables.
wind side of the open manhole chimney. Once the man-
Like pipe-type cable installations, the duct requires the
hole has been opened and ventilated, it should be
insertion of a fish line from manhole to manhole by
checked for flooding. In most cases, a flooded manhole
blowing a mouse through the duct using compressed air
is filled with water; however, if in doubt, the fluid should
for propulsion. Once the fish line has been inserted, a
be analyzed, and, if it is other than water, the cognizant
circular wire brush, mandrel, and cotton cloth swab
government agency should be informed and guidance
should be attached along with the 0.375-in. (0.95-cm) or
requested. If the fluid is water, then the crew can usually
larger wire rope and pulled back into the cable feed-in
pump it into the nearest storm sewer. Acid, oil, gasoline,
manhole for attachment to the power cable for installa-
or other fluids must be discharged into a refuse tanker
tion. The circular wire brush will scrape and remove any
for removal by a licensed carrier for proper disposal.
duct joint material to permit a smoother cable pull.
13-27
In the interest of installation economics, the cleaning, ney edge permitting the cable to leave the reel and enter
mandrelling, and swabbing of the ducts and the placing directly into a flexible steel cable-feed tube. Since these
of a fish line in each duct by a crew working several are single-conductor cables, only one cable will be posi-
weeks in advance of the cable-pulling operation could tioned and pulled at a time. It is always best to feed the
prove beneficial. This crew would, after checking the cable from the manhole chimney furthest from the
integrity of the duct, leave in each duct a synthetic rope mouth of the duct being filled. All rigging for the cable-
having sufficient strength to pull the steel pulling rope feed guidance system should be well anchored and
into the duct. Steel rope is generally the preferred pull- devoid of any sharp protrusions that could damage the
ing rope since the metal dissipates heat generated by the cable during or after installation. The cable reel will be
frictional forces during pulling. Synthetic rope, which supported by either jacks or a cable reel trailer, which
does not dissipate the heat built up during pulling, will serves as a transporter and reel support platform for
either melt the duct, the pulling rope fibers, or both, and cable pulling. A jacket integrity test should be per-
damage the duct system. An inventory of long lengths formed before the cable is pulled.
of fish line (pilot line) rope is required for insertion into
the ducts before the cable-pulling operation begins. Since these cables are sheathed and jacketed, inclement
Some utilities require that the conduit be pre-lubricated weather conditions, other than extremely cold tempera-
by placing a quantity of lubricant in front of the final tures of less than 20°F (-7°C), do not impede the plan-
swab before it is pulled through the completed conduit. ning and pulling schedule. However, like pipe-type
cable, these cables should be protected and kept in warm
The positioning of the cable-pulling machine and the storage until the pulling date. To ease cable-pulling ten-
rigging of the manhole is done in the same fashion as sions, which are much less than for pipe-type cable pulls,
for pipe-type cable installations. The pulling line should a suitable pulling lubricant should be applied to the
exit the duct, traveling as straight as possible to the cable jacket when entering the duct. The lubricant
far wall to allow as much slack in the manhole for splic- should be carefully evaluated and selected to prevent
ing as possible, then it is routed up and out of the man- any sheath or jacket deterioration during aging in the
hole chimney guided by a series of sheaves onto the duct. The cable manufacturer should be consulted on
pulling machine. the proper lubricant to use, and the conduit may be pre-
lubricated as described above.
Figure 13-14 shows a typical extruded-dielectric cable
installation set up at the feed-in manhole. At the feed-in After the cable pull has been completed, another jacket
manhole, the cable reel is placed at the manhole chim- integrity test should be performed, the cable should be
cut to the necessary length to facilitate splicing, hermet-
ically capped to preclude the entrance of water, and
trained or placed in a position to protect the cable from
any additional work within the manhole. If the cable has
a lead sheath, a splicer or lead wiper should form and
wipe a lead cap over and onto the sheath for the greatest
protection. If no lead sheath is used, as on some
extruded-dielectric cables, then commercially available
heavy-wall heat-shrink caps should be installed.
13.8.3 Pipe-type Cable Installation

Since pipe-type cable has no cable sheath other than the
pipe into which the cables are pulled, weather condi-
tions are a major factor in the day-to-day pulling sched-
ule. Cable pulls must not be performed in inclement
weather. If the weather changes and becomes inclement
during the cable pull, both the pulling and cable feed
equipment must be covered with a large tent or canopy
to prevent any free water from falling onto the cable
and/or cable-pulling rope. Any cable pull already started
should continue nonstop until completion, regardless of
Figure 13-14 Installation of extruded-dielectric the hour of the day. Covered reel carriers are an option
cable (with permission of PG&E). that the contractor or owner may consider.
13-28
The placement of the feed and pulling equipment is for possible moisture and excessive dirt, which would
determined by the direction of the pull. The direction of indicate the need for additional swabbing. At this point
pull must be followed unless specifically approved by the in the construction schedule, dirt and moisture should
design engineers. not be present if quality control has been maintained
during pipe installation.
To start a pipe-type cable pull, the pipe must be
“moused” (i.e., a light steel wire rope is installed into the The next step is to rig both the feed-in and pull-out
pipe, which in turn is attached to a larger wire rope for manholes. The pull-out manhole requires the anchoring
attachment to the power cables). Observing all safety and bracing of a sheave approximately 24 in. (0.6 m) in
precautions, this crew enters the manhole, attaches a diameter in the manhole to guide the steel pulling rope
high-pressure hose to the valving on the pressure plate exiting the pipe upward to the pulling machine. Also, to
bolted to the pipe reducer, and discharges the nitrogen protect both the wire rope and the power cables exiting
gas contained in the pipe to the atmosphere. During the the pipe, a curved pipe shoe, similar in shape to a shoe-
dumping of the nitrogen gas, all personnel must be out horn, is installed on the pipe reducer to guide and sup-
and away from the manhole. Remember, nitrogen alone port the cables.
can kill.
The pulling winch (machine) used for pipe cable is sized
Once the gas is removed, another specially designed and rated for about 100,000 pounds (45,300 kg),
pressure plate is placed onto the pipe to blow a mouse although the majority of cable pulls rarely exceed a
(piston) from the feed-in manhole to the pull-out man- designed maximum pulling tension of 60,000 pounds
hole. This plate permits the attachment of a nitrogen gas (27,200 kg) depending on conductor material and size.
hose to blow the mouse while trailing the 0.375-in. The steel wire rope used for pulling is a compact-
(0.95-cm) wire rope. The pressure plate on the pipe in stranded wire rope, which minimizes any cutting or saw-
the pull-out manhole must be removed before blowing ing action on the pipe wall while pulling the power
the mouse through the pipe. While blowing the mouse, cables. To assure a proper cable installation, the pulling
the internal pipe surface is lubricated with cable dielec- winch (machine) is equipped with instrumentation to
tric fluid. This is accomplished by pouring the dielectric provide a recorded log of length pulled, pulling speed,
fluid into the pipe in front of the piston (or in the case of and pulling tension at any given footage count.
a series of pistons, between each) during insertion. The
basic purpose of lubricating the pipe wall is to reduce At the feed-in manhole, the cable reels should be
the friction of the mousing wire rope while pulling the arranged in tandem fashion (in line, one after the other)
large steel wire rope back to the feed-in manhole. The and the three single-conductor cables pulled into the
mousing wire rope can do more damage to the pipe due pipe simultaneously. Figure 13-15 shows the arrange-
to the sawing action than the cable pulling rope would ment of cable reels and a guide wheel at the manhole
ever do. entrance. The cable reels are supported on cable jacks
(cradles) or, as in many cases today, either on a reel
All personnel must exit both manholes before and dur-
ing the mousing operation. Both manholes must be ven-
tilated thoroughly and checked for oxygen content prior
to reentering after the mousing operation.
At the cable pull-out manhole, the pulling winch (or

machine) is placed in a position to allow personnel
access into the manhole for rigging and pulling control
and to permit sufficient cable to be pulled into the man-
hole for splicing. The pulling machine is normally
placed at the end of the manhole away from the pipe
being cabled. Once the pulling machine is in position
and safety barriers, signs, and cones have been placed
strategically, the pulling rope can be pulled into the pipe
by the mousing rope. At this time, one last quality-con-
trol check of the pipe is made by attaching a pipe man-
drel and clean, lint-free cloth swabs between the
mousing and pulling ropes. The swabs must be checked Figure 13-15 Typical manhole-to-manhole HPFF cable pull
(with permission of Northeast Utilities).
13-29
trailer handling all three reels or on individual reel trail- 40.6-cm) diameter carbon-steel pipe approximately 3 ft
ers for each cable reel. The latter design is more eco- (0.9 m) long welded to the cable pipe. Three smaller
n o m i c a l a n d ap p l i c abl e fo r b o t h m a n h o l e a n d stainless-steel pipes are welded to the other end of the
termination pulls. Once the cable reels are in position at termination assembly. They progress upwards above
a feed-in manhole and the wood lagging has been ground into a steel support structure designed to hold
removed, each single-conductor cable is pulled from the the cable terminators (potheads). A typical termination
top of each reel to the front of the cable reel lineup so cable pull set (for upward movement) places the reels
the cable pulling bolts can be attached to a three-phase side-by-side centered on each pothead position. Each
cable pulling yoke. The three cables in triangular config- cable is then pulled upward onto and over a large sheave
uration will then be pulled over a minimum 6-ft (1.8-m) (also called a bull-wheel) or, as an alternate, onto a
diameter guide wheel (or feed chute) into a rigid pre- semicircular-shaped device with rollers and conveyor
bent or flexible steel tube having the same diameter as belt, permitting a smooth directional guide for the cable
the pipe in which the cable is being installed. The steel entering each stainless-steel riser pipe. The cables are
tube is placed and rigged (supported and bolted to the gathered at the trifurcator, attached to the cable-pulling
splice reducer) to provide a large-radius bend to feed yoke, and pulled as previously described. Some utilities
and install the cables into the pipe without damage. use a belowground trifurcating joint to transition to the
Some manhole designs provide a sloping buried casing three stainless steel pipes. The cables are pulled individ-
extending out from the manhole end wall upward at ually down the pothead stub into the trifurcating joint
approximately a 15° angle terminating beneath the or they may be pulled individually from the trifurcating
street subsurface paving. This casing is an access hole joint to the top of pothead elevation if there is not
that allows the flexible steel feed-in tube to be inserted enough room in the substation to set up the cable reels.
and attached to the splice reducer, thus permitting a Figure 13-16 illustrates how the cable reels are stationed
shallow feed-angle cable pull into the pipe rather than at ground level, and the bull wheels direct the cables into
the normally designed and practiced 90° angle profile the risers.
cable feed. During the cable pull, cable pipe-filling
dielectric liquid is slowly poured onto the cable to lower When the cable pull is completed, the cable must be
the pulling tension. inspected for possible damage, cut to a designated
length, identified and marked as to phasing, and, lastly,
During the cable pull, there will be workers in the feed- night-capped. A night-cap is a slightly curved pipe with
in and pull-out manholes performing inspection and the same diameter as the cable pipe. It is curved just
control functions. All personnel in manholes require enough to avoid the next section’s night-cap, and it has a
forced air ventilation, and if the blower equipment flange on one end to attach to the cable pipe reducer
ceases to function, the cable pull must be stopped, and and a pipe fitting and valve on the other end. Its length
the workers must exit the manholes until the equipment is approximately equal to or longer than that of the
is functioning again. designed splice length. Night-caps for the termination
ends are basically straight pipe lengths with flanging to
When the pulling equipment is in position and the
cables are attached, the cable pull should proceed at a
pulling speed of approximately 30 ft/min (9 m/min).
This speed permits visual inspection of each single con-
ductor and permits a safe reel-braking operation if the
cable pull must be stopped. However, pulling speed may
be varied in an attempt to control the amount of surging
(start-stop movement of the cable). No cable should
touch the ground or be allowed to form short loops of
less than 3 ft. (0.9 m) in radius during the pull or while
braking or stopping.
The termination cable pull requires more setup time and

equipment than a manhole-to-manhole pull. Therefore,
these pulls are basically scheduled based on a weather
forecast of several dry days in a row. The trifurcator or
spreaderhead termination assembly for a pipe-type
cable system consists of a large 10- to 16-in. (25.4- to Figure 13-16 HPFF aboveground termination pull.
13-30
fit the pothead base plates and valving for evacuating 13.9 SPLICING
and pressurizing the system after the cable pull. After Highly trained and skilled personnel are required to
the cable pull and night-capping, the section should be properly splice and terminate transmission cables. All
evacuated to a pressure of 250 μm or less of mercury as splicers should be familiar with the design, technique,
indicated by the vacuum gauge attached to the night- required material, tools, and equipment to accomplish a
cap at the other end of the pipe section in the adjacent quality-controlled splice (or termination). If there is any
manhole. On reaching the designated pressure, the vac- question as to the workers’ abilities, the utility should
uum pump should continue pumping for at least a one- require schooling, which can be provided by the cable or
hour period, then stopped and isolated from the pipe accessory manufacturers or experienced consulting
section by closing all valves. A pressure-drop test should companies. Whether trained splicers are on hand, for
then be performed for one-half hour to assure the dry- the best quality work many utilities require splicer
ness of the section. The allowable pressure rise should schooling either every year or for each cable project
not be more than 250 μm. If the pressure-drop test fails, being installed. Many utilities require that factory splic-
the vacuum should be reapplied and the testing repeated ers or factory representatives be present to inspect the
until the requirements are met. When the pipe section cable splicing and terminating for 138-kV through
has successfully passed the pressure-drop test, the vac- 345-kV installations.
uum should be broken with dry nitrogen gas and main-
tained under a minimum positive pressure of 5 psi The splicing of any cable type should be performed in
(34 kPa) until splicing operations are started. accordance with the cable and/or accessory manufac-
turer’s drawings and procedural specifications. This
When excellent weather conditions prevail and the includes the use of all specified materials, tooling, and
cable pull has taken more time than usual, the vacuum equipment required for each type of splice.
test may be postponed until the following day. In this
case, the pipe section is purged with dry nitrogen gas To ensure a quality splice, the following setup and work
displacing about 1.5 to 2 times the pipe volume, and procedures should be practiced regardless of the type of
then pressurized. On the next day, the gas should be cable system:
exhausted, the pipe evacuated and drop-tested, and the
section repressurized. • All splices in permanent or temporary manholes
should be protected from dirt, debris, inclement
The most efficient and portable vacuum pump sug- weather, and humidity (e.g., all splicing work should
gested for use in the field is a 130–150 ft 3 /min (3.68– be done in a clean and dry controlled environment).
4.25 m3/min) rotary piston vacuum pump with a 1600 • The splicing area should be thoroughly cleaned and
ft3/min (45.3 m3/min) booster (blower) driven with an all dirt and debris should be removed to prevent any
air-cooled engine. All hoses should be rated for vacuum possible contamination of the splice regardless of
(50 μm) and pressure (100 psi [680 kPa]) and feature whether it is a taped or molded splice.
snap-lock connectors for quick and easy assembly • If a temporary housing or manhole is used, the hous-
and disassembly. ing should be designed to prevent any free water from
entering the splice area and a water pump should be
13.8.4 Self-Contained Fluid-Filled Cable on hand at all times.
Installation
• If permanent manholes are used, the manhole
The installation of self-contained fluid-filled cables is
entrances at street level should be provided with a
very similar to extruded-dielectric cable, with the excep-
water barrier to prevent free water from flowing into
tion that the cable has dielectric fluid installed under
the manhole during the splicing operation.
pressure. After the cable is installed, the fluid reservoir is
removed from the tail end of the cable and installed on • Although pipe-type cable specifications require that
the front end of the cable. This allows the cable to have splicing be done under a temperature-and humidity-
constant pressure/fluid flow when the tail end of the controlled atmosphere with proper dehumidifiers
cable is cut. After the cable is cut, a lead sleeve is wiped and air-conditioning equipment, all other types of
onto the end of the cable to protect it until the splice cables would benefit from being spliced under well-
is started. controlled conditions. The basic atmospheric man-
hole conditions for pipe-type cable splicing range
from 68–72°F (20–22°C) temperature and 10–35%
relative humidity, depending on the voltage rating
13-31
and splice design. A common practice to help main- and spares should be purchased for future cable installa-
tain manhole atmospheric control is to paint the tions and possible repairs.
manhole to retard moisture penetration through the
concrete and also to line the manhole with plastic Refer to Chapter 8 for a discussion of splice connectors.
sheeting whether the manhole is permanent or tem-
porary so that, in essence, the splice is made in The following paragraphs describe the principal differ-
a bubble. ences in splicing among pipe-type, self-contained fluid-
• All standard tools and equipment should be on filled, and extruded-dielectric cables. Refer to Chapter 8
hand before the start of the splice. In addition, any for details of splice design for each cable type.
special equipment required for a particular splice
design should be at the splice location along with 13.9.1 Extruded-Dielectric Cables
spare parts in case of wear or damage during the This type of cable must be trained and racked in the
splicing operation. manhole before starting the splice operation. Since this
• For quality work on HPFF and HPGF cables, once cable is rather stiff, the bending and training can be
facilitated by warming the cable for an extended time.
the splice construction has started, work should con-
One method is to wrap heat tapes around the cable,
tinue until the splice has been completed. However, if
perhaps after strapping the cable to a length of angle
the splicing work must be interrupted, a planned
iron. Once the cable is fully warmed and trained, the
emergency procedure should be instituted to protect
heat is removed, and the cable is allowed to stiffen in its
all exposed cables against possible damage.
final position. Be careful not to let the maximum tem-
• Since most splice locations are on public rights-of- perature exceed 212°F (100°C) or the manufacturers’
way (e.g., streets and sidewalks), all precautions recommendation.
should be practiced to protect the workers, pedestri-
ans, and vehicular traffic. Barriers must be placed Since most extruded-dielectric cables have a metallic
around the work area to create a safety island. These sheath, a splice sleeve is applied over the splice as
barriers should be illuminated at night. In addition to with self-contained fluid-filled cables. With this type of
the created safety island, police protection may be cable system, grounding and cross-bonding is important
required because of the location of the cable route and must be done in accordance with the cable system
and manhole. If the cable route is through an unsafe layout design.
neighborhood, then permanent, rather than tempo-
rary, manholes should be used. The extruded cable typically has substantial movement
with load changes. This movement requires that the
Regardless of the type of cable being installed, splicing cable be adequately racked/supported in the manholes
should be done in a clean and dry area. The easiest and and on the termination structures. Figure 13-18 shows
most economical way to provide these conditions is to two phases of a 69-kV cable system. The center phase
place a portable work housing over the top of the man- splice has been completed, and the bottom phase splice
hole entrances. Most contractors have designed a trailer has just begun.
or truck just for this function. This vehicle may have its
own power supply and atmospheric control equipment. 13.9.2 High-Pressure Pipe-Type Cable Splice
In addition, the vehicle should have a complete inventory
Prior to starting a HPFF or HPGF pipe-type cable
of standard small tools, equipment, and material. It is
splice, the night caps (protective steel pipe sealing end
still necessary to provide water barriers to keep street sur-
caps bolted to the pipe reducers) should be removed
face water from entering the manhole from the chimneys.
after the nitrogen atmosphere has been vented and the
manhole atmosphere tested so it is safe for the workers
The cable and splice manufacturers have different splice
to enter and begin work. Before placing the cables into a
kits available for both paper and extruded-dielectric
splicing position, the steel pipe sleeves are slid back over
cables. These kits include insulating tapes, factory- or
the reducer and away from the splice area. Upon com-
field-molded dielectric materials, or a combination of
pletion, the steel sleeves are pulled back over the splice
molded material and tape. The splice designed for each
area and welded. Welding should be done intermittently
type of cable system should be thoroughly evaluated
to minimize heat buildup in any one area so the cables
and discussed with each cable or splice manufacturer
are not damaged. If the welder moves back and forth,
before a selection is made. Also, if special equipment is
welding in different locations, the splice sleeves should
required to construct the splices, then this equipment
remain at a safe temperature to prevent cable damage.
13-32
Lastly, once the casing has been completed and sealed, sured and marked. Once both cable ends are cut and the
the welds are soap tested, and, if there are no leaks, all conductors prepared for joining, the cables are brought
bare metal is taped with an approved anticorrosion tape into a horizontal position and a connector is pressed
and tested with a holiday detector to ensure the integrity onto the butted conductors. During this operation,
of the coating. Figure 13-17 shows pipe-type cables cable-filling liquid is made to flow slowly from remote
being spliced. reservoirs to preclude any air from entering the hollow
core so that the core is completely filled with liquid at all
13.9.3 Self-Contained Fluid-Filled Cables times. As the splice insulation is applied, this liquid flow
This type of cable requires careful training and racking from the copper conductor strands will slow and eventu-
in the manhole before splicing begins. Figure 13-18 ally stop. On completing the splice insulation, the splice
shows cable splicing/racking in a manhole. The splice sleeves are pulled into position over the splice and
sleeve is slid back over the cable before the actual sealed. This can be done by lead wiping, soldering,
cable cutting and splicing is started. Since the single- or welding.
conductor cable used at transmission voltages has a hol-
low core in the conductor center to maintain a flow of Once the splice sleeve is racked in place, a vacuum is
cable-filling fluid, the cable ends must be raised into a applied, and the sleeve is filled with insulating liquid. If
vertical position for cutting after they are properly mea- piping connections from fluid reservoirs are required,
they should be attached and checked for leaks, the
grounding and the cross-bonding system should be con-
nected, and the entire assembly should be wrapped with
corrosion tape. Some manufacturers use a premolded
corrosion cover over the splice, which is then filled with
a bitumastic compound.
13.9.4 Link/Grounding Box

The link/grounding box is typically provided by the
cable manufacturer. These boxes are normally water-
proof enclosures that are mounted on the manhole wall
or termination structures. When mounting the boxes in
manholes, the boxes should typically be installed as high
on the manhole wall as practical to keep the boxes
above the ground water level.
13.10 TERMINATIONS
Figure 13-17 HPFF semi-stop splice (with permission Terminating practices for any cable system are basically
of ComEd).
an extension of the splicing procedure. The same care
that is taken during the splicing activity should be taken
during termination installations. Terminations should
be constructed in a clean, dry, and dust-free atmosphere.
Many utilities will erect a temporary housing over the
terminations to prevent dust from flying into the work
area and for emergencies such as a sudden rain or snow
shower. Normally, utilities schedule the construction of
the cable terminations around a favorable weather fore-
cast so that the possibility of inclement weather is less
probable. In some areas of the country, where the
humidity is always rather high (above 50%), it is advis-
able to install air-conditioning units with heat-strip
resistors to lower the relative humidity to approximately
35–40%. A climate-controlled enclosure is normally
used for voltages of 345 kV and higher, regardless of
Figure 13-18 Cable splicing/racking in manhole (with outside humidity
permission of SMECO).
13-33
As with the splicing procedure, all tools, equipment, and time of year, terminations are often performed outdoors
material should be at the job site in clean, operating without erecting an enclosure around the work area.
condition before any work proceeds. An inventory of all This is especially true for the lower transmission cable
parts, including even the smallest of items (nuts and systems that are installed on tall riser structures. The
bolts), should be done before the scheduled workday. actual installation of the terminations, especially at
Each splicer should have at hand drawings of the termi- 115 kV and above, should be performed by the cable
nation to be constructed; the unit must be constructed manufacturer, however at lower voltages some utilities
in accordance with the drawings and specifications. If choose to perform the installation of the terminators
the splicers are not familiar with the termination design, and splices themselves, under the supervision of the
training should be scheduled to review any new meth- cable/terminator/splice manufacturer’s representative. It
ods, tools, or materials to be installed. Most termination is common practice for the cable manufacturer to pro-
suppliers provide a complete and detailed installation vide an inspector/representative during the construction
specification for the construction of their terminations. of the terminations. Although utility personnel are
It is very important to ensure coordination between sometimes trained by the cable manufacturer to install
cable and termination suppliers, beginning in the initial or assist with the splicing and terminating activities,
thawing-approval stages. they normally do not have the years of experience that
the manufacturer’s technicians have. This becomes
13.10.1 Extruded-Dielectric Cables extremely important at the higher voltages, as tolerances
The installation setup for extruded cables is very similar and dimensions become more and more critical. Most
to the HPFF cables in that a clean dry work area must cable manufacturers will not warrant a cable system
be maintained. However, the need for humidity control unless they have their technicians performing the splic-
is not as critical for extruded dielectric cables as it is for ing and terminating.
paper insulated cables. Depending on the weather and
The cable manufacturer should provide with their termi-
nations any specialized tools that are specified in the
installation instructions as being required to make the
terminations. Each manufacturer normally requires that
only their tools be used to make the terminations. These
tools are typically not readily available in the United
States unless they are ordered with the terminations.
Figure 13-19 shows a typical self-supporting steel riser
structure with two partially completed terminations
(without the porcelain installed yet). Notice how the
cable is supported below the terminator arms.
Extruded-dielectric-insulated cable terminations are

sometimes purged with dry nitrogen and filled with a
di electri c f lu id imm edi at ely on co mp leting the
installation.
13.10.2 Pipe-Type Termination

On completion and mounting of a high-pressure pipe-
type termination, a gentle purge of nitrogen gas should
be performed to maintain a dry atmosphere within the
termination until a vacuum can be applied and the ter-
mination filled with the insulating liquid. Some termina-
tion designs are pre-filled with dielectric liquid, some of
which is displaced as the termination is lowered over the
Figure 13-19 Installing terminators on a steel riser structure
(with permission of Platte River Power Authority).
13-34
cable. Figure 13-20 illustrates a typical work enclosure specify the fluids that are required for the different
around a high-pressure pipe-type termination. HPFF and SCFF cable systems.
Figure 13-21 shows a high-pressure pipe-type cable
being terminated within an enclosure. 13.11.1 HPFF Cables
After all of the cable has been installed and the splices
13.10.3 Self-Contained Fluid-Filled Termination and terminations have been completed, the system is
The construction of a self-contained fluid-filled termi- ready for the final vacuum over the cable, and filling the
nation is performed similar to that of a high-pressure system with dielectric fluid. The engineer and/or the
pipe-type termination. Upon completion of the stress contractor should review elevations of the pipes at each
cone, a self-contained liquid-filled pothead is immedi- of the manholes and the termination locations. The pre-
ately put under vacuum and filled with liquid. A degas- ferred setup is to install the fluid from the location with
ifier is typically used to treat the dielectric fluid and fill the lowest elevation. The vacuum pumps should be set
the terminations. up at the manhole locations where the pipes are at their
highest elevations. The number of vacuum pumps is
13.11 DIELECTRIC FLUID INSTALLATIONS determined based on the length of the circuit and the
capacity of the pumps.
Dielectric fluid is typically installed in HPFF and SCFF
cable systems. The cable manufacturer will typically
Based on the size of the project, the fluid can be shipped
from the manufacturer by either rail car or truck. Prior
to arranging shipment, the contractor should investigate
the fill and staging sites to make sure there is adequate
space available.
The fluid is normally tested after arriving on site

and prior to being installed in the pipe. The owner/engi-
neer typically requires a dielectric strength and power
factor test.
The fluid is typically pumped from the tanker at a

rate not to exceed 50 gallons per minute. The contrac-
tor must ensure that the pump being used for the instal-
lation has not been contaminated with fluid from
another application.
Figure 13-20 High-pressure pipe-type cable termination 13.11.2 SCFF Cables

work enclosure (with permission of ComEd). SCFF cable systems use substantially less fluid than
HPFF cable systems. Their reservoirs also are substan-
tially smaller than the HPFF systems. SCFF cables
are typically shipped from the factory or stored with
fluid in the core and under pressure. Therefore, when
filling SCFF cables, smaller volumes of fluid are han-
dled. The fluids are typically pumped by means of a
degasifier unit.
13.12 NITROGEN GAS INSTALLATION

The installation of ASTM Type III dry nitrogen gas on
HPGF systems is performed after all of the cable has
been installed and when the splices, terminations, and
vacuum over cable is completed. The nitrogen is deliv-
ered to the job site by a tanker. Since the nitrogen is nor-
mally delivered in the liquid state, a heat exchanger is
required to covert the nitrogen from the liquid state to a
Figure 13-21 High-pressure pipe-type cable termination gas state prior to being installed into the pipe.
(with permission of ComEd).
13-35
13.13 INSTALLATION OF PRESSURIZATION Typical pressurization plants contain a fluid tank,

PLANTS control panel, motors and pumps, valves, piping,
Pressurization plants, also referred to as pumping and nitrogen bottles to provide the required operations
plants, are manufactured by a relatively small number of of the plant. Figure 13-23 shows the control panel
companies and are normally delivered to the site assem- for an HPFF pumping plant. See Chapter 9 for addi-
bled and ready for installation, or with a small number tional information.
of pieces to assemble. High-pressure fluid-filled pump-
ing plants generally consist of a large fluid reservoir The foundation for an HPFF pressurization plant is
tank, pumps, motors, valves, nitrogen bottles, and a designed to support the plant on its skids, which run
control panel with alarms and gauges. These plants lengthwise on each side of the unit, and sometimes pro-
come preassembled on a skid that can be off-loaded vides a below-grade concrete well between the skids
from a flat bed trailer and installed directly on a founda- designed to contain the volume of fluid in the event of a
tion. These pressurization systems are designed to be tank rupture.
highly reliable and have the ability to communicate the
cable system status to either a system operator at a Because the pressurization plants are self-contained and
remote location or to a local operator. High-pressure preassembled, the connections of the plant to the pipe
gas-filled pressurization plants also are preassembled as cable systems are very simple. Generally, three connec-
self-contained units, requiring only the connection of tions are required to connect the unit to the system:
local power, the supply of nitrogen, and the piping to (1) a 2-in. (5.08-cm) fluid line, (2) a 240-volt three-phase
the cable system. power source, and (3) any supervisory control and data
acquisition (SCADA) or communications connections
13.13.1 High-Pressure Fluid-Filled Pressurization from a terminal unit in the pressurization plant to an
Plants outside fiber-optic cable or other means of communica-
tion source. Normally, the fluid is connected to the cable
Pressurization plants used for HPFF cable systems
pipe by a 2-in. (5.08-cm) pipe run from the underside of
come from the manufacturer preassembled, skid
the pressurization plant underground to the spreader-
mounted, and ready for setting on a concrete founda-
head. Because of pressure loss, it is generally preferred
tion. Although the size of each plant varies according to
to limit the length of 2-in. pipe to no more than 300 ft
the size of the fluid tank and the associated equipment,
(90 m).
the dimensions for a commonly sized plant are roughly
13 ft (4.0 m) wide by 12 ft (3.7 m) high by 40 ft (12.2 m)
The manufacturer’s technical representative should
long. These units are shipped to the site on a flatbed
be contracted to provide onsite services during off-
trailer and are off-loaded by crane onto the concrete
loading, setting, tank filling, systems checkout, and star-
foundation. Figure 13-22 shows a pumping plant
tup activities.
being offloaded onto a foundation pad. The dry
weight of a typical pressurization plant is approximately
40,000 pounds (18,000 kg).
Figure 13-22 HPFF pressurized plant being installed Figure 13-23 Typical HPFF control panel (courtesy W. A.
(courtesy W. A. Chester LLC). Chester LLC).
13-36
13.13.2 Gas Cabinets for High-Pressure Gas-Filled 13.14 AFTER INSTALLATION TESTS
Systems This section addresses only the installation require-
Pressurization plants for HPGF cable systems, also ments of the after-installation tests. Chapter14, Testing,
called gas cabinets, are significantly less complicated, provides a full discussion of cable testing and standards
much smaller, and significantly less expensive than pres- for testing.
surization plants for HPFF cable systems.
After installation, tests (commonly referred to as field
The main components of a gas cabinet are the control acceptance tests) are performed to prove the condition
panel, piping and valving, and two or more nitrogen of the cable circuit prior to energization. Specifically,
bottles connected together through a header system. All these tests look for defects or damage in the cable,
are attached to a supporting structure that is free-stand- splices or terminations that may have resulted from the
ing and requires a space of approximately 6 ft (1.8 m) installation process. It is generally assumed that the rou-
long by 3 ft (0.9 m) deep. Because the gas cabinets are tine factory tests would have identified any defects or
physically small and simple, they can easily be installed damage as a result of the manufacturing process. Suc-
inside a substation control building or a separate enclo- cessful completion of the after-installation tests are
sure; however; it is not uncommon for them to be required before the cable system can be connected to the
designed for installation outside. Occasionally, a small utility’s electrical grid and operated commercially.
roof will be added to the support structure to shelter the
control panel from direct exposure to the elements. In the past, the common industry practice was to per-
form a high-voltage dc test on the installed cable and
Connecting the gas cabinet to the cable system requires accessories prior to energization, whether the cable was
an ac power source (assuming the cabinet will contain extruded or paper-insulated. After many years of investi-
alarms), a 1-in. (2.5-cm) x 1.5-in. (3.8-cm) gas line from gation, the detrimental effects of high-voltage dc testing
the gas cabinet to the spreaderhead, and any additional on high voltage extruded-dielectric cables have been rec-
communications connections desired. ognized. Research has shown that a 24-hr ac soak test at
rated voltage is much more effective in revealing incipi-
As with the HPFF pressurization plant, the manufac- ent faults in accessories without the potentially damag-
turer’s technical representative should be onsite to over- ing effects of dc testing. As a result, dc testing on
see the installation and connection of the gas cabinet to extruded cables as an after installation test is no longer
the pipe system. considered an accepted practice. However, it is still con-
sidered common industry practice to perform high-volt-
13.13.3 Self-Contained Fluid-Filled Pressurization age dc testing on paper-insulated cables. Recently,
Cabinet variable frequency resonant testing is being considered
Pressurization systems for SCFF cables can be either as a viable alternative test in some applications. Field test
low- or medium-pressure systems. The pressurization sets, capable of providing the required over-voltages, are
units are typically located at or on the termination available for the lower to mid transmission voltages, but
structures and in manholes along the route. Small 25- to are in limited supply at the higher transmission voltages.
30-gallon (95- to 114-liter) reservoirs, one for each
phase, are used to supply the fluid under gravity pres- 13.14.1 Extruded-Dielectric Cable Tests
sure or under a nitrogen blanket to the cable system. Generally, two tests are performed prior to energization
Depending on the length and size of the cable system, of a high-voltage extruded-dielectric cable system. These
larger reservoirs may be needed. Fluid level is easily tests are the jacket integrity test and 24-hour soak test.
checked at the termination locations through a sight
glass. The reservoirs are usually mounted directly to one The jacket integrity test confirms that the cable sheath is
of the support beams of the termination structure, one fully insulated from the jacket and that the jacket has no
next to each termination. The fluid piping system is punctures or scrapes that would weaken the manufac-
connected to each phase through a header system and tured electrical isolation between the jacket and the
valves. Aside from SCADA requirements and area light- cable sheath. This test requires that the jacket have
ing, no local power supply is required. The equipment at either a graphite coating applied in the factory or that
each termination location is assembled onsite. Reservoir the jacket is manufactured with a semiconductive coat-
tanks located in manholes are mounted on the walls of ing. This test is performed by applying a dc voltage of
the manhole. 10 kV between the cable sheath and the outer jacket for
a specified period of time (frequently performed for
10 minutes; however, required only for 1 minute) and is
13-37
specified by IEC standards. The applied voltage is pro- terminations by measuring the PD that is produced by
vided by a field-portable high-voltage dc test set, which imperfections in the accessories as a result of workman-
is readily available for this purpose. ship. (Manufacturing defects are assumed to have been
identified by testing in the manufacturing facility).
This test is normally performed one or more times dur- These small imperfections, such as small particles, dust,
ing the course of cable installation and splicing and ter- droplets, and moisture could possibly cause a defect in
minating. For instance, it is frequently performed before accessories to reduce the lifetime of the cable system or
and after the cable is installed, prior to installation of potentially lead to a system breakdown. The PD mea-
the accessories. This has the benefit of ensuring at each surement can be performed while the cable system is
stage of cable installation that the cable was not dam- connected to the electrical grid using the system voltage
aged in the process of pulling in. Many utilities require or while the cable system is de-energized by using sepa-
the cable jacket to be tested again as a complete system, rate testing equipment. The actual procedure used to
including the splices and terminations. It will be neces- measure the PD level should be agreed upon between
sary to disconnect sheath voltage limiters in the link the purchaser and the installer.
boxes and reconnect links to provide a continuous
sheath path and avoid grounds. Ideally, it is preferred to use an increased system-rated
voltage for the PD measurement. However, measure-
Preparation of the cable for this test consists of strip- ments while the cable is de-energized can be made using
ping the jacket away from one end of the cable for a variable frequency resonant test system, which is made
approximately 6 to 12 in. (15 to 30 cm). The graphite up of an exciter transformer, a resonant reactor, and a
coating or semiconducting coating is removed from the control unit. There are limitations on what cable sys-
jacket for a distance of 6 in. (15 cm) from the exposed tems can be tested using PD measurement. For instance,
shield. One lead from the test set is attached to the at extra-high voltage (EHV) levels or on long cable sys-
sheath, and one end is connected to the jacket and is tems, a resonant test set having the required capacity
grounded. The setup and teardown time for this test is may not be available, or a system may require two or
typically less than an hour. Care should be taken to more resonant test sets in parallel. Availability of test
ensure a safety zone is clearly identified around the sets with this high capacity are limited.
cable ends during this test. The test is successful if no
voltage breakdown or flashover is observed. If a break- 13.14.2 High-Pressure Pipe-Type and SCFF Tests
down occurs, the cable manufacturer should be con- Pipe-type cable systems and SCFF cable systems typi-
sulted to recommend an appropriate repair to the cally have two after-installation tests that are per-
jacket. Frequently, depending upon the magnitude of formed: the dc HiPOT test and the 24-hour soak test.
the damage, a heat-shrink repair sleeve will be applied Although the 24-hour soak test is not specified by the
over the damaged jacket area after it is exposed, and the AEIC, it is commonly performed on pipe-type systems.
test will be performed again. Unlike extruded cables, the paper-insulated cables that
are used in pipe-type and SCFF cable systems are not
The soak test is performed by connecting only one end damaged by conventional high-voltage dc testing.
of the completed cable system to line voltage for a speci- Therefore, for pipe-type cable systems, the dc HiPOT is
fied time, usually 24 hours. The utility connects jumpers an accepted after-installation test. The specifications for
to the terminators on one end of the line and leaves the this test are clearly spelled out in the AEIC standards
jumpers at the terminators at the other end off. Safety with regard to time and voltage requirements. When
zones are established at each end of the cable to protect testing pipe-type cables, however, this test is performed
personnel from dangerous open-circuit voltages at the only after the cable system is spliced, terminated, and
cable end. Some utilities require longer soak periods, pressurized. This test is performed by connecting one
such as 48 or 72 hours. During this test, the cable manu- lead of the test set to the conductor at the terminator
facturer and the installation contractor are either onsite and the other lead to ground. The voltage is effectively
or nearby in case of a fault on the system. After success- applied between the conductor and the shield. This test
ful completion of this test, the cable system is ready for is successful if no flashover or insulation punctures
commercial operation. occur. Setup and teardown for this test is normally 8 to
10 hours.
An optional test being considered by some utilities prior
to energization is the partial discharge (PD) measure- After successful completion of the dc HiPOT test, the
ment test. The purpose of this test is to make a determi- 24-hour soak test is performed, similar to the process
nation of the “goodness” of the installed splices and described for extruded-dielectric cables.
13-38
13.15 INSTALLATION OF PIPE FREEZES look for opportunities to maximize pulling lengths,
On HPFF systems, it is often required to either tie a new using the “Figure-8” technique of pulling in opposite
circuit into an existing circuit or make a repair to an directions from a central manhole.
existing HPFF circuit. Under these circumstances, the
fluid in the existing pipes has to be solidified or, as When the fiber-optic cable is to be installed in a duct
known in the industry, a “freeze” has to be made. system, a common procedure is to install the fiber-optic
cable in either a 2-in. (5-cm) PVC conduit or in a 3-in.
A freeze is made by either wrapping the pipe with coils (7.6-cm) PVC conduit containing two or three inner
of copper tubing or installing freeze jackets around the ducts. Fiber cable that is to be used for temperature
pipe. After the coil tubing is wrapped around the pipe, monitoring purposes needs to be installed in between
liquid nitrogen is pumped through the copper tubes. two of the power cable ducts to provide accurate tem-
Thermocouples are installed on the outside of the pipe perature results. Fiber cable ducts that are intended only
adjacent to the copper coils. Once the desired tempera- for communications purposes can be installed anywhere
ture is obtained, a valve is opened to determine whether within the duct run but are usually located on the out-
the freeze is “holding.” side of the power cable ducts.
If the freeze is only needed for a short time, then dewars Many utilities prohibit communications-/temperature-
(nitrogen bottles) can be used to make the “freeze” monitoring cables from being spliced or even run
instead of liquid nitrogen tube trailers. Typically, two to through the same manhole with the power cables
four dewars are used in a 24-hour duration. If the because of the requirement to de-energize the power
cutover will take more than a couple of days, the con- cables before allowing personnel into the manhole to
tractor will arrange for a liquid nitrogen tube trailer to perform any work. When this is the case, utilities often
be dropped at the site. require the fiber cables to be separated from the power
cable ductbank at the manhole locations. In this case,
Making a freeze and handling liquid nitrogen should the fiber cable duct will be diverted away from the main
only be performed by experienced personnel since certain run, routed around the manhole, and then routed back
safety precautions are required handling the nitrogen. into the duct run after passing by the manhole. A shal-
low hand hole is often installed in a convenient location
adjacent to the manhole to be used for pulling and/or
13.16 INSTALLATION OF FIBER-OPTIC CABLES
splicing operations. Some utilities also include a small
Sophisticated digital temperature-monitoring systems local terminal station for connecting a portable device
and fault location systems being used on transmission to read local temperatures. These are small waterproof
cable systems often require a fiber-optic cable installed enclosures commonly seen for communications fiber-
as part of the project design. The installation of fiber- optic splices.
optic cables is, therefore, becoming increasingly more
common. This section discusses the installation of fiber- An alternative to separating the fiber cable ducts from
optic cables that are not an integral part of the power the power cable ducts is to route the fiber cables through
cable but are installed in a separate conduit within the the manhole in a conduit attached to one wall of
duct bank system or pipe-type cable trench or alongside the manhole.
the direct buried power cable.
13.17 SUBMARINE CABLE INSTALLATION
The requirement to include fiber-optic cables in the
same trench with the power cables requires the designer This section addresses submarine cable installations of
to consider (1) how the fiber-optic cable will be installed relatively short length and relatively shallow water. Typ-
and (2) whether the fiber cable will be routed through ical installations falling into this category would be
the power cable manholes or installed in separate adja- crossings of rivers, lakes, channels, or bays, having
cent handholes. lengths of between several hundred feet at the short end
to several thousand feet, or a few miles, at the longer
Fiber-optic cables used with power cable systems gener- end of the range. Longer crossings of many miles in
ally have a 600-lb maximum pulling tension limit. length and significantly deeper water are addressed in
Although this is a low threshold, it is sufficient for these Chapter 7, Section 7.3.
cables to be pulled through many of the power cable
manholes without needing to be spliced, thus reducing The design of various types of submarine cables is dis-
the number of splices and, hence, losses in the fiber cussed in Chapter 12.
cable circuit. Wherever possible, the designer should
13-39
The information presented in this section applies to the allow even a barge to operate, in which case the cables
installation of extruded-dielectric, high-pressure pipe- may have to be installed using a marsh buggy. Short
type, and self-contained fluid-filled submarine cables. submarine cable projects described in this section that
Direct current (dc) cables are normally not installed in are to be buried will most likely be installed from a
short lengths and, thus, are discussed in Chapter 7. barge, using either jetting, plowing, or trenching proce-
dures. Figure 13-24 shows a typical arrangement of
Submarine cables are required to operate in both wet equipment and cable reels on a spud barge. Alterna-
and dry environments. Most submarine cable installa- tively, if burial is not required, the cables are laid directly
tions transition through five different environments: on the bottom.
(1) an underground section beginning at the riser struc-
ture (in air), (2) a land/water interface, also referred to Horizontal directional drilling (HDD) is becoming
as the shore fall, (3) the underwater section, (4) another more common for cable crossings of rivers and small
shore fall section on the other side of the water crossing, lakes. As of 2006, the longest extruded-dielectric trans-
and (5) another underground section terminating at the mission cable crossing is approximately 4000 ft
riser structure (in air). Different installation techniques (1200 m), and the longest pipe-type cable crossing is
are applicable to various environments. The designer about 8000 ft (25,000 m).
must fully understand the nature of each of the environ-
ments in which the submarine cable will have to operate. 13.17.1 Installing Extruded-Dielectric and Self-
Contained Fluid-Filled Cables
Submarine cable installation is a highly specialized con- Extruded-dielectric cables and SCFF cables can be
struction activity that should only be performed by installed using similar methods, although SCFF cables
qualified and experienced marine contractors. The con- require more specialized handling and care during the
tractors that perform this work must have specialized process so as not to damage the fluid containment
equipment and experienced crews to address the various within the cable.
unexpected conditions that frequently accompany most
submarine cable installation projects. Submarine cable One common technique for installing submarine cables
installation can require a significant crew size, a signifi- is the jetting process. Jetting is a process that can simul-
cant amount of equipment, and a marine vessel, typi- taneously trench, lay, and embed the cable with one
cally a barge for short water crossings. device. This method is used where the sediments are suf-
ficiently soft, without significant rocky material, and
A cable-laying ship is typically required for large lake local environmental regulations permit the disturbed
crossings, as well as for the longer crossings described in material to be naturally re-deposited into the trench. In
Chapter 7. All of this means that daily standby charges this method, a rubber-tired or skid-mounted plow is
for delays in the work resulting from either unexpected
site conditions, weather, or design changes will be signif-
icant. Therefore, it is prudent for the designer to become
familiar with the process of submarine cable installation
to be well educated as to the potential risks and costs of
this work.
Various methods are available for the installation of

submarine cables. These methods employ different types
of specialized equipment specifically designed for sub-
marine cable installation. The selection of the appropri-
ate installation method for any given project depends on
many factors, but primarily on the length and depth of
the underwater route and whether the cables are to be
embedded into the bottom sediments or just laid on the
bottom. Normally, regardless of the installation method
chosen, a barge or boat will be required.
Short crossings of streams and rivers can usually be Figure 13-24 Typical arrangement of equipment,
accomplished without the need for a specialized cable- cable reels, and jet plow on a spud barge (courtesy
laying ship. Sometimes the water depth is too shallow to Caldwell Marine).
13-40
pulled along the bottom behind the barge. Figure 13-25 The speed of the cable as it is being payed out over the
shows a skid-mounted jet plow being lowered into back of the barge is controlled by a device called a linear
the water. tractor, also called a caterpillar. This device has a series
of grippers that squeeze the cable as it is passed though
High-pressure water from barge mounted pumps is the caterpillar, from the cable reel. The grippers are on a
injected into the sediments through nozzles situated hydraulically driven track that moves the cable forward
along the plow, causing the sediments to momentarily or back, as the control operator deems necessary to con-
fluidize, creating an open trench. As the plow is pulled trol the cable installation. The amount of gripping
along the route behind the barge, the cable is laid into length for a particular installation is based on the maxi-
the opened trench through the back of the plow. mum cable tension expected, the allowable sidewall
Depending on the nature of the soils, the trench is back- pressure, and its tension shear limit. These devices were
filled by the water current and the natural settlement of once used only on deeper submarine cable installations,
the suspended material. Depth of burial is controlled by but are now used on most installations.
adjusting the angle of the plow relative to the bottom.
Figure 13-26 shows a tire-mounted jet plow installing Various umbilical cords connect the submerged jet plow
cable near the shore. The high pressure water line for the to the barge control equipment to allow the operators to
jet nozzles can be seen at the black hose at the rear of monitor and control the installation process and make
the jet plow. adjustments to speed and alignment as the installation
proceeds across the water. In the proper soils, this
method is quick, and any turbidity created is generally
short-lived. The speed of cable installation using this
method is slower, as several simultaneous activities are
occurring at once. Although the cable is immediately
protected, once it is in the water, this process is slow and
subjects the barge and crew to the potential for bad
weather longer.
Another installation method used frequently involves

laying the cables on the bottom first, then jetting them in
place afterwards. This post-embedding process is limited
to soft bottom materials and requires either divers using
hand-held water jets or special embedding machines.
Another method of submarine cable installation is called

plowing. Plowing is similar to the jetting method in that
Figure 13-25 Skid-mounted jet plow being lowered it simultaneously opens a trench and lays the cable.
into water (courtesy Caldwell Marine). However, high-pressure water is not used to fluidize the
sediments. Instead, the plow is lowered into the sediment
to the appropriate depth, and the trench is opened as the
plow arm is pulled through the material, much like an
aboveground plow. This process is used where the sedi-
ments are too hard and compacted for jetting, but are
not hard enough to require trenching. This method
tends to reduce the amount of sediments that are tempo-
rarily suspended in the water at the insertion point.
When the bottom sediments are too hard for jetting or

plowing, trenching may be required. In this method, it is
customary for the trench to be opened first, using con-
ventional marine trenching equipment, and the cable is
laid directly in the trench in a second operation using
divers or submersibles to direct the cable into its final
Figure 13-26 Rubber-tired jet plow (with permission position. Sometimes a small metallic guide wire is
of SMECO).
13-41
placed in the trench and used for guidance purposes, 13.17.3 Repair of Damaged Cables
with the guide wire being removed as the power cable is Repairing a damaged submarine cable is an expensive
installed. Sometimes the spoils from the trenching pro- and lengthy process. The time required to repair an
cess must be removed and taken to a disposal site, underwater cable must be realistically considered in the
depending on environmental regulations. Trenching is system design in terms of required cable system reliabil-
usually a slower and more costly technique compared to ity and availability (e.g., how many spare circuits or
jetting and plowing. phases, what spacing is required between phases, and
what embedment depth). The outage time due to a fail-
13.17.2 Installing Pipe-Type Cables ure frequently requires that a spare phase or spare cir-
In the case of a new submarine installation, the pipe sec- cuit be installed.
tions are usually welded on shore or on a barge, then
floated to the desired location. The ends of the pipe are The first step is to locate the damage, which requires
always sealed to keep water out of the pipe during boats, divers, and fault location equipment to be mobi-
installation. If the pipes are to be embedded in a trench, lized to the site. Mobilizing the specialized contractor(s)
the trench would be opened first using methods and equipment to the site often takes several days to
described previously, then the pipe either pulled directly several weeks. After the location is determined, the
into the trench or floated to a position over and then retrieval and repair process can be evaluated and appro-
submerged into the trench. Due to buoyancy concerns priate equipment mobilized to the site. If the failure
when installing an empty pipe, the pipes are frequently occurs during bad weather, weather conditions could
coated with a heavy concrete coating, called heavy coat. delay repairs for extensive periods of time. The depth of
This heavy coat is in addition to the normal corrosion the cable and whether the cable was embedded or just
protection coating applied to pipe-type cable systems. laid on the bottom have a significant impact on the cost
and schedule of the repair. The designer should work
with the appropriate authorities and contractors spe-
cialized in submarine cable repair to develop a proce-
dure to follow in the event of a cable failure.
13-42
REFERENCES EPRI. 1992. Electric Power Research Institute.

AEIC. 2005. Association of Edison Illuminating Com- Underground Transmission Systems Reference Book.
panies. AEIC CG3-05. “Guide for Installation of Pipe Palo Alto, CA. TR-101670.
Type Cable Systems.” Second Edition. Alabama.
IEEE. 2005. Institute of Electrical and Electronics
AEIC. 1997. Association of Edison Illuminating Engineers. Std. P1120. IEEE Guide for the Planning,
Companies. AEIC CG4-97. “Guide for Installation of Design, Installation, and Repair of Submarine Power
Extruded Dielectric Insulated Power Cable Systems Cable Systems. IEEE PES-ICC Committee. March.
Rated 69 kV Through 138 kV.” Second Edition.
Alabama. NACE. 2004. National Association of Corrosion Engi-
neers. Standard RP 02 74. “High Voltage Electrical
API. 1994. American Petroleum Institute. Standard Inspection of Pipeline Coatings Prior to Installation.”
No. 1104. “Standard for Welding Pipe Lines and Revision 4.
Related Facilities.” Eighth Edition. March.
NASTT. North American Society for Trenchless
EEI. 1957. Edison Electric Institute. Underground Sys- Technology. Website address: http://trenchless-
tems Reference Book. technology.org/replace.html#Anchor-Mic-11377.
13-43
CHAPTER 14 Testing
Author: John H. Cooper, Power Delivery Consultants, Inc.
Reviewer: Ernesto Zaccone, Prysmian Cables and Systems
Testing transmission cables and accessories is vitally important in designing new compo-
nents, qualifying components for utility acceptance, and installing, operating, and main-
taining them once they are supplied to the user.
Various test standards and procedures have evolved to accommodate the unique charac-
teristics of different cable systems. Many were developed by utility/industry organiza-
tions, such as the Association of Edison Illuminating Companies (AEIC), the Insulated
Conductors Committee (ICC) of the IEEE, the International Electrotechnical Commis-
sion (IEC), and the Insulated Cable Engineers Association (ICEA). Others originated
from research organizations, individual utilities, consultants, or manufacturers to address
specific concerns with individual cable components.
Although the electrical tests required for cables are similar to those for other types of
electrical equipment, few high-voltage laboratories other than at cable factories are prop-
erly equipped to perform most of the tests required for transmission cables. This is
because of the relatively high capacitance of the cable samples, the non-self-healing fail-
ure mechanisms, the very high ratios of impulse to ac breakdown voltage, and the need
for specially skilled technicians for jointing and terminating cable samples.
This chapter describes the overall principles of cable and accessory testing, defines gen-
eral categories of testing, and summarizes the applicable standards, guides, and proce-
dures that are commonly accepted by the cable industry. It addresses specific test
procedures, laboratories, and equipment for ac, impulse, dc, and thermomechanical tests,
and describes diagnostic procedures employed in the laboratory and in the field.
John Cooper is a principal engineer with Power Delivery Consultants,

Inc. (PDC). He received a BSEE degree from Texas A&M University
(1967) and an MSEE degree from the University of Pittsburgh (1968).
He joined Westinghouse Electric in 1967, where he initially performed
power systems planning studies. He later held various engineering and
technical management positions at the Waltz Mill Underground Trans-
mission Test Facility. Mr. Cooper then worked for Power Technologies,
Inc. (PTI) from 1988 to 1991 as a consultant to electric utilities on the
design, specification, and installation of transmission cable systems. He has worked at
PDC since it was founded in 1991. Mr. Cooper is a Fellow of the IEEE Power Engineer-
ing Society, a Voting Member of the IEEE Insulated Conductors Committee (ICC), and
a member of CIGRE. He is a past Chairman of the ICC Cable Systems Subcommittee.
He is a registered professional engineer in the states of Pennsylvania, Georgia, New Mex-
ico, and Texas.
14-1
Chapter 14: Testing EPRI Underground Transmission Systems Reference Book
14.1 INTRODUCTION • Maintenance or Proof Tests

Many tests, covering the various aspects of cable design, • Special-purpose Tests
quality control, and maintenance, have been devised
over the years for cables and their accessories. These These test categories sometimes overlap and sometimes
tests can be grouped into the following general catego- are referred to by different names. To add to the confu-
ries for discussion purposes: sion, there are several organizations that write transmis-
• Development Tests sion cable test standards, specifications, and guides, and
a number of these overlap as well.
• Long-term Prequalification Tests
• Type or Qualification Tests Test procedures for each category can consist of ac,
• Factory Production (Routine) Tests hot impulse, switching surge, dc, and load-cycle tests,
as well as mechanical tests that can be applied to any
• After-installation Tests
test category.
Section 14.2 describes typical high-voltage electrical

tests including ac voltage withstand and breakdown
tests, dc hi-potential tests, impulse and switching surge
tests, insulation dissipation factor tests, and factory par-
tial discharge measurements. Section 14.3 reviews type
or qualification tests and factory production tests for
different types of cables, joints and terminations. Sec-
tion 14.4 discusses after-installation tests. Section 14.5
outlines maintenance or proof tests. Section 14.6
describes special-purpose tests, including long-term
accelerated life tests, mechanical bending tests, thermo-
mechanical bending tests, and longitudinal water pene-
tration tests.
Figure 14-1 illustrates the relationship between recom-

mended tests and the stages of cable development, pro-
duction, installation, and operation.
14.1.1 General Testing Categories

Development Tests
Development tests, sometimes called design tests, refer
to the electrical and mechanical tests conducted during
the development of new cables or cable accessories.
They usually consist of the basic laboratory test proce-
dures, described in Section 14.2, which allow the results
of the tests to be compared to those for similar equip-
ment or industry standards. However, they may also
include unique tests specific to the type of equipment
being developed and technical issues that must be
resolved. The reader is encouraged to review the docu-
ments cited in the references (Allam 1986; Bolin 1982;
Blais 1972; Auerbach 1980) for specific examples of
design tests. Typical design tests are ac breakdown tests,
mechanical bending tests, impulse breakdown tests, dis-
sipation factor measurement, partial-discharge mea-
surements, etc.
Development tests are, in principle, carried out privately

by the cable system manufacturer before all the other
Figure 14-1 Tests during cable system life cycle.
14-2
EPRI Underground Transmission Systems Reference Book Chapter 14: Testing
tests. These tests may be recommended by some stan- XLPE cable systems with rated voltages of 150 kV and
dards or specifications. Their details and procedures are less. A recent CIGRE working group (CIGRE 2006)
generally left to the discretion of the manufacturer, but recommended that PQ tests be performed on extruded-
should preferably include the following: dielectric cables systems with voltages up to 150 kV for
• Evaluation of materials and processes electrical gradients higher than 8 kV/mm at conductor
screen and 4 kV/mm at the insulation screen.
• Evaluation of Weibull parameters (CIGRE 1989)
Type or Qualification Tests
• Determination of the voltage versus time breakdown
Type tests or qualification tests are performed to dem-
exponent “n” (see Section 14.6.1).
onstrate the adequacy of designs, manufacturing, and
Long-term Prequalification Tests materials of a cable system or component to achieve the
The International Electrotechnical Commission (IEC) desired performance characteristics. The term “qualifi-
defines prequalification (PQ) tests as “a test made cation test” is commonly used in Association of Edison
before supplying of a general commercial basis of a Illuminating Companies (AEIC) and Insulated Cable
cable system in order to demonstrate satisfactory long Engineers Association (ICEA) cable specifications and
term performance of the complete cable system. The PQ standards. IEC cable standards contain similar tests, but
test need only be carried out once unless there is a sub- they are called “type tests” in IEC documents. The qual-
stantial change in the cable system with respect to mate- ification test series—consisting of mechanical, electrical,
rial, manufacturing process, design and design levels.” and load cycle tests—is typically performed to qualify a
cable design, materials, and manufacturing process of a
Generally, the cable system, or at least part of the cable particular manufacturer. A type or qualification test is
system, is installed in an environment that duplicates the generally required when there is any significant change
installation conditions where the cable will be applied. in materials or manufacturing process even when there
For example, if the cable is to be installed in a concrete- is a history of successful operating experience with a
encased duct bank, then the prequalification tests are cable of similar design. These tests are described more
performed with the cable in a duct bank. In many cases, fully in Section 14.3.
the prequalification test facility includes multiple instal-
Factory Production (Routine) Tests
lation conditions (i.e., duct bank, direct-buried, and tun-
Two categories of tests are performed on each manufac-
nel installation segments). Figures 14-2 and 14-3 show
turing order for transmission cable. The first category,
the CESI (Centro Elettrotecnico Sperimentale Italiano)
sample tests, consists of destructive tests that are per-
prequalification test facility used for 400-kV XLPE cable
formed on short samples of the cables that were manu-
system tests and a schematic of the test facility.
factured for the order. The sample tests typically include
dimension checks, material property tests, and electrical
The requirement for long-term PQ tests is generally lim-
property tests. The second category of tests, called tests
ited to recently developed cable systems that have little
on shipping lengths, are nondestructive tests performed
or no commercial operating experience. For example,
on each shipping reel of cable and prefabricated acces-
IEC 62067 requires PQ tests for EHV XLPE cable sys-
sories, which serves as a final quality control check.
tems, whereas IEC 60840 does not require PQ tests for
These tests typically include a high-voltage withstand
test, partial discharge or insulation dissipation factor
measurement, conductor resistance measurement,
jacket integrity test, etc. These tests are described more
fully in Section 14.3.
Figure 14-2 CESI prequalification test area (courtesy CESI). Figure 14-3 Schematic of CESI prequalification test area.
14-3
After-installation or Commissioning Tests cable accessories are capable of withstanding the joule
After-installation tests, also called commissioning tests heating from fault currents and to determine if mechan-
or acceptance tests, are tests that the cable system must ical restraints are capable of withstanding the mechani-
pass after installation and before it is placed in service. cal forces produced by fault currents.
This type of test is intended to detect shipping or instal-
lation damage, and to show any defects or errors in Materials compatibility tests are often performed dur-
workmanship on splicing and terminating the cable. A ing the development test phase or when materials in an
document was issued by CIGRE based on the present existing design are replaced with new materials. These
utility practice regarding the maintenance and failure tests are to determine that the different materials in a
modes for the usual cable systems and the related diag- cable system component do not significantly affect the
nostic methods. The document is also recommending electrical or mechanical properties of components fabri-
the right procedures for effective and efficient mainte- cated from other materials. Special-purpose tests are
nance (CIGRE 2005). These tests are described more described more fully in Section 14.6.
fully in Section 14.4.
14.1.2 Test Standards and Guides
Maintenance or Proof Tests
Maintenance tests are defined (IEEE 2001) as a field Several levels of guides or standards may determine how
test made during the operating life of a cable system. It a specific test is performed. For example, the AEIC
is intended to detect deterioration of the system and to qualification tests commonly require a hot impulse test.
check the serviceability so that suitable maintenance The AEIC specification, in turn, refers to IEEE Stan-
procedures can be initiated. The term proof test is not dard 82, “Test Procedure for Impulse Voltage Tests on
defined in IEEE Std. 400, but it is commonly used to Insulated Conductors” (IEEE 2002). This standard
mean the same as maintenance test. Maintenance tests refers to another IEEE/ANSI (American National Stan-
are less severe than after-installation tests. Guidance for dards Institute) standard, “IEEE Standard Techniques
performing maintenance tests is provided by the IEEE, for High Voltage Testing” (IEEE 1995), to describe basic
the ICEA, and the AEIC. These tests are described test terminology and standard impulse waveforms. In
more fully in Section 14.5. general, a standard is a document that describes specific
mandatory procedures for testing certain equipment,
Special-purpose Tests while a guide provides suggestions or recommendations
As indicated by the name, special-purpose tests include for approaches to good testing practices.
a variety of special tests that have been devised to
address certain cable system problems or to obtain spe- 14.1.3 Standards Organizations
cific information not covered by the categories of tests The organizations responsible for most of the cable test
described above. standards and guides used in North America are AEIC,
IEEE, and ICEA. These standards and guides com-
Many mechanical bending tests are commonly classified monly refer to basic test standards written by the Amer-
as special-purpose tests. These tests are seldom per- ican Society for Testing and Materials (ASTM). The
formed routinely, but are included with other tests as AEIC is an organization composed of electric utilities,
specifically agreed upon in advance by the purchaser generating companies, transmitting companies, distrib-
and manufacturer. An example of this type of test is the uting companies (including investor-owned, federal,
thermomechanical bending (TMB) testing of pipe-type state, cooperative, and municipal systems from both
cables and joint restraint systems. TMB tests were within as well as outside the United States), and entities
devised and refined in the late 1970s and 1980s responsible for technical research and for promoting,
(Bankoske et al. 1977; Aabo 1988) to learn more about coordinating, and ensuring the reliability, integrity, and
damage that may occur to HPFF cables as a result of efficient use and operation of the bulk power supply sys-
bending from load cycling. An extensive series of tem. The AEIC has a cable engineering section that
mechanical tests was developed by CIGRE Study Com- writes and updates cable industry specifications with
mittee 21 (CIGRE 1972) for submarine cables to deter- consultation from cable manufacturers and technical
mine if a cable design will withstand the bending that advisers. AEIC cable specifications include qualification
occurs during cable laying and retrieval, and by ocean tests and routine tests on cable components and com-
currents. Additional details of submarine cable tests are pleted cables. The IEEE standards relating to transmis-
given in Chapter 7. sion cables are written by working groups in the IEEE
Insulated Conductors Committee (ICC). The ICC stan-
Short-circuit tests are performed at a high-power testing dards are consensus standards and guides, because
laboratory to determine if the cable shield/sheath and agreement must be reached among ICC members con-
14-4
sisting of manufacturers, users, and independent con- tests and the special requirements that may apply to
sultants before they are accepted. cable testing.
The ICEA prepares specifications for cable measure- 14.2.1 AC Voltage Withstand and Step-Breakdown
ments and test procedures that are commonly refer- Tests
enced in the AEIC specifications. IEC standards are The most common types of ac tests performed on cable
specified if bids are solicited internationally or if no samples are ac-withstand tests and ac step-breakdown
applicable national standard is available from AEIC or tests. The allowable range of frequency for ac tests is 49
IEEE. to 61 Hz, because the frequency in the testing labora-
tory may be different from the power frequency in the
The IEC standards are consensus standards based on country where the cable will be installed. The ac-with-
the studies developed by working groups of the CIGRE stand, or high-voltage-time, test is a pass/fail test per-
study committee B1 (formerly 21) and must be approved formed by applying a specified ac withstand voltage to
by the national technical committees of the countries the cable for a specified time period, usually with the
that actively participate in IEC. cable at ambient temperature. The specified ac test volt-
age is generally a compromise between stressing the
14.2 HIGH-VOLTAGE TESTS AND cable insulation hard enough to detect defects while not
MEASUREMENT PROCEDURES doing damage to good insulation. The specified ac test
High-voltage tests for transmission cables and cable voltage for a particular cable type is chosen, based on
accessories are similar to those for other types of power previous test experience. Ac-withstand voltages required
system equipments in that ac, lightning impulse, switch- for AEIC/ICEA qualification tests and IEC type tests
ing surge, and sometimes HVdc sources provide the
basic electrical stresses that are used to perform the tests
required in the applicable test standards. The primary
difference between transmission cable tests and those
for other types of power system components is that the
cable tests generally require more time and are more
expensive to prepare for testing. It is common for a
transmission cable system type test to cost more than
$100,000. Since it is often not economically feasible to
perform the numerous, repetitive tests with the statisti-
cal confidence that is possible with self-restoring insula-
tion or distribution-class equipment, special non-
destructive measurements, such as partial-discharge and
insulation dissipation factor measurements, have been
devised to measure insulation characteristics before
electrical breakdown occurs.
Due to the special skills and equipment required to

prepare transmission cables for testing, and the rela-
tively high capacitance of transmission cables, few high-
voltage laboratories are capable of performing these
tests. These tests are most frequently performed in a
cable manufacturer’s laboratory or, in some cases, at an
independent testing laboratory such as KEMA, CESI,
CREEPI, etc. Figure 14-4 shows the interior of a cable
manufacturing laboratory during type testing of a
345-kV XLPE transmission cable system.
The following subsections describe the high-voltage

tests and measurements that are normally used for cable Figure 14-4 Test setup for 345-kV XLPE cable type test
(courtesy SILEC, Northeast Utilities).
14-5
are given in Table 14-1. The AEIC/ICEA qualification two requirements often lead to the use of resonant ac
test ac-withstand voltages are from AEIC CS2 (HPFF), test sets for cable tests rather than conventional step-up
CS4 (SCFF), CS9 (XLPE), and ICEA S-108-720 transformers. Figure 14-5 shows such an ac resonant
(extruded insulation). The IEC type test ac withstand test set. In this type of circuit, the fault causes loss of
voltages for SCFF cables are from IEC 141-1. IEC circuit resonance, and the test voltage collapses almost
60840 and 62067 (extruded insulation) type test proce- immediately after a cable breakdown. The adjustable
dures do not include an ac-withstand test other than a inductors (shown as X S1 and X S2 in Figure 14-6) also
15-minute withstand test after the hot impulse test. limit the fault current to several amperes or less.
The ac step-breakdown test is, in principle, a develop- Numerous methods have been used to reduce the time
ment test and consists of applying increasing voltage and cost of terminating cables for ac high-voltage tests.
levels to the cable sample until a breakdown occurs. The This is especially true for extruded-dielectric cables,
increases in the test voltages are typically fixed percent- because they commonly have very high ac breakdown
ages of the preceding test voltage. Because of the dielec- voltages when they are new, and it is not economically
tric losses, heating may cause thermal runaway for some practical to design commercial terminations for such
cable types. The duration of the test at each voltage high breakdown voltages. A special test termination for
level, in this case, may decrease as the voltage is raised. extruded-dielectric cables, called a water termination, is
The primary objectives of ac step-breakdown tests are
to obtain relative ac strengths for variations in cable
construction and/or to estimate the amount of deterio-
ration in breakdown strength. The collection of the ac
step-breakdown tests, together with the long-duration
voltage tests, is used to determine the statistical Weibull
parameters and the evaluation of the expected life of a
relative process or material.
The major test equipment considerations with ac high-

voltage transmission cable tests are the relatively high
reactive power requirement, which is a result of the large
cable capacitance, and the desire to limit destructive
energy in the vicinity of a cable breakdown. Minimizing
energy to the fault greatly increases the possibility of
determining the cause of an electrical breakdown. These
Figure 14-5 Ac resonant test set (center), voltage divider
(left) and standard capacitor (right) (courtesy Southwire).
Table 14-1 AEIC/ICEA and IEC Ac High-Voltage Test
Requirements for Transmission Cables (Phase-to-Ground)
Rated Voltage Cable Test Duration and Voltage
(L-L kV) Type AEIC/ICEA IEC
138 HPFF 24 hr 200 kV --
138 Extruded 2 hr 200 kV --
138 SCFF 6 hr 217 kV 24 hr 190 kV
230 HPFF 24 hr 332 kV --
230 SCFF 6 hr 327 kV 24 hr 320 kV
345 HPFF 24 hr 500 kV --
345 SCFF 6 hr 445 kV 24 hr 430 kV
500 HPFF 24 hr 600 kV --
500 Extruded -- --
500 SCFF 2 hr 577 kV (24 hr 600 kV)a
Figure 14-6 Schematic of a cascade series resonant ac
a. For information only. high-voltage test set.
14-6
shown schematically in Figure 14-7. In this test terminations, dc high-potential (Hi-Pot) tests are only used for
tion, the cable sample is prepared by removing the after-installation testing of laminated dielectric trans-
jacket and insulation shield to a prescribed length. The mission cable systems.
water, which surrounds the cable insulation, is continu-
ously deionized to give a high resistivity to provide 14.2.3 Impulse and Switching Surge Tests
stress relief by resistively dividing the test voltage along Cable lightning impulse tests are performed to confirm
the length of the termination. Even with continuous the impulse-withstand capability of a particular cable as
deionization of the water in the termination, heat from well as to measure the intrinsic dielectric strength of the
resistive losses in the termination must be removed by cable. At transmission voltage levels, impulse tests have
circulating the water through an external cooling unit. been used for a long time as design tests, and have been
Water terminations are commercially available with included in U.S. qualification tests only recently. (The
rated ac voltages up to approximately 800 kV. hot impulse test was included in the AEIC pipe-type
cable specifications [AEIC 1997] for the first time in
Test terminations for impregnated-paper-insulated cables 1982.) This is partially because of the belief that the
are typically commercial terminations with a higher volt- impulse-breakdown level of a well-made, impregnated-
age rating than that of the cable to be tested. Special ter- paper-insulated cable is predictable. Further, it is expen-
minating procedures have also been devised for research sive to conduct impulse-breakdown tests for transmis-
projects; however, they are justified only where a large sion cables, and many cable engineers thought that the
number of cable breakdown tests are performed. ac high-voltage-time tests gave a satisfactory indication
of the quality of a cable.
Test terminations for compressed-gas-insulated cables
normally utilize special test fixtures that, in most cases, Switching-surge tests have not been included in the cable
are available only at the manufacturer’s laboratories. industry testing specifications in the United States, pri-
marily because many power systems engineers now hold
14.2.2 Dc Hi-Potential Tests the opinion that lightning impulses are more severe than
Dc testing of ac transmission cables is more commonly the switching surges that cables experience. This opinion
employed as an installation or maintenance test, rather may change with the improved lightning overvoltage
than as a production test, because other test methods protective levels afforded by modern surge arresters at
have been found to be more suitable in detecting manu- cable system voltages at and above 500 kV. A switching-
facturing imperfections that may occur in ac transmis- surge withstand test has been included in purchase
sion cables. In addition, space charges can take as long specifications by at least one major utility, and others
as one day to dissipate in low-loss insulations such as may follow.
XLPE or compressed-gas-insulated systems, which
would cause overstressing of portions of the insulation Impulse test procedures for cables are specified in IEEE
during subsequent ac or impulse tests. With rare excep- Standard No. 82 and IEC Standard 60230. These
standards call for 10 voltage impulses at each polarity.
Basically the impulse-withstand and breakdown tests
differ only by stopping the withstand test before a
breakdown occurs followed by a short-time (15-minute)
ac-withstand test to demonstrate withstand at a given
impulse level. The additional ac-withstand test is to
demonstrate that there is no incipient failure produced
by the application of impulses up to the required with-
stand level. Most cable impulse tests must be performed
with the cable conductor at maximum normal operating
temperature, because breakdown strength decreases
with temperature, and lightning overvoltages may occur
when the cable is at its maximum temperature.
A more recent variation of the impulse tests is the

dc-biased impulse test recommended for testing dc cable
(CIGRE 1980). In this type of test, some form of
Figure 14-7 Cable water termination for extruded-dielectric
dc-blocking impedance is required to isolate the
cables.
14-7
impulse generator from the dc voltage source until the 14.2.4 Insulation Dissipation Factor Test
impulse is applied. In addition to the high-voltage tests described above,
two types of measurements are used to determine the
As in the case of ac step-breakdown tests, one of the condition of the cable insulation. The first of these non-
most difficult tasks in performing cable impulse-break- destructive measurements is called the insulation dissi-
down tests is to terminate the cable so that the termina- pation factor, or tan δ, measurement. This measurement
tion will not break down before the cable does. is sometimes called an insulation power factor measure-
ment, but this is not technically correct, as described
This is particularly true for extruded cables, for which below. The quantity measured by this test is the tangent
the impulse breakdown voltage may be as much as three of the angle between the actual cable charging current
times its Basic Impulse Insulation Level (BIL). With that is flowing through the cable insulation and the
impregnated-paper-insulated transmission cables, quadrature current that would flow in a perfect or loss-
impulse breakdowns can usually be obtained with the free dielectric (see Figure 14-8). Tan δ is almost numeri-
same type of terminations as for ac tests. With extruded- cally equal to the cosine of the angle (power factor)
dielectric cables, one of two types of test terminations is between the voltage across the insulation and the cur-
commonly used. If the resistivity of the insulation semi- rent flowing through the insulation for practical cable
conducting screen is sufficiently low, an impulse termi- dielectrics. The dissipation factor is important, because
nation may be obtained by simply removing the metallic it is proportional to the total amount of loss in the cable
insulation shield and using the insulation semiconduct- insulation. Dipole losses normally constitute the major-
ing layer for resistive voltage grading. A second, simple ity of these losses, but conduction and ionization losses
termination is similar to the ac water termination, but are measured as well. The magnitude of the dissipation
when terminating for impulse tests, the resistivity of the factor is important, as well as the changes of the magni-
water should be much lower than for the ac tests; in tude at specific temperatures and test voltages. An
many cases, tap water is sufficient. The primary draw- increase in the dissipation factor at the same test condi-
back of both of these simple impulse terminations is tions is one indication of cable insulation aging. This
that they may not be used for the subsequent, short- increase in dissipation factor for impregnated-paper
term, ac-withstand test. insulation is believed to be caused by an increase in the
moisture in the insulation. This occurs because water is
For compressed-gas-insulated cables, the same termina- a high-temperature decomposition product of paper.
tion is almost always used for impulse and ac-withstand Figure 5-9 in Chapter 5 shows a typical variation of
tests; the peak impulse strength of compressed-gas-insu- insulation dissipation-factor with moisture and with
lated cables is not much greater than the peak voltage of temperature for oil-impregnated paper.
the sine wave for ac-withstand tests. Table 14-2 desig-
nates typical impulse test requirements. In summary, the tan δ measurement has been found to
give the best indication of the quality of cables with
impregnated-paper insulation. These cables must be
Table 14-2 AEIC, ICEA, and IEC Impulse Test Requirements made with high-quality paper, and correctly taped in a
for HPFF, SCFF, and Extruded-Dielectric Cables (10 shots low-humidity environment to meet the dissipation-factor
both polarities unless otherwise noted) requirement of industry specifications. However, it is not
Switching necessarily the best measurement for extruded-dielectric
Rated Cable Lightning Impulse (kV) Impulse (kV)
Voltage Type AEIC ICEA IEC IEC
138 kV HPFF ± 650 ± 650 ± 520 N/A
138 kV Extruded ± 650 ± 650 ± 650 N/A
138 kV SCFF N/A N/A ± 650 N/A
230 kV HPFF ± 1050 N/A ± 1050, 950 N/A
230 kV Extruded ± 1050 ± 1050 ± 1050 N/A
230 kV SCFF ± 1050 N/A ± 1050, 950 N/A
345 kV HPFF ± 1300 N/A ± 1050, 950 N/A
345 kV Extruded ± 1300 ± 1300 ± 1175 ± 950
345 kV SCFF ± 1300 N/A ± 1125, 1050 N/A
500 kV Extruded N/A N/A ± 1550 ± 1175
500 kV HPFF ± 1500 N/A N/A N/A
Figure 14-8 Dissipation factor in ideal and practical
500 kV SCFF ± 1500 N/A
dielectric insulations.
14-8
cables such as XLPE, because it gives the average condi- standard for making tan δ measurements, although the
tion or losses in a length of cable and does not detect more modern transformer ratio-arm bridge is com-
localized imperfections. Partial discharge measurements monly used in high-voltage laboratories because it is
are generally considered to give a better indication of easier to balance and is less susceptible to certain mea-
insulation defects for extruded-dielectric cables. surement errors.
Two similar bridge-type circuits used to measure dissi- 14.2.5 Factory Partial-discharge Measurements
pation factor are shown in Figures 14-9 and 14-10. The A second, frequently used, cable insulation measure-
first circuit, called a Schering bridge, was developed ment is called a partial-discharge or PD measurement.
approximately 80 years ago to measure cable dielectric This measurement was devised to detect the current
losses. The second circuit is called a transformer ratio pulses in picocoulombs (pC) that are generated when
arm bridge. Both of these circuits require a high-voltage, there is a partial breakdown or discharge in high-volt-
standard capacitor (see Figures 14-9 and 14-10) with a age insulation. It is important to detect the presence of
voltage rating equal to the highest ac test voltage that is these discharges because they cause deterioration of the
specified for tan δ measurements. This very stable stan- insulation in the location of the void, which can cause a
dard capacitor has a loss-free, gas dielectric. cable failure.
In both circuits, the phase angle of the quadrature cur-

rent from the standard capacitor is shifted by a measur-
able amount until it has the same phase angle with
respect to the test voltage as that flowing through the
cable insulation. The Schering bridge is the recognized
Figure 14-9 Schematic diagram of Schering bridge. Figure 14-10 Schematic diagram of transformer ratio-
arm bridge.
14-9
Partial-discharge Measurement Circuits twice the power frequency (i.e., once each for the posi-
Figure 14-11 shows an extruded-dielectric cable with a tive and negative polarity of the sine wave), if the
void in the high-voltage insulation and an equivalent applied voltage is high enough. When the void breaks
circuit used to describe the partial-discharge phenome- down, it effectively shorts out the small capacitance, C1,
non (CIGRE 2006). that is in series with the solid cable insulation. This par-
tial discharge causes the following results.
If a gas-filled void occurs in the high-voltage insulation,
there will be electrical breakdowns across the void at
Figure 14-11 Partial discharge in cable insulation void.
Figure 14-12 Schematic diagrams for straight-type, partial-discharge detectors.
14-10
1. The insulation capacitance increases in the area of except that it detects the current pulses in a discharge-
the partial discharge because part of the dielectric is free coupling capacitor. This second circuit is used when
shorted out. it is inconvenient or impossible to isolate the cable
2. A pulse of current flows into the faulty section of sheath from ground.
insulation to charge the increased capacitance.
In both of these circuits, it is difficult to relate the mag-
3. Traveling-wave current pulses travel in opposite direc-
nitude of the voltage pulses across the detection imped-
tions to the ends of the cable.
ance directly to the magnitude of the partial discharges
4. There is a small step decrease in the voltage across occurring in the cable insulation. For this reason, a low-
the insulation because of charge redistribution until voltage calibration signal is connected across the cable
there is enough time for current to flow into the cable sample terminals prior to each measurement. The cali-
from the external voltage source. brator injects pulses with known charge into the cable,
and the gain of an amplifier in the partial-discharge
All of the above changes take place in a very short detector is adjusted until the meter on the instrument
period of time, less than a microsecond. The reference indicates the magnitude of the calibration pulse. The
An Introduction to High-Voltage Experimental Technique primary drawback of these straight detector circuits is
gives an excellent description of the partial-discharge that they are very susceptible to interference signals
process and relates the measured or apparent discharge from many sources.
to the discharge at the insulation voids (Kind 1978).
Test experience (EPRI 1975) has shown that partial- Some of the methods used to reduce the effects of the
discharge measurements give a better indication of interference signals are:
incipient faults in extruded-dielectric cables than dissi-
pation-factor measurements. This is because the partial- • A partial-discharge filter may be used to block
discharge measurements detect local imperfections, such corona signals that originate in the voltage supply or
as voids, which are the most common failure mechanism the high-voltage connection.
in extruded-dielectric cable. Dissipation-factor measure- • The measurements can be made in a shielded room
ments, on the other hand, give an indication of the aver- or laboratory to reduce radiated signals such as
age condition of the insulation. Unfortunately, partial- radio-frequency transmissions.
discharge measurements are difficult to make because of • The center frequency of the detector amplifier can be
the minute amount of electrical charge in the individual changed to a frequency where there is less noise from
discharges as well as the electrical interference that external sources.
occurs from many extraneous sources.
Even using these methods to reduce noise levels, it is
Field measurements are difficult because of the interfer- often difficult to measure discharges with a sensitivity of
ence present in outdoor substation environments. less than 5 pC, which is required by test standards for
Partial-discharge measurements with the required sensi- extruded-dielectric cables. This difficulty led to the
tivity level (typically 5 pC or less) are also difficult in a development of the balanced-bridge, or differential par-
factory environment because of interference caused by tial-discharge, detector shown schematically in
manufacturing equipment. Figure 14-13. This type of circuit reduces the effects of
interference signals by using two test samples to cancel
Many types of partial-discharge measurement circuits out common-mode noise. If an external noise source is
have been devised (Tanaka 1993; Nattrass 1988); how-
ever, two types of circuits, called straight detectors and
balanced-bridge detectors, are commonly used for labo-
ratory or factory production tests. Schematic diagrams
of two variations of the straight detector circuits are
shown in Figure 14-12. The circuit shown in Schematic
A detects the presence of partial discharges in the cable
insulation by detecting the high-frequency current
pulses that flow through the cable sample to ground.
The current pulses cause small oscillating voltages (frac-
tions of a millivolt) to appear across the detection
impedance. These voltage pulses are amplified by a
frequency-selective amplifier in the discharge detector.
The circuit in Schematic B works on a similar principle, Figure 14-13 Differential partial-discharge detector.
14-11
present, and the two high-voltage samples are of similar On-line vs. Off-line PD Measurements
capacitance, similar noise-voltage signals will appear at PD measurements can be performed when the cables are
points A and B on either side of the bridge. The input energized from the power system (on-line measure-
stage of the discharge detector is a differential amplifier, ments) or when they are energized by a separate voltage
which removes common-mode noise signals. If dis- source (off-line). On-line PD measurements are the sim-
charges occur in one of the two cables, there is no simul- plest and lower-cost method of making PD measure-
taneous canceling signal at the opposite side of the ments, because there is less setup (preparation) time
bridge; therefore, there is an amplified output signal required to perform the measurements and no addi-
from the differential amplifier in the discharge detector tional equipment to energize the cable. The primary
indicating that discharges are present. In practical situa- advantage of off-line PD measurements is that the cable
tions, this type of bridge yields a noise-rejection ratio of and accessories can be stressed with a higher-than-rated
somewhere from ten to one up to as much as several voltage, which, in general, makes it possible to detect
hundred to one. smaller defects compared to the on-line measurements.
A number of electronic signal-enhancing techniques A number of different types of voltage sources have
have been applied to both types of discharge detectors been used to perform off-line PD measurements. The
in an effort to improve the noise-rejection performance different types of excitation voltage sources that have
of the instruments. Some of these enhancement tech- been used for commissioning tests are:
niques that have been employed in commercially avail- 1. Power Frequency Series Resonant Test Equipment. In
able, partial-discharge detectors are signal-averaging, this case, an adjustable high-voltage reactor is placed
digitally activated shutters that block noise signals syn- in series with one phase of the cable system. The
chronous with the power frequency, and circuits that series inductor is then adjusted so that the series L-C
distinguish the direction of current pulses. circuit resonates at power frequency (i.e., 50 or
60 Hz). In this case, the reactive vars to energize the
Even if noise-rejection problems are solved, the inter- cable are supplied by the high-voltage reactor rather
pretation of partial-discharge measurements is not than the power supply to the test equipment. The
always straightforward. First, it is difficult to relate the input power to the test equipment is just the real
magnitude of the measured charge pulses and the power losses in the high-voltage reactor and the cable
amount of potential damage to the cable insulation. For capacitance. For commercially available series reso-
example, discharges of several hundred picocoulombs nant test equipment and XLPE cables, the input volt-
can occur under certain circumstances in the insulation amps (VA) to the test equipment is typically less than
semiconducting layer, producing minor, short-term con- 2% of the VA required to energize the cable system.
sequences. Failure may occur in a short period of time if Additional details of this test equipment are con-
the same level of discharges is present in the insulation tained in Section 14.2.1 (“AC Voltage Withstand and
wall. Oscilloscopes are typically used with discharge Step-Breakdown Tests”).
detectors to display the phase relationship of the current
2. Variable Frequency Resonant (VFR) Test Equipment.
pulses in relation to the applied voltage. This helps to
In this case, a high-voltage reactor is placed in series
determine the source and location of the discharges
with one phase of the cable system, and the frequency
(Nattrass 1988).
of the voltage supplied to the series L-C circuit is
adjusted until the capacitance of the cable resonates
14.2.6 Field Partial-discharge Measurements
with the series reactor (see Figure 14-14). The reso-
During the past decade a number of significant develop- nant frequency, given by Equation 14-1, is typically
ments have occurred in the equipment and methods for in the range of 30 to 300 Hz. The variable frequency
performing partial-discharge measurements in the field.
Field PD measurements are frequently used to perform
measurements on distribution cables, and to a lesser
extent, on transmission cables. The IEEE ICC Cable
Systems Subcommittee has developed a guide (Std.
400.3) to assist users in understanding the many com-
mercial equipment and measurement services currently
available (IEEE 2001). The following different types of
measurements are currently being used. Some of the
more significant factors addressed in this guide are sum- Figure 14-14 Schematic of variable frequency resonant
marized in the following sections. test equipment.
14-12
voltage source is supplied by power electronics equip- This circuit requires the same test equipment as for
ment. Figure 14-15 shows VFR test equipment for high-pot tests, with the exception of the inductor, L1,
conducting commissioning tests on a 400-kV XLPE and sphere gap. The cable is charged with dc voltage
cable system. Example 14-1 illustrates the fundamen- until a breakdown of the sphere gap occurs. The
tal calculations for VFR after-installation tests for a series cable and inductor circuit produce an oscilla-
230-kV XLPE cable system. tory test voltage with a frequency (Equation 14-2) of
approximately 1 to 10 kHz. The OW tests have pri-
1
f = marily been used for distribution class cables due to
2π L C 14-1 test equipment limitations, but testing of transmis-
Where: sion cables up to 150 kV have been performed in the
f = frequency of resonance, Hz. Netherlands (Gulski 2003).
L = inductance of series reactor, Henrys. 1
C = capacitance of cable between high voltage f =
conductor and sheath, Farads. 2 π L1 C 2 14-2
3. Very Low Frequency (VLF) Test Equipment. VLF test Where:

equipment applies a 0.1 Hz test voltage to the cable f = frequency of resonance, Hz.
system. The primary advantage of this type of test L1 = inductance of series air core reactor,
equipment is that the reactive MVA required to ener- Henrys.
gize the cable system is much lower at 0.1 Hz com- C2 = capacitance of cable between high voltage
pared to power frequency. Although VLF testing is conductor and sheath, Farads.
relatively common for distribution cables, it has been Example 14-1—Variable-frequency Resonant Test
used on a limited basis for lower-voltage transmission Requirements
cables due to the maximum-voltage test equipment The VFR test equipment shown in Figure 14-15 uses a
that is currently available. 250-kV high-voltage reactor with a 1-hour current rat-
4. Oscillatory Wave (OW) Test System. A damped oscil- ing of 83 A. The inductance of the reactor is 16.2 H.
latory test voltage is produced with the test circuit Determine if it is capable of testing a 5-mile (8-km)-long
shown schematically in Figure 14-16 (Gulski 2003). 230-kV XLPE transmission line for 1 hour at 140% of
rated voltage. The 230-kV XLPE transmission cable
parameters are given in Example 11-1 of Chapter 11
(Ampacity).
The cable capacitance per unit of length is given by

Equation 14-3:
16.95 ε 10 −12 16.95 ∗ 2.3

C′ = =
⎛ Di ⎞ ⎛ 3.464 ⎞
ln ⎜⎜ ⎟⎟ ln ⎜ ⎟
⎝ Dcs ⎠ ⎝ 1.652 ⎠
Figure 14-15 Variable frequency resonant test equipment
(courtesy HV Technologies, Inc.). = 52.65 × 10−12 F / ft (172.7 F / m) 14-3
The total capacitance of the 5-mile (8-km) line is:

ft
C = 52.65 × 10 −12 F × 5 mi × 5280
mi
= 1.39 × 10 −6 F 14-4
The resonant frequency of the high-voltage reactor and

the cable capacitance is:
1
f = = 33.54 Hz
2 π 16.2 H × 1.39 × 10 −6 F
Figure 14-16 Schematic diagram of oscillatory wave
test system. 14-5
14-13
The cable charging current at 140% of rated voltage and Termination discharges can be identified by time-
33.54 Hz is: domain location methods and by means of special
probes, although some instruments have been developed
I c = 1.4 × El − g ω C = to isolate these discharges by means of very-high-
frequency amplifiers/filters. These would detect signals
emanating only from within short distances, since cables
230 kV
= 1.4 × × 2 × π × 33.54 × 1.39 × 10 −6 rapidly attenuate very-high-frequency signals.
3 Frequency-domain PD Measurements
Frequency-domain testing is based on measuring the
= 54.5 A 14-6
frequency components of the PD pulses and can be per-
formed either off-line or on-line. The frequency spec-
The above calculations indicate that the VFR test equip-
trum of the PD pulses can be obtained either by
ment will be able to provide the required test voltage
computation (e.g., Fast Fourier Transform method) or
(186 kV) without exceeding the 1-hour maximum cur-
through the use of analog swept filters as in a spectrum
rent rating of the high-voltage reactor.
analyzer. For both methods, the measurement result
depends on the setting of the measuring device used
The test equipment will supply 10.1 MVAR (186 kV *
(parameters such as sample rate in the case of digital
54.5 A) to the cable during the test. If the quality factor,
conversion, filter bandwidth and sweep rate in the case
Q, of the high-voltage reactor is 80, then the three-phase
of analog conversion).
480-V supply to the test equipment must be capable of
supplying 126 kVAR (10.1 x 106/80).
In general, the detected frequency components of a PD
Time-domain vs. Frequency-domain PD signal are in the range of a few hundred kHz up to
Measurements 1 GHz, depending on the location of the PD with
Two significantly different methods have been used to per- respect to the measuring point. However, a frequency
form field PD measurements. These are the time domain range of a few hundred kHz to 300 MHz is more practi-
and the frequency domain measurement approaches. cal due to limitations imposed by the frequency band-
width of the PD detection sensors. Frequency-domain
Time-domain PD Measurements
testing normally is conducted in both full- and zero-
Time-domain PD tests are usually performed off-line;
span modes. In the full-span mode, the frequency range
although on-line measurements with multiple sensors
scanned can be adjusted to examine signals in narrow-
are also possible. Individual pulses can be measured if a
frequency bands as well as wide-frequency bands. Mea-
wide bandwidth detector of at least several tens of mega-
surements made using narrow-frequency bands signifi-
hertz is used. Lower-bandwidth detectors may result in
cantly enhance the resolution in locating the PD sites.
the superposition of pulses. When using low-bandwidth
PD energy vs. frequency is a strong function of distance
detectors, for example, as per IEC 60270, the superposi-
from the PD source. The zero-span mode is used to
tion of pulses can be checked by using a double-pulse
examine PD pulses in a time domain as measured in a
calibrator with an adjustable-pulse sequence from 1 μs
relatively narrow bandwidth at a center frequency,
to 100 μs according to IEC 60885-3. The phase posi-
which provides good detection sensitivity (i.e., within a
tions, magnitudes, and repetition rates of the pulses can
frequency range where little interference occurs). The
be measured.
zero-span mode is used to find PD pulses occurring at
one or more cycles of the operating voltage of a specific
A common method of estimating the location of a PD
frequency. The signal from the zero-span mode is often
in a cable uses the principle of reflectometry, as illus-
coupled to a pulse-phase analyzer. The pulse-phase ana-
trated in Figure 14-17. The excitation and measurement
lyzer is capable of recording PD pulses sorted by their
are implemented from one cable end (near end). Each
partial-discharge creates current pulses that travel in
opposite directions. The direct signal traveling toward
the near end is recorded first as pulse 1. The signal trav-
eling in the opposite direction is reflected at the remote
end and travels back to the near end, where it is
recorded as pulse 2. Using the difference, Δt, between
the arrival times of these two pulses and the velocity of
pulse propagation, the PD location can be estimated.
Figure 14-17 Time-domain reflectometry method of
PD location.
14-14
phase angle and magnitude relative to the power fre-

quency excitation.
The frequency-domain technique does not provide a

direct basis for calibration of the resulting data in terms
of apparent charge (pC). PD sources are recognized on
the basis of their spectral characteristics and through
the use of the zero-span mode of the spectrum analyzer
to provide a correlation between the signal in the
selected spectral region and power frequency. The rela-
tionship between the measured signal and the “severity”
of PD source is based on experience and information
from laboratory testing.
PD location is estimated by measuring the energy vs.

frequency of the PD pulses. The accuracy of the loca-
tion depends on two factors: Figure 14-18 High-frequency current transformers for PD
measurements (courtesy author).
• The frequency range of the pickup sensors
• The distance between two consecutive test points
PD location is often judged from the frequency content

o f t h e P D - i n d u c e d s i g n a l i n c o m b i n at i o n wi t h
knowledge of the cable type (i.e., the high-frequency
attenuation of the cable). If the PD-induced signal is suf-
ficiently large, the location can be determined using
time-domain methods.
Cable System vs. Cable Accessory PD
Measurements
Most field PD measurements detect the partial dis-
charges in splices or terminations where the instrumen-
tation is attached to the cable system (i.e., splice vaults or
cable termination structures). In this case, one of the fol-
lowing types of sensors are connected to the cable system
where the PD measurements are being performed.
• High-frequency current transformers are connected
to the cable system bonding leads (Figure 14-18). Figure 14-19 Temporary capacitive electrodes for PD
measurements (courtesy author).
• Temporary capacitively-coupled electrodes are
attached to the outside of the cable (Figure 14-19).
• Factory-installed and calibrated capacitive taps are
installed in the cable splices and terminations
(Figure 14-20).
Some commercial field PD testing services have devel-

oped equipment to perform field PD measurements of
complete cable systems using high-voltage coupling
capacitors in series with detection impedances that are
attached to the cable system at one or both ends of the
cable circuit. In this case, maximum circuit length is typ-
ically limited to several miles because of attenuation of
the partial-discharge pulses as they travel to the ends of
the cable circuit. Figure 14-20 400-kV XLPE joint with integral PD sensor
(Kaumanns 2003).
14-15
14.3 TYPE/QUALIFICATION TESTS AND nation breakdown or flashover, a special test termina-
FACTORY PRODUCTION TESTS tion is used to terminate the cable. The advantage is that
The most commonly used qualification tests for trans- the cable breakdown voltage indicates how much mar-
mission cables in North America are those written by gin there is in the cable design; the disadvantage is that
the AEIC and ICEA. The standards used for cable ter- there is no indication of how well commercial termina-
mination and joint qualification tests are those written tions and joints will work with the cable being tested.
by the Insulated Conductors Committee of the IEEE.
IEC testing standards for transmission cable systems are Production tests are tests that are performed on every
commonly used outside of North America; however, a manufacturing order before the material is shipped to
number of countries have developed national standards the purchaser. These tests vary with cable types but gen-
(i.e., JEA [Japan Electric Association], VDE [Associa- erally include conductor resistance, ac high-voltage, and
tion of German Electric Engineers], etc.). ionization or partial-discharge measurements. Qualifi-
cation, sample, and production tests for the different
North American qualification tests evolved as separate cable types are summarized in the following sections.
tests for cables, terminations, and joints; in other Only sample and production tests are described for
words, components are tested rather than systems. SCFF cables, because AEIC specifications for this cable
However, the recently released AEIC CS9-06 addresses type currently do not include qualification tests.
the complete cable system. International qualification
or type tests, such as those provided by the IEC, are Submarine cable electrical and mechanical tests are
typically written for complete cable systems. There are described in Chapter 7 (DC and Submarine Cables)
obviously advantages and disadvantages associated
with these two different approaches. 14.3.1 Qualification/Type Tests for Extruded-
Dielectric Cables (ICEA S-108-720-2004)
The primary difference between the two is in breakdown The testing requirements for AEIC CS9-93 and ICEA
tests. To ensure a cable breakdown rather than a termi- S-108-720-2004 are summarized in this section for
extruded-dielectric cables for the following reasons.
ICEA S-108-720-2004 covers XLPE and EPR transmis-

sion cable systems up to 345 and 138 kV, respectively.
Qualification tests for extruded-dielectric cables (ICEA

S-108-720) differ significantly from those for pipe-type
c able s ( A E I C C S2 ). T h e q u a l i fi c at i o n t e s t ( s e e
Figure 14-21) consists of electrical tests as preceded by
mechanical bending and load-cycle-precondition condi-
tioning procedures. Cable samples are conditioned
by cable bending and 20 thermal cycles. Electrical
tests include an ac voltage withstand test and a hot
impulse test with dissipation-factor (tan δ) and partial-
discharge measurements. Semiconducting shield (con-
ductor and insulation) volume resistivity-stability tests
are also required.
Load-Cycle Tests
Load-cycle tests are performed on two cable samples
that are subsequently subjected to the high-voltage-time
tests and hot impulse tests. The load-cycle tests are spec-
ified primarily because extruded-dielectric cables have a
much greater expansion than oil-paper insulation and
are therefore subject to significant mechanical stress
Figure 14-21 Flowchart for ICEA qualification tests during cyclic loading up to the cable emergency conduc-
of extruded-dielectric cables.
14-16
tor temperature. Each of the 24-hr load cycles must cable’s impulse-withstand capability after it has been
comply with the requirements in Table 14-3. subjected to normal aging conditions for several years.
The hot impulse test is not necessarily the most difficult
Ac-withstand Test part of the qualification test for extruded-dielectric
The ac-withstand test, or high-voltage-time test, is con- cables, because new extruded-dielectric cables tend to
ducted at room temperature with the cable energized at have a very high impulse breakdown level in relation to
2.5 times rated voltage for a duration of 2 hr. the cable BIL requirements.
Partial-discharge Test
Resistivity-stability Test
A partial-discharge test is performed on the cable sam-
Volume resistivity measurements are required on sam-
ple at room temperature after it has completed the load
ples of the conductor and insulation shields to demon-
cycle and ac-withstand test. The cable sample must have
strate that the resistivity does not increase significantly
less than 5 pC of partial discharge at test voltages up to
at high temperatures.
twice rated voltage.
Insulation Dissipation Factor Test 14.3.2 Production Tests for Extruded-Dielectric
Dissipation factor measurements are performed after Cables (AEIC CS9-06 and ICEA S-108-
completion of the partial-discharge tests at the cable 720-2004)
emergency operating temperature. The maximum allow- Shipping Length Tests
able dissipation factor values are 0.1% and 1.5% for Shipping tests for XLPE transmission cables consist of
XLPE and EPR insulations, respectively. electrical-withstand tests, shown in Table 14-4, and par-
Hot Impulse Test tial-discharge measurements. The full-reel, ac-withstand
Hot impulse tests for extruded-dielectric cables are tests may be conducted using the 15-, 30-, or 60-minute
included for the same reasons as they are for taped- test shown in this table.
insulation cables—i.e., to demonstrate that the cable will
withstand standard lightning impulses at the required The partial-discharge test requires that each shipping
BIL and maximum conductor temperature. The tests length of cable have a partial-discharge level of 5 pC or
are similar to those for pipe-type cables, except that the l e s s at vo l t ag e s u p t o t wi c e t h e r at e d vo l t ag e.
cable sample must withstand three negative impulses Figure 14-22 shows the equipment for performing par-
approximately 25% above the nominal BIL. The tial-discharge measurements in a shield room to attenu-
impulses are applied with the cable operating at its ate electrical noise from cable manufacturing equipment.
emergency operating temperature (105 to 130°C).
This requirement was included because test experience

indicates that there may be a significant decrease in the
Table 14-3 Heating Cycle Requirements

Requirement Condition
Duration of heating cycle 8 hr
Time at emergency temperature 2 hr
Duration of cooling cycle At least 16 hr
Test voltage Twice phase-to-ground
Table 14-4 Ac Voltage Withstand Tests for Extruded-

dielectric Transmission Cables (Phase-to-Ground, kV)
System 15-Minute 30-Minute
Voltage 60-Minute Ac Ac Voltage Ac Voltage
(kV) Voltage (kV) (kV) (kV)
69 80 120 100
115 135 200 160
138 160 240 200
161 185 280 230
230 265 N/A 330
Figure 14-22 PD tests on a 138-kV XLPE
345 400 N/A N/A
cable in shielded room (courtesy Southwire).
14-17
The dc conductor resistance of each reel of cable is mea- 62067 standards for extruded cable systems having rated
sured, and temperature corrections are applied to give voltage >150 up to 500 kV have common or similar test
the resistance per 1000 ft at 25°C. The dc resistance at procedure but with different requirements and test val-
25°C cannot exceed the nominal conductor resistance ues and the following major differences:
by more than 2%. • IEC 60840 includes requirements for testing the
complete cable systems (i.e., cable, terminations,
A jacket integrity test is performed on each reel of cable
and joints) as well as the cable alone and the acces-
to determine if any significant imperfections occurred
sory alone.
during manufacturing or handling of the cable. A dc test
voltage ranging from 10 kV to a maximum of 25 kV is • IEC 62067 includes requirements for testing the com-
applied to the jacket for 1 minute. The jacket integrity plete cable system. It also includes a requirement for
test is a pass/fail test. switching impulse tests for cable systems with rated
voltages above 230 kV.
A sample of cable approximately 1 ft (0.305 m) is cut • IEC 62067 includes a long-term prequalification test
f ro m b o t h e n d s o f e a c h s h i p p i n g re e l t o c h e c k for the cable system.
the dimensions of the various cable layers and to check • The IEC documents do not include a short-term
for moisture.
emergency temperature rating for extruded-dielectric
Sample Tests transmission cable and, therefore, do not include any
Table 14-5 summarizes some of the production sample test requirements for operation of the cable system at
tests required by AEIC CS9 specifications and ICEA S- conductor temperatures higher than 5 to 10°C above
108-720 for extruded-dielectric transmission cable. This the maximum continuous conductor temperature rat-
table also contains a brief explanation about the pur- ing (90°C).
pose of the tests.
Routine Tests
Routine tests are the tests made by the manufacturer on
14.3.3 Routine, Special, Type, and
each manufactured component (length of cable or pre-
Prequalification Tests for Extruded Cable
fabricated accessory) to check that the component
Systems (IEC 60840 and IEC 62067)
meets the specified requirements. Uo in the IEC specifi-
The IEC 60840 standards for extruded cable systems cations is defined as the rated line-to-ground voltage of
having rated voltage >30 up to 150 kV and the IEC the cable system.
Table 14-5 AEIC CS9 and ICEA S-108-720 Sample Tests on Extruded-dielectric Transmission Cables
Test Description Purpose of Test
Required at least each extrusion run
Hot creep Measure elongation and set of insulation at 150°C and Determine if insulation crosslinking is adequate
ambient temperature
Wafer boil Wafer of insulation and semiconducting shields boiled Determine if crosslinking of semiconducting material
in solvent is adequate
Min. required at least each 50 K ft
Dimensional stability Measure insulation shrink-back for a cable sample Determine that the cable insulation will not have
(not required by heated for 20 hr at 121°C ±2 C excessive shrink-back, which may cause problems at
ICEA S-108-720) accessories
Shrink-back Measure insulation shrink-back on five cable samples Additional test to quantify insulation shrink-back
that have been temperature-cycled between ambient
and 50°C
Conductor and Measure the volume resistivity of the semiconducting Check that the volume resistivity does not increase
insulation shield layers at 90C and emergency temperature above acceptable limits at high operating temperatures
volume resistivity
Require on each length (reel) or at least 10 K ft
Void and contamination Measure the size and number of voids, foreign material, Determine that insulation is free of voids, foreign mate-
determination and and shield protrusions from wafers of cable core rial, and semiconducting shield protrusions that would
shield protrusion lead to failure
Insulation eccentricity Measure maximum deviation in insulation thickness as Check centering of conductor in cable high-voltage
a percentage of maximum insulation thickness insulation
Cable core dimension Check that the minimum point and average thicknesses Check for compliance with specified values
checks of the cable shields, insulation, sheath, and jacket are
within specified tolerances
14-18
The following tests are required: The frequency of the tests in items 9 and 11 shall be
• Partial-discharge test. The partial-discharge test shall according to the agreed quality control procedures indi-
cated in the relevant standard.
be carried out in accordance with IEC 60885-3 for
cables, except that the sensitivity, as defined in IEC Type Tests
60885-3, is 10 pC or better. The test voltage is raised Type tests are tests made before supplying, on a general
gradually to and held at 1.75 Uo for 10s and then commercial basis, a type of cable system covered by the
slowly reduced to 1.5 Uo. No detectable discharge is IEC standard, in order to demonstrate satisfactory per-
allowed above the declared test sensitivity level from formance characteristics to meet the intended applica-
the test object at 1.5 Uo. tion. Once successfully completed, these tests do not
• Voltage test. The voltage test is made at ambient tem- need to be repeated, unless changes are made in the
perature using an alternating test voltage at power cable or accessory materials, or design or manufacturing
frequency. The test voltage is raised gradually to the process, which might change the performance character-
specified value and held for the specified time istics. The relevant IEC standards also specify the range
between the conductor and metallic screen/sheath of type approval for the cable and the accessories (joints
according to the tables given in relevant IEC stan- and terminations). This is necessary in order to clarify
dard for the specific class of voltage. the cases where the previous performed type tests shall
be considered applicable or shall be repeated.
• Cable jacket integrity test. A cable jacket integrity test
is required. The test procedure is specified in The type test sequence consists of electrical tests on the
IEC 60229. cable system and nonelectrical tests on cable compo-
Sample Tests nents.
These are tests made by the manufacturer on samples of Electrical-type Tests
completed cable or components taken from a completed The electrical-type tests are carried out on a cable sam-
cable, at a specified frequency, to verify that the finished ple at least 10 m long out of terminations. The test cir-
product meets the specified requirements. cuit, cable and accessories, are arranged according the
requirements of the relevant standard. The test sequence
The following tests are carried out on samples which, is the following:
for the tests in items 2 and 7, may be complete reel
lengths of cable, taken to represent batches: 1. Bending test followed by installation of the accesso-
ries and a partial-discharge test at ambient tempera-
1. Conductor examination for compliance with the ture.
cable manufacturer’s drawing of the cable (i.e., num-
2. Insulation Tan δ measurement.
ber of strands, etc.).
3. Heating cycle voltage test (20 cycles at 2Uo, temper-
2. Measurement of electrical resistance of conductor.
ature from 5 to 10°C above maximum normal oper-
3. Measurement of thickness of insulation and over- ating temperature).
sheath.
4. Partial-discharge tests (with a sensitivity as defined
4. Measurement of thickness of metallic sheath. in IEC 60885-3, 5 pC or better) at ambient tempera-
5. Measurement of diameters, if required. ture, and at high temperature. No detectable dis-
6. Hot set test for XLPE and EPR insulations. charge is allowed above the declared test sensitivity.
The tests are carried out after the final cycle of
7. Measurement of capacitance.
step 3 above or, alternatively, after the lightning
8. Measurement of density of PE insulation. impulse voltage test in step 6 below.
9. Lightning impulse voltage test followed by a power 5. Switching impulse voltage test (required for
frequency voltage test (for IEC 62067 only). Um ≥ 300 kV).
10. Water penetration test, if applicable. 6. Lightning impulse voltage test, followed by a power
11. Tests on components of cables with longitudinally- frequency voltage test.
applied metal foil. 7. Partial-discharge tests, if not previously carried out
in step 4 above.
The sample tests in items 1 to 8 shall be carried out on
8. Tests of outer protection for direct-buried joints.
one length from each batch of the same type and cross
section of cable, but shall be limited to not more than 9. Examination of the cable system with cable and
10% of the number of lengths in any contract. accessories shall be carried out after completion of
the tests above.
14-19
10. Resistivity of semiconducting screens (measured on complete test period of 8760 hr (1 year). The cycles of
a separate sample). heating (temperature from 0 to 5°C above the maxi-
mum normal operating temperature), and cooling is
Nonelectrical-type Tests
carried out at least 180 times.
The nonelectrical-type tests shall be carried out on sam-
ples of cable or components and are the following: 2. Lightning impulse voltage test on cable samples.
1. Check of cable construction. 3. Examination of the cable system after completion of
the tests above of the cable system.
2. Tests for determining the mechanical properties of
insulation before and after aging. The test arrangement must be representative of
3. Tests for determining the mechanical properties of the installation design conditions—e.g., rigidly fixed,
oversheaths before and after aging. flexible, and transition arrangements, underground and
4. Aging tests on pieces of complete cable to check in air. In particular, special attention is given to thermo-
compatibility of materials. mechanical aspects of accessories.
5. Loss of mass test on PVC oversheaths.
The prequalification test needs only to be carried out
6. Pressure test at high temperature on oversheaths. once unless there is a substantial change in the cable sys-
7. Tests on PVC oversheaths at low temperature. tem with respect to materials, manufacturing process,
8. Heat shock test for PVC oversheaths. design, and design stress levels. A “substantial change”
9. Ozone resistance test for EPR and HEPR insula- is defined as a change that might adversely affect the
tions. performance of the cable system.
10. Hot set test for EPR, HEPR, and XLPE insulations. 14.3.4 Qualification Tests for Pipe-type Cables
11. Measurement of density of PE insulation. (AEIC CS2-97)
12. Measurement of carbon black content of black PE The more salient parts of the AEIC qualification tests
oversheaths. and the objective for including each test as part of the
13. Shrinkage test for PE, HDPE, and XLPE insulations. qualification test series are summarized below. Current
14. Shrinkage test for PE oversheaths. specifications should be consulted for the exact test
requirements for each voltage class and conductor size.
15. Determination of hardness of HEPR insulation.
16. Determination of the elastic modulus of HEPR The qualification test series consists of mechanical
insulation. conditioning of the cable sample, dielectric measure-
17. Test under fire conditions. ments, an ac high-voltage withstand test, and a hot
18. Water penetration test. impulse test.
19. Tests on components of cables with a longitudinally Mechanical Conditioning
applied metal foil. The mechanical conditioning test is specified to demon-
strate that the cable has adequate electrical strength
The test requirements are given in the relevant IEC after it goes through the mechanical-bending process
60840 or 62067 standards. The test methods and proce- necessary to install the cable. This portion of the test
dures are given in the IEC 60811 standards series. consists of three reverse bends of the cable on a reel with
a diameter 25 times that of the cable core plus the con-
Prequalification Test ductor diameter.
The prequalification test is a test made before supplying,
on a general commercial basis, a type of cable system Following the mechanical conditioning, the cable sam-
covered by the IEC 62067 standard, in order to demon- ple is pulled into a pipe, terminated, and pressurized to
strate satisfactory long-term performance of the com- rated pressure (200 psig) before any of the following
plete cable system. tests are performed.
The prequalification test consists of electrical tests on Dissipation-Factor and Ionization-Factor
the complete cable system with approximately 100 m of Measurements
full-sized cable including accessories. The normal The dielectric measurements consist of ionization-factor
sequence of tests is: and dissipation-factor (tan δ) measurements at three
1. Heating cycle voltage test: a voltage of 1.7 Uo and different temperatures. The dissipation-factor measure-
heating cycles are applied to the assembly during the ments indicate if the cable insulation is relatively free of
contamination or other gross defects. The ionization-
14-20
factor measurement indicates if the cable is free from (5.9 kV/mm) for 69-kV to 100 V/mil (3.9 kV/mm) for
partial discharges at its normal operating voltage; it is 500-kV system voltage be applied to the cable for
actually the change in dissipation factor between two 15 minutes with the cable at atmospheric pressure and
different levels of voltage stress. A high ionization factor room temperature. The ac test voltages are approxi-
combined with a normal dissipation factor at the low mately one-half of those for other cable types, because
stress indicates that there is some form of ionization the cable cannot be pressurized on the reel at its normal
occurring in the cable insulation. operating pressure.
Ac-withstand Test The ionization factor, which must be less than 0.1% for
The ac-withstand test, also referred to as a high-voltage- HPFF and medium-viscosity HPGF cables, and 2.0%
time test, requires that the cable withstand approxi- for high-viscosity HPGF cables, is determined by the
mately 2.5 times the rated line-to-ground voltage for a difference in the cable dissipation factor at low- and
period of 24 hr for system voltages up to 345 kV. Cables high-voltage stress levels.
with system voltages greater than 345 kV are tested at
1.73 times the rated phase-to-ground voltage plus Sample Tests
100 kV. The cable is tested at rated dielectric-fluid pres- The sample tests required by AEIC specifications for
sure and at ambient temperature plus rise due to dielec- pipe-type cables consist of mechanical integrity, delami-
tric heating only. Test experience has shown that nation (for laminated paper-polypropylene, LPP, cable
localized manufacturing defects, such as conductor or only), bending, and electrical tests.
shield defects, may result in cable failure under the high-
voltage-time test. The objective of the mechanical integrity checks are to
determine that the cable was manufactured with the
Hot Impulse Test
proper insulation and shielding thicknesses, that the
The objective of the hot impulse test is to demonstrate
conductor is not out of round, and that there are no
that the cable will withstand standard lightning impulses
tears or excessive wrinkles in the insulating tapes. Dur-
at the required BIL while it is operating at maximum-
ing this inspection, the shield, insulation, and binder
rated conductor temperature. This is an important test
tapes are removed from samples of the cable and
for impregnated-paper-insulated cables, because it is
inspected for any abnormalities.
usually the most difficult part of the qualification test
for the cable to pass. The hot impulse test specifies that
The objective of the delamination test is to determine if
the cable withstand 10 positive and 10 negative lightning
the kraft paper and polypropylene layers of the insulat-
impulses while the conductor is maintained at a temper-
ing tapes are adequately bonded together. The insulat-
ature of 85ºC. The hot impulse test is passed if the cable
ing tapes are removed from an 18-in. (0.46 m) length of
withstands the applied impulses at rated BIL followed
cable, inspected for delamination, coiled in a roll, reap-
by an ac-withstand test for 15 minutes at 1.67 times
plied to the cable core, and inspected a second time for
rated voltage-to-ground plus 10 kV. The AEIC qualifi-
delamination after removal from the cable core.
cation test also specifies continuing the hot impulse tests
up to cable breakdown for information purposes only. The cold bend test is intended to demonstrate that the
The next higher voltage class of termination is often cable will not be damaged by bending during installa-
used to terminate the cable for these tests in an attempt tion even at very low ambient temperatures. Basically,
to reach cable breakdown before a termination flash- this test consists of taking approximately 15 ft (5 m) of
over or breakdown occurs. cable, subjecting it to a temperature of minus 10°C for a
period of at least 2 hr and bending the cable four times
14.3.5 Production Tests for Pipe-type Cables with reverse bends around a cylinder whose diameter is
(AEIC CS2-97) 12 to 20 times the diameter of the cable, depending on
Shipping-length Tests the insulation thickness. The cable sample is then dis-
Shipping-length tests for pipe-type cable consist of a sected to determine if there has been any damage to the
conductor-resistance measurement, an ionization test, cable that would affect its electrical performance or
and an ac high-voltage test. The conductor-resistance cause cable installation problems.
measurement is made to determine that the resistance of
the cable conductor, corrected to 25°C, does not exceed The electrical tests consist of dissipation-factor tests and
the limits in the AEIC specification. ionization tests with a sample of the cable pressurized at
normal working pressure and a high-voltage-time test
The full-reel, ac high-voltage tests requires that a volt- with the cable sample operating substantially below its
age corresponding to a stress ranging from 150 V/mil rated gas or fluid pressure (100 psig). The cable’s dissi-
14-21
pation factor must be below limits specified for the volt- tion of the specified ac withstand voltage for 1 minute
age class of the cable and insulation type (paper or on shipping sections.
LPP), at temperatures of 25, 80, and 90°C. The cable
sample must also have an ionization factor of no more 14.3.7 Tests for Cable Joints
than 0.1% for cables rated at 161 kV and below. The IEEE Standard 404, “IEEE Standard for Extruded and
ionization factor must be no more than 0.05% for cables Laminated Dielectric Shielded Cable Joints Rated
rated above 161 kV. 2500 V–500000 V” (IEEE 2000) is a comprehensive doc-
ument that establishes electrical ratings and test require-
The high-voltage-time test is performed with HPFF and ments of cable joints used with extruded and laminated
medium-viscosity HPGF cables pressurized at 100 psig dielectric shielded cables. IEEE 404 includes short-time,
(700 kPa) or at 150 psig (1050 kPa) for high-viscosity high-current tests, joint shield conductivity tests, and
HPGF cables. Cable samples must withstand an ac test connector thermal and mechanical tests in addition to
voltage corresponding to an average voltage stress of high-voltage dielectric tests. IEC requirements for trans-
430 V/mil (16.93 kV/mm) for 6 hr to pass the test. A sub- mission cable joints are included in the same standards
sequent 3-hr test is also specified with the cable energized as for the same type of transmission cable.
at an average stress of 520 V/mil (20.47 kV/mm), but this
part of the test is for informational purposes only. The electrical test requirements for joints are generally
similar to test requirements for the cable that they are
14.3.6 Qualification Tests for Compressed-Gas- designed to join. The test levels or number of impulse
Insulated Cables repetitions vary between joint and cable test standards,
The only domestic standard applicable to this type of primarily because the standards were written indepen-
cable is IEEE/ANSI document C37.122, “Gas Insulated dently. The major difference in the test requirements for
Substation—Standards, Recommended Practices, and joints compared to those for cable is that joints must be
Guides,” which was prepared by the Substation Com- tested for connector problems or problems due to
mittee of the IEEE (IEEE/ANSI 1993). Since this docu- mechanical shifting of components during thermal
ment pertains to substation switchgear as well as gas- expansion and contraction. These potential problem
insulated bus (or cable), some of the tests do not apply areas led to the inclusion of high-voltage, cyclic-load
to cables. The standards section of this document con- tests for all joint types and connector thermal and
tains what are referred to as “design tests” and “produc- mechanical tests.
tion tests.”
Tables 14-6 and 14-7 summarize the test requirements
The design tests that are applicable to cables are: specified by IEEE 404 for extruded-dielectric and lami-
• Tests to demonstrate that the cable will carry nated-dielectric cable joints.
rated current without the components exceeding
The IEEE 404 design tests for extruded-dielectric cable
specified temperatures.
joints are considerably more complex than those speci-
• Dielectric-withstand tests to demonstrate the power fied for the laminated-dielectric cable joints. The design
frequency, impulse, and switching-surge dielectric- tests include requirements for multiple samples, wet-and
withstand levels of the cable. Partial-discharge tests dry-load cycle tests, and short-time current tests. Chap-
are also required to demonstrate that the insulation ter 8 provides additional information on testing require-
system is free of partial discharges at 150% of the ments for cable joints and terminations.
maximum rated ac voltage.
• Short-circuit tests to demonstrate that the device will
not be damaged by specified short-circuit currents. Table 14-6 Summary of IEEE Std. 404 Cable Joint
Test Requirements for Extruded-dielectric Transmission
• Insulator tests to demonstrate that the enclosure will Cable Joints
operate satisfactorily during thermal cycling. AC Voltage AC HV
• Pressure tests to demonstrate that the enclosure will Withstand Time Test DC 15 min
withstand gas pressures up 1.5 times maximum oper- System BIL Time Voltage Time Voltage Withstand
Voltage (kV) (min) (kV) (h) (kV) (kV)
ating pressures.
138 650 5 200 6 200 325
161 750 5 230 6 232 375
The primary production or routine test that applies to
230 1050 5 330 6 332 525
compressed-gas-insulated cables consists of the applica-
345 1300 5 500 6 500 650
500 1550 5 720 6 725 775
14-22
Design Tests Production Tests

Extruded dielectric joints must complete the following All premolded transmission cable joints must pass the
sequence of design tests to be in compliance with following production tests.
IEEE 404. 1. Partial-discharge. Less than 5 pC at 150% of rated
1. Partial-discharge voltage level phase-to-ground voltage.
2. Ac-withstand voltage 2. Ac-withstand or full-wave impulse-withstand volt-
3. Dc-withstand voltage age. At voltages shown in Table 14-6 or 14-7.
4. Impulse-withstand voltage at 25°C
14.3.8 Tests for Cable Terminations
5. Impulse-withstand voltage at emergency tempera-
Test requirements for cable terminations are more com-
ture
plex than for any other component of a cable system. It
6. Partial-discharge voltage level is necessary to demonstrate that a termination meets the
7. Cyclic aging in-air, 30 cycles requirements of open-air substation equipment or riser
8. Cyclic aging in-water, 30 cycles termination and performs its cable-sealing and voltage-
grading functions. Additionally, cable terminations have
9. Partial-discharge voltage level
a wider variety of constructions than the cables that
10. High-voltage time they terminate. As a result, cable-termination test
11. Impulse-withstand voltage at 25°C requirements combine cable tests with termination insu-
12. Partial-discharge voltage level lator-air insulation tests.
13. Sectionalizer test (if applicable)
Cable-termination test requirements differ from cable or
14. Shield restoration: two short-time current tests joint tests in the following ways:
(if applicable)
• Adjustments must be made to some of the electrical
15. Jacket restoration: cyclic aging in-water, 10 cycles
test values to compensate for variations in air temper-
16. Short-time current ature, barometric pressure, and humidity.
17. Ac withstand voltage • Some of the tests must be performed with water spray
18. Shielding to simulate rain.
19. Connector thermal and mechanical • Radio influence voltage (RIV) is a design consider-
ation.
IEEE 404 requires the following sequence of design tests • Some electrical breakdowns on the outside of the ter-
for laminated-dielectric cable joints.
mination are allowed.
1. Ac-withstand voltage
2. Dc-withstand voltage The primary standard for testing transmission cable ter-
minations in North America is IEEE Standard 48-1996,
3. Impulse-withstand voltage at emergency tempera-
“IEEE Standard Test Procedures and Requirements for
ture
Alternating-Current Cable Terminations 2.5 kV Through
4. Ionization 765 kV” (IEEE 1996). The standard covers test require-
5. Cyclic aging in-air ments for extruded-dielectric terminations up to 230 kV
and laminar-dielectric terminations up to 765-kV system
voltage. Tables 14-8 and 14-9 summarize the test require-
ments specified by IEEE 48 for extruded-dielectric and
Table 14-7 Summary of IEEE Std. 404 Cable Joint laminar-dielectric transmission cable terminations.
Test Requirements for Laminar-dielectric Transmission
Cable Joints Design Tests
AC Voltage Withstand Extruded- and laminar-dielectric cable terminations
System BIL DC 15 min
Voltage (kV) Time (h) Voltage (kV) Withstand (kV) must pass the following sequence of tests to comply with
138 650 24 200 325 IEEE 48 design test requirements.
161 750 24 230 375 1. Power frequency voltage dry withstand test—
230 1050 24 330 525 1-min duration
345 1300 24 500 650 2. Power frequency voltage wet withstand test—
500 1550 24 720 775 10-sec duration
14-23
3. Power frequency voltage dry withstand test— 14.4 AFTER-INSTALLATION TESTS

6-hr duration After-installation or commissioning tests have been a
4. Partial-discharge (extruded-dielectric) or ionization part of cable industry standards for many years. These
factor test (laminated-dielectric) tests are included to determine if there has been any
5. RIV test damage or defect created in a cable circuit as a result of
installing, terminating, and splicing the cable. It is
6. Lightning impulse withstand test
assumed that routine factory tests have detected and
7. Switching impulse wet withstand test (EHV voltages eliminated any manufacturing defects.
only)
8. Direct voltage dry withstand test—15-min duration The ideal after-installation test would reveal defects in
9. Cyclic aging test the cable system but not cause damage to the cable insu-
lation. Unfortunately, the severity of the after-installa-
Pressurized laminar-dielectric cable terminations must tion test requires a trade-off between these two
pass a pressure leak test in addition to the above dielec- requirements. With the possible exception of the
tric tests. impulse test, it is generally acknowledged that the most
suitable after-installation tests would be a power fre-
Production Tests quency voltage somewhat higher than the normal oper-
All cable terminations must pass an ac voltage-withstand ating voltage. In fact, the SCFF cable test specification
test for 1 min under dry conditions. Tables 14-8 and 14-9 includes ac test voltages as the primary or preferred
show the required ac withstand voltage test requirements. acceptance test. Until recently, it was not economically
feasible to test transmission cable circuits with any mag-
Extruded-dielectric cable terminations with premolded nitude of power frequency voltage other than rated volt-
stress cones must pass the partial-discharge require- age. This was because of the large size and weight of ac
ments shown in Table 14-9. field test equipment that could supply the high charg-
ing-current requirements of most transmission cables.
Pressurized laminar-dielectric cable terminations must A practical compromise to this problem has been the
pass a pressure leak test in addition to the ac voltage use of dc voltage rather than ac voltage for the after-
withstand test.
Table 14-8 Summary of IEEE Std. 48 Extruded-dielectric Cable Termination Test Requirements
Ac Voltage Withstand Cyclic Aging Radio PD Extinction Lightning 15 min Dc
System 1 min Dry 10 s Wet 6 hr Dry Dry Influence Dry Voltage Impulse Dry Voltage
Voltage (kV) (kV) (kV) (kV) (μV) (kV) (kV) Withstand (kV)
69 175 145 120 80 300 60 350 245
115 205 190 160 133 400 80 450 275
138 310 275 210 160 500 120 650 355
161 365 315 250 186 500 140 750 395
230 460 445 330 265 -- 200 1050 510
Table 14-9 IEEE Std. 48 Laminar-dielectric Cable Termination Test Requirements

Ac Voltage Withstand Switching
Cyclic Aging Radio Lightning Impulse Wet 15 min Dc
System 1 min Dry 10 s Wet 6 hr Dry Dry Influence Dry Impulse Dry (Dry) Voltage
Voltage (kV) (kV) (kV) (kV) (μV) (kV) (kV) Withstand (kV)
138 310 275 210 160 500 650 -- 355
161 365 315 250 186 500 750 -- 395
230 390 380 320 265 500 900 -- 450
230 460 445 320 265 500 1050 -- 510
345 520 -- 440 300 500 1175 (900) 555
345 575 -- 440 300 500 1300 825 600
345 575 -- 440 300 500 1300 900 600
500 575 -- 440 435 500 1300 (1100) 600
500 690 -- 440 435 500 1550 1050 700
500 750 -- 575 435 500 1675 1110 745
500 750 -- 575 435 500 1675 1175 745
14-24
installation test. If the dc test voltage is raised slowly to • AEIC CS9-06 (69 to 345 kV). This EPR and XLPE
the desired test level, a portable dc test set with a current cable specification does not provide any specific after-
rating of 20 to 100 milliamperes can be used to energize installation tests. Rather, the recommendation is for
a cable up to several times its rated voltage. The cable the purchaser and manufacturer to decide upon a
conduction, or leakage, current varies significantly with mutually acceptable after-installation test.
temperature but is fractions of a milliampere per thou- • ICEA S-108-720-2004 (69 to 345 kV). The standard
sand feet (300 m) at 25°C. Unfortunately, dc Hi-Pot
for XLPE and EPR transmission cables recommends
tests are not very effective in detecting many installation
one of two ac voltage withstand tests. The first option
defects, and there is evidence that high dc voltage can
is to perform a 1-hr ac withstand test of 1.4 to 1.7
damage extruded-dielectric cable systems under some
times rated phase-to-ground voltage. The specified
circumstances. The high dc voltage injects an electric
frequency of the ac test voltage must be in the range
charge into the extruded-dielectric insulation, which is
from 20 to 300 Hz, which includes the most common
trapped in the insulation for some period of time after
range for variable-frequency series resonance tests.
the dc voltage is removed. This space charge will
increase local electrical stress concentrations if the cable The second option is to energize the cable system for
is energized with ac voltage before the trapped space at least 24-hr. This second option is normally called a
charge dissipates. 24-hr rated voltage “soak test.”
• IEC 62067 (161 to 500 kV). This standard for XLPE
During the past five to ten years, significant advances transmission cables provides three alternatives for the
have been achieved in the commercialization of ac field after-installation ac withstand test. The first option is
test equipment for transmission cables. The commercial to perform a 1-hr test with an applied voltage of 1.7
development of high-voltage, transportable variable- times nominal phase-to-ground voltage.
frequency-resonant test equipment has changed commis- The second option is to perform a 1-hr test with an
sioning test practices in Europe and other parts of the applied voltage shown in Table 14-10.
world and is starting to be utilized in North America.
The third option is to perform a 24-hr rated voltage
“soak test.”
After-installation or commissioning test practices vary
significantly around the world. The following sections
The most common practice in North America has been
summarize commissioning test practices for the differ-
to perform a 24-hr rated voltage soak test. However, one
ent types of transmission cables.
of the first major 230-kV XLPE transmission cable cir-
cuits in the United States (Cooper 2002) performed a
14.4.1 Extruded-dielectric Cable After-installation
1-hr ac voltage withstand test using a variable-frequency
Tests
series resonant test set at a voltage of 1.4 times rated
After-installation tests for extruded-dielectric transmis- phase-to-ground test voltage (186 kV).
sion cables vary significantly around the world (CIGRE
2002). The most common approach in most countries is Diagnostic Tests after Installation
to perform a jacket integrity test with dc voltage to A number of countries perform partial-discharge
detect cable damage during installation, followed by an measurements on the joints and terminations using
ac voltage test of the cable system. However, several either an on-line test (rated voltage) or off-line testing
countries in Asia still perform a dc Hi-Pot on the cable with an ac test voltage ranging from 1.25 to 1.7 times
system high-voltage insulation. rated line-to-ground voltage. In North America, PD
testing of XLPE transmission cables has generally been
Jacket Integrity Tests limited to EHV cable systems. IEEE Standard 400.3
A 1-min dc Hi-Pot test is normally performed with a (IEEE 2006) provides guidance in performing field
test voltage of 10 to 25 kV. ICEA S-108-720 recom- PD measurements.
mends a dc test voltage of 150 V/mil (6 kV/mm) of the
average thickness of the cable jacket with a maximum of
24 kV between the metallic shield/sheath and the semi- Table 14-10 IEC 62067 Ac Test Voltages After Installation
conducting outer coating. IEC 60229 standard has simi- Test Voltage
lar requirements. Nominal System Voltage (kV) (phase-to-ground kV)
220 to 230 180
Ac Voltage Withstand Tests 275 to 287 210
The AEIC, ICEA, and IEC recommendations for the 330 to 345 250
after-installation ac withstand test are: 380 to 500 260
500 320
14-25
Other After-installation Tests and HPGF cable systems in service. The results of these
One or more of the following tests are frequently per- tests provide a baseline for future diagnostic tests. The
formed prior to the ac voltage withstand test. fluid quality tests generally consist of the following
• Baseline distributed temperature sensor (DTS) mea- diagnostic tests:
surements using multimode optical fibers built into • Fluid dielectric strength (ASTM D-877)
the cable or a separate fiber optic cable. This test is • Fluid dissipation factor at 100°C (ASTM D-924)
performed to locate and identify any external heat
sources before the cable system is loaded. Chapter 12 • Moisture content (ASTM D-1533)
(Operation and Installation Design) provides addi-
tional details of DTS temperature monitoring. 14.4.3 SCFF Cable System After-installation Tests
• Measurement of bonding lead currents with a hand- After-installation tests for SCFF cable systems are simi-
lar to those for XLPE transmission cable systems,
held clamp-on current transformer. This measure-
except that a dc voltage Hi-Pot test is performed in place
ment is made to determine if the cable system bond-
of an ac voltage withstand test.
ing is functioning correctly.
Jacket Integrity Tests
14.4.2 HPFF and HPGF Cable After-installation The jacket integrity test is performed with a dc voltage
Tests Hi-Pot test of 10 to 25 kV for 1 min. This is a pass/fail
type of test.
The after-installation test procedures of HPFF and
HPGF transmission cables are well established because DC Hi-Pot Tests
of their long service history in North America. The Table 14-12 shows the recommended dc test voltages
after-installation test procedures are in AEIC CS2, recommended by IEEE Standard 400.1 for laminated-
AIEC CG3, and IEEE Standard 400.1. dielectric cables. The cable system should be at ambient
ground temperature before the Hi-Pot test is started.
Pipe Corrosion Protection
Making sure that the pipe corrosion protection is in
good condition before placing HPFF and HPGF cable Table 14-11 AEIC CS2 Hi-Pot Test Voltages for HPFF and
systems in service is imperative for long-term reliability HPGF Cables
of pipe-type cable systems. This generally consists of Dc Test Voltage (kV)
checking the pipe potential with respect to a copper- System 4-Hr 15-Min
copper sulfate cell and checking the pipe-coating resis- Voltage BIL Acceptance Acceptance
tance is within acceptable limits. Chapter 10 (Grounding (kV) (kV) Test Test Proof Test
and Cathodic Protection) covers the details of these 69 350 125 165 145
after-installation tests. 115 550 190 255 225
120 550 200 265 235
Dc Hi-Pot Tests 138 650 225 300 265
Dc Hi-Pot tests have historically been the most accepted 161 750 260 350 310
after-installation test for HPFF and HPGF transmis- 230 1050 340 455 400
sion cables. The dc voltage hi-potential (Hi-Pot) test is 345 1300 465 625 550
performed with the cable pressurized at nominal work- 500 1550 500 670 595
ing pressure. It is also important that the pipe pressure 765 2050 720 965 850
has been maintained at nominal pressure for at least
24 hr prior to conducting the Hi-Pot test. Experience Table 14-12 IEEE 400.1 Hi-Pot Test Voltages for Laminated-
has shown that approximately 24 hr at full pressure is dielectric Cables
required to make sure that any voids in the cable insula- System BIL 15-min Acceptance 15-min Maintenance
tion have been eliminated by the pipe fluid pressure. Voltage (kV) (kV) Test (kV) Test (kV)
Table 14-11 contains AEIC CS2-97 recommendations 69 350 175 130
for after-installation (or acceptance test) and proof test 115 550 275 205
dc test voltages. 138 650 325 245
161 750 375 280
Diagnostic Tests during Commissioning
230 1050 525 395
Many utilities also collect pipe fluid samples from all
345 1300 650 490
sampling ports and perform dissolved gas analysis
500 1550 775 580
(DGA) and fluid quality tests prior to placing HPFF
14-26
14.4.4 Sheath-bonding Verification Tests 14.4.5 Series Impedance Measurements

Some utilities perform tests on single-conductor cable Cable system positive- and zero-sequence series imped-
systems with cross-bonded or single-point-bonded cable ances (see Chapter 16) are measured in some cases
sheaths to verify that they are functioning as intended (Kaumanns 2003) prior to cable system energization to
(i.e., to minimize induced sheath currents). These sheath- verify calculated impedance values for relay settings and
bonding verification tests are performed as follows. for system planning studies. In general, the procedure
for impedance measurements is to short the three cable
Cross-bonded Systems terminations together at one end of the cable circuit and
Once the installation of the cross-bonded system is com- then apply balanced three-phase currents from a low-
plete, the correctness of the bonding connections shall voltage source for positive-sequence impedance mea-
be demonstrated by applying a three-phase current of surements (see Figure 14-23).
approximately 100 amperes to the phase conductors by
shorting the three cable terminations together at one The zero-sequence impedance measurements are per-
end of the cable circuit and applying adjustable three- formed by connecting the three cable terminations to
phase low voltage at the opposite end. The resulting ground and applying a single-phase voltage from a low
sheath currents and voltages with respect to the local voltage source to the three cable terminations at the
ground are then recorded for the following conditions. opposite end of the circuit (see Figure 14-24).
1. The bonding links in their correct position.
2. The bonding leads repositioned to simulate errone- The test equipment used to apply the positive- and zero-
ous bonding. sequence voltages and currents depends on the length of
the transmission cable circuit, but it is generally
3. The bonding links returned to their correct position.
designed to produce 25 to 100 A RMS of current in the
cable system. The following types of instruments can be
The sheath currents and voltages are then compared
used to perform the impedance measurements.
against the values from design calculations to determine
if the cross-bonding system and connections are func-
tioning as expected.
Table 14-13 Maximum Contact Resistance
Single-Point-Bonded Systems Maximum Contact
Sheath-bonding verification tests for single-point or Contact Resistance (μΩ)
multiple, single-point-bonded cable systems may be con- Link contact 20
ducted in a similar manner as for cross-bonded cable SVL terminal connection 50
systems. The sheath currents should be less than 1 A Joint lug and bonding connector 20
with 100 A of balanced three-phase currents flowing in Earth connection 50
the cable high-voltage conductors.
The sheath-bonding verification tests can also be per-

formed after the cable system has been placed in service
by installing split-core current transformers around the
cable-bonding leads and measuring the magnitude of
the currents.
Contact Resistances
The contact resistance of all earthing and sheath-
bonding connections are measured using a calibrated Figure 14-23 Schematic diagram of circuit for positive-
digital micro-ohmmeter. The contact resistance between sequence impedance measurement.
each lug attached to the joint sleeve of a sectionalized
joint and the corresponding bonding lead connector
should be measured prior to installation of the outer
protective covering.
The contact resistance shall not exceed the levels shown

in Table 14-13.
Figure 14-24 Schematic diagram of circuit for zero-
sequence impedance measurement.
14-27
1. Voltmeters, ammeters, and phase-angle meters Zero Sequence Impedance (Zo)

2. Voltmeters, ammeters, and wattmeters Single-phase currents are injected into the cables con-
nected in parallel, with the test current returning
3. Three-phase power analyzer. This type of instrument
through the cable sheaths, earth, and ground continuity
contains analog-to-digital converters for digitizing
conductors, if present.
the three-phase voltages and currents and accurately
calculating three-phase power and phase angles from
Measurements of current, voltage, and either power or
these digitized signals.
phase angle are used to calculate the zero-sequence
impedance as follows:
All sheath bonding and earthing connections are con-
nected as for normal operation. When there is more than Wattmeter Measurements Phase-Angle Measurements
one cable per phase, the conductors and sheaths are also 3V 3V
connected as for normal operation and the measure- Zo = Zo =
ments described below carried out on the total circuit. I I
3W 3V
Particular care is taken to ensure that contact resis- Ro = Ro = cos θ
tances are small compared with the resistance of the I2 I
conductor. At the time of the tests, an estimate of the 3V
sin θ
2 2
cable temperature is noted so that allowance can be Xo = Zo - R o Xo =
made for the temperature dependence of the resistances. I
Positive Sequence Impedance (Z1) Where:
The positive-sequence impedance measurements are V = measured conductor to earth voltage (V).
performed with the magnitudes of the three phase cur- I = measured total current (A).
rents as close to equal as possible. Measurements of W = measured total power (W).
current, voltage, and either power or phase angle are θ = measured phase angle (degrees).
made on each of the single-core cables. From these mea-
surements, the positive-sequence impedance is calcu- 14.5 FIELD MAINTENANCE/PROOF TESTS
lated as follows. Field maintenance test practices for transmission cables
Wattmeter Measurements Phase-Angle Measurements are not well defined. They generally consist of the fol-
Vρ Vρ lowing types of tests.
Zρ = Zρ =
Iρ Iρ 14.5.1 Electrical Tests
Wρ Vρ Electrical maintenance tests, in turn, consist of the fol-
Rρ = 2
Rρ = cos θ lowing tests.
Iρ Iρ
Hi-Pot Tests
Vρ
X ρ = Zρ 2 - R ρ 2 Xρ = sin θ Dc voltage Hi-Pot tests are generally performed on
Iρ laminated-dielectric cables after repairs or major main-
tenance have been completed and prior to returning the
Σ Zρ cable circuit to normal operation. See Tables 14-11 and
Z1 = ohms/phase
3 14-12 for the maintenance or proof test voltage levels.
Variable-frequency series resonant ac voltage mainte-
Σ Rρ nance tests have been performed in a few cases
R1 = ohms/phase
3 (Motallebi 2005), but there are no industry guidelines
for performing ac voltage maintenance tests on
Σ Xρ laminated-dielectric transmission cables.
X1 = ohms/phase
3
Dc voltage Hi-Pot tests are not recommended for
Where: extruded-dielectric transmission cables because of the
Vρ = measured phase voltage (V). reasons that are summarized in Section 14.2.2. Series
Iρ = measured phase current (A). resonant ac voltage tests have been performed as a post-
Wρ = measured phase power (W). repair maintenance test in some cases for extruded-
θ = measured phase angle (degrees). dielectric cables after repairs. However, a 24-hr soak test
14-28
is the most common procedure prior to returning a 14.6.1 Long-Term Accelerated Life Test
repaired extruded-dielectric cable system to service. The long-term accelerated life test, or Waltz Mill-type
Jacket Integrity Tests test, is a test procedure that was started at the Cornell
Most cable manufacturers recommend a jacket integrity 345-kV cable test project (Bertolett and Schifreen 1962;
test on extruded-dielectric and SCFF cable circuits at Rosson 1962) and continued on subsequent prototype
least every five years to determine if there has been any cable systems at EPRI’s Waltz Mill Underground Cable
jacket damage. The jacket integrity test is performed on Test Facility (Burrell and Young 1971). Basically, this
an out-of-service cable system by removing jumpers in two-year test on relatively long cable systems subjects
the appropriate link boxes and applying a dc voltage of the cable system components to higher-than-normal
5 to 10 kV for 1 min. voltages and temperatures to simulate a 40-year operat-
ing life. The objective of this type of test is to gain confi-
PD Tests dence before a new type of cable system is installed in
Some utilities have performed either on-line of off-line electric utility systems that it will operate reliably for
PD testing of the joints and terminations of extruded- long periods of time. Important considerations in meet-
dielectric cable systems after repairs. ing this objective are the following:
Dissipation Factor Tests • The cable system should be complete with accessories
Field dissipation factor measurements have been per- and should be long enough to represent initial utility
formed at rated system voltage as a maintenance test on installations.
some laminated-dielectric transmission cables to assess • The installation should reflect typical field condi-
the condition of the cable insulation or after major tions, and the installation personnel should possess
repairs (Motallebi 2005). The rated voltage dissipation qualifications similar to those required for commer-
factor measurements have been made using instrumen- cial installations.
tation developed by EPRI research (Cooper 1993).
• The test program should reflect conditions that are
14.5.2 Other Maintenance Tests considered to be the primary modes of aging for
potential causes of degradation.
Other maintenance tests on laminated-dielectric trans-
mission cables consist of fluid quality measurements Similar types of test facilities and programs have been
and DGA testing. Chapter 15 provides additional implemented in several European countries and Japan.
details concerning diagnostic tests performed during An example of the accelerated life test program devised
transmission cable maintenance. for 500-kV pipe-type cables is shown in Table 14-14.
The underlying basis for specifying this test program
14.6 SPECIAL-PURPOSE TESTS was to subject the cables to conditions that would be
Numerous special-purpose tests have been devised to representative of at least 40 years of service on a typical
address concerns for unusual applications of conven- utility system.
tional cable systems, to evaluate the performance of
cables with changes in construction, or to evaluate new Table 14-14 Test Program for 500-kV HPFF Transmission
cable systems. These special-purpose tests are not nor- Cable Samples at Waltz Mill
mally applied on a routine basis, but some are included Test Voltage Cond.
Test
in test standards, while others are not utilized frequently Period % L-L/L-G kV Temp. °C Operating Mode
enough to be included in industry standards. The former 1 110 550/317 75 Continuous
case includes special mechanical tests for submarine 2 110 550/317 75 Cyclic
cables that are part of IEC standards. An example of the 3 125 625/360 75 Continuous
latter type of special-purpose test is the thermomechani- 4 125 625/360 90 Continuous
cal bending (TMB) test devised to analyze the mechani- 5 125 625/360 90 Cyclic
cal performance of 345-kV pipe-type cables. Some of 6 125 625/360 90 Continuous
these special-purpose tests were developed by a particu- 7 125 625/360 100 Continuous
lar manufacturer, utility, or research organization to 8 125 625/360 100 Cyclic
address specific problems and have evolved into de facto 9 125 625/360 100 Continuous
test standards. The EPRI Waltz Mill accelerated life test 10 138 690/400 75 Continuous
is an example of this type of test. Some of the more com- 11 138 690/400 90 Continuous
mon special-purpose tests are described below. 12 138 690/400 100 Continuous
14-29
The following empirical equation (Equation 14-7), relat- One of the tests is a three-conductor test using three
ing the test voltage and conductor temperature to loss 20-ft (6-m) long cables installed in the size of pipe to be
of life for laminar impregnated-paper insulation, was used in the field. The cable sections are subjected to
used to specify the test voltages and current (Allam et 20,000–30,000 cycles, causing them to bend and slide
al. 1986). against the pipe wall, simulating the forces developed
n during load cycling. This test procedure was used during
⎡E ⎤
T = ⎢ p ⎥ Tp e α (t −85) yr an EPRI-sponsored project (Aabo et al. 1984).
⎣ Eo ⎦ 14-7
The other test, also developed during the EPRI project,
Where:
subjects one 10-ft (3-m) section of cable to the motions
Eo = rated test voltage, V.
that it will experience during load cycling. This test is
Ep = applied test voltage, V.
simple and reproducible, and will cause deterioration in
Tp = length of a test period, yr.
inferior cables.
t = temperature of a test period, °C.
T = total equivalent service, yr.
14.6.4 Longitudinal Water Penetration Tests
α = temperature aging exponent.
(ICEA T-34-664)
Exponent n is assumed to be in the range of 10 to 12 Many XLPE transmission cables have been manufac-
for impregnated paper. The temperature aging expo- tured with water-blocking materials in the high-voltage
nent, alpha, is assumed to be in the range of 0.07 to conductor and/or between the cable core and the metal-
0.087. This range of temperature exponents is equiva- lic sheath. Several industry standards have been devel-
lent to assuming that the rate of aging is doubled for oped to test the effectiveness of conductor and cable
each 8–10°C increase of the insulation temperature core longitudinal water blocking. ICEA publication
above its maximum continuous operating temperature, T-34-664 (Guide for Conducting Longitudinal Water
85°C. If Equation 14-7 is summed for the 12 test peri- Penetration Resistance Tests on Longitudinal Water
ods, the total equivalent test time is in excess of 50 years. Blocked Cables) is an example of this type of test proce-
The cyclic test periods were included to simulate dure (ICEA 1996).
mechanical cycling of the cable system components, and
the continuous load periods were included to provide The ICEA water penetration resistance testing guide
thermal aging of the insulation. contains both qualification test procedures as well as
production test procedures for extruded-dielectric
Cable systems subjected to the two-year test program cables.
consist of approximately 1000 ft (300 m) of single-phase
cable, two field-installed joints, and two commercial ter- The qualification test procedure consists of the follow-
minations. The condition of the cable system compo- ing two parts:
nents is monitored during the test program by making • Part A. Requires testing of two samples of cable that
insulation dissipation-factor measurements as a func- have been preconditioned with mechanical bending.
tion of voltage and conductor temperature several times One of the 36-in. (914 mm) long cable samples is
during the test periods. This type of test has been used further preconditioned by placing the sample in a
for pipe-type, compressed gas insulated, and extruded- 100°C oven for 4 hr. The second sample is precondi-
dielectric cable systems. tioned by placing the sample in a cold chamber
(-10°C) for 4 hr. Both of these cable samples are then
14.6.2 Mechanical Bending Tests subjected to a water pressure of 5 psig for 1 hr at
Several types of mechanical bending tests have been room temperature.
included in test standards, ranging from relatively sim- • Part B. Also preconditions a 10-ft (3 m) sample of the
ple cable sample preconditioning to elaborate, long- cable with three reverse bends. The cable sample is
term tests devised for submarine cables (see Chapter 7). then load-cycled 20 times with the peak conductor
temperature reaching the cable emergency operating.
14.6.3 Thermomechanical Bending Tests The exposed cable core and cable conductor (if appli-
One of two TMB tests may be specified to further cable) are subjected to water at 5 psig pressure during
ensure that pipe-type cables will withstand thermal the load cycling.
cycling over the life of the cable.
No water leakage at the cable ends opposite from the
pressurized water source is allowed in Part A or Part B.
14-30
The production test procedure is similar to Part A of the the transmission cables will be installed in a location
qualification test procedure, except that the sample is where they will be exposed to water with an appreciable
not preconditioned at the high/low temperatures, and salt content.
the duration of the test is 15 min.
IEC 60840 and 62067 both contain a longitudinal
It should be noted that the water-swellable materials water-blocking test that is similar to the ICEA T-34-664
that are used for water blocking in most transmission test procedure. The ICEA standards for extruded-
cables are not as effective when exposed to saltwater dielectric transmission cables require that the water-
(compared to freshwater). Therefore, the ICEA water- blocking test be performed as part of the type test, if
blocking test procedures should use saltwater if applicable to the cable design.
14-31
REFERENCES Blais, L. D. 1972. “Development and Installation of

138-kV Cable for Tests at EEl Waltz Mill Station—
Aabo, T., J. A. Moran, and J. F. Shimshock. 1984. Cable 22.” IEEE Transactions on Power Apparatus
“Thermo-Mechanical Bending Effects in EHV Pipe- and Systems, Vol. PAS 91. No. 4, July/August.
Type Cables.” IEEE/PES 1984 T & D Conference. pp. 1448–1456.
Kansas City, Missouri. April 29–May 4.
Bolin, P. C. 1982. “Mechanical Design and Test of 1200-
Aabo, T., and J. A. Moran. 1988. Thermomechanical kV Semi-Flexible SF6 Insulated Transmission Line.”
Bending Effects on Extra-High Voltage Pipe-Type IEEE Transactions on Power Apparatus and Systems.
Cables. EPRI Report EL-5880. Vol. PAS 101. No. 7. pp. 2194–2202.
AEIC. 1993. AEIC Specification CS4-93. Specifications Burrell, R. W., and F. S. Young. 1971. “EEl—Manufac-
for Impregnated-Paper-Insulated Low and Medium turers 500/550 kV Cable Research Project, Waltz Mill
Pressure Self-Contained Liquid Filled Cable. 8th Edi- Tests.” IEEE Transactions on Power Apparatus and
tion. Association of Edison Illuminating Companies. Systems. Vol. PAS-90. No. 1. January/February.
Alabama. pp. 180–190.
AEIC. 1997. AEIC Standard CS2-97. Specification CIGRE. 1972. CIGRE Study Committee 21 WG 06.
for Impregnated Paper and Laminated Paper Polypropy- “Recommendations for Mechanical Tests on Subma-
lene Insulated Cables, High-Pressure Pipe-Type. rine Cables.”
6th Edition. Association of Edison Illuminating Com-
panies. Alabama. CIGRE. 1980. CIGRE Study Committee 21 WG 01.
“Recommendations for Tests of Power Transmission
AEIC. 2005. AEIC Guide CG3-05. Guide for Installa- DC Cables for a Rated Voltage Up to 600 kV.”
tion of Pipe-Type Cable Systems. 2nd Edition. Associa- ELECTRA. No. 72. pp. 105–114.
tion of Edison Illuminating Companies. Alabama.
CIGRE. 1989. CIGRE Study Committee 21 WG 09.
AEIC. 2006. AEIC Specification CS9-06. Specification “Application of Weibull Distribution to the Study of
for Extruded Insulation Power Cables and Their Power Cable Insulation.” ELECTRA. No. 127.
Accessories Rated Above 46 kV through 345 kVac.
1st Edition. Association of Edison Illuminating Com- CIGRE. 2002. CIGRE TF 21.05. “Experiences with
panies. Alabama. AC Tests after Installation on the Main Insulation of
Polymeric (E)HV Cable Systems.” ELECTRA. No. 205.
Allam, E. M., J. H. Cooper, and J. F. Shimshock. 1986.
Development and Long-Term Testing of a Low-Loss CIGRE. 2005. CIGRE WG B1.04. “Technical
765-kV High Pressure Oil-Filled Pipe Cable. Brochure 279 Maintenance for HV Cables and Accesso-
Report 21-06, CIGRE Paris Meeting. ries.”
Auerbach, R. H., et al. 1980. “525 kV Self-Contained CIGRE. 2006. CIGRE Study Committee D1 WG 2.05.
Oil Filled Cable Systems for Grand Coulee Third Power “Practical Aspects of the Detection and Location of
Plant—Design, Test Manufacture, and Installation.” Partial Discharges in Power Cables.” ELECTRA.
IEEE Transactions on Power Apparatus and Systems. No. 226.
Vol. PAS 99. No. 5. pp. 1722–1733.
CIGRE. 2006. CIGRE WG B1.06. “Technical Brochure
Bankoske, J. W., et al. 1977. “Mechanical Effects of 303—Revision of Qualification Procedures for High
Load Cycling On Pipe-Type Cable.” IEEE PES Summer Voltage and Extra High Voltage AC Extruded Under-
Meeting F77. pp. 559–8. ground Cable Systems.” Technical Brochure 303.
Bertolett, A. D., and C. S. Schifreen. 1962. “Cornell Cooper, J. H. 1993. Field Measurement of Cable Dissipa-
EHV Cable Testing Station: Equipment and Design.” tion Factor. EPRI Report TR-102449.
Transactions of the American Institute of Electrical Engi-
neers. Vol. 81. pp. 496–508.
14-32
Cooper, J. H. 2002. Demonstration of Commissioning IEEE. 1975. EPRI High-Voltage Solid Dielectric Oper-
Tests for Extra-High Voltage Cross-Linked Polyethylene ating Committee. “Status of 138-kV Solid Dielectric
Cable Systems at Los Angeles Department of Water and Cables Being Evaluated at Waltz Mill.” IEEE Confer-
Power. EPRI report 1001856. ence Paper. C75 007-0. Winter Power Meeting.
Gulski, E., F. J. Wester, et al. 2003. “Transmission IEEE/ANSI. 1993. Institute of Electrical and Electron-
Power Cables PD Detection at Damped AC Voltages.” ics Engineers/American National Standards Institute.
JICABLE ’03—International Conference on Insulated Substation Committee of IEEE. Document C37.122.
Power Cables. Gas Insulated Substation—Standards, Recommended
Practices, and Guides.
ICEA. 1996. Insulated Cable Engineers Association.
Publication T-34-664. Guide for Conducting Longitudi- IEEE. 1995. Institute of Electrical and Electronics
nal Water Penetration Resistance Tests on Longitudinal Engineers. Standard 4-1995. Standard Techniques for
Water Blocked Cables. High-Voltage Testing.
ICEA. 2004. Insulated Cable Engineers Association. IEEE. 1996. Institute of Electrical and Electronics
Publication S-108-420-2004. Standard For Extruded Engineers. Standard 48-1996. IEEE Standard Test Pro-
Insulation Power Cables Rated Above 46 Through cedures and Requirements for Alternating-Current Cable
345 kV. Terminations 2.5 kV Through 765 kV.
IEC. 1966. International Electrotechnical Commission. IEEE. 2000. Institute of Electrical and Electronics
“Standard 60230 Impulse Tests on Cables and Their Engineers. Standard 404-2000. Standard for Extruded
Accessories.” and Laminated Dielectric Shielded Cable Joints Rated
2500 V–500000 V.
IEC. 1982. International Electrotechnical Commission.
“Standard 60229 Tests on Cable Oversheaths Which IEEE. 2001. Institute of Electrical and Electronics
Have a Special Protective Function and are Applied Engineers. Standard 400-2001. IEEE Guide for Field
by Extrusion.” Testing and Evaluation of the Insulation of Shielded
Power Cable Systems.
“Standard 60811 Common Test Methods for Insulating IEEE. 2002. Institute of Electrical and Electronics
and Sheathing Materials of Electric Cables.” Engineers. Standard 82-2002. Standard Test Procedure
for Impulse Voltage Tests on Insulated Conductors.
“Standard 60885-3 Electrical Test Methods for Electric IEEE. 2006. Institute of Electrical and Electronics
Cables—Part 3: Test Methods for Partial Discharge Engineers. Standard 400.1-2006. IEEE Guide for Partial
Measurements on Lengths of Extruded Power Cables.” Discharge Testing of Shielded Power Cable Systems in a
Field Environment.
“Standard 60060 High Voltage Test Techniques.” Kaumanns, J., E. Plieth, and R. Plath. 2003. “On-Site
AC Testing and PD Measurements of 345 kV/2500
IEC. 1999. International Electrotechnical Commission. mm2 XLPE Cable Systems for Bulk Power Transmis-
“Standard 60840 Power Cables with Extruded Insula- sion.” JICABLE International Conference on Insulated
tion and Their Accessories for Rated Voltages above Power Cables.
30 kV (Um = 36 kV) up to 150 kV (Um = 170 kV)—
Test methods and Requirements.” Kind, D. 1978. An Introduction to High-Voltage Experi-
mental Technique. Braunschweig/Wiesbaden: Friedrich
IEC. 2001. International Electrotechnical Commission. Vieweg & Sohn. pp. 60-64.
“Standard 62067 Power Cables with Extruded Insula-
tion and Their Accessories for Rated Voltages above Motallebi, H. 2005. “Toluca-Van Nuys Cable C, 230-kV
150 kV (Um = 170 kV) up to 500 kV (Um = 550 kV)— HP Pipe-Type Restraint Installation.” ICC Subcommit-
Test Methods and Requirements.” tee C Minutes. November 1.
14-33
Nattrass, D. A. 1988. “Partial Discharge Measurement Steennis, E. F., and A. M. F. J. Van De Laar. 1989.
and Interpretation.” IEEE Electrical Insulation. Vol. 4. “Characterization Test and Classification Procedure for
No. 3. May/June. Water-Tree Aged Medium Voltage Cables.” ELECTRA.
No. 125. July. pp. 90–97.
Rosson, J. L. 1962. “Extra-High-Voltage Research Sta-
tion at Cornell—Techniques and Procedures of Mea- Tanaka, T., and A. Greenwood. 1983. Advanced Power
surements.” Transactions of the American Institute of Cable Technology: Basic Concepts and Testing. Vol. 1.
Electrical Engineers. Vol. 81. December. pp. 508-512. Boca Raton: CRC Press. pp. 173–180.
14-34
CHAPTER 15 Operation and

Maintenance
Authors: Stephen E. Turner, Underground Systems Inc. (USi)
Reza Ghafurian, Consolidated Edison Company of New York, Inc.
Thomas B. Welles, Advanced Testing Systems Inc.
Reviewers: Frank L. Kuchta, Prysmian Cables and Systems
Reza Ghafurian, Consolidated Edison Company of New York, Inc.
Proper system operations and maintenance procedures are vital to long-term, trouble-
free operation of underground cable systems. Cable systems can operate satisfactorily for
many years without any maintenance, and will show no obvious signs of deterioration.
When trouble does occur, however, the repairs can be very expensive and time consum-
ing, and the availability of the circuit may be adversely affected.
Although it is not possible to provide specific procedures suitable for all utilities, this
chapter does present routine operations and maintenance procedures for all types of
cable systems, and it describes procedures for locating and repairing electrical faults and
fluid leaks. Individual utilities are encouraged to prepare and follow procedures specific
to their cable systems.
Stephen E. Turner is Vice President of Underground Systems, Inc.

(USi). His responsibilities center on engineering and consulting in the
field of medium- and high-voltage underground power transmission
systems, with special emphasis in the area of accessory material for
those systems. Prior to joining USi in 1979, Mr. Turner spent 20+ years
with The Pirelli Cable Corporation (formerly General Cable Corpora-
tion), where as Chief Engineer, he was responsible for design, engineer-
ing, and installation of laminar-insulated transmission cable systems.
Over the years he has presented numerous talks on cable system installation, conducted
“hands-on” cable splicing schools for installation contractors and utility companies, and
has been in charge of cable system design and installation for more than 30 utility compa-
nies, the U.S. government, and private contractors. Mr. Turner has authored a number of
publications on underground transmission and is a member of the Institute of Electrical
Engineers, the Insulated Conductors Committee (IEEE), and CIGRE. He was past
Chairman of the ICC Working Group that developed submarine power cable design and
construction standard IEEE STD 1120.
Reza Ghafurian received his MSEE and Ph.D. degrees from the Univer-
sity of London, Imperial College of Science and Technology in the United
Kingdom. He joined the R&D Department of the Consolidated Edison
Company of New York in 1985. Prior to that, he was Associate Professor
of Electrical Engineering at the Iowa State University, in Ames, Iowa.
From 1985 to 1999, he held positions of Principal Research Engineer and
Manager of Transmission Research Program. In 1999, he joined the Elec-
trical Engineering Department of Con Edison as Project Manager, and in
2002, he became Section Manager, Transmission Feeders Engineering. In October 2005,
he was assigned as Project Leader for the development of new designs of the future substa-
tion and distribution systems. In March 2006, he was recognized as the company’s Techni-
cal Leader for Underground Transmission. Dr. Ghafurian is an IEEE Fellow and Con
Edison’s representative on the Cable Engineering Committee of the AEIC. From 1999 to
2004, he served as Chairman of the EPRI Underground Transmission Task Force.
15-1
Chapter 15: Operation and Maintenance EPRI Underground Transmission Systems Reference Book
15.1 INTRODUCTION especially submarine sections, should not be reclosed

This chapter provides a general overview on issues due to the possibility that the problem may be in the
related to the operation and maintenance (O&M) of submarine portion of the circuit. Such circuits should
underground transmission systems. It includes routine incorporate Distance to Fault Relay protection.
O&M procedures for all types of cables. Section 15.2
identifies the aspects of system operation that should be 15.3 ROUTINE MAINTENANCE
covered by written procedures. Section 15.3 describes Reliable operation of any underground transmission
routine maintenance for the different areas of the cable system is a function of the maintenance programs devel-
system and for the different cable components—includ- oped and implemented for each type of system. Earlier
ing the cable, dielectric fluid, joints, terminations, and installed pipe-type (high-pressure fluid-filled [HPFF]
pressurization systems. Sections 15.4 through 15.6 dis- and high-pressure gas-filled [HPGF]) and self-contained
cuss procedures for electrical fault location, electrical fluid-filled (SCFF) cable circuits were normally
failure repair, and fault repair material. Section 15.7 equipped with remote alarms designed to provide
describes techniques and tools for fluid leak detection, advanced warning of impending problems. However,
location, and prevention. due to the importance of the ancillary fluid support sys-
tems, only HPFF pipe-type cable systems traditionally
15.2 SYSTEM OPERATIONS invoked a disciplined maintenance program at least for
Individuals responsible for overseeing the reliable opera- the pressurizing end of the circuit. For HPFF cable sys-
tion of underground cable systems should have available tems, utility personnel were required to visit the pressur-
written procedure covering all operating aspects of each ization units to change charts on the pressure recorders.
cable system within their jurisdiction. Procedures should Such visits, though not considered maintenance in the
cover the following areas: true sense of the word, did afford the operator the
opportunity to inspect the pressure profiles and fre-
• Ampacity Rating: For paper-insulated cables, the rat- quency of pump operations over the previous week. Fre-
ings should cover the normal 85°C conductor tem- quent, unexplainable pump operations may indicate a
perature rating, and emergency ratings including leak in the circuit or a malfunctioning component
both the 100-hour 105°C and 300-hour 100°C rat- within the pressurization unit. At the same time, the ter-
ings, based upon reasonable knowledge of the overall minations and all exposed piping could be inspected for
thermal characteristics of the cable environment in fluid leaks or signs of damage, which could indicate
both summer and winter operation. A more detailed impending problems. The cathodic-protection rectifiers
description of cable rating is presented in Chapter 11. using independent anode beds could be checked and
• De-energizing and Energizing Circuits: Procedures for compared with previous levels, provided that the units
de-energizing and energizing circuits under emer- indicated leakage current and not impressed current
gency or fault conditions and for routine system across a dropping resistor. Excessive leakage current
maintenance should be established and executed, may indicate damage to the pipe coating. Cursory
when needed. inspections of this type, though important to the opera-
• Fluid Pressurized Systems: Procedures should be tion of the circuit, are considered part of system opera-
established and available to quickly respond to circuit tion, rather than maintenance.
fluid pressurizing problems. An increasing number of
new installations have provisions for remotely moni- Recently installed pipe cable circuits incorporate remote
toring the status of these pressurizing systems, monitoring systems, which monitor the operating
whether they are fluid pumping plants or gas pres- parameters of the pressurization plant, including system
sure control cabinets used on pipe-type cable systems pressures, fluid storage tank levels, pump operations,
or reservoirs used on self-contained fluid-filled and cathodic-protection rectifier current and voltage.
(SCFF) cable systems.
Cable circuits consist of cables installed in pipe or duct
Solid dielectric circuits may include fluid-filled termi- or directly buried in the ground. Fluid-filled systems
nations, which may or may not be connected to reser- require pressurization units or reservoirs to maintain
voir systems that may also be monitored and positive system pressures. Nitrogen gas-pressurized cir-
alarmed. Procedures should be established to cuits normally have gas cabinets, which need to be
respond to low fluid pressures in the termination inspected and pressure alarms tested (see Figure 15-5).
fluid system. In addition, circuits include manholes containing splices
• Reclosing Procedures: Overhead transmission cir- and splice casings, terminations at the cable end points,
cuits, which incorporate underground cable sections, cathodic protection units, and cross-bonding and
15-2
EPRI Underground Transmission Systems Reference Book Chapter 15: Operation and Maintenance
grounding systems, etc. All require periodic inspection, As an active part of a maintenance program, frequent
testing, and, when required, repair or replacement, if the observation of construction activities over or adjacent
circuit is to provide reliable, uninterrupted service. The to pipe or duct systems can greatly reduce emergency
following sections outline recommended maintenance repairs resulting from external damage caused by such
procedures for cable system components. projects. Monitoring the One Call system for areas/loca-
tions where buried circuits are located may provide the
15.3.1 Cable Installed in Pipe or Duct utility with the opportunity to work with the construc-
tion company to avoid underground circuits.
Pipe-type Cables (HPFF and HPGF)
The pipe portion of a pipe-type cable system permits 15.3.2 Pipe Coating Damage
minimal pipe maintenance because the amount of
exposed pipe is small and is normally limited to the Of all the components of a pipe cable system, the pipe
manholes. Those areas observed to be rusted or cor- and pipe coatings are most susceptible to undetected
roded during routine inspections should be scheduled damage. The pressurization unit or remote pressure-
for immediate reparation/maintenance. The pipe must monitoring system readily detects moderate to major
be cleaned and wire-brushed to bare metal before apply- damage resulting in the loss of pressurizing medium.
ing new protective material. Pipe that is installed in tun- However, minor damage to the pipe coating only may go
nels, on bridges and or in manholes, which cannot be undetected for months or even years, if the circuit is not
cathodically protected, are susceptible to periodic flood- equipped with a remote monitoring system or the opera-
ing or seepage, and therefore require more frequent tor does not periodically monitor the cathodic protec-
inspections than pipe in relatively dry environments. tion system. Left unprotected, the exposed metal will
Failure to properly inspect and maintain pipes in such eventually corrode to the point where the pressurizing
locations can lead to extensive repair operations at a fluid or gas begins to leak through pinholes in the pipe.
later date and may result in fluid loss and a substantial Initially, the leak rate is small, making detection
cleanup operation. extremely difficult without the aid of a leak detection
program as outlined in Section 15.7, “Leak Manage-
The choice of materials used to repair deteriorated pipe ment Techniques for Fluid-Filled Feeders.” The presence
coatings depends on the pipe location and environment. of coating damage of this type can sometimes be
Pipes installed above grade and exposed to weather are detected before leaks develop by periodically performing
normally painted or covered with a material resistant to overall coating resistance checks or close interval poten-
the sun’s ultraviolet rays such as epoxy enamel paints. tial surveys on the circuit pipes. Schedules for pipe coat-
Taped coatings tend to have poor weathering character- ing inspections of this type are generally based on
istic and are normally not used. In wet or humid envi- operating experience, the extent of the system, and avail-
ronments such as tunnels and manholes, moisture tends able manpower. However, inspections should be per-
to permeate most repair coatings. Wax-impregnated formed, at a minimum, at least once every two years.
fabric tape has proven to provide good protection in More frequent inspections are recommended when there
these environments. is active construction in the area.
Cables in Duct Coating-Resistance Tests

Cabled duct systems require little, if any, maintenance Under normal circumstances, the pipe is isolated from
unless the ducts are intended to operate in a sealed con- ground by either polarization cells or electronic isola-
dition. In that case, the duct mouth seals must be tors. A dc potential can be intermittently applied
inspected periodically to ensure that cable load cycling between the coated pipe and ground, as shown schemat-
has not disrupted or damaged the seals, which would ically in Figure 15-1. The change in voltage, V, of the
permit manhole water to enter the duct. Maintenance of pipe, as referenced to a remote ground electrode, divided
this type is normally done in conjunction with routine by the current, I, gives an average resistance to earth of
manhole inspection, cleaning, and repair operations. the pipe. Where an appreciable voltage drop is indicated,
Seldom are ducts inspected or cleaned unless a new the voltage and current change should be measured at
cable installation is scheduled or the ducts have become several locations, and the average value should be used
contaminated with materials that may damage the cable to calculate the coating resistance. This resistance, when
or cable jacketing material. When bonding/grounding multiplied by the total surface area of the pipe section
conductors or ribbon anodes used for cathodic protec- under test, gives an indication of the effective coating
tion must be replaced, the ducts may require cleaning resistance in ohms per unit area of pipe surface. The
before installing the new components. value of pipe coating resistance obtained depends on the
type of pipe coating, the resistance of the surrounding
earth, the moisture content of the coating, and the
15-3
Figure 15-1 Coating resistance test.
amount of bare pipe exposed to the earth (see repaired by hot-molding pieces of asphalt material over
Table 15-1). Changes from earlier coating resistance the primed pipe surface or covering the primed area
measurements should be used to determine the status of with a field jointer filled with hot coal-tar enamel.
the coating. Today split heat-shrinkable sleeves or self-bonding tapes
are used for making most coating repairs for all pipe-
Refer to AEIC CG3-2005, Guide for Installation of coating materials.
Pipe Type Cable Systems (2nd Edition), January 2005.
Note: Older coal-tar enamel wrap-type coatings con-
When lower-than-expected values of coating resistance tained a layer of asbestos pipeline felt, which must be
are encountered, an over-the-ground survey may be use- carefully removed and disposed of as hazardous material.
ful to identify the area of concern. An intermittent dc
voltage is applied between the coated pipe and a suitable When the pipe has sustained severe corrosion or minor
low-resistance ground, as shown in Figure 15-2. Using a physical damage, it may be necessary to weld steel
high-resistance voltmeter, the local pipe-to-soil poten- patches over the damaged areas before reconstituting
tial difference between two reference electrodes is mea- the coating. All coating repairs must be electrically
sured at spaced intervals along the pipe route. An tested at the appropriate voltage levels before backfilling
intermittently applied voltage is used (rather than a con- the repair areas.
stant voltage) to identify the test signal from other stray
current sources. When a defect is approached, the mag- 15.3.3 Manhole and Joint Casings
nitude of the deflection increases and peaks directly Manholes are considered Confined Space. As a result,
over the defect. This method is often modified to reflect all individuals that will be associated with any work in
the change in potential along the cable route. and around manholes must have completed a Confined
Space Training Course prior to entering manholes. At
Pipe to soil potential surveys are further discussed in the very least, an indicating device for oxygen deficiency
Chapter 10, Grounding and Cathodic Protection. and combustible gas must be present and in working
Pipe Coating Repairs order in the manhole, at all times, when individuals are
Methods used to repair defective or damaged pipe
coating depend on the original coating material. Gener-
ally, repairs are made by exposing and cleaning the
defect area. All disbonded coating is removed, and the
edges of the sound coating feathered. In the past,
asphalt and coal-tar enamel wrap-type coatings were
Table 15-1 Pipe Coating Resistance Measurement Values

Coating Resistance
Pipe Coating Material (MΩ-ft2 [MΩ−m2])
Coal-tar Enamel Wrap-type 2
Asphalt Somastic 10
Polyethylene Polypropylene Type 10
Figure 15-2 Over-the-ground survey.
15-4
in the manhole. In congested areas, it has been recom- buckets. The removed materials are deposited in tankers
mended that manholes be monitored for carbon dioxide for proper disposal. Such cleanings are normally recom-
and carbon monoxide in addition to oxygen. mended before major manhole maintenance or repairs
are undertaken. Manholes that consistently flood should
Inspection of transmission cable manholes should be have the seals on all penetrations carefully examined.
performed on a yearly basis, unless prior inspections dic-
tate either more or less frequent inspection intervals. Some utilities have incorporated the EPRI or other
Manholes that tend to flood should be scheduled for manhole cover restraint system. If present, the compo-
routine pumping. Fluids removed from the manholes nents of this system should also be inspected.
must be disposed of in accordance with local regulations.
Manhole inspections should cover the physical condition 15.3.4 Cable
of all components. These include covers, entrance chim- Cable maintenance varies with the type of system
neys, ladders, ceiling, walls, floor, cables, cable accesso- installed and the method of installation (e.g., directly
ries (e.g., link boxes), and the internal environment, all of buried, installed in duct and manholes or, in the case of
which may affect the safety of personnel, continuity of pipe-type cable, installed in pipe). Problems, such as
service, and the life of the cable system. The area around cable movement resulting from load cycling, are com-
the manhole should be checked for settling, which may mon in duct installations, but historically they were not
indicate pending soil instability problems. considered when designing early pipe cable systems.
Today, pipe-type cable movement, and the problems
On pipe-type cable circuits, the condition of the pipe, associated with that movement, is well known to cable
end-wall seals, joint casings, and associated hardware, engineers as Thermomechanical Bending (TMB).
such as valves and data-access ports, should be exam-
ined with repairs scheduled as required. If test stations Pipe-Type Cable
are present, the continuity of all leads should be The cable portion of HPFF and HPGF pipe-type cable
checked, including reference anodes. On SCFF or systems is basically maintenance-free due to the cable’s
extruded-dielectric cable circuits, inspections should confinement in a continuous, welded steel pipe. Experi-
include corrosion on all exposed surfaces, the degree of ence has shown that cable movement resulting from
deterioration of the corrosion protective jackets, cable expansion and contraction appears to be the main cause
sheaths and joint casings, arc proofing, cable neutrals, of cable-related problems. TMB has occurred in the
and bonding systems including the condition of the pipe adjacent to splices that were not properly sup-
sheath voltage limiters (if present). The cable-support- ported. The flexing or bird caging of the cables nor-
ing devices should be inspected for missing or broken mally controls cable expansion rather than axial
support insulators. In addition, cables should be exam- movement along the pipe. The degree of cable bending
ined for movement in and out of the ducts, which can associated with bird-caging is controlled by the pipe
result in kinked or damaged areas that may have diameter and was thought to be well within the design
occurred due to unexpected lateral movement of the constraints of the cable. However, where circuits are
cables. Based on the findings, remediation may be installed on inclines, the cable will tend to move down
required and should be scheduled for repair at a time the incline due to load cycle ratcheting and traffic flow,
that is consistent with the associated risk. unless proper restraining methods are implemented to
anchor the cable at splice locations. Recently, several
Manholes must be inspected for unsafe conditions. cases of TMB have occurred in flat cable sections
When necessary, corrective steps should be taken to between manholes; the reason for this phenomenon is
ensure the safety of personnel before any maintenance not fully understood.
work is started. This includes removing and properly dis-
posing of any hazardous fluids or materials found within At joint locations, where the pipe no longer confines the
the manhole. Materials that are deleterious to the cable cables, excessive cable slack may develop, which is free
or cable jacketing material must be removed, and the to bend and twist, if the cables are not properly
manhole and ducts cleaned. In some areas where man- restrained and supported. This uncontrolled thermome-
holes are located, bees and spiders are a potential risk. chanical bending results in soft spots developing within
the insulation structure adjacent to the splice. If left
As part of normal maintenance procedures, many utili- uncorrected, this can lead to cable failure. Similar cable
ties schedule regular manhole cleaning programs in movement and resulting TMB can be experienced on
which the interiors are hosed down and the accumulated cable circuits installed under roadways or along rail-
material and debris removed using suction pumps and roads as a result of traveling ground vibrations, caused
by traffic, in conjunction with load cycling. The cables
15-5
tend to move in the direction of traffic flow. As part of supporting hardware and insulators, joint sheath insula-
any cable maintenance program, joints susceptible to tors, connections for cross-bonding and grounding, and
movement should be periodically x-rayed for signs of associated sheath voltage limiters.
cable distortion resulting from TMB or joint movement.
Cases have been found where the joints were wedged Cables and joints should be properly anchored in the
into the joint reducer, representing movement of more manhole to accommodate mechanical forces generated
than 1 ft (30 cm). by load cycling. Cables that are prone to movement due
to the circuit profile should be anchored using split bas-
When TMB or excessive cable movement is encountered ket grips located near the duct mouth, or in the duct,
within joint enclosures, repair procedures are based on when possible. Movement is only a problem if the cable
the degree of total movement and the amount of dam- is not racked appropriately. Cables must be examined
age to the cable or joint. Where excessive joint move- for possible damage at the duct mouth, resulting from
ment without adjacent cable damage is encountered, the cable movement due to load cycling. Where damage is
joint reducer can be moved back to provide clearance present, repairs must be made and protective devices
between the joint and the reducer face. In addition, must be installed to prevent further damage.
tapered joint restrictors can be installed to limit future
Directly Buried Installations
joint movement, as shown in Figure 15-3. Intermediate
Directly buried cable systems do not lend themselves to
cable supports must be added to the original support
an inspection and maintenance program due to the
system, and the entire assembly should be bound at
nature of the installation. Circuits that are cross-
intervals from reducer face to reducer face to prevent
bonded may have surface-mounted link boxes, which
future flexing within the joint enclosure. The enlarged
are accessible for routine inspection in the manhole or
opening is closed using a split coupling and sleeve
on above grade structures. In this case, bonding connec-
assembly in combination with the original sleeves.
tions and associated sheath voltage limiters should be
Where minor cable damage is present, it may be possible
checked periodically.
to disassemble the splice and re-lay the cable insulation
in the damage area. If this approach cannot be used, the
15.3.5 Dielectric Fluid
damaged section of the cable must be replaced using a
section of new cable and two joints, as discussed in Sec- Until the early 1970s, maintaining dielectric fluids used
tion 15.5. in cable systems consisted of periodically sampling the
fluid for laboratory measurements of dissipation factor,
Duct Installations dielectric strength, and color. With the development of
Cable maintenance on duct-installed extruded-dielectric DGA (Dissolved Gas Analysis) in the 1970s to deter-
and self-contained cable circuits is minimal; it is nor- mine the condition of transformer fluids, cable engi-
mally performed in conjunction with routine manhole neers are now, on a regular basis, using the same
inspections. Components that must be maintained in technology to examine the condition of impregnated-
proper condition to ensure reliable cable service include paper HPFF, HPGF, and SCFF cable circuits. Under
the cable’s outer protective jacket, cable fireproofing, elevated temperatures, the cable insulation paper tends
to decompose, generating carbon monoxide (CO), car-
bon dioxide (CO2), and water (H2O). Dielectric fluids,
on the other hand, emit hydrogen (H2) and methane
(CH4) at moderate temperatures, and ethane (C2H6),
ethylene (C2H4), and propane (C3H8) at higher tempera-
tures. Under arcing conditions, acetylene gas (C2H2)
is formed.
A “fingerprint” for the circuit can be established by

using a gas chromatograph to measure the relative con-
centrations of the above gases contained in samples of
fluid removed from terminations and joint locations.
Future changes in concentrations or ratios of concen-
trations of these gases may provide insight into pending
problems or may indicate accelerated aging of circuit
components. Investigations of circuits with high com-
bustible-gas contents have resulted in findings of arcing
Figure 15-3 Tapered joint restrictor (courtesy USi). and incipient cable failures within those circuits, indicat-
15-6
ing that dissolved gas analysis is an important mainte- tions. Where surge voltage limiters form a part of the
nance tool for determining the status of fluid-filled and cross-bonding scheme, the units should be checked for
gas-pressurized cable circuits. proper operation, and components should be replaced,
when necessary. The same is true for the open end of
Fluid samples should be taken using glass tubes or single-point-grounded sections of a circuit.
syringes terminated with a luer lock fitting to which is
attached a three-way stopcock. Fluid flow control and Neutral and bonding conductors must be checked for
pressure regulation are accomplished by adapting a corrosion and loose connections. Where carbon track-
small needle valve to the joint or termination valve ing of sheath insulators is suspected based on manhole
assembly. About one quart of the fluid should be care- observations, the insulators must be cleaned and
fully flushed out to clear the valve assembly. The syringe inspected, and necessary repairs must be made before
is attached to the valve using a section of clear tygon recovering the insulator with an insulating, protective
tubing. All components involved in sampling must tape. Problems with the shield interrupts on extruded-
be thoroughly flushed to remove all air bubbles before dielectric cable joints most likely require replacing the
taking a final sample for test. However, gas bubbles that entire joint. The joints on buried cable circuits seldom
appear after extracting the sample must not be removed. required maintenance, with the exception of bonding
The bubbles represent dissolved gas that has come connections, and then, only when such connections are
out of solution, which is critical for accurate analysis of accessible. It is recommended that joint locations be
the sample. inspected periodically to avoid problems from external
sources, such as construction activities adjacent to
Explicit instructions for taking DGA samples are con- the easement.
tained in the ASTM Specification D-3613-98 (ASTM
2000). Guidance on interpreting the results of the dis- 15.3.7 Terminations
solved gas analysis can be found in IEEE Std. 1406 Terminations on underground transmission cable sys-
(IEEE 1998). tems require little maintenance. After the initial installa-
tion, bolt torque values should be checked once during
An extracted sample should be carefully packaged for the yearly maintenance inspection to ensure proper seat-
shipment to the testing laboratory, avoiding exposure to
sunlight when possible.
Figure 15-4 shows cable fluid sampling for DGA by the

EPOSS (EPRI Pressurized Oil Sampling) method devel-
oped by Detroit Edison Co. under EPRI sponsorship
and the syringe method.
15.3.6 Joints
Maintenance of underground transmission system joints
is performed in conjunction with cable and manhole
maintenance based on scheduled system inspections.
Maintenance of pipe cable joints, as covered in Section
15.3.3, involves repairing deteriorated or damaged cor-
rosion-protective coatings and repairing leaking or dam-
aged valves and access ports. Based on the joint location
and circuit loading, periodic x-rays and DGA sampling
are recommended to assess the condition of the cable
and joint for potential problems resulting from move-
ment or TMB damage associated with movement. Duct-
installed, self-contained, and extruded-dielectric cable
joints require periodic testing of the bonding and ground
systems, in addition to the maintenance of the corrosion-
protective coverings. DGA sampling should also be done
on any fluid-filled components. Single-conductor cir-
cuits operated in a cross-bonded configuration must
remain insulated from ground at the cross-bonding loca- Figure 15-4 Field sampling for dissolved gas
analysis tests (courtesy USi).
15-7
ing of components with gaskets. At other times, the A removed termination can be refurbished for use as
bolts should be retorqued only when found to be loose. a spare by replacing the seals and refilling with the rec-
Over-torquing, in the case of flat gaskets, results in con- ommended compound. Many solid dielectric termina-
tact between mating parts, which will lead to weeptype tions are equipped with monitoring systems and
leaks that cannot be stopped by tightening. The same fluid reservoirs.
applies to O-ring-type seals, where metal-to-metal con-
tact is initially made during installation. Further tight- During yearly inspections, the terminations should be
ening is not possible, and the leak can only be examined for damaged or broken sheds. If the damage is
eliminated by replacing the O-rings. Leaks developing extensive, the termination should be scheduled for
after years of service that cannot be stopped by tighten- replacement. The porcelain surface should also be exam-
ing require gasket and/or termination replacement, if ined for signs of surface tracking. Replacing the termi-
the leak is located at the cemented joint between the nations should be considered depending on the degree
metal flange and the porcelain. For HPFF cable termi- of tracking.
nations, replacement is accomplished by “freezing” the
riser pipe or, when the terminations are located at the Terminations in highly contaminated locations that
high point of the circuit, reducing the termination pres- have been coated with silicon grease to improve flash-
sure to zero and draining the fluid below the level of the over resistance should be scheduled for cleaning and
termination base plate. The termination can then be recoating based on the degree of contamination in the
lifted and all gaskets replaced. area. In such areas, terminations of the next higher rat-
ing are often used, and silicone coating is not required.
On three-conductor, self-contained cables where the Periodic high-pressure washing is normally sufficient to
spreaderhead contains a semi-stop in each of the riser maintain the integrity of the porcelain surface. Some
tubes, the bypass valves in the piping around the semi- utilities report that longer-lasting coatings such as sili-
stop can be closed and the termination drained without cone room temperature vulcanizing (RTV) coatings
affecting the system pressure. To change gaskets on a reduce flashovers and also reduce washing and grease-
single-conductor cable, the fluid feed to the termination replacement work.
must be closed and a section of the cable below the ter-
mination frozen. The same procedure is followed to 15.3.8 Pressurization System
repair leaking lead wipes at the base of the terminations. Pipe-type and self-contained cable systems are designed
to operate under positive fluid pressure at all times. On
A repaired termination is filled only after the termina- pipe-type circuits, pumping plants are used to maintain
tion has been vacuum-treated and drop-tested to ensure the fluid pressure within a preset operating range. When
the integrity of the repair. HPFF terminations can be the system pressure drops below or rises above the preset
filled separately from a fluid treatment unit, by melting range, fluid is introduced into or withdrawn from the cir-
the freeze or raising the pressure on the circuit. SCFF cuit. Though the pressurization unit is designed for auto-
terminations can be filled separately from a fluid treat- matic operation, periodic maintenance and testing are
ment unit, by opening the bypass valve on the spreader- required to ensure that all components are operating
head semi-stop or melting the freeze on the riser cable. properly. Older pressurization units were equipped with
Fluid should be flushed from the termination top until pressure recorders, which may require chart changes
acceptable electrical characteristics are met (i.e., dielec- once a week. During these regular visits, the operator
tric strength and/or dissipation factor). Section 15.5.2 should not only examine the charts for unusual pump or
provides more detail on the freezing procedure. relief valve operations but also visually inspect the area
for potential problems and arrange for necessary repairs.
Dielectric fluid samples should be taken from sampling Such inspections will only note the more obvious prob-
valves on the top of the termination, provided the valve lems. Units having fluid-filtering systems not equipped
is accessible without disconnecting from the overhead with differential pressure alarms to indicate when filters
bus or from the bottom valve, if so equipped, at the are clogged should have their filters changed periodi-
same time that samples are taken from splices, following cally, based on experience. Once a year, each unit on the
the same procedures. system should be scheduled for routine inspection and
component testing. Specific maintenance requirements
Leaks of filling compound on extruded-dielectric and testing procedures are detailed in the operating
terminations are not easily repaired in the field. When instructions for each unit, and these instructions should
problems develop on these terminations, the entire be adhered to for proper testing of all unit components.
termination must be removed and a new one installed.
15-8
Newer pressurizations units are equipped with remote for proper operation. Remote monitoring systems are
monitoring systems, which monitor most, if not all, of also available for HPGF circuits. GA (Gas Analysis)
the critical components of the pressurizing system. sampling should also be done on these circuits.
However, these systems do not provide visual inspec-
tions, which still need to be done periodically. SCFF cable circuits operate at pressure levels based on
the hoop stress of the cable’s sheath or the design of the
The following areas should be addressed during any sheath reinforcement. Maintenance procedures for these
pressurization unit inspection and testing program: systems vary depending on the type and location of the
• Exterior and interior portions of the unit should be pressurizing units. Most cable circuits rely on reservoirs
connected at the termination ends and at intermediate
visually inspected for damage or fluid leaks, and
locations as required to maintain positive fluid pressure
repairs made as required.
throughout the entire circuit under all operating condi-
• Pressure gauges, recorders, and switches should be tions. Nonreinforced, lead-sheathed constructions have
checked against a standard gauge for proper calibra- a maximum pressure rating on the order of 15–19 psi
tion, and switch trip points should be checked for (103–131 kPa).
proper pressure settings.
• All annunciator alarm points and remote alarms Reservoir types used include the gravity feed (Type CC),
should be functionally tested. which provides constant pressure based on the elevation
• Fluid pressurizing and circulation pumps should be of the reservoir with respect to the cable. Also, pressure
reservoirs (Type AC) maintain the pressure between 1
tested, including checking settings of weep relief,
and 15 psi (7 and 103 kPa) above atmospheric pressure
normal relief, and safety relief valves. Older units
depending on fluid content of the reservoir. The bal-
that use solenoid valves for pressure control should
anced-pressure reservoir (Type DC) gives a reservoir
be tested for proper, leak-free operation and hold-in
pressure that is relatively flat, with a pressure range and
timer settings that maintain desired pump-on times.
profile based on the auxiliary gas pressure and volume,
• Fluid storage fluid level should be checked and com- and is used where the pressure normally does not exceed
pared with the level-indicating device and previously 15 psi (103 kPa). Reinforced sheath constructions can be
recorded levels. The high-low fluid-level alarms, operated at pressures on the order of 220 psi (1510 kPa),
along with the nitrogen pressure alarms, should be but normally operate below 100 psi (690 kPa). Reser-
tested. On units where the fluid tank is maintained voirs used on these cables are of the pre-pressurized type,
under vacuum, as is the case for units associated with with factory-set pressure levels based on circuit length
SCFF submarine cable circuits and some older pipe-
type cable circuits, the vacuum and alarm systems
should be tested.
• Strainers should be cleaned, and where fluid-filtering
units are installed, the filters changed.
• Heating, ventilating, and dehumidifying equipment
should be checked for proper operation.
• Emergency power source and automatic transfer
switches, including loss-of-power alarms, should
be checked.
• Circuits incorporating forced-cooling equipment
should be tested following the manufacturer’s recom-
mended procedures.
Most high-pressure, gas-pressurized pipe-type cable cir-

cuits incorporate gas supply cabinets (see Figure 15-5)
for maintaining proper pressure levels. The cabinets
contain a pressure gauge (1), high-low pressure alarm
switches, nitrogen bottles (5), and a pressure regulator
(2). The bottles are not normally connected to the cir-
cuit (3), but are there in the event of a low-pressure
alarm. Once a year, the alarm circuits should be tested Figure 15-5 Nitrogen supply cabinet for HPGF cable
system (courtesy USi).
15-9
and profile. Of these reservoirs, only the Type CC reser- Fault location can generally be accomplished in about
voir has the outer shell vented to the atmosphere—a one working day. However, in some instances, particularly
condition that, in time, may lead to contamination of the with pipe-type cable, several days can be required to pin-
immersion fluid. This requires flushing the shell, filling it point the failure. The equipment used is a matter of
with new f luid, and recalibrating the reservoir. choice, but a suggested complement includes:
Section 9.4.3 provides additional details on reservoir • DC Insulation Tester and AC or DC Hi-Potential Test
construction.
Set. VLF test sets, at present, are used only on distribu-
tion circuits
Yearly maintenance includes inspecting reservoirs and
piping for corrosion, checking piping and valve panels • Digital Time Domain Reflectometry (TDR) with
for leaks, testing fluid-line insulators, and checking res- high-voltage coupling unit for measurement
ervoir fluid/pressure levels. High-low fluid or pressure • Arc Reflection
alarm contacts should be tested electrically and checked • Impulse Current
for accuracy of calibration. Reservoir fluid/pressure lev-
els can be calculated knowing the reservoir, air, and • Low voltage TDR
ground or sheath temperatures. In addition, the cable • High-Voltage Bridge
current loading must be known. It is also recommended • High-Voltage Impulse Fault Locator with fault burn-
that both temperatures and current loadings be rela-
ing capability
tively constant for a number of hours prior to taking
readings to avoid errors resulting from transient condi- • Directional Acoustic/Magnetic Detector
tions. The ideal condition is when the circuit has been • Cable tracing equipment and cable phase tracing
out of service for several days and the cable is at ambi- equipment
ent temperature. Most manufacturers furnish graphs
from which reservoir fluid and pressure levels can be These functions may all be incorporated in one truck or
read directly, knowing the above temperatures and cur- trailer-mounted test system.
rent loadings.
Proper preparations can greatly reduce fault location
15.4 ELECTRICAL FAULT LOCATION time. Therefore, it is important to have all pertinent
information readily available. This includes cable char-
15.4.1 Electrical Faults on Underground acteristics (conductor size, insulation type, voltage rat-
Transmission Circuits ing, etc.), and accurate cable maps showing manholes,
Electrical faults on underground transmission circuits splices, and cable lengths. All protective relay operations
are infrequent, as are failures in other elements of the at the time of the fault and digital recordings, if avail-
bulk power system. Even though most bulk power sys- able, should be provided. Previous TDR traces can be
tems are redundant, an underground failure seems to cause useful for comparative purposes. Anecdotal information
unusual concern until the fault is pinpointed and repair is usually not reliable, but, nonetheless, should be pro-
schedules can be made. This is due, in part, to the fact that vided. After an analysis of this information, actual fault
location of the failure in the overhead and station equipment locating can commence utilizing the recommended
is soon known, which allows repair schedules to be initiated methods as discussed below.
quickly. Fault location usually takes much longer for
underground cable circuits. 15.4.2 General Location
After normal safety procedures have been completed
An experienced and well equipped fault-location crew is and the circuit is ready for fault location, personnel and
necessary for promptly locating electrical faults on equipment should be available at both the test end and
underground transmission circuits. Major transmission- far end of the circuit to ensure that:
cable-using utilities have the required crews, expertise, • All aerial connections are removed from the termina-
and equipment. Several of these utilities, as well as pri- tions.
vate contractors, now offer fault-location service. If a
utility has underground transmission, but does not have
• Cathodic protection, if any, is disconnected.
the expertise and equipment to find faults, arrange- • Pipes are directly grounded with polarization cells out
ments should be made before trouble occurs so that of the ground path.
working relationships are well established and the fault- • Forced-cooling or fluid circulation pumps are off (but
locating utility or contractor knows the cable system. pressure should not be reduced).
15-10
• Any special bonding insulators are shorted and bond- 15.4.3 Pinpointing
ing transformers disconnected. Once prelocation has been completed, the fault can be
pinpointed utilizing a capacitor discharge (“thumper”)
Personnel should be available at both ends to make con- fault locator at the test end and a field crew(s) with
nections as necessary for testing. directional acoustical/magnetic detector. In areas where
the pipe or cable sheath is available, the sheath-drop
The circuit is now ready for prelocation of the fault. method can also be utilized to pinpoint the fault. It is
Proper prelocation is essential to minimize not only the important that clear communication is available
time to pin-point the fault but also the potential stress between the field crew(s) and personnel at both ends of
to the cables. The following procedure should be used: the feeder.
1. Perform an insulation resistance or hi-potential test
(excluding extruded-dielectric cables) on all phases, Typically, the fault can be located with good accuracy
and identify the faulted phase(s) and breakdown volt- from the surface. However, excavation is necessary to
ages. precisely pinpoint faults on many installations, particu-
larly pipe-type cable.
2. If necessary, reduce the breakdown voltage of the fault
to levels within the range of your equipment by utiliz-
15.4.4 Variations
ing the fault burn mode of the test equipment (not for
use with extruded-dielectric cables). Sections 15.4.2 and 15.4.3 apply to all transmission sys-
tems, although some precautions should be noted for
3. Connect a faulted phase to a good phase on the far end
systems other than HPFF.
with a suitable “jumper test lead,” and perform a high-
voltage bridge measurement. Repeat with the other
With HPGF systems, the initial TDR reading (Step 4
phases as necessary.
above) is usually not productive, since less carbonized
4. Utilizing low-voltage TDR, compare a faulted phase to material stays in the fault than with fluid-filled systems.
a good phase and to previous TDR traces, if available. However, the high-voltage TDR (Step 5 above) should
If necessary, perform TDR measurements with the far produce results.
end of the cable open circuit and grounded to accu-
rately verify the end of the cable. SCFF cable is treated like any other lead-covered cable.
5. Utilize the TDR equipment in conjunction with the
high-voltage coupling unit to perform arc reflection, Extruded-dielectric cable systems require a modified pro-
surge pulse, or impulse current measurements on the cedure. DC hi-potential testing should not be employed
faulted phase(s). because it can degrade otherwise healthy insulation.
6. Analyze the results of these measurements, and if
there is a significant uncertainty in the results, repeat Faults on SF 6, systems cannot be located in the usual
these measurements from the far end of the feeder. way, because the breakdown voltage remains high, and
the fault impedance cannot be reduced; therefore, the
Figure 15-6 shows the configuration for performing a user should proceed directly to the thumper and acous-
high-voltage bridge measurement on a faulted cable. tic pickup with several listeners stationed along the cir-
Figure 15-7 shows the configuration for performing cuit. This is truly “hunt-and-peck” fault location.
TDR measurements by arc reflection, surge pulse, or
impulse current.
Figure 15-6 Bridge measurement configuration (courtesy Figure 15-7 Configuration for TDR measurements by arc
Advanced Testing Systems Inc.). reflection, surge pulse, or impulse current. (courtesy
Advanced Testing Systems Inc.)
15-11
15.5 ELECTRICAL FAILURE REPAIR repair may preclude future removal of the cable without
Repair procedures for electrical failures on transmission reopening the duct. As with all cable faults, the repair
cables depend to a great extent on cable type, failure method depends on the extent of cable damage and,
location, and, in the case of pipe cable, the number of most importantly, the condition of the cable and cable
phases damaged. Since repair methods are cable-type shielding adjacent to the fault and the economics of the
dependent, each cable type is addressed individually. repair options. Cable shield damage has occurred
remote from the cable failure location. Unless the cable
Repair times can range from a few days to many months is of a water-block construction, water present at the
(if new cable must be manufactured or replacement ter- time of failure may enter the conductor as the cable
minations ordered). Typical repair time is 2–4 weeks, cools, migrating back along the conductor strands. The
provided repair components consisting of cable, splicing water must either be removed or the cable must be cut
material, joint casings and, terminations, if required, are back until dry cable is located before a repair is made.
available from stock. If cable having the same or larger Even by cutting back to what appears to be dry conduc-
conductor size and voltage rating or higher is not avail- tor, there is no guarantee that all traces of moisture have
able from stock or another utility, the cable has to be been removed. A piece-out is then installed, or the cable
ordered from a manufacturer, which can take weeks or may even have to be replaced manhole-to-manhole.
months if the manufacturer is not making similar cable
of equal or higher voltage. New terminations can take Unlike joints on laminar-dielectric cables, which can be
up to 6 months if not available from another utility. disassembled in the event of a failure and the cable
Figure 15.8 shows a single-phase failure in a 345-kV reused using a normal or extended length connector,
HPFF pipe-type cable. Note the damage on the adja- pre-molded or hand-taped joints, where the tapes have
cent phase. amalgamated, may be difficult to re-enter.
A fluid-filled cable failure near or adjacent to the sub- With taped or premolded slip-on joints, it may be possi-
station can result in internal damage to the terminations ble to break down the joint and incorporate a portion of
due to high-pressure waves. Figure 15-9 shows damage the remaining cable in the replacement joint. The
to a termination semi-stop gland at the base of a termi- removal of all or a portion of the joint in the manhole
nation that was located 1000 ft from the cable failure. may not leave sufficient cable remaining to make a joint.
In such cases, it will be necessary to reshape the remain-
15.5.1 Extruded-Dielectric ing cable or extend the manhole. Some utilities loop the
cable around the manhole on the original installation to
Extruded-dielectric cables are installed in ducts or
provide spare cable for future repairs.
directly buried. Repair procedures are, therefore, a func-
tion of the installation method and failure location.
The types of repair joints for extruded-dielectric cable
With a multiple cable duct bank with installed circuits,
can be hand-taped, hand-taped and molded, molded
where the ducts are not accessible, a cable failure
using preformed sections, or premolded slip-on. Where
between manholes is repaired by removing the failed
section and installing a replacement cable, provided the
damaged cable can be removed from the duct. There
have been cases where faulted cables have fused to the
polyvinyl chloride (PVC) or polyethylene ducts, making
removal extremely difficult. The alternative is to open
the duct bank at the fault location and make the neces-
sary repairs. Depending on the duct size, this method of
Figure 15-8 Pipe-type cable electrical failures Figure 15-9 Failure of termination semistop assembly
(courtesy USi). (courtesy USi).
15-12
traces of moisture may be present, the field-molded With the fault location pinpointed through the use of
design should be avoided to prevent porosity from fault-locating equipment, or the discovery of a fluid
developing in the insulation adjacent to the conductor leak resulting from pipe burn-through or external
during high-temperature curing of the insulating mate- mechanical damage, the pipe must be exposed and
rial. For repair joints, the hand-taped splice can be opened for inspection. The extent of the damage must
designed to include an elongated connector or two con- be assessed before repair procedures can be established
nectors and a section of conductor when the fault dam- and material requirements determined. Where a fluid
age is minimal. Two joints and a section of cable are leak is present, the leak must first be stopped or the rate
required when using molded or premolded designs. reduced before proceeding with repairs. This is normally
accomplished using a split repair clamp. Before opening
Cable or joint failures within the confines of the man- the pipe, the fluid in the fault area must be isolated from
holes are repaired based on the amount of remaining the remainder of the circuit by freezing on one or both
undamaged cable. If failure damage is minimal, a hand- sides of the fault. Freezing the fluid requires reducing its
taped joint using either an elongated connector or two temperature to a level low enough to solidify a fluid slug
connectors and a piece of conductor is used. Otherwise, capable of withstanding a pressure differential in the
a two-joint repair, with the replacement cable looping order of 60 psi (414 kPa) depending on the elevation of
the manhole, is used. the circuit. Higher differential pressures may require
multiple freezes. Low-viscosity fluids like DF 100
Repair procedures for directly buried cables depend on require temperatures on the order of -240°F (-151°C).
the extent of damage; normally a two-joint repair is
required. The exception with buried cables occurs when Faults located near the termination end of a circuit may
the cable route has been overbuilt and the cables are no be frozen on the line side of the fault and the fluid
longer accessible. In this situation, the entire circuit may drained from the termination back to the fault. Faults
have to be replaced along a different route. located at or adjacent to the highest point of a circuit
can normally be repaired without freezing, provided the
Termination Failures hydraulic fluid head throughout the remainder of the
Depending on the type of termination failure, it may be circuit is adequate to maintain positive pressure at all
possible to remove the termination, clear the failure, locations to exclude water ingress. Where this is the case,
and install a new termination. Termination failures that the fluid is drained from the fault area and repairs are
result in cable damage require replacing the termination made. Normally, however, a freeze is required on both
and the termination riser cable back to a convenient sides of the fault. On single circuits pressurized from one
jointing location, such as a manhole or a buried splice at end only, a temporary fluid source is required on the
the base of the structure. The termination is constructed nonpressurized side. This source must be capable of pro-
in accordance with the manufacturer’s instructions, and viding adequate fluid for flushing as well as maintaining
the joint is based on the user’s standard practices. positive pressure for the duration of the repair and
allowing for expansion and contraction of the fluid. One
15.5.2 Pipe-Type method that has been used is to install a temporary
Fluid loss resulting from either electrical failure or bypass around the proposed freeze location. This can be
mechanical damage mandates immediate attention with accomplished by installing a Mueller-type fitting com-
the urgency dictated by the rate of fluid loss. Where the plete with the valve on the pipe at a location that is away
loss rate is severe (greater than 10 gal/hr (38 l/hr), the from the cables. An X-ray of the pipe may be required to
system pressure may be lowered, if the feeder can be determine the location of the cables. The bypass can also
taken out of service, to reduce the leak rate while main- be used to flush fluid from the failure area and maintain
taining positive pressure throughout the entire circuit, fluid pressure on the nonpressurized side during the
including terminations, to exclude water entrance. Pres- repair after the freeze has been established. It is recom-
surization units are normally equipped with reduced- mended that, prior to establishing the freezes, an attempt
pressure provisions specifically designed for operation be made to remove any contaminated fluid from the
under leak situations or repair process. Units not so fault area by flushing from both ends of the circuit.
equipped require that the relief valves be reset manually
to reduced-pressure levels and the pump operating Pipe Freezing
mode be changed from “Auto” to “Manual.” The oper- Freeze pits, located approximately 50 ft (15 m) from the
ating instructions for the particular unit should be work area, are dimensioned on the basis of fluid viscos-
referred to for reduced-pressure operation. ity. Higher-viscosity fluids such as the old Sun #6 (no
longer available) required a relatively short freeze
length, on the order of at least 3 ft (0.9 m). Low-
15-13
viscosity fluids such as alkylbenzene/polybutenes used Two coils are recommended so that, on completion of
in forced-cooled systems require longer freeze lengths of the repair, the freeze coils closest to the repair can be
up to 10 ft (3 m) depending on circuit profile. Although melted, and any moisture that has collected on the
freezes had been made using dry ice and acetone in the freeze face removed during the vacuum treatment of the
past on Sun #6 fluid, liquid nitrogen is used exclusively repair area prior to fluid filling.
for freezes today and is the only medium that will freeze
low-viscosity fluids. The fluid can be frozen using cop- When rag barriers are used to seal the pipe ends, along
per tubing or specially made freeze jackets. When using with a dry air feed into the pipe instead of sealing the
copper tubing to freeze pipe-filling fluids, two freeze pipe with neoprene barriers with a dry air feed, small
coils are normally recommended on each side of the amounts of moisture may penetrate through the rag
fault. Each coil is approximately 15 in. (38 cm) in barrier into the pipe ends and freeze on the freeze face.
length, formed from closely spaced turns of soft copper The two-coil freeze approach has greatly reduced vac-
tubing placed on the pipe after all pipe-coating material uum time in case moisture has entered the pipe during
has been removed. A bedding of copper mesh braid can the repair operation.
be applied over the pipe before the coil is applied to pro-
vide better heat transfer. A 6-in. (15-cm) spacing is nor- Liquid nitrogen is permitted to completely fill the coils
mally used between coils. Thermocouples, placed at the or jackets until droplets of liquid are discharged from
center of the freeze between the coils and at either end the vent along with vapor. Nitrogen flow is then reduced
of the freeze, are used to monitor freeze temperatures until only vapor is discharged. This condition is main-
and determine flow rates for the liquid nitrogen. Several tained until the temperature at the outer edges of
layers of urethane foam or fiberglass insulation cover the freeze reaches -50°C for high-viscosity fluid or
the entire freeze coil area. An overall covering of a mois- -150°C for low-viscosity fluid. This is typically a period
ture-impervious material is recommended to prevent of 3–6 hours, but longer periods are required when fluid
moisture from condensing and freezing in the insula- is flowing due to circuit-cooling or eddy-current fluid
tion, a condition that will lower the insulating value of flow resulting from pipe elevation changes and fluid
the insulation. Figure 15-10 depicts a typical fluid-freez- temperature differences in the pipe section being frozen
ing arrangement using copper tubing. or due to a leak. The preferred method of checking the
integrity of the freeze is by increasing the pressure on
Figure 15-10 Typical fluid freezing arrangement using copper tubing (courtesy USi).
15-14
one side of the freeze and monitoring the pressure on joint opening, shifting the connector may necessitate
the other side. If the pressure does not change, the freeze moving the joint reducer. Otherwise, the joint is con-
has been established. An alternate procedure is to install structed in accordance with the original design.
a valve and pressure gauge between the two freezes. A
freeze is normally achieved when the pressure begins to Valves must be installed on the top and bottom of repair
drop below the initial level. If this method cannot be sleeves for pressure testing and evacuation of the section
used, a small hole (1/8 in., 3 mm diameter) can be drilled between freezes after repairs are completed. If more
in the top of the pipe. Even though a freeze has been than one freeze coil has been used, the flow of nitrogen
achieved on both sides, there may be a momentary spurt to the freeze coils closest to the repair is stopped to per-
of fluid from the hole before all flow stops. It is mit the front face of the freeze to melt, which releases
recommended to wait at least 15 minutes before testing any moisture that became trapped on the freeze face,
the freeze. along with a portion of the contaminated fluid. This
material is drained from the pipe, when possible, before
With the integrity of the freeze confirmed, fluid is vacuum is applied.
drained from the area and a 120° window, 12–18 in. (30–
45 cm) long, is removed from the top of the pipe by The repaired section should be evacuated to 150
grinding or cutting with a circular saw. This provides an microns and evacuation continued for an additional 8
opening through which the failure location can be hours before conducting a 1-hour vacuum drop test.
confirmed and the extent of damage assessed. Based on Maximum pressure rise is not to exceed 100 microns
the information obtained, a repair procedure is estab- above the starting value. Using this procedure, accept-
lished, the repair pit is completed, and repair materials able vacuum levels can be achieved prior to filling. The
are procured. When making the window, care must be evacuated section can be filled from an external fluid
taken to minimize the amount of metal shavings that source, from the bypass if used, or by permitting one of
enter the pipe. Any material that falls into the pipe must the freezes to melt. In the latter case, the fluid pressure
be removed. on the backside of the freeze being melted should be
reduced to a minimum to prevent the freeze plug from
Repair Procedures moving and damaging the cables when it breaks free of
Cable failures resulting from external mechanical dam- the pipe. Fluid is drained from the repaired section until
age seldom result in extensive cable damage. Depending the dielectric strength of the drained fluid is acceptable.
on the extent of damage, the cable can be repaired with The second freeze is then melted and fluid flushed from
a splice using an elongated connector or two standard the opposite direction until acceptable dielectric
connectors and a section of conductor. The splice is strength and power factor values are obtained. To facili-
constructed with crepe paper, hard paper, or LPP tape, tate a rapid thaw, all insulation should be removed from
depending on the cable type used on the circuit. When the freeze coils. An external heat source may be used,
the failure is of unknown origin or cable damage is provided that it is gradually applied.
extensive, a section of cable containing the failure or
damaged area is normally removed for failure analysis. With the repair completed, the repair sleeves and
The repair is made using two normal joints and a sec- exposed pipe in the freeze pits are coated with a corro-
tion of replacement cable. The completed repair is sion-protective material, which is compatible with the
closed using either split reducers and split repair sleeves existing pipe coating.
and couplings (single cable repair), or when all three
phases have been removed or cut, standard reducers and In all cases where the pressure has been reduced to 50
repair sleeves. psi (345 kPa) or lower for more than a day, it is recom-
mended that the line be repressurized slowly at a rate on
Joint failures or cable damage within joint enclosures the order of 25 psi/hour (173 kPa/hour) and held at full
resulting from thermomechanical bending (TMB), pressure for at least 24 hours before testing or voltage is
which cannot be repaired due to space limitations applied. It is also recommended that the terminations
within the manhole, are repaired by constructing tempo- and joints be checked for gas and bled, if needed, before
rary splice pits adjacent to the manhole and replacing applying the 24-hour soak test.
the section of cable through the manhole. When damage
is restricted to one side of the joint, only the cable on the In the event that a section of new cable has been
damaged side needs to be replaced. The existing splice is installed, it is recommended that the line be repressur-
torn down, and the connector is shifted by half its ized in steps of 25 psi (173 kPa) and held for a minimum
length toward the good end, to permit pressing the new of 8 hours between steps. The circuit should be at rated
connector on undisturbed conductor. Depending on the
15-15
pressure for at least 24 hours before testing or voltage to be replaced. In lieu of cutting the cable, the cable
is applied. could be frozen until the fault can be exposed and the
loss of fluid stopped. Here also the unpressurized section
(Note: It is recommended that the cable manufacturer may be lost due to moisture entrance.
be consulted regarding pressurization steps and soak
periods.) Failures between manholes are repaired either by
replacing the entire manhole-to-manhole section with
Termination Failure new cable, or by opening the duct bank in the fault area
Termination failures on pipe-type cable circuits normally and making the necessary repairs. The latter method is
require replacing the cable back to the trifurcator, where seldom used due to the difficulty of the repair, especially
a single-phase joint is made. If the trifurcator is not large when the cable occupies an inside duct. When a spare
enough to accept a joint, either the enclosure must be duct is available, restoration time can be greatly reduced
enlarged to permit a splice to be made, or a buried joint by installing the replacement cable in the spare duct and
using split reducers and sleeves must be installed adja- splicing the cable before removing the damaged section.
cent to the trifurcator enclosure. In the case of an above- Without a spare duct, the cable must be removed, the
ground spreaderhead, a buried joint is required at the duct must be inspected, and the necessary repairs must
bottom of the 90° pipe bend, and only the failed phase be made before installing the replacement cable.
needs to be replaced, provided the failed cable can be
removed. Otherwise all three terminations and cable Replacement cable is spliced-in by disassembling the
have to be replaced. The spreaderhead has to be opened existing joint and splicing-in the new cable. Depending
to assist in removing and replacing the failed cable. on the connector type (soldered or indented), it may be
necessary to shift the connector centerline to expose
In some cases a new joint can be avoided. If the cable in undisturbed conductor to install the new connector. On
the termination is not too badly damaged, a few layers three-conductor stop-joints, only one side of the joint
of cable tape can be removed, new layers applied, and needs to be disassembled. However, depending on the
the termination rebuilt. joint design, it may be necessary to hydraulically isolate
the opposite side at the joint semistop and drain the
15.5.3 Self-Contained Fluid-Filled joint casing.
SCFF cables have traditionally been installed in duct
banks to provide ease of access in the event of problems, Repair methods for failures occurring within a manhole
and because duct systems are more practical to install in depend on cable type (single- or three-conductor), joint
congested areas. As a result, electrical failures or design (normal or full-stop), failure location, and cable
mechanical damage occurring between manholes may and space availability. If sufficient undamaged cable is
require replacing an entire manhole-to-manhole cable available, and space permits, it may be possible to
section. This is not to say that repairs cannot be made replace a failed joint with an elongated joint. On three-
by opening the duct bank at the failure location, repair- conductor cable, repairs can be made using elongated
ing the cable, and reconstructing the duct bank over the connectors and an extended splice of a design similar to
repair. However, depending on the diameter of the the existing design. Where conductor damage is more
repair in relation to the inside diameter of the duct sys- extensive, two connectors and a short piece of conduc-
tem, this method may prevent future cable removal. tor can be used to repair each phase conductor. In either
case, an elongated repair requires the use of special
Fluid loss should be stopped or the flow rate reduced to sleeves. The alternative is to use two joints and a section
conserve reservoir fluid until repair operations can be of connecting cable, which loops the manhole. On sin-
undertaken. In accessible areas, temporary repair clamps gle-conductor cables, where fluid flows through the core
can be applied over the failure area after first removing of the conductor, special connectors containing fluid-
any protective covering applied over the metallic sheath. control valves have to be constructed if an elongated
Where damage is extensive, preventing the use of a repair connector was used for repair. For this reason, the two-
clamp, the cable should be cut clear of the fault area and joint repair method is preferable to an elongated splice.
temporary caps applied. For fluid loss originating from a
failure within the duct where access to the cable is not Where damage is extensive, a section of spare cable,
possible, control is best accomplished by cutting and which loops the manhole, can be inserted using the
capping the cable in adjacent manholes. However, with appropriate joints, one of which will be a normal joint.
this procedure, the faulted section is no longer protected The procedure of looping the manhole requires modify-
against moisture ingress, and the section most likely has ing the supporting structure within the manhole. In
15-16
addition, temporary reservoirs are required to maintain An equilibrium gas content of 0.05% or lower is consid-
pressure on the spare cable while making the first joint. ered acceptable for service.
Directly buried cables are considerably easier to repair 15.6 FAULT REPAIR MATERIAL
once the fault is located. Cables are exposed over a suffi-
cient distance to properly space them for repair. The 15.6.1 Extruded-Dielectric Cable
repair is made by splicing-in a section of spare cable
In North America, extruded-dielectric cables are gener-
using two normal joints. When cutting back the failed
ally either duct installed or directly buried. With duct
or damaged cable, it is recommended that the conductor
installed cables, spare replacement cable must be equal
shield and insulation immediately surrounding the fluid
to the longest manhole section length in the circuit.
channel on single-conductor cable, or the filler material
Spare splice materials may have a limited shelf life, and
and surrounding insulation on three-conductor cable, be
as a result, records must be kept and periodically
tested for moisture using a petroleum-based compound
reviewed to prevent a problem when the material is
heated to approximately 150°C. “Sizzling” indicates
needed. With transmission-class XLPE cables, when at
moisture. If moisture is present, the cable must be cut
all possible, the spare material should be from the same
back until dry cable is located. A review of the plan and
manufacturer as the originally installed cable system.
profile drawings for low points in the failure area pro-
Fault repair materials include:
vide an indication of how far water may have traveled.
• Cable of required conductor sizes and voltages
A termination failure on self-contained cable normally • Splice kits of required conductor sizes and voltages
requires replacing the termination riser cable as well as
the termination. On three-conductor, duct-installed
• Terminations for each voltage class
cable, this generally means replacing the cable back to • Fluid for terminations, if fluid filled
the first joint or installing a new manhole and replacing
the cable back to that location. All three terminations 15.6.2 Pipe-type Cable
have to be reconstructed. With buried cable, a buried In the past, when utilities had few high-voltage under-
joint can normally be installed near the termination ground transmission cable systems, fault repair material
structure. On single-conductor cable, only the cable was limited to reels of cable for the longest manhole sec-
associated with the termination failure needs to be tion length, several splice kits complete with joint cas-
replaced; this can be accomplished with a joint near the ings, and several spare terminations complete with paper
base of the structure. rolls for the voltage and conductor size of the circuits.
Unlike pipe-type cables, where an adequate supply of The spare cable, normally having a temporary lead
reserve fluid is available from the pressurization unit, sheath for protection during storage, was based on the
the liquid for self-contained cables is supplied by reser- longest section length. Due to drainage of the impreg-
voirs having limited reserves. In the event of a leak, res- nating fluid, the reels should be periodically rotated
ervoirs must be continuously refilled until repairs are one-half turn, reversing rotation direction each time to
made or the cable is sealed. Where a portable treating prevent the fluid from traveling under the temporary
unit is available, the unit can be used to supply fluid lead sheath towards the cable ends. In addition, the
directly to the damaged cable until repairs are com- cables should be stored indoors, out of the weather, to
pleted. Reservoirs that have either lost all fluid or have limit temperature excursions.
become contaminated with gas from the cable core must
be re-treated and calibrated before being placed back in Splice kits, containing fluid-impregnated insulating tape
service. Also, sections of undamaged cable that are con- in sealed cans, should be good for many years, provided
taminated with gas from the failure must be flushed the cans do not rust and develop leaks—in which case,
until the fluid meets the required electrical characteris- the tape should not be used.
tics. The gas content of the section must also be checked
before splicing-in replacement cable or reconnecting res- Today, utilities have many circuits with different con-
ervoirs. The section under test must be isolated from all ductor sizes and voltage ratings. This becomes a major
pressure sources, and the absolute pressure change must problem when addressing repair material requirements.
be noted when a measured quantity of fluid is with- Recently, utilities have started substituting other cable
drawn. With this data and knowledge of the total fluid sizes and, in some cases voltage levels, to make repairs,
content of the section, including joints and termina- provided the replacement cable fits into the existing pipe
tions, the gas content of the section can be calculated. and the conductor size is equal or larger than the cable
15-17
being repaired. For example, a 2500 kcmil (1267 mm2) cally read and compared with pressure/temperature
138-kV full-wall, kraft paper-insulated cable installed in charts for that specific reel to determine if there is a leak
an 8-5/8 in. (21.9 cm) diameter pipe could be repaired in the cable or reservoir system. Indoor storage of the
using the same conductor size 345-kV cable insulated reels greatly reduces pressure excursion.
with 0.600 in. (15.24 mm) of LPP insulation. If the
replacement cable had a larger conductor but would fit As with pipe-type cable splice kits, the insulating mate-
into the pipe, only transition connectors would be rial is contained in fluid-filled cans and, provided that
required, along with a revised splice design for the the cans remain leak-free, should be good for many
respective cables. years. Fault repair materials include:
• Cable of required conductor sizes and voltages
Substituting cable sizes and voltages could greatly
reduce the spare cable requirements. Fault repair materi- • Splice kits
als include: • Terminations for each voltage class
• Cable of required conductor sizes and voltages • Reservoirs
• Splice kits • Dielectric fluid
• Reducers and joint sleeves • Fluid treatment equipment, batch-type unit
• Normal and Transition connectors, Indent and FM
type depending on voltage 15.7 LEAK MANAGEMENT TECHNIQUES FOR
• Terminations for each voltage class FLUID-FILLED FEEDERS
• Dielectric fluid. Large quantities may have to be pur- 15.7.1 Background

chased. Fluid is shipped from the supplier in tank
Dielectric fluid is the main component of high-pressure,
trucks under a blanket of nitrogen and does not
fluid-filled, pipe-type, and self-contained feeders. Leaks
require treatment prior to filling the circuit.
have been a major concern since widespread use of these
• Fluid-treating equipment, if fluid is not purchased. feeders began in the 1940s. The utility experience to date
Batch-type units for small quantities or trailer- indicates that the cable itself, which was originally con-
mounted units capable of treating and filtering on the sidered to have a lifetime of 40 years, generally does not
order of 10 gal (37.58 l) per minute under a vacuum show significant signs of aging. On the other hand,
of less than 300 microns. The treated fluid is pumped localized deterioration of the steel pipe, which houses
directly into the circuit. the cables, has resulted in dielectric fluid leaks. Dielec-
tric fluid leaks, as well as being an environmental con-
15.6.3 Self-Contained Fluid-Filled Cable cern, could also affect the integrity and useful life of the
SCFF cables are generally duct installed with splices cable system. In case of a large leak, the pumping plant
located in manholes. As a result, the reel lengths of may not be able to maintain the required operating pres-
spare replacement cable must be equal to the longest sure. Loss of pressure may result in gas evolution in the
manhole section in the circuit. If the circuit consists of terminations and elsewhere along the feeder, causing
one three-conductor cable, it is possible to replace only electrical breakdown and failure. Loss of pressure may
the failed portion of the cable by breaking into the duct also result in the intrusion of water into the dielectric
at two locations, one on either side of the fault and fluid at a leak site, which otherwise should be free of
removing the failed section. Before any repairs are moisture. Moisture accelerates the aging process and
made, the dielectric fluid must be flushed from both reduces the dielectric strength of the fluid.
cable ends; the fluid tested for acceptable levels of power
factor, dielectric strength, and moisture; and DGA con- In general, feeder leaks may result in feeder outages,
ducted on the removed fluid. This repair method and are costly to pinpoint and repair. To address these
requires only a short piece of cable and two repair concerns, fluid-filled feeders should include a leak man-
splices. In addition, dielectric fluid and equipment for agement system consisting of the following elements:
treating the fluid are required. • Leak Detection
Spare cable must be maintained under a small positive • Leak Location
fluid pressure at all times to prevent moisture or air • Leak Prevention
ingress. Fluid is supplied to the cable by pressurized res-
ervoirs mounted within the reel. The reservoirs are These methods are described in the following sections.
equipped with pressure gauges, which must be periodi-
15-18
15.7.2 Causes of Leaks vations and, if warranted, leaks may be verified by pres-
The main causes of pipe leaks are corrosion, contractor sure-drop or other applicable tests. Detecting leaks by
damage, and water impingement. The water impinge- this method depends largely on the experience of the
ment occurs when a water main, in proximity to the observer. To help the process, utilities can install pump-
feeder pipe, develops a leak and causes the sand to act ing plants that are equipped with a leak warning system
as a sand-blasting agent and to drill a hole in the feeder that monitors the tank level and pump operation and
pipe. A number of feeder leaks associated with water compares it to the historical performance of the system
impingement have been experienced by utilities. In such and alarms abnormal conditions. More sophisticated
cases, the feeder fluid should be tested for moisture or systems are described in the following section.
free water before proceeding with the repair process. On-Line Leak Detection
In 1986, EPRI funded the Dynamic Rating and Under-
Corrosion leaks usually start at a low rate, and over ground Monitoring System (DRUMS) program, which
time, if not repaired, could become larger. However, was designed to improve HPFF cable system operation
leaks caused by the contractor damage or water and solve cable problems through a distributed system
impingement are usually large and could adversely of data acquisition equipment monitoring a variety of
affect the feeder pressure. cable system parameters in real time and communicat-
ing data to a central processor (Engelhardt and Purn-
15.7.3 Leak Detection Techniques hagen 1991).
Rapid detection of a fluid leak reduces the amount of
fluid loss. Leak detection is especially important for Consolidated Edison Company of New York, Inc. (Con
HPFF cables because of the large quantity of fluid Edison) continued the funding of this project by focus-
involved. A number of leak detection methods are coming the research on the development and demonstration
mercially available or are being developed. Some of of an on-line leak detection system for force-cooled
these techniques, applicable to fluid-filled feeders, are feeders. Results of this work successfully culminated in
described in the following section. The leak detection the development of a system with sensitivity for detect-
system for HPGF feeders relies on the monitoring and ing leaks down to 0.05% of total feeder volume, with a
alarming of the gas pressure. A low-pressure alarm minimum leak rate of 1 gal/hr (3.8 l/hr). The system has
could be set and monitored as part of the nitrogen cabi- subsequently been installed on several major transmis-
nets, which are installed at both ends of the feeder. The sion feeders and successfully detected leaks under actual
termination low-pressure alarm could also complement field conditions. The on-line leak detection system con-
this and provide a backup alarm. tinuously performs a dynamic mass balance everywhere
along the cable system. Leak detection is accomplished
Conventional Leak Detection
when there is a deficit between predicted fluid entering
Conventional leak detection is based on one or all of the
the cable system and the corresponding measured quan-
following indications:
tity. Figure 15-11 represents an actual leak alarm on a
• Low-reservoir-level alarm 345-kV feeder. As shown, the separation between the
• Frequent-pumping alarm
• Low-pressure alarm
These alarms are only sensitive to gross leaks. A large

fluid loss may be required to set off a low-reservoir-level
alarm. A typical frequent-pumping alarm installed on a
system containing 100,000 gal (378,500 l) may require a
leak of about 100 gal/hr (379 l/hr) to trigger it. A low-
pressure alarm could require a leak in excess of the
pump flow rate to be activated. Furthermore, if the
cable is in a heating cycle due to loading or earth ambi-
ent changes, then the leak tends to be masked and sensi-
tivities could be reduced. Leaks of lower rates are
usually detected by inspecting the pumping plant charts
weekly, or more frequently if warranted. Experienced
personnel can often flag suspicious behavior on a cable
system. This is followed by more frequent chart obser- Figure 15-11 Online leak detection alarm (courtesy of USi).
15-19
two graphs, which represents the total volume of fluid pipe and soil temperatures, fluid pressure, and conduc-
leaked, becomes larger as the leak continues. tor current. The remaining two algorithms mimic opera-
tion of the pressurizing pump and the pressure relief
The basic hardware components of an on-line leak valve that keep the dielectric fluid pressure within pre-
detection system consist of the central processing unit scribed limits. All three algorithms use the same data
(CPU), the remote telemetry units (RTUs), and the obtained from the sensors. Since real-life data is very
flow-metering unit (FMU). The CPU gathers analog “noisy,” it requires extensive preprocessing. Systems
data from local inputs and remote data from the RTUs installed at Con Edison on a static-pressure feeder and
distributed throughout the system. The CPU performs at Hydro One have shown that, under actual field opera-
dynamic modeling and leak detection/location function, the system is capable of continuously monitoring
tions, stores this information, and provides an interface the feeder without creating false alarms. Simulated leak
to a remote user access workstation. RTUs located tests have been successfully carried out. However, more
throughout the system gather analog data, which is data, in particular detection of actual leaks, is needed to
made available to the CPU upon request. Local and establish the true reliability and accuracy of the system.
remote leak alarms are initiated by the CPU. An FMU
is located between the pumping plant and the cable sys- 15.7.4 Leak Location Techniques
tem. Its principal purpose is to accurately meter the flow Following the confirmation of a leak on a feeder, the
in and out of the cable system, but it also supplies tem- location of the leak needs to be rapidly pinpointed and
perature and pressure information to the CPU. Its main repairs carried out. Considering that a feeder can be as
component is a precision bidirectional positive displace- long as 20 miles (32 km), there are two types of tech-
ment flow meter, which has been extensively calibrated. niques that can be employed to pinpoint the location of
A current metering system provides an analog signal a leak: sectionalizing and pinpointing. The sectionaliz-
proportional to cable load current to the CPU. ing methods can be employed to narrow the leak
location between two points along the feeder length.
A PC-based workstation can be used to access and pro- The pinpointing methods are used to accurately locate
cess information available on the CPU by calling it over the leak.
dial-up lines, leased lines, fiber optic, or any other com-
munication link. Its various functions are monitoring in Sectionalizing Methods
real time, and consist of unloading data, database man- Routine Inspection
agement, graphic and text displays, and graphic and text In the event of a leak (especially a large leak), the crew
printout. To accomplish the accuracy and reliability of should drive along the route in search of new construc-
the system, extensive and accurate data must be gath- tion, lift manhole covers at joints, and also lift the cov-
ered in real time and processed. This requires an exten- ers of neighboring utilities. A review of cathodic
sive investment in reliable communication systems, protection surveys may also provide useful location
precision instrumentation, and maintenance. information by flagging areas where the pipe coating is
suspected to be poor.
Development of an alternative, on-line leak detection
system that may rely on limited instrumentation and Freezing
data requirement is being pursued by EPRI and Con Leaks not located by obvious inspection techniques
Edison. In addition, references (Anders et al. 1999; have historically been located by freezing, using a binary
Anders et al. 2006) describe a novel, neural-network- search technique. First, a freeze is applied at the center
based method. This method uses three different markers point of the de-energized cable, and then a pressure-
for leak detection to characterize the operation of a drop test is performed on either side of the freeze to
pipe-type cable system. These are: (1) temperature- determine which section is leaking. Then a freeze is
pressure marker that relates the changes in the liquid applied at the midpoint of the leaking section, and so
pressure to the temperature changes, (2) pump opera- on, until the leak has been isolated to a small enough
tion marker that counts the number of the pressurizing section that can be bar-holed, excavated, or pinpointed
pump operations, and (3) valve operation marker that using techniques described in the following subsection.
monitors the operation of the pressure relief valve. The The freeze method requires a feeder outage, is very time-
main algorithm uses a neural network model that learns consuming, and requires multiple excavations to install
the relationship between the four measured quantities: the freeze jacket on the pipe. A picture of a freeze site is
15-20
shown in Figure 15-12. If full stop joints are installed on 15.7.5 Pinpointing Methods
the feeder, they can be used in a similar fashion to a During the past decade, an extensive research effort by
freeze point in the leak search. EPRI and Con Edison has been carried out to develop a
Thermal Probes reliable method for pinpointing leaks. References (Will-
In the 1950s, Con Edison developed a thermal probe iams et al. 1983; Williams 1982) provide a summary of
that could sectionalize leaks without freezing by deter- the work performed by EPRI. Some of these techniques
mining flow direction at manholes. The instrument con- were further pursued by Con Edison (Ghafurian et al.
sists of a heater and temperature sensor, which can be 1999b). A brief summary of these techniques is pre-
inserted in the flow stream, as shown in Figure 15-13. sented below.
When power is applied to the heater, the downstream Subsurface Radar
temperature sensor responds, indicating the direction of Use of subsurface radar for leak location has been
fluid flow, which is presumed to be toward the leak. addressed by EPRI and others (ESEERCO 1996)).
Local thermal flows could result in an erroneous read- These studies have produced marginal results. This is
ing. This technique requires that the feeder be de-ener- due to the fact that various conditions—such as when a
gized for a significant period of time. In general, the leak has not created a pool or if the leak is masked under
probe method may be successful for leaks larger than the pipe, or if ground is saturated with moisture—would
15 gal/hr (57 l/hr). inhibit the process. If improvements are made in the
future, subsurface radar could be a complementary tool
to be used with other techniques for pinpointing leaks.
Acoustic Method
An acoustic method, based on the detection of sounds
generated by the pressurized liquid escaping from the
pipe, has been tried in actual field conditions. Due
mainly to the low signal level, signal attenuation, and
effect of background noise, the method has not pro-
duced successful results.
Scent -Trained Dogs
EPRI and Con Edison pursued this approach with lim-
ited success. Among the parameters adversely affecting
this method are the presence of noisy traffic, which is a
distraction to the dogs, weather conditions of high
humidity and heat, which fatigue the dogs, and difficulty
with distinguishing between vapors of an old leak and a
Figure 15-12 Freeze jacket (courtesy Consolidated Edison new one.
Company of New York, Inc.).
Hydrocarbon Sensor Wire
Several types of sensor wires have been developed that
can be laid parallel to the cable and react to the presence
of hydrocarbons. The basic principle involves a swelling
or dissolution of the plastic insulating material when
exposed to hydrocarbons, which then causes contact
between internal metallic sensor wires. Locating the leak
is accomplished by using a method to determine the
point of contact, such as a resistance bridge measure-
ment. The concept is mainly applicable to new construc-
tion or in areas where the cable is readily accessible,
such as in tunnels. As indicated in reference (Ghafurian
et al. 1989), the method was applied in a limited demon-
stration project.
15.7.6 PFT Leak Location Method

Figure 15-13 Thermal probe (courtesy Consolidated Based on a technology developed at the Brookhaven
Edison Company of New York, Inc.). National Laboratory (BNL), Con Edison and EPRI
15-21
have successfully developed a method for pinpointing are extremely sensitive to measurement and, due to
leaks. This new technique is called the perfluorocarbon extremely low ambient background levels, can be
tracer (PFT) method, which relies on the injection of detected with the PFT analyzer. With this, a general
PFT, which is a nonpolluting and safe fluid, into the location of the leak, within a few blocks, is identified.
pipe and subsequent measurement of the tracer above • Bar-hole sampling: After the general area of the leak
ground. A detailed description of the method is pre- is localized, within a few blocks, that area is narrowed
sented in references (Ghafurian et al. 1999a; Garcia down by making about seven or eight bar-holes at
1993; Dietz 1987). about 40-ft intervals in the surface of the street, and
air samples from the soil are tested. Further bar-hol-
The following is a detailed description of the steps ing at shorter intervals will finally pinpoint the loca-
involved in this technique. tion of the leak. The bar-hole with the highest
• Preparation of tagging fluid: Since PFT dielectric liq- concentration is over the leak site. Figures 15-14 and
uid is heavier than conventional dielectric fluid, with 15-15 show the PFT Leak Location/Bar-Hole Con-
only limited solubility, it does not easily mix with centration Map and Leakmobile.
dielectric fluid. Care is taken during the mixing to
preclude gases; this is especially important for self- 15.7.7 Field Trial Results
contained systems. A small mixer is used to produce The PFT method has been successfully employed by
a uniform mixture of primary tagging fluid in the utilities in pinpointing leaks on pipe-type as well as self-
dielectric fluid to around the 1% range. The mixture contained feeders. The leak rates have ranged from a
is called the “tagging fluid.” Subsequently this is small one-gallon per day (3.8 l/day) rate up to very large
extensively diluted to the final concentration required leaks. The method has proven to be very accurate and
in the feeder. extremely reliable in pinpointing the location of leaks.
• Injection and circulation: The tagging fluid is diluted Recently PFT has been successfully used to locate leaks
into the pipe using an appropriate injection system. on HPGF pipe-type circuits.
Injection is made at one or more points along the
feeder having access via valves. Online dilution is
readily achievable for feeders with high-rate circula-
tion (force-cooled feeders). However, for static feed-
ers, other methods—such as removing the dielectric
fluid from the far end to create circulation or return-
ing dielectric fluid along a parallel feeder—need to be
implemented to ascertain that tagged fluid has
reached the point of the leak and diffuses through the
soil. The PFT is then transported by conventional
driving forces (barometric pumping, wind-induced
pressure gradients, etc.), ultimately venting into the
air above the street. This venting generally occurs not
too far from the subsurface leak location. Usually the
Figure 15-14 PFT leak location. (courtesy Consolidated
emissions of PFT vapors into the air reach steady Edison Company of New York, Inc.)
state in less than 24 hours.
• Survey of feeder route: After allowing time for the
PFT tagged fluid to reach to the leak area and diffuse
through the soil to the surface, the route of the feeder
is traversed with the Leakmobile (i.e., a vehicle
equipped with PFT-sensing equipment) and air sam-
ples are automatically tested for traces of PFT. Based
on the tagging levels used, the magnitude of the leak,
the prevailing meteorology, and a number of other
site conditions (e.g., depth to the feeder and the
street/above-ground interface conditions), the typical
PFT concentrations in the air above the street at a
leak site are in the few to several hundred parts-per-
Figure 15-15 PFT Leakmobile (courtesy Consolidated
quadrillion (10-15) range. The resulting PFT vapors Edison Company of New York, Inc.).
15-22
15.7.8 Leak Prevention Methods replace a section of an HPFF feeder that experiences
The following techniques may be employed to mitigate high leaks. Detail of the joint is described in reference
the impact of dielectric fluid leaks or to prevent the (Chu et al. 1992).
occurrence of a leak. Manhole Inspection
Cathodic Protection System Since cathodic protection current cannot reach the pipe
The most effective way of preventing a leak is to main- in the manhole, corrosion could occur. Routine inspec-
tain adequate cathodic protection system on the pipe. tion of the manhole, including testing of the coating,
This is discussed in Chapter 10. could prevent future leaks.
Termination Rupture Detection of Corrosion and Coating Disbonding

If a transmission feeder has a very high elevation When coating is not adhered to the surface of the pipe,
difference between the two station ends, a rupture of a water can penetrate under it and support galvanic cor-
termination at the low end of such a feeder could result rosion of the pipe surface. This condition can occur on
in the loss of a large amount of dielectric fluid. To pre- both buried pipes as well as on above-ground pipes such
vent such a condition, the following two methods may as pipes in manholes. Disbonded coating on buried
be implemented: pipes cannot be mitigated by the application of cathodic
protection because the coating prevents cathodic protec-
• Full-Stop Joint. A full-stop joint provides for com- tion current from reaching the pipe surface, resulting in
plete hydraulic isolation, when the bypass valve, galvanic corrosion. The only way to address the prob-
across the joint, is manually or automatically closed. lem is to be able to find the area(s) of disbondment,
• Riser Crusher. A riser crusher may be installed on remove the old coating, repair the pipe surface, if neces-
each riser. It consists of a pair of dies and two sary, and reapply new coating.
hydraulic cylinders. The hydraulic cylinders push the
dies into the riser pipe and squeeze the pipe shut At present, no known methods have been shown effec-
around the cable in the event of a termination rup- tive in finding disbonded coating. However, R&D
ture. This method, if applied, results in the destruc- projects, such as the use of thermography and spectros-
tion of the cable and the riser pipe. copy to inspect pipes inside manholes, are being pur-
sued to develop a possible solution.
Section Replacement
A transition joint, which provides for splicing a
solid dielectric cable to a HPFF cable, can be used to
15-23
REFERENCES Garcia. F. 1993. Perfluorocarbon Tracers for HPOF

Aabo, T. and J. A. Moran, Jr. 1988. Thermomechanical Leak Location. Brookhaven National Laboratory. EPRI
Bending Effects on Extra-High Voltage Pipe-Type Project RP 7905.
Cables. July. EPRI Research Project 7873-1. Report No.
EL-5880. Ghafurian, R., H. Chu, J. Holmes, and H. Elbadaly.
1989. “Detection and Location of Dielectric Fluid
Anders, G., J. M. Braun, S. Rizzetto, M. Vainberg, Leaks on Pipe-Type Cables.” IEEE Transactions on
R. Ghafurian, and L. Tang. 1999. “Leak Detection in Power Delivery. Vol. 4. July. pp. 1499-1506.
High Pressure Pipe-Type Cables Using Artificial Neural
Network.” 5th Annual Conference on Insulated Paper Ghafurian, R. and R. Mauro. 1994. “Locating Leaks
Cables (JICABLE). June. Paris, France. and Faults on Pipe-type Cables: Challenges/New Meth-
ods.” International Electric Research Exchange Work-
Anders, G., R. Ghafurian, and W. Tylman. 2006.“A shop. June 13-16. Toronto, Ontario.
Novel Leak Detection System for Pipe-Type Cable
Installations.” IEEE Transactions on Power Delivery. Ghafurian, R., R. N. Dietz, J. Dominguez, N. Tai, and
Vol. 21. July. T. Rodenbaugh. 1999a. “Leak Location in Fluid-Filled
Cables Using the PFT Method.” IEEE Transactions on
ASTM. 2000. ASTM Specification D-3613-98. Stan- Power Delivery. January. Vol. 14. pp. 18-23.
dard Practice for Sampling Insulating Liquids for Gas
Analysis and Determination of Water Content. May. Ghafurian, R., J. Dominguez, A. Santini, and C. Sobel.
1999b. “New Advances in Mitigating the Environmental
Chu, H., R. Ghafurian, K. Maguire, B. Parmiagiani, Impact of Pipe-Type Cables.” IEEE Transactions on
and J. Walker. 1992. “Development of 138 kV Transi- Power Delivery. April. Vol. 14. pp. 314-319.
tion Joint Between HPPT Cable and Extruded Cable.”
IEEE Transactions on Power Delivery. Vol. 7. October. IEEE. 1998. IEEE Std. 1406. Guide to the Use of Gas-
in-Fluid Analysis for Electric Power Cable Systems.
Dietz, R. N. 1987. “Perfluorocarbon Tracer Technol-
ogy”. In Regional and Long Range Transport of Air Pol- Nigel, O. 1970. “Hydraulic Methods for Locating Oil
lution. Sandroni, S., Ed. pp. 215-247. Elsevier Science Leaks in Underground Cables.” IEEE Transactions on
Publishers B. V. Amsterdam. The Netherlands. Power Apparatus and Systems. Vol. PAS-89. No. 7. Sep-
tember/October. pp. 1434-1439.
EEI. 1957. Edison Electric Institute. Underground Sys-
tems Reference Book. Williams, J. A. 1982. Development of a Leak Location
System for Use on Underground Electric Power Trans-
Engelhardt, J. S. and D. W. Purnhagen. 1991. Dynamic mission Cable. October. EPRI Research Project
Rating and Underground Monitoring System, Feasibility EL-2679. Report No. EL-2679.
Study. Underground Systems, Inc. July. EPRI Report
No. EL-7341. Williams, J. A., S. Kozak, and T. J. Rodenbaugh. 1983.
“Leak Location Methods for HV Underground
ESEERCO. 1996. Empire State Electric Energy Cables.” IEEE Transactions on Power Apparatus and
Research Corporation. Improved Dielectric Fluid Leak Systems. Vol. PAS-102. No. 7. July. pp. 2029-2037.
Detection. ESEERCO Final Report. March.
15-24
CHAPTER 16 Cable System Considerations

Author: John H. Cooper, Power Delivery Consultants, Inc.
Reviewers: Allen MacPhail, BC Hydro
Dennis Johnson, POWER Engineers, Inc.
Underground transmission cables have significantly different electrical characteristics

than overhead lines, and these differences must be taken into account during cable system
planning, design, and operation. The higher shunt capacitance of underground cables
generates significant charging currents and higher inrush currents during energization,
affects the system voltage profile, and results in high surge arrester energy dissipation
requirements. Shunt reactors may be required to mitigate these effects. The lower series
reactance can result in higher-than-desired power flow on a cable installed in an overhead
transmission network. This chapter presents the equations and procedures for calculation
of cable system symmetrical component series impedances and shunt admittances
required for system planning studies. Cables generally have a lower forced-outage rate,
but outage duration can be much longer than for an overhead line. Restoration proce-
dures for a system containing underground cables can be more difficult than for an all-
overhead system.
A detailed evaluation of the effects of underground cables on a specific utility system

requires sophisticated load-flow, transient-stability, short-circuit, and transient-overvolt-
age calculations, all of which need advanced computer programs. This chapter describes
the general theories and potential difficulties arising from cable integration on a utility
system. The chapter also presents a summary of inductive coordination guides and stan-
dards that are applicable to underground transmission lines. The calculation of above-
ground magnetic fields produced by underground cable systems and management meth-
ods is also covered.
John Cooper is a principal engineer with Power Delivery Consultants,

Inc. (PDC). He received a BSEE degree from Texas A&M University
(1967) and an MSEE degree from the University of Pittsburgh (1968).
He joined Westinghouse Electric in 1967, where he initially performed
power systems planning studies. He later held various engineering and
technical management positions at the Waltz Mill Underground Trans-
mission Test Facility. Mr. Cooper then worked for Power Technologies,
Inc. (PTI) from 1988 to 1991 as a consultant to electric utilities on the
design, specification, and installation of transmission cable systems. He has worked at
PDC since it was founded in 1991. Mr. Cooper is a Fellow of the IEEE Power Engineer-
ing Society, a Voting Member of the IEEE Insulated Conductors Committee (ICC), and
a member of CIGRE. He is a past Chairman of the ICC Cable Systems Subcommittee.
He is a registered professional engineer in the states of Pennsylvania, Georgia, New Mex-
ico, and Texas.
16-1
Chapter 16: Cable System Considerations EPRI Underground Transmission Systems Reference Book
16.1 INTRODUCTION To illustrate the difference in electrical constants for

The characteristics of transmission cables are signifi- overhead and underground transmission lines, Table
cantly different from those of overhead lines, and these 16-1 lists a number of the parameters for typical 230-kV
differences must be taken into account when integrating overhead and underground transmission lines and indi-
cables into a transmission system composed primarily cates the related system planning considerations.
of overhead lines (Stewart et al. 1986). The following is
a list of some of the most important effects and special The following sections of this chapter summarize the
requirements that should be considered from a systems major items that should be considered when evaluating
point of view when integrating transmission cables into the design and operational effects of an underground
a utility system or calculating their relationships with cable system. Procedures for calculating the electrical
other components of the system. parameters of the most common types of cable circuits
are also summarized. Section 16.2 explains impedance
• Cable reactive-compensation requirements calculations and provides worked examples for them.
• Effects on power flows Section 16.3 identifies considerations for integration of
• Effects on switching devices cables into the utility system. Section 16.4 describes cal-
culations and planning for magnetic fields. Section 16.5
• Effects on surge-protective devices describes design considerations for inductive coordina-
• Steady-state voltage effects tion and provides worked examples for them. Section
• Impact on system parallel harmonic resonance fre- 16.6 lists the nomenclature relevant to this chapter.
quency
• System restoration procedures 16.2 IMPEDANCE CALCULATIONS
A cable circuit appears as series and shunt impedances
• System reliability
to the rest of the utility system, as shown in the π-equiv-
• Short-term overload characteristics alent circuit of a single cable in Figure 16-1. The series
• Losses impedance portion consists of the conductor resistance
and inductive reactance, while the shunt portion con-
• Special requirements for hybrid lines
sists of the capacitive reactance and the resistance that
• System fault protection (i.e., relaying) considerations represents dielectric losses. Although dielectric losses
• Environmental considerations (inductive coordina- are important in evaluating cable system losses, the
tion and magnetic fields) shunt resistance is insignificant from a system stand-
point when compared to the system loads. However, the
An in-depth analysis of these topics requires sophisti- series impedances can have an effect on system voltages
cated load-flow, transient-stability, short-circuit, and and angles and are normally included to represent
transient overvoltage calculation computer programs. cables in load-flow and transient-stability studies.
General theory and potential problem areas that should
be investigated are discussed below. Since the phases of a three-phase transmission line nor-
mally interact through magnetic coupling, the cables
The majority of the systems planning considerations are must be represented on a three-phase basis rather than
a consequence of the two characteristics of underground the simple equivalent circuit of a single cable shown in
cables that differ most significantly from overhead lines: Figure 16-1. This interaction is usually represented in
the lower series inductance and the much higher shunt system planning studies by using symmetrical compo-
capacitance of underground cables. The series induc- nent circuit analysis techniques. This chapter assumes
tance of cable circuits is typically one-half to one-third that the cable engineer or system planner has a basic
that of overhead lines, primarily because the phase con- understanding of these techniques, which are covered in
ductors are much closer. The difference in shunt capaci- detail in numerous references (Glover and Sarma 2002;
tance is even more dramatic, typically 30 to 40 times Westinghouse 1964). This chapter summarizes the pro-
higher than that of an overhead transmission line. This is cedures to calculate the symmetrical component imped-
because the high-voltage conductor is much closer to the ances for transmission cables.
ground-potential conductors (i.e., insulation metallic
shield), and the relative permittivity (also called dielec- Positive-sequence impedance is the ratio of the normal
tric constant or SIC) of the electrical insulation is, in fundamental frequency sinusoidal voltages (equal in
most cases, several times that of air. magnitude and 120° out of phase) across power system
components to the positive-sequence currents flowing
16-2
EPRI Underground Transmission Systems Reference Book Chapter 16: Cable System Considerations
Table 16-1 Comparison of Typical Electrical Characteristics for 230-kV Overhead and Underground HPFF/XLPE
Transmission Lines (60 Hz)
Underground Effects and
Parameter Overhead HPFF XLPE Units Consideration
Shunt Capacitance 0.015 0.43 0.259 μF/mi Inrush current, charging current
0.009 0.27 0.162 μF/km
Series Inductance 2.0 0.7 1.0 mH/mi Inrush current
1.24 0.43 0.61 mH/km
Series Reactance 0.77 0.26 0.37 ohm/mi Load sharing
0.48 0.16 0.23 ohm/km
Charging Current 1.4 21.7 12.8 A/mi Losses, maximum power transfer
0.87 13.5 8.0 A/km
Dielectric Loss – 21.4 0.5 kW/mi Losses, ampacity
– 13.3 0.32 kW/km
Reactive Charging 0.3 8.6 5.1 MVA/mi Voltage profile
0.19 5.34 3.2 MVA/km
Capacitive Energy 0.26 7.6 4.6 kJ/mi Surge arrester energy dissipation
0.16 4.72 2.86 kJ/km
Surge Impedance 375 40 52 ohms Transient overvoltages
Surge Impedance Loading 141 1300 1020 MW Transient stability
Note: Reactive charging and surge impedance loading are three-phase quantities. The remaining quantities are single-phase
quantities.
through the components. Positive-sequence impedances sequence is reversed. The negative-sequence impedance
are of primary importance in calculating normal steady- equals the positive sequence for a cable or overhead line.
state conditions in a power system. A cable’s positive-, negative-, and zero-sequence imped-
ances are calculated using methods similar to proce-
Zero-sequence impedance is the ratio of the equal and dures used for calculating the sequence impedances for
in-phase voltages applied to all three phases of a power overhead transmission lines. The following sections
system component to the in-phase currents that flow summarize the calculation procedures for positive- and
through the components. This quantity is of primary zero-sequence impedances.
importance in calculating the currents or voltages in a
power system during an abnormal condition such as a 16.2.1 Shunt Impedance
line-to-ground fault. As with overhead lines, the resistive part of the shunt
impedance, the resistance, is so high compared to the
The negative-sequence impedance is calculated the same capacitive reactance that it is assumed to be infinite.
way as the positive sequence, except that the phase Thus the calculation of shunt impedances for cables is
usually just the calculation of the shunt capacitance.
Calculating the shunt capacitance for transmission
cables is straightforward, because each of the high-volt-
age conductors is completely surrounded by a conduct-
ing material (the insulation shield), which is at ground
potential. There is, therefore, no coupling between the
voltages of the three phases, and the capacitive reac-
Figure 16-1 Simplified cable equivalent circuit. tance for each of the sequence shunt capacitances is
equal.
16-3
The following equation may be used for calculating the E0 = operating voltage, line-to-ground, kV.
capacitance between each high-voltage conductor and IC = charging current, A/1000 ft (or A/km).
the grounded insulation shield. kVAR = reactive power per phase.
7.354 ε 16.2.2 Series Symmetrical Component

C= x10−3 μF /1000 ft
log10 Di / Dc Impedances
Since there is magnetic coupling among the phase cur-
24.13 ε
= x10−3 μF / km rents, and in some cases among currents in the cable
log10 Di / Dc sheaths or shields, calculating the series symmetrical
16-1
Where: component impedances for cable circuits is not as sim-
ε = insulation dielectric constant, dimension- ple as calculating the shunt impedance quantities. In
less. general, a set of simultaneous equations must be solved
Di = diameter over cable insulation, in. for the voltage drop in each of the current-carrying con-
(or mm). ductors, including metallic shields and ground continu-
DC = diameter over conductor shield, in. ity conductors, to calculate each of the series sequence
(or mm). impedances. Fortunately, single-conductor, three-phase
cable circuits (excluding pipe-type cable) impedances
The numerator and the denominator of the log10 argu- may be calculated with good accuracy using approxi-
ments must have the same units (i.e., inches or millime- mate formulas for some installation conditions. For
ters). cable systems with multiple cables per phase, the more
complex method of solving simultaneous equations
Table 16-2 shows typical values for the dielectric con- must be used to accurately calculate the sequence
stant (specific inductive capacitance, or SIC, in IEEE impedances when the circuits are close enough to have a
Standard 835), for each of the insulating materials com- significant amount of magnetic coupling. Calculating
monly used in transmission cables. series impedances for pipe-type cables are even more
complex, because the set of equations that relates the
Both the positive- and zero-sequence shunt reactances, series voltage drop along the cable to the currents flow-
per unit of length, are then calculated by the following ing in each of the conductors becomes a set of nonlinear
equation. simultaneous equations due to the nonlinear magnetic
characteristics of the steel pipe. Because of this diffi-
1 culty, the series impedances of pipe-type cables are usu-
X1 = X0 = x l06 Ω/l000 ft (or Ω/km) 16-2 ally estimated from semiempirical equations derived
2π f C
from measurements that were performed by Neher
(Neher 1964).
Frequently-used equations for the cable charging cur-
rent, IC, and reactive charging power, kVAR, which are
16.2.3 Single-Conductor Cable Symmetrical
direct applications of the above equations, are:
Component Series Impedance Formulas
IC = 2πf C E0 x 10-3 A/1000 ft (or A/km) 16-3
Impedance formulas have been developed for the sym-
2
kVAR = 2πf C E o x 10-3 kVA/1000
ft (per phase) 16-4 metrical component sequence impedances of several rel-
Where: atively simple single-conductor cable systems. Typically,
f = frequency, Hz. these formulas for single-conductor cable systems are
C = cable capacitance, μF/1000 ft (or μF/km). based on the assumption that the three cables are
located at the vertices of an equilateral triangle. How-
Table 16-2 Dielectric Constants for Insulation Materials
ever, calculations have shown that the simplified formu-
Dielectric Constant las give accurate results when the cables are installed in
Insulation Material Range Typicala configuration other than an equilateral triangle by using
Impregnated Paper 3.5-3.7 3.5 the geometric mean distance (GMD) between the cables
Laminated Paper-Polypropylene 2.7-2.9 2.8 as the equivalent spacing between cables.
Crosslinked Polyethylene 2.3-2.4 2.3
The concept for these calculations and the nomencla-
Ethylene-Propylene Rubber 2.5-4.0 3.0
ture for the formulas for the sequence impedances are
Electronegative Gas/Spacer 1+ 1 shown in Figure 16-2.
a. IEEE Standard 835.
16-4
Single-Point Bonded Cable Sheaths without Where:

Ground Conductor dij = Distance Between the Center of Conduc-
For cables with the metallic sheaths grounded at a single tors i and j, in. (mm).
point (Figure 16-2 A), the positive-sequence inductive dii = Geometric Mean Radius for Strand i, in.
reactance at 60 Hz is calculated using Equation 16-5. (mm).
= 0.779 x the Radius of Strand i, in. (mm).
GMDφ
X1 = 0.0529 log10 Ω /1000 ft
GMR φ The GMR for a cylindrical conductor (e.g., GIS cables)
with outside and inside diameters of r1 and r2, respec-
⎛ GMDφ ⎞
= ( 60 ) ( 2 π ) ( 2 ×10−7 ) (1000 ) Ln ⎜
⎜ GMR ⎟⎟
⎝ φ ⎠
⎛ GMDφ ⎞
= 0.0754 Ln ⎜ Ω / km
⎜ GMR ⎟⎟
⎝ φ ⎠
16-5
Note: The metric units equation uses the natural log (Ln)
term to be consistent with the inductance equation pre-
sented in Chapter 3; however, the following equations for
impedance calculations use log10 to be consistent with the
references (Westinghouse 1964 and Lewis and Allen
1978) that are the basis for the impedance calculation
procedures in this chapter.
Where:
GMDφ =Geometric mean distance between the cen-
ters of the cable high-voltage conductors,
in. (mm).
= 3 d ab d bc d ac (see Figure 16-2 B)
GMRφ = Geometric mean radius (see following sec-
tion) of the high-voltage cable conductors,
in. (mm).
Conductor Geometric Mean Radius
The geometric mean radius of a solid round or compact
1
− Figure 16-2 Actual and equivalent circuits for calculation
stranded conductor, GMRφ, is equal to e 4 (or 0.779) of sequence impedances.
times the actual radius of the conductor. For concentric
stranded conductors, the GMR may be calculated using
Equation 16-6 and Figure 16-3. Another alternative is
to obtain the GMR for a specific conductor from con-
ductor tables such as those in Chapter 3. Values for the
GMR are most often obtained from conductor tables
because the individual strands deviate from the circular
cross section shown in Figure 16-3.
GMR = n 2 d11d12 d13d14 d1n d 21..d ij ..d nn Figure 16-3 Geometric mean radius of a conductor
, in. (or mm) composed of n strands.
16-6
16-5
tively (Figure 16-4) is calculated using Equation 16-7 Zss = Self-impedance of metallic sheath with
(Calabrese 1959). earth return.
r14 ⎛3 ⎛ r ⎞⎞ De
− r12 r22 + r24 ⎜⎜ + Ln ⎜⎜ 1 ⎟⎟ ⎟⎟ Zss = R s + R g + j 0.0529 log10 (Ω /1000 ft)
4 rsm
ln (GMR) = ln (r1 ) − ⎝4 ⎝ r2 ⎠ ⎠
(r1
2
− r22 )
2
= R s + 0.05922 + j 0.1736 log10
De
(Ω / km)
16-7 rsm
16-11
The positive sequence resistance, Rφ, is the ac resistance Rs = Resistance of cable metallic shield at oper-
of the cable high-voltage conductor at operating tem- ating temperature, Ω/1000 ft (or Ω/km).
perature taking skin effect and proximity effects into rsm = Radius from center of cable phase conduc-
account. Chapter 11 (Ampacity) contains the equations tor to center of metallic shield, in. (mm).
for calculating the ac resistance of cable conductors. Zab = Mutual impedance between phase conduc-
Combining the positive sequence resistance and inductors with earth return.
tive reactance, the positive sequence impedance for sin-
De
gle-conductor cables with single-point grounded cable Zab = R g + j 0.0529 log10 (Ω /1000 ft)
sheaths, is shown in Equation 16-8. GMD φ
GMD φ De
Z 1=Rφ + 0.0529 log10 Ω /1000 ft = 0.05922 + j 0.1736 log10 (Ω / km)
GMR φ GMD φ
16-12
GMD φ
Z 1=Rφ + 0.1736 log10 Ω / km Zas = Mutual impedance between phase conduc-
GMR φ tor and metallic sheath with earth return.
16-8
De
Zas = R g + j 0.0529 log10 (Ω /1000 ft)
A formula for the zero-sequence impedance, Z 0 , is rsm
shown in Equation 16-9.
De
(Z + 2Zab ) 2 = 0.05922 + j 0.1736 log10 (Ω / km)
Z0 = Zaa + 2Zab − as Ω /1000 ft (or Ω / km) rsm
Zss + 2Zab 16-13
16-9 Rg in the preceding equations is the frequency-dependent

Where: resistance of the earth return path, which may be approx-
Zaa = Self-impedance of phase conductor with imated at low frequencies by Equation 16-14 (Meliopou-
earth return. los 1988).
De
Zaa = R φ + R g + j 0.0529 log10 Ω /1000 ft Rg ≅
2 π f μ0
Ω / m = 0.05922 Ω / km at 60 Hz
GMR φ 8
De = 0.01807 Ω /1000 ft at 60 Hz
= R φ + 0.05922 + j 0.1736 log10 Ω / km 16-14
GMR φ Where:
μ0 = Magnetic permeability of free space (4 π x
16-10
10-7 h/m).
De in the preceding equations is the mean depth of the

current that returns in the earth.
De ≅ 399.6 ρ / f (m)
= 25918 ρ / f (in.)
16-15
Figure 16-4 Geometric mean radius of a cylindrical

conductor.
16-6
ρ = electrical resistivity of native soil, Ω-m (see GMRgcc = geometric mean radius of ground conti-
Table 16-3). nuity conductor in. (or mm).
Multi-point Bonded Cable Sheaths
The zero-sequence impedance varies significantly with the soil
For cables with the metallic sheaths grounded at multi-
electrical resistivity, ρ, and the presence of any other conductors
ple points, the positive-sequence reactance is calculated
in the ground, such as ground wires or water pipes.
as shown in Equation 16-19.
Single-Point Bonded Cable Sheaths with Ground
Continuity Conductor (Zas − Zab ) 2
Z1 = Zaa − Zab − Ω /1000 ft
As explained in Chapter 10 (Grounding and Cathodic Zss − Zab (or Ω / km)
Protection), it is recommended that a ground continuity 16-19
conductor be installed in the same trench as cables with Where:
single-point bonded sheaths. Exceptions to this recom- Zaa, Zab, Zas, Zss
mendation are single-point bonded cable systems are the same as shown above (in the section
installed in substations with a ground grid. The ground entitled “Single-Point Bonded Cable
grids in substations provide a low impedance path for Sheaths without Ground Conductor”).
zero-sequence currents. Equations 16-16 and 16-17
show the equations for the positive and zero-sequence The zero-sequence impedance for cable systems with
impedance values for single-point bonded cable systems multi-point bonded cable sheaths is shown in Equation
with a ground continuity conductor (Figure 12-24 b). 16-20.
Z1 = Z aa− Z ab Ω/1000 ft (or Ω/km) 16-16

(Zas + 2 Zab ) 2
Z0 = Zaa + 2 Zab − Ω /1000 ft
Z an2 Zss + 2 Zab (or Ω / km)
Z 0 = Z aa + 2 Z ab − 3 Ω/1000 ft (or Ω/km) 16-17
Z nn 16-20
Where:
Zaa and Zab are the same as defined above. Cross-Bonded Cable Sheaths
Znn = Self-impedance of ground continuity A common sheath-bonding method to minimize
conductor with earth return. induced currents is cross-bonding (see Chapter 12). In
this case, the induced sheath currents are negligible if
De
Znn = R gcc + Rg + j0.0529 log10 Ω /1000ft the minor section lengths are approximately equal. If
GMR gcc the induced sheath currents are negligible compared to
De the phase currents, the positive-sequence impedance for
= R gcc + R g + j0.1736 log10 Ω / km cross-bonded cable systems can be calculated using
GMR gcc Equation 16-16 (single-point bonded sheaths). The
16-18 zero-sequence impedance for cable systems with cross-
Where: bonded cable sheaths is the same as for multi-point
Rgcc = resistance of ground continuity conduc- bonded cable sheaths and Equation 16-20 applies.
tor Ω/1000 ft (or Ω/km).
Circuit Impedances for Multiple Installation
Conditions
Table 16-3 Typical Values of Soil Electrical Resistivity In general, cable circuits do not have exactly the same
(EEI 1957; IEEE Std. 80 2000) installation geometry and sheath-bonding configura-
Soil Conditions Resistivity Ω-m tions for the entire length of the circuit. In this case, the
Average of a large number of determina- circuit is segregated into segments with different instal-
100
tions lation parameters, and the impedances calculated for
Sea water 0.1—1.0 each of the line segments using the applicable formulas.
Swampy ground 10—100 The total impedance of the line is then the sum of the
Dry earth 1,000 complex impedances for all of the segments.
Bedrock 104
Pure slate 107 16.2.4 Generalized Calculation Method for
Sandstone 108 Symmetric Component Series Impedance
Note: These values are typical, but soil resistivity can In more complicated transmission cable systems (i.e.,
vary by more than an order of magnitude with tem- multiple cables per phase and/or multiple ground conti-
perature and moisture content. nuity conductors), standardized formulas for calculating
symmetrical component impedances are generally not
16-7
available. In these cases, generalized numerical calcula- The elements of the impedance matrix may be calcu-
tion procedures must be used to calculate the symmetrical lated as follows, using the geometric notations shown in
component series impedances. Figure 16-5 shows a gen- Figure 16-2 A.
eral three-phase cable circuit, which is the basis for matrix
Equation 16-21 to relate the series voltage drop along the For the self-impedances, Zaa, Zbb, Zcc, Zss, and Znn with
cable conductors to all of the currents flowing in the high- earth return, see previous Equations 16-16 to 16-18.
voltage conductors, the cable sheaths, and ground conti-
nuity conductor. This figure and the corresponding set of For the mutual impedance between the phase conductor
equations are presented to demonstrate a method that and the cable metallic shield with earth return, Zas, see
may be used to calculate cable circuit impedances for a previous Equation 16-14.
variety of cable system configurations. This more elabo-
rate method is not required for common single-circuit For mutual impedance between phase conductors (e.g.,
cables, but it may have to be extended to calculate imped- Zab) and between cable sheaths (e.g., Z s1s2) with earth
ances for more complex cable systems. return:
Since the objective of the calculations is to determine De

Z ab = Z s1−s2 = Rg + j 0.0529 log10 Ω /1000 ft
the series voltage drops along the high-voltage conduc- d ab
tors in terms of positive- or zero-sequence current, the
De
high-voltage conductors may be assumed to be = Rg + j 0.1736 log10 Ω / km
grounded at one end to simplify the equations. d ab
16-22
⎡Z Z Z Z Z Z Z ⎤ ⎡I ⎤
⎡ Ea ⎤ ⎢ aa ab ac as1 as2 as3 ag ⎥ ⎢ a⎥ For mutual impedance between phase conductors and
⎢ E ⎥ ⎢ Z ab Z bb Z bc Z bs Z bs Z bs Z bg ⎥ ⎢Ib ⎥ ground continuity conductor (e.g., Z ag ) with earth
⎢ b⎥ ⎢ 1 2 3
⎥ ⎢ ⎥ return and between cable sheaths and ground continuity
⎢ Ec ⎥ ⎢ Z ac Z bc Z cc Z cs1 Z cs2 Z cs3 Z cg ⎥ ⎢Ic ⎥ conductor (Zs1g) with earth return:
⎢ ⎥ ⎢ ⎥ ⎢Is ⎥
⎢0 ⎥ = ⎢ Z as1 Z bs1 Z cs1 Z ss Z s1s2 Z s1s3 Z s1g ⎥ ⎢ 1⎥ De
⎢0 ⎥ ⎢ Z Z Z Z Z Z Z ag = Zs1g = Rg + j 0.0529 log10 Ω /1000 ft
Z ⎥ ⎢ I s2 ⎥ d an
⎢ ⎥ ⎢ as2 bs2 cs2 s1s2 ss s2s3 s 2 g ⎥ ⎢ ⎥
⎢0 ⎥ ⎢ Z as Z bs Z cs Z s s Z s s Z ss Z s 3 g ⎥ ⎢ I s3 ⎥ De
⎢0 ⎥ ⎢ 3 3 3 1 3 2 3 = Rg + j 0.1736 log10 Ω / km
⎣ ⎦ ⎥ ⎢I ⎥ d an
⎣ Z ag Z bg Z cg Z s1g Z s 2 g Z s 3 g Z gg ⎦ ⎣ g⎦ 16-21 16-23
Where:
dan = distance between a phase conductor or
sheath and ground continuity conductor,
in. (or mm).
In all cases, the numerator and the denominator of the

log10 arguments must be in the same units (i.e., in. or mm).
The matrices of Equation 16-21 may be rewritten using

submatrices as follows to facilitate calculations for the
cable sequence impedances:
⎡ E p ⎤ ⎡ Z pp Z pn ⎤ ⎡ I p ⎤
⎢ 0 ⎥ = ⎢Z | Z ⎥⎢I ⎥
⎣ ⎦ ⎣⎢ np nn ⎥
⎦⎣ n ⎦ 16-24
Where:
[Ep] = column matrix of voltage applied to cable
high-voltage conductors, V.
[Ip] = column matrix of currents in the high-volt-
age conductors, A.
Figure 16-5 Cable circuit for positive- and zero-sequence [In] = column matrix of the currents in the cable
impedance calculations. sheaths and ground continuity conductor,
A.
16-8
[Zpp] = square (three rows by three columns) The complex operator a is defined as:
matrix of the self- and mutual impedances
for the high-voltage conductors, Ω/1000 ft 1 3
a=− −j
(or Ω/km). 2 2
[Zpn], [Znp]
= rectangular matrices of the conductor-to- 5. The diagonal elements of the [Zseq] matrix are the
sheath and conductor-to-ground mutual zero (Z0), positive (Z1), and negative sequence (Z2)
impedances, Ω/1000 ft (or Ω/km). impedances, and the off-diagonal elements are the
[Znn] = square matrix of the self and mutual mutual impedances between the zero, positive, and
impedances of the cable sheaths and negative symmetric component networks. That is:
ground continuity conductor, Ω/1000 ft (or ⎡ Z0 Z 01 Z0 2 ⎤
Ω/km).
[0] = column matrix of zeros, (V/1000 ft or [Z ]
seq
⎢
= ⎢ Z1 0 Z1
⎥
Z1 2 ⎥ Ω/1000 ft (or Ω/km)
V/km). ⎢Z 2 0 Z 21 Z 2 ⎥⎦
⎣
16-31
Equation 16-24 may now be rewritten as the two follow-
ing matrix equations: If the cables are single-point bonded, the above proce-
[Ep] = [Zpp] [Ip] + [Zpn] [In] 16-25 dure is simplified, because the sheath currents in step 1
are known to be zero. The effect of a ground continuity
[0] = [Znp] [Ip] + [Znn] [In] 16-26
conductor, if present, must, however, be included.
The procedure to calculate the positive sequence imped-
For cables with cross-bonded sheaths, Equation 16-21
ance for the three-phase cable circuit in Figure 16-5 is:
must be rewritten for each of the three sheath sections
1. Solve for the sheath and ground continuity conduc- where the sheaths are transposed, to reflect the different
tor currents [In] in terms of the phase currents, [Ip], magnetic coupling between the individual sheaths and
using Equation 16-27. the phase conductors. An “exact” calculation would
[In] = -[Znn]-1 [Znp] [Ip] 16-27 then require the solution of eighteen simultaneous equa-
Where: [Znn]-1 is the inverse of matrix [Znn] tions. The positive- sequence impedance can be approxi-
mated by assuming that there is no sheath current.
2. Substitute the solution for the matrix [In] into matrix
Equation 16-25
The above procedure is not practical to perform as a
[Ep] = ([Zpp] -[Zpn] [Znn]-1[Znp]) [Ip] 16-28
manual calculation, but is easily implemented using a
3. The matrix of equivalent phase impedances for the personal computer and matrix inversion and multiplica-
three-phase cable circuit is then the ratio of the volt- tion techniques.
ages applied to the conductors, [Ep], divided by the
phase currents, [Ip]. The equivalent phase impedance 16.2.5 Worked Examples—Single-Conductor
matrix, [Zphase], is a 3 x 3 matrix with the self- and Cable Sequence Impedances
mutual impedances between the three phases taking The following example calculation demonstrates the use
the mutual coupling to the cable sheaths and ground of the equations for calculating the positive- and zero-
continuity conductors into account. sequence impedances of a cable circuit and compares
[Zphase] = [Zpp] -[Zpn] [Znn]-1[Znp] 16-29 the results of the approximate Equations 16-16 and
4. The equivalent phase impedance matrix, [Zphase], can 16-19, and with the impedance values obtained using
be transformed into the symmetric component the generalized impedance calculation method.
impedance (zero, positive, and negative) matrix,
[Zseq], using the following matrix equation (Glover Worked Example 1: Single Conductor Cable
and Sarma 2002). Symmetrical Component Impedances Using
Formulas
[Zseq] = [A] [Zphase] [A]-1 16-30
Calculate the positive- and zero-sequence impedances of
Where:
a 230-kV XLPE cable installed in a vertical configura-
⎡1 1 1⎤ tion with phase spacings of 10 in. (25.4 mm) between the
[A] is the matrix: ⎢1 a 2 a ⎥⎥ centers of adjacent cables (see Figure 16-6). The cable
⎢ sheaths are single-point bonded to a 250 kcmil (125
⎢⎣1 a a 2 ⎥⎦ mm2), concentric round, copper ground continuity cable.
The 230-kV XLPE cables have a 1200 mm2 (2400 kcmil)
compact segmental copper conductor. Additional details
16-9
of the transmission cable and ground continuity conduc- The radius from the center of the cable conductor to the
tor constructions are given in Table 16-3. middle of the sheath is:
1.526 0.095
Other parameters needed for calculating the sequence rsm = + 0.063 + 0.906 + 0.067 +
2 2
impedances are shown in Table 16-4.
= 1.85 in. (47 mm)
The geometric mean distance between the transmission
cable phase conductors from Equation 16-5 is: The self-impedance of a conductor with earth return is:
GMDφ = 3 10 x10 x 20 = 12.6 in. (320 mm) Z aa = 0.00611 + 0.01807
⎛ 33500 ⎞
+ j 0.0529 log10 ⎜ ⎟ Ω /1000 ft
The geometric mean distance between the phase con- ⎝ 0.623 ⎠
ductors and the ground continuity conductor is: = 0.02418
GMDφ − gcc = 3 14.14 x 10 x 14.14 = 12.6 in. (320 mm) + j 0.250 Ω /1000 ft (0.0793 + j 0.820 Ω / km)
The equivalent depth of the earth return current from The mutual impedance between phase conductors is:
Equation16-15 is: ⎛ 33500 ⎞
Zab = 0.01807 + j 0.0529 log10 ⎜ ⎟
De = 25918
100
= 33500 in. (850,900 mm)
⎝ 12.6 ⎠
60 = 0.01807
+ j 0.1811 Ω /1000 ft ( 0.05928 + j0.5942 Ω / km )
The self-impedance of the ground continuity conductor

(Equation 16-17) with earth return is:
Znn = 0.0489 + 0.01807
⎛ 33500 ⎞
+ j 0.0529 log10 ⎜ ⎟ Ω /1000 ft
⎝ 0.221 ⎠
= 0.067 + j 0.274 Ω /1000 ft ( 0.212 + j0.899 Ω / km )
The mutual impedance between phase conductors and

the ground continuity conductor is:
⎛ 33500 ⎞
Figure 16-6 Example 1: 230-kV XLPE duct bank cross Zan = 0.01807 + j 0.0529 log10 ⎜ ⎟ Ω /1000 ft
section. ⎝ 12.6 ⎠
= 0.01807 + j 0.1811 Ω /1000 ft
Table 16-4 Parameters for Example 1 Impedance
Calculations ( 0.05928 + j0.5942 Ω / km )
230 kV Cable Parameters
Conductor diameter: 1.526 in. (38.8 mm) Substituting the above impedances into Equations 16-16
Thickness of conductor shield: 0.063 in. (1.6 mm) and 16-17 gives the following positive and zero-sequence
AC Resistance at 90°C: 6.11 μΩ/ft (20.06 μΩ/m) impedances:
Metallic shield resistance at
80°C: 27.34 μΩ/ft (89.7 μΩ/m) Z1 = 0.0061 + j 0.0691 Ω/1000 ft
Metallic shield mean diameter: 3.693 in. (93.8 mm) (0.0200 + j 0.2266 Ω/km)
Earth electrical resistivity: 100 Ω-m Z0 = 0.0747 + j 0.2604 Ω/1000 ft
Insulation thickness: 0.906 in. (23 mm) (0.2451 + j 0.8543 Ω/km)
Thickness of insulation shield: 0.067 in. (1.7 mm)
Geometric mean radius of
conductor: 0.623 in. (15.8 mm)
Ground Continuity Conductor Parameters
Conductor diameter: 0.575 in. (15.6 mm)
Conductor GMR: 0.221 in. (5.61 mm)
AC Resistance at 60°C: 48.9 μΩ/ft (160.4 μΩ/m)
16-10
Worked Example 2: Single Conductor Cable Symmetrical Component Impedances,

Generalized Calculation Method
Calculate the positive- and zero-sequence impedances of the 230-kV
XLPE cable system in Example 1 using the generalized impedance cal-
culation method (Section 16.2.4).
230-kV XLPE Transmission Cable System Data:
f = 60 Power system frequency (Hz)
Rdc25 = 0.00456 Resistance of HV conductor (25°C), Ω/1000 ft
Rac = 0.0061 ac conductor resistance at 90°C, Ω/1000 ft
Dc = 1.526 Diameter of 1200 mm2 high-voltage conductor, in.
GMRφ = 0.623 GMR of high-voltage conductor, in.
Rs = 0.02734 Resistance of cable sheath at 80°C, Ω/1000 ft
Dsm = 3.693 Mean Diameter of cable sheath, in.
Ground Continuity Conductor Data:
Rgcc = 0.0489 Resistance of 250 kcmil ground continuity conductor (60°C), ohms/1000 ft
GMRn = 0.221 GMR of neutral conductor, from conductor handbook, in.
The x coordinates are inches from left column of duct bank.
Installation Data:
The y coordinates are inches below the surface of the ground.
x1 = 0 y1 = −65 Phase A horizontal and vertical locations, in.
x2 = 0 y 2 = −75 Phase B horizontal and vertical locations, in.
x3 = 0 y 3 = −85 Phase C horizontal and vertical locations, in.
x 4 = 10 y 4 = −75 Ground continuity conductor (GCC) horizontal and vertical locations
Other System Data:

ρelec = 100 Earth electrical resistivity (assumed/typical), ohm-meter
ρelec Depth of earth return and resistance of earth return at

D e = 25918· Rg = 0.01807
60 Hz (Ω/1000 ft)
f
Constants and arrays used for calculations:
⎡1 1 1 ⎤
3 ⎢ ⎥
a = − 0.5 − j A = ⎢1 a 2 a ⎥
Symmetrical component transformation matrix
2 ⎢1 a a 2 ⎥
⎣ ⎦
Calculate distances between cables and between cables and ground continuity conductor.
d ab = (x 2 − x1 )2 + (y 2 − y1 )2 dab = 10 in. (0.254 m)
d bc = (x 3 − x 2 )2 + (y 3 − y 2 )2 dbc = 10 in. (0.254 m)
d ac = (x 3 − x1 )2 + (y 3 − y1 )2 dac = 20 in. (0.508 m)
d an = (x 4 − x1 )2 + (y 4 − y1 )2 dan = 14.142 in. (0.359 m)
d bn = (x 4 − x 2 )2 + (y 4 − y 2 )2 dbn = 10 in. (0.254 m)
d cn = (x 4 − x 3 )2 + (y 4 − y 3 )2 dcn = 14.142 in. (0.359 m)
16-11
Calculate Self- and Mutual Impedance Terms

⎛ De ⎞
Z aa = Rac + Rg + j 0.0529 log10 ⎜ ⎟ Self-impedance of cable high-voltage conductor with earth return
⎜ GMR ⎟ (Ω/1000 ft)
⎝ φ ⎠
⎛ De ⎞
Z nn = Rgcc + Rg + j 0.0529 log10 ⎜⎜ ⎟⎟
⎝ GMRn ⎠
Self-impedance of GCC with earth return (Ω/1000 ft)
⎛D ⎞ Mutual impedance between Phase a and Phase b

Z ab = Rg + j 0.0529 log10 ⎜⎜ e ⎟⎟
⎝ d ab ⎠
with earth return (Ω/1000 ft)
⎛D ⎞ Mutual impedance between Phase b and Phase c

Z bc = Rg + j 0.0529 log10 ⎜⎜ e ⎟⎟
⎝ d bc ⎠
⎛D ⎞ Mutual impedance between Phase a and Phase c

Z ac = Rg + j 0.0529 log10 ⎜⎜ e ⎟⎟
⎝ d ac ⎠
⎛D ⎞ Mutual impedance between Phase A conductor

Z an = Rg + j 0.0529 log10 ⎜⎜ e ⎟⎟
⎝ d an ⎠
and GCC with earth return (Ω/1000 ft)
⎛D ⎞ Mutual impedance between Phase B conductor

Z bn = Rg + j 0.0529 log10 ⎜⎜ e ⎟⎟
⎝ d bn ⎠
⎛D ⎞ Mutual impedance between Phase C conductor

Z cn = Rg + j 0.0529 log10 ⎜⎜ e ⎟⎟
⎝ d cn ⎠
Setup self- and mutual impedance matrix, Zpp

⎡ Z aa Z ab Z ac ⎤
⎢ ⎥
Z pp = ⎢ Z ab Z aa Z bc ⎥ Matrix with self- and mutual impedances between phase conductors
⎢⎣ Z ac Z bc Z aa ⎥⎦
⎡0.0242 + j 0.2502 0.0181 + j 0.1864 0.0181 + j 0.1705 ⎤
Z pp = ⎢⎢0.0181 + j 0.1864 0.0242 + j 0.2502 0.0181 + j 0.1864 ⎥⎥ Values of self- and mutual impedances,
⎢⎣0.0181 + j 0.1705 0.0181 + j 0.1864 0.0242 + j 0.2502⎥⎦ (Ω/1000 ft)
Setup Z np and Z pn matrices – Mutual impedance terms between the HV conductors and shields are eliminated
because of single-point bonding.
[
Z np = Z an Z bn Z cn ] Matrix with mutual impedances between ground continuity conductor and phase
conductors
Z np = [0.0181 + j 0.1785 0.0181 + j 0.1864 0.0181 + j 0.1785 ] Values of Znp matrix elements, (Ω/1000
ft)
Matrix with mutual impedances between phase conductors and ground continuity
Zpn = ZnpT
conductor (equals transposed Znp matrix)
Zn = Znn Self-impedance of ground continuity conductor with earth return
Zn = 0.0670 + j 0.2740 (Ω/1000 ft)
16-12
Calculate Zphase matrix (equivalent self- and mutual impedance for phase conductors).
Zphase = Zpp – Zpn·Znn –1·Znp
⎡0.0285 + j 0.1362 0.0231 + j 0.0674 0.0224 + j 0.0565 ⎤
Z phase = ⎢⎢0.0231 + j 0.0674 0.0299 + j 0.1260 0.0231 + j 0.0674⎥⎥ (Ω/1000 ft)
⎢⎣0.0224 + j 0.0565 0.0231 + j 0.0674 0.0285 + j 0.1362⎥⎦
Transform the equivalent phase impedance matrix, Zphase, to a symmetrical component impedance matrix, Zseq.
Zseq= A·Zphase·A-1
⎡ 0.0747 + j 0.2604 − 0.0002 − j 0.0007 − 0.0006 + j 0.0005⎤
Z seq = ⎢⎢− 0.0006 + j 0.0005 0.0061 + j 0.0690 − 0.0093 + j 0.0053 ⎥⎥ (Ω / 1000 ft )
⎢⎣− 0.0002 + j 0.0007 0.0092 + j 0.0054 0.0061 + j 0.0690 ⎥⎦
Zero-sequence impedance (Ω/1000 ft) is first diagonal element
Z0 = Zseq[0,0] Z0 = 0.0747 + j 0.2604
of Zseq + matrix
Positive-sequence impedance (Ω/1000 ft) is second diagonal
Z1 = Zseq[1,1] Z1 = 0.0061 + j 0.0690
element of Zseq matrix
Negative-sequence impedance (Ω/1000 ft) is third diagonal
Z2 = Zseq[2,2] Z2 = 0.0061 + j 0.0690
element of Zseq matrix
The positive- and zero-sequence impedances calculated Another problem with calculating the zero-sequence
in using the generalized impedance calculation method impedance for pipe-type cables is that the impedance
give the same results as the impedance values obtained varies with the magnitude of the zero-sequence current.
by the solution of Equation 16-21. This means that short-circuit studies may need an itera-
tive process if the fault location is primarily supplied by
16.2.6 Symmetrical Component Series pipe-type cable circuits.
Impedances for Pipe-Type Cables
Although pipe-type cables still comprise the majority of The following paragraphs summarize the method devel-
existing transmission cable systems in North America, oped by Neher to calculate the positive- and zero-
series impedance calculation methods for this type of sequence impedances of pipe-type cables. The general
cable are the least accurate. This is because the nonlin- formulas in Neher’s paper for zero-sequence impedances
ear permeability and losses in the steel pipe make it dif- include terms, such as the thickness of the pipe wall,
ficult to calculate the flux linkage within the wall of the which are applicable to rarely used nonmagnetic pipe. As
pipe and external to the pipe. explained by Neher, the zero-sequence current is concen-
trated toward the inner surface of a steel pipe due to its
The most commonly used method (IEEE Relaying high magnetic permeability Therefore, the effective
Committee 1996) to calculate impedances for this type thickness of the pipe wall is small for steel pipe, and this
of cable is based on an analysis of pipe-type cable term has been omitted in the following equations.
impedances by Neher (Neher 1964). Neher developed
semiempirical formulas based on laboratory measure- The semiempirical formula in Equation 16-32 for calcu-
ments on short sections of pipe-type cables installed in lating the positive-sequence inductive reactance of pipe-
carbon steel and nonferromagnetic pipes. These formu- type cables is relatively simple. As may be determined by
las are of questionable accuracy, especially for zero- comparing Equation 16-32 to the positive-sequence
sequence impedance, but there are no methods currently impedance equation for cables without a steel pipe
available that have been demonstrated to be more accu- (Equation 16-5), the presence of the steel pipe increases
rate. In one example in Neher’s paper, the discrepancies the inductive reactance of the cable circuit by approxi-
between measured and calculated values for zero- mately 15%. The constant 2.57 arises from the use of Dc
sequence resistance and reactance values were as high as rather than GMRφ in the logarithmic term.
19% and 35%, respectively.
16-13
The calculation of positive-sequence inductive reac-

tance, X1, is given in Equation 16-32. Note that values in
R 'p = effective pipe resistance, μΩ/ft (or μΩ/m).
Equations 16-32 through 16-36 are given in microhms/ft Xc = conductor ac inductive reactance, μΩ/ft (or
( o r m i c ro h m s /m ) r at h e r t h a n o h m s / 1 0 0 0 f t ( o r μΩ/m).
ohms/km). Common usage for cable engineering appli- X 'p = pipe effective inductive reactance, μΩ/ft (or
cations, such as Chapter 11 (Ampacity) use micro-
hms/ft, and these equations refer to that earlier usage. μΩ/m).
System engineering applications typically use the unit
ohms/1000 ft. Conductor ac resistance includes skin and proximity
effects as described in Chapter 11 (Ampacity). The
⎡ 2.57S ⎤ effective resistance (R'p) and inductive reactance (X'p) of
X1 = 60.9 log10 ⎢ ⎥ µΩ / ft the cable pipe are determined from empirical formulas,
⎣ DC ⎦ Equations 16-36 and 16-37, developed by Neher from
laboratory test data.
⎡ 2.57S ⎤
= 200 log10 ⎢ ⎥ µΩ / m
⎣ DC ⎦ μ (H)
16-32 R p = 29.9 μ Ω / ft (D p in inches)
Where: Dp
S = axial spacing between phases, in. (or mm).
Dc = diameter over cable conductor, in. μ(H)
(or mm). = 2475 μ Ω / m (D p in mm)
Dp
For cables in a triangular configuration, S is simply the 16-36
separation between cable centers.
For cables in a cradled configuration, S may be calcu-
μ (H )
lated from Equation 16-33.
X p = 18.1 μ Ω / ft ( D p in inches)
2
Dp
⎡ DS ⎤
S = 1.26 DS 6 1 − ⎢ ⎥ μ (H )
⎣ DP − DS ⎦ in. (or mm) = 1508 μ Ω / m ( D p in mm)
16-33 Dp
Where:
16-37
Ds = diameter over skidwires, in. (or mm).
Dp = pipe inside diameter, in. (or mm).
The quantity of μ (H ) in Equations 16-36 and 16-37
The resistive part of positive-sequence impedance, R1, is the square root of the relative magnetic permeability
for pipe-type cables is calculated as shown in Equation of the steel pipe as a function of magnetizing force (H).
16-34. Figure 16-7 from Neher’s paper contains a plot of the
R1 = Rdc (l +Yc+YS+Yp) μΩ/ft (or μΩ/m) 16-34
Where:
Rdc = conductor dc resistance, μΩ/ft (or μΩ/m).
Yc = increment in losses due to conductor ac
effects, dimensionless.
YS = increment in losses due to shield ac effects,
dimensionless.
Yp = increment in losses due to pipe ac effects,
dimensionless.
Neher’s formula for the zero-sequence impedance of

pipe-type cables, Z0, is given in Equation 16-35.
Z 0 = 3 [ RC / 3 + R ′p + j ( X C + X ′p )] μΩ/ft (or μΩ/m)

16-35
Where:
Figure 16-7 Curve of μ (H ) versus Ip/Dp for carbon
Rc = conductor ac resistance, (1 + Yc)Rdc' μΩ/ft
steel pipe (Neher 1964).
(or μΩ/m).
16-14
6. Calculate the zero-sequence impedance from the fol-

quantity μ ( H ) as a function of magnetizing force lowing formula.
(Ip/Dp= 3 I0/Dp).
⎡R ⎤
Z0 = 3 ⎢ c + R ′p + j (X c + X′p ) ⎥ x 10−3 Ω /1000 ft
Conductor inductive reactance, XC, is calculated from ⎣ 3 ⎦ (or Ω / km)
Equation 16-38.
16-40
7. If short-circuit calculations indicate that the phase-
⎡ D 3p ⎤ to-ground fault current for the pipe-type cable circuit
X c = 17.6 log10 ⎢ 2 ⎥
μΩ/ft is significantly different from the assumed value in
⎣⎢ 3.115 D c S ⎦⎥ step 1, then the above calculations must be repeated
using the improved estimate for pipe fault current.
⎡ D 3p ⎤
= 57.7 log10 ⎢ 2 ⎥
μΩ/m It should be noted that the zero-sequence impedance
⎢⎣ 3.115 D c S ⎥⎦ 16-38 value calculated using the Neher method assumes that
the pipe magnetic characteristics shown in Figure 16-7
The effective pipe impedance, Z 'p, taking into account apply to all pipe-type cables. However, the magnetic per-
the division of fault current between pipe and cable meability of steel pipe varies depending on the carbon
shield/skid wires, is: content and the manufacturing process (see Section
16.4.2, Cables with Ferromagnetic Components).
2
⎡ R p + j(X m + X p ) ⎤⎦
Z′p = R ′p + jX′p = R p + jX p − ⎣ Worked Example 3: Symmetrical Component
Rs Impedance Calculations for HPFF Cable
+ R p + j(X m + X p ) Calculate the positive- and zero-sequence impedances
3
for a 345-kV cable that is in a cradled configuration in a
16-39
Where: 8-in (200 mm), pipe, as used for the Worked Example 3
in Chapter 11 (Ampacity). Assume that the cable con-
(D p + t p )3 ductor temperature is 85°C and that the cable has the
X m = 17.6 log10 μΩ/ft following construction:
4 D′s S2
(D p + t p )3 Conductor 2500 kcmil (1268 mm2) compact seg-
= 57.7 log10 μΩ/m
4 D′s S2 size; type: mented, copper
Shielding: Two intercalated stainless-steel shield
Rs = Resistance of metallic shield and skidwire in tapes
parallel (see Chapter 11), μ/ft (or μ/m).
Skidwires: Two 0.1 in. (2.54 mm), stainless-steel
In summary, the procedure for calculating pipe-type “D” skidwires
cable zero-sequence impedance is: Insulation 0.600 in. (15.2 mm) of LPP insulation
thickness:
1. Assume a value for the magnitude of the pipe phase-
to-ground fault current, which is three times the zero- Cable Diameter: 3.194 in. (81.1 mm)
sequence fault current (3 Io). L-G Fault 10,000 A, and 50,000A
Current
2. Determine μ (H ) from Figure 16-7 and the pipe
fault current divided by Dp. The positive-sequence impedance is calculated as fol-
3. Calculate the pipe zero-sequence resistance, Rp, and lows:
inductive reactance, Xp, using Equations 16-36 and The positive-sequence resistance, R1, is determined from
16-37, respectively. Equation 16-34.
4. Calculate the conductor inductive reactance, Xc , R1 = Rdc (1 +Yc+Ys+Yp)
using Equation 16-38. = (5.459) (1 + 0.144 + 0.0015 + 0.469)
5.Calculate the effective pipe impedance, Z 'p, using = 9.054 μΩ/ft @85°C
= 0.0091 Ω/1000 ft (0.030 /km)
Equation 16-39.
16-15
The equivalent conductor spacing, S, and inductive 9.81

reactance are: R p = 29.7 = 35.86 μΩ/ft (117.7 μΩ/m)
2
8.125
⎛ 3.194 ⎞
S = 1.26(3.194) 6 1 − ⎜ ⎟ = 3.676 in.(93.4 mm)
⎝ 8.125 − 3.194 ⎠ 9.81
X p = 18.1 = 21.86 μΩ/ft (71.7 μΩ/m)
8.125
And the positive-sequence inductive reactance, X 1 , is
calculated from Equation 16-32. The effective pipe impedance is calculated from Equa-
tion 16-39.
⎡ 2.57(3.676) ⎤ Z′p = R ′p + jX′p
X 1 = 60.9 log10 ⎢ = 43.5 μΩ/ft (142.7 μΩ/m)
⎣ 1.824 ⎥⎦
[35.85 + j(8.18 + 21.86)]
2
And therefore = 35.86 + j21.86 −

Z1 = 0.0091 + j 0.0435 Ω/1000 ft (0.0295 + j 0.1427 Ω/km) 36745
+ j(8.18 + 21.86)
3
The zero-sequence impedance is calculated as follows: = 35.78 + j21.39 µΩ / ft (117.7 + j 70.2 µΩ / m)
The conductor, shield, and skidwire resistances are:

The zero-sequence impedance (for 10 kA fault current)
Rc = (1 + Yc) Rdc= (1 + 0.144)(5.343 μΩ/ft) is calculated from Equation 16-40.
= 6.00 μΩ/ft (19.7 μΩ/m)
Rshield = 10,407 μΩ/ft (34,144 μΩ/m) ⎡ 6.0 ⎤
Z0 = 3 ⎢ + 35.78 + j(14.86 + 21.39 ⎥ 10−3
Rskidwire = 79,068 μΩ/ft (259,410 μΩ/m) ⎣ 3 ⎦
= 0.1166 + j0.1093 Ω /1000 ft
1
Rs = = 0.382 + j0.359 Ω / km
2 2
+
R shield R skidwire
For a phase-to-ground fault current of 50,000 A, the
1 above calculation procedure gives a zero-sequence
=
2 2 impedance of:
+
10, 407 79, 068 Z0 = 0.0567 + j0.0737 Ω /1000 ft
= 36, 745 µΩ / ft (120,550 µΩ / m) = 0.1840 + j 0.2410 Ω / km
From Equation 16-38, 16.3 INTEGRATION OF CABLES INTO THE

⎡ (8.125) ⎤ 3 UTILITY SYSTEM
X c = 17.6 log10 ⎢ 2 ⎥
⎣ (3.115)(1.824)(3.194) ⎦ 16.3.1 Capacitance and Reactance Effects
= 14.86 µΩ / ft (48.8 µΩ / m) The high capacitance of cable circuits may have a signif-
icant effect on steady-state voltages throughout a power
system. Since a capacitive current flowing through an
Calculate the quantity, Ip/Dp, and determine μ (H ) inductance causes a voltage rise across the inductor
(Ferranti effect), the charging current created by the
from Figure 16-7.
capacitance of a cable circuit can cause high system
voltages by two related phenomena. The first, and most
Ip 10000 common situation, is for the cable charging current to
= = 1, 230 A / in. (48, 480 A / m) cause voltage rises across the inductive impedances
Dp 8.125 external to the cable circuit. This situation, which is
worst during light-load or load-rejection conditions, is
illustrated in Figure 16-8. In this circuit, the reactances,
From Figure 16-7, μ (H ) = 9.81 XL and XT, represent the reactances of transmission
lines and transformer impedances between the genera-
tor and the cable circuit. Since the voltage at the load
must be held near 100%, the system voltages will rise
16-16
Figure 16-8 Example of voltage rise due to cable charging current.
significantly above rated voltage as the cable charging can be estimated by multiplying one-half the total
current flows through the inductive impedances, XL and charging current times the positive-sequence reactance.
XT, to the generator. During light-load conditions, this
voltage rise may be more than 105% of the nominal rat- 16.3.2 Reactive Compensation
ing for which most power system components are In addition to the steady-state overvoltage problems
designed. described in the previous section, the high capacitive
charging current inherent in cable circuits may cause a
When this condition exists, several measures can be reduction in the amount of real power that can be trans-
taken to reduce system voltages to acceptable levels. mitted through a cable circuit, and may cause apprecia-
First, voltages can be regulated within certain limits by ble energy losses. As the cable circuit length is increased,
adjusting the voltage magnitude at the generator by low- the capacitance and, therefore, the charging current
ering the field excitation, or by lowering voltage taps on increase linearly. In the worst case, the magnitude of the
transformers. The second approach is to provide shunt charging current may increase until it is equal to the
compensation of the cable’s reactive charging current by cable ampacity, at which point no real power may be
placing shunt reactors at one or both ends of the cable transmitted without overheating the cable. This length is
circuit (Dougherty and Shifreen 1962). This latter commonly called the critical cable length.
ap p ro a ch ha s num e ro us a dva n t ag e s, wh i ch a re
described in the following section. During normal- and Fortunately, charging current is in quadrature with the
high-load conditions, the high-voltage problem caused current associated with real power transfer so that avail-
by cable charging current is reduced, because the lag- able ampacity for real power transfer is more than the
ging current drawn by most loads and the voltage drops algebraic difference between the cable ampacity and the
caused by the inductive load current offset the effects of cable charging current. Figure 16-9 shows a plot of the
the cable charging current. real power that may be transferred across 2000 kcmil
(1000 mm2) copper conductor, HPFF and XLPE trans-
The second steady-state voltage-rise phenomenon, mission cables at different system voltages. The power
called Ferranti voltage rise, is caused by the voltage rise
along an unloaded transmission line due to its capaci-
tive current flowing through its own series inductance.
The Ferranti voltage rise occurs when the cable is ener-
gized from one end only. It is not a problem for most
cable circuits, because they are usually relatively short,
and their positive-sequence impedance is low. However,
some cable circuits warrant evaluation during surge
arrester selection. Long single-conductor submarine
cables with wide phase spacings are usually the only
cable circuits where the Ferranti effect is significant
(Cowley and Foxell 1983). If the cable insulation has a
high dielectric constant and wide phase spacings (which
increase the positive-sequence reactance), the voltage
may rise high enough at the open end of the cable to
cause failure of the surge arresters, which are normally
connected near the cable terminations. This voltage rise Figure 16-9 Maximum real power transfer for XLPE and
HPFF cables.
16-17
transfer curves for HPPF cables assume LPP insulation system impedances) at the cable location. The reactors
for system voltages of 230 kV and above. This figure can either be directly connected at line voltage, or they
gives an estimate of the maximum circuit lengths for can be connected to the power system through tertiary
uncompensated cables at various system voltages, with windings of a substation three-phase transformer or
all of the reactive charging current supplied from one single-phase transformers. The majority of installed
end of the cable circuit. A more realistic maximum cable shunt reactors at 60 kV and above are in the 30 to
circuit length is when the real power transfer capability is 300 MVA (three-phase) range (IEEE Std C37.015 1993).
reduced to 80% of the thermal capacity of the cable. The largest single-phase units have ratings of 125 MVA
Inductive VARs consumed by the cable, even during or 375 MVA as a three-phase group. Tables 16-5 shows
heavy load periods, are not sufficient to compensate typical characteristics for EHV iron-core reactors that
capacitive VARs, so a cable never approaches surge- are connected to the power system at line voltage. Table
impedance loading. 16-6 shows medium-voltage reactor characteristics,
which are typically air-core and usually connected to
Shunt Reactors transformer tertiary windings.
Shunt reactors are often placed at one or both ends of
long cable circuits (Figures 16-10 and 16-11) to limit the Shunt reactors may be directly connected to the substa-
voltage rise during light-load conditions, especially tion bus or termination station, or may be connected
where the local power system is relatively weak (high with load break switches. The latter allows the reactor to
be removed from the system when the system can safely
absorb the reactive current during high-load periods.
Shunt reactor switching imposes a unique and severe
duty on the connected system and circuit breaker. Par-
ticularly at high voltages, the current to be interrupted is
generally less than 300 A. However, successful current
interruption is a complex interaction between the circuit
breaker and the circuit. This interaction can result in
significant overvoltages. IEEE Standard C37.015,
Application Guide for Shunt Reactor Switching, should
be reviewed for additional information concerning
shunt reactor switching. Shunt reactors are also avail-
Table 16-5 EHV Reactor Electrical Characteristics (IEEE

Std. C37.015-1993)
System Normal
Voltage Rating Frequency Rated Induc-
(kV) (MVAR) (Hz) Current (A) tance (H)
765 150 – 300 60 113 – 226 5.17– 10.35
525 135 60 148 5.43
Figure 16-10 138-kV air-core shunt reactor (courtesy 400 120 – 200 50 173 – 289 2.55 – 4.25
author).
236 125 60 306 1.18
132 55 50 240 1.0
115 25 60 126 1.4
Table 16-6 Medium-Voltage Reactor Characteristics (IEEE

Std. C37.015-1993)
System Normal
Voltage Rating Frequency Rated Induc-
(kV) (MVAR) (Hz) Current (A) tance (H)
36 35 – 100 50/60 560 – 1600 98 – 34
24 35 – 100 50/60 840 – 2400 44 – 15
17.5 40 – 80 50/60 1400 – 2600 20 – 10
12 40 – 75 50/60 1900 – 3600 10 – 5
Figure 16-11 138-kV, 30-MVAR, iron-core shunt reactor
(courtesy author).
16-18
able with under-load tap changers that make it possible allowable loading of the cable. The heating effect and
to vary the shunt compensation over a range from losses from the reactive current can be reduced by plac-
approximately 50 to 100%. ing reactors at both ends of the circuit and, ideally,
selecting their relative sizes so that the reactive current
A relatively recent development in shunt compensation flows approximately equally out of both ends of the
is the advent of static VAR compensators (SVC). These cable circuit (Figure 16-12 B) for normal load condi-
are power electronic devices that work in combination tions. The contribution of the reactive current to heating
with reactors to provide a continuously adjustable reac- at the hottest spot is reduced by a factor of four when
tance source. Their main advantages are that they can the reactive current flows equally out both ends of the
be used to vary the amount of cable compensation as cable, as compared to all of the charging current flowing
system load fluctuates, and the changes can be accom- out of one end. In preliminary planning studies, 60%
plished in a fraction of a second. Since the power elec- shunt compensation is often used for cost estimation.
tronics devices are most economical at high currents Some utilities install 100% shunt compensation.
and lower voltages, SVCs are connected to the power
system through transformer tertiaries or dedicated
transformers. Table 16-7 summarizes the advantages
and disadvantages of the various shunt-compensation
methods.
The optimum amount and location of shunt compensa-

tion must be determined by running load- flow cases at
different load levels and studying the effects of future
system expansion. In systems having a significant
amount of cables, system power factor can be influenced
at light-load periods to the extent that the system could
go leading if not corrected by reducing other capacitive
VAR sources. In general there are definite advantages in
reducing losses by having one reactor per phase at each
end of the cable circuit. This is because the reactive cur-
rent flowing through the cable causes I2 R losses in the
cable (Figure 16-12 A). Reactive current loss generated
by the cable capacitance is not uniform, and generally
causes the ends of the cable to operate hotter than the
center. Because heating will be greatest at the end, and
because the thermal environment is usually more or less
uniform throughout the cable length, this potential “hot Figure 16-12 Shunt compensation of cable circuits.
spot” at the end of the cable circuit will dictate the
Table 16-7 Comparison of Shunt Compensation Schemes

Type of Switching
Location Device Device Advantages Disadvantages
Fixed Value Iron-Core Lower cost, less
None
Transmission Line Reactor or complications for system Not controllable
(Direct or Disconnect)
Air-Core Reactor restoration
On/off controllable, better
Transmission Line Iron-Core Reactor Circuit Switcher Higher cost
system reliability
Variable Iron-Core Reactor Wide range of
Transmission Line Circuit Switcher Higher cost
(Tap Changer) compensation
Not as effective as HV
connection because of
Transformer Tertiary Air-Core Reactor LV Circuit Breaker Lowest cost transformer impedance.
Induction heating in nearby
metallic equipment
Static VAR Compensator High cost,
Transformer Tertiary LV Circuit Breaker Continuously controllable
(SVC) more complicated
16-19
Although shunt reactors can be used to reduce the adjustments and transformer voltage taps. Shunt reac-
losses in a transmission cable circuit, they have iron core tors or capacitors can also be added where necessary to
and copper losses that should be taken into consider- maintain a fairly flat voltage profile across the system
ation in an economic analysis. Typical losses in EHV under normal load conditions.
shunt reactors are approximately 0.15% of the MVAR
of the reactor. For example, a 150 MVAR reactor would The distribution of real power flow is a function of the
have losses of approximately 250 kW. relative voltage angles around the system and the inter-
connecting impedances. Unfortunately, it is not as easily
Keeping the reactive current flowing equally out both controlled as reactive power flow because the circuit
ends of a cable circuit typically requires more than just impedances are fixed, and it may not be economical to
placing equal-sized shunt reactors at each end of the control power flows by adjusting angle (and therefore
cable. If system voltages are lower at one end of the the generated power) at the various generating units.
cable, the cable capacitive current will flow toward that Hence, where network branch loadings are not approxi-
end, resulting in greater heating there. In addition to the mately equal to the thermal capacities of the circuits, the
reactive current generated by the cable, there may also power transfer cannot be increased once a circuit is
be reactive current from other parts of the ac system loaded to its thermal limit even though other circuits
flowing on the cable. Because of its low reactance, the may not be very heavily loaded.
cable tightly couples the two parts of the system that it
interconnects, so that even a small difference in voltage The flow of real power in an ac power system can be cal-
between the two systems can result in very large reactive culated by the following relationship:
currents flowing from generators in one area to genera-
V1 V2
tors or loads in another area. This system reactive cur-
P12 = sin δ MVA
rent also causes heating. Any such current contributes X 12 16-41
to heating of the cable, and limits the real power that the
cable can carry. This dynamically changing situation is Where:
one reason for using adjustable reactive power devices P12 = power flow between points 1 and 2, MVA.
for cable shunt compensation, such as reactors with V1 = magnitude of voltage at point 1, kV.
underload taps or SVCs. V2 = magnitude of voltage at point 2, kV.
X12 = reactance between the two points, Ω.
For long cable circuit lengths, it may become necessary δ = angle between the voltage phasors at the
to place shunt compensation at some intermediate point two locations, degrees.
or points along the cable circuit to reduce the maximum
charging current flowing in the cable, and therefore bet- If a cable and an overhead line are in parallel, as illus-
ter utilize the cable’s ampacity for real power transfer trated in Figure 16-13, the voltage magnitudes and
(Dougherty and Shifreen 1962; Cowley and Foxall angles are the same at the ends of the cable and over-
1983). The cost of the added land, equipment, and reac- head line. Therefore, the power that flows through the
tor losses must be compared to the cost of a larger cable cable and the overhead lines is inversely proportional to
conductor and higher cost of cable losses to determine if the reactance of each of the circuits. Since a cable typi-
and where such intermediate shunt compensation is
warranted. As with many engineering trade-offs, eco-
nomics usually influences the decision rather than the
theoretical limits shown in Figure 16-9.
16.3.3 Load Sharing and Power Flows

Another commonly encountered problem with the inte-
gration of transmission cables into power systems is
uneven load sharing between overhead lines and under-
ground lines. This “load hogging” by underground
cables is a result of the significantly lower series reac-
tance of cables compared to overhead lines.
The reactive loading of various system branches is con-

Figure 16-13 Load sharing between parallel
trolled by the magnitude of the voltages across the sys- overhead and underground transmission lines.
tem and can be adjusted by generator excitation
16-20
cally has a series reactance which is one-half to one- input and output voltages. A phase-angle regulator
third that of an overhead line, the power flowing over may be used to balance loading between two cables that
the cable is approximately twice to three times that of have different lengths, ampacities, designs, or any other
the overhead line. If the cable and the overhead line have difference that requires control of the load sharing
about the same ampacity, the cable will be severely over- between them.
loaded before the power flow in the overhead line
reaches half of its rating, assuming that the circuits are Air-core reactors (see Figure 16-15) can be placed in
approximately the same length. series with a cable circuit to reduce the amount of power
flow over the cable.
Even two cable circuits that are not identical will not
share the total power equally. Because the reactance of
both cables is small, even an incremental difference will
cause some difference in the load sharing between the
two circuits. Such an imbalance would be acceptable
only if the ampacities of the cables happen to match the
natural imbalance in power loading. Significant imbal-
ances can result from parallel cables of unequal length.
The higher total series reactance of the longer cable will
cause it to carry less power than the shorter cable.
Calculating the load sharing between an overhead line

and a cable between the same substations is straightfor-
ward using Equation 16-41; however, actual networks
are much more complex, and a load-flow program must
be used to calculate possible cable overloading. If load-
flow studies reveal load-sharing problems with a
planned cable circuit, the following methods could be
used to redistribute the power flows.
• Use capacitors in series with overhead lines to reduce
the total reactance of the line
(Figure 16-14 A)
• Use a phase-angle regulating (shifting) transformer
to artificially decrease the angle across the cable cir-
cuit (Figure 16-14 B) Figure 16-14 Methods to improve load sharing
• Install a reactor in series with the cable circuit (Fig- between overhead and underground transmission
ures 16-14 C and 16-15) circuits.
• Complete the connection between the two transmis-

sion circuits at the distribution voltage rather than at
the transmission voltage (Figure 16-14 D)
Continuous or step adjustment of series capacitors is

not currently practical; they are expensive since they
must be mounted on insulated platforms at line poten-
tial. Therefore, they are not typically used for any over-
head circuits other than very long ones. Phase-angle
regulators are most widely used in transmission cable
circuits because motor-operated winding taps in the
phase-angle regulators (or phase shifters as they are
sometimes called) can provide an adjustable phase shift
to compensate for changing conditions.
Figure 16-15 Series reactors for load flow control
A phase-angle regulator is a transformer that is wound (courtesy author).
in a manner to create a change in the angle between the
16-21
Connection at distribution voltage may or may not be whether a transition station is required, underground
possible, depending on the power system where the lines need to be constructed in areas that have access to
cable is being applied. It is sometimes possible to supply trucks, trailers, and other equipment used to support
one (or more) substation transformers with the cable the duct, manhole, and cable installation.
and a second transformer with the overhead line where
the two circuits are only connected through the rela- If a transition station is needed in an urban or suburban
tively high impedance of the distribution network. If area, then noise produced by equipment in the transi-
this is done, there are usually disconnect switches, which tion station (i.e., shunt reactors or forced-cooling equip-
are normally open but may be closed during emergency m e n t fo r H P F F c abl e s y s t e m s ) m ay n e e d t o b e
conditions. addressed. Noise barriers can be installed to reduce the
impact of noise from a shunt reactor or mechanical
16.3.4 Hybrid Lines refrigeration equipment; however, there may still be
Hybrid overhead/underground lines are relatively com- some level of noise at the perimeter of the station.
mon in power systems but have special design issues that Protection and Control
must be considered. Overhead-to-underground transition stations can cause
Transition Stations a number of difficulties that must be addressed when
Whenever a transmission line transitions from overhead designing transmission line relay protection and control
to underground, a transition station or transition pole is systems. Depending on system requirements, the transi-
required (See Figure 16-16). If no switches or reactors tion station could be as simple as a single pole used to
are required, and the cable has extruded-dielectric insu- connect the overhead line to the underground cable
lation, then steel poles are often used to make the over- (Figure 16-16) or as complicated as a substation with
head-to-underground transition, as shown in Figure circuit breakers, shunt reactors, protective relaying, ac
16-16. Riser pole OH/UG transitions have been used for and dc power supplies and communication facilities.
HPFF and HPGF transitions in a few cases; however, Each transition station design provides its own relaying
HPFF and SCFF OH/UG transitions typically require concerns and design considerations.
a transition station (Figure 16-17) at one or both ends
of the line. Some of the main points to consider concerning protec-
tion and control of lines with transition stations include
Transition stations can occupy several acres of land. the following:
Depending on the voltage of the line, the height of the • Variations in the overall line impedance caused by
structures within a transition station can reach in excess having one or more parallel cables out of service for
of 100 ft (30 m). If shunt reactors are required to com- maintenance while the remaining cable(s) and over-
pensate the underground line, then the transition station head portions remain in service.
needs to be located where very large trucks can access it
• Impedance mismatch between the overhead and
to deliver and someday replace the reactors. This
underground sections, preventing the use of power
requires either the construction of a road or location of
line carrier equipment.
the transition station close to a road. Regardless of
Figure 16-16 Typical 115/138-kV XLPE OH/UG transitions Figure 16-17 Typical 230-kV HPFF OH/UG transition
(right, courtesy Georgia Power). station (courtesy author).
16-22
• Shunt reactors, if located at transition station, Another issue for high-speed reclosing is the effect of
require local fault protection or transfer tripping of trapped charge on the magnitude of switching overvolt-
the line end breakers. If local fault protection is used, ages. Special circuit breakers with preinsertion resistors
normal power supply redundancy (e.g., dual ac and may be required to control switching surges if automatic
dc systems) is required. reclosing is employed. This is because transmission
• HPFF or SCFF fluid pressurization units located in cables can retain trapped charge for a long period of
time, resulting in significantly higher switching surge
transition stations must also have redundant power
overvoltages compared to overhead lines. Potential
sources to run the pressurizing pumps and alarm
transformers, if present, will discharge the trapped
annunciation/telemetry for one or more days.
charge within several cycles.
• Fault detection and/or location equipment at the
transition station is desirable to determine if the fault 16.3.5 System Restoration
is in the cable section to enable or prohibit automatic
System restoration can present problems for utilities
reclosing (see “Hybrid Line Protection” in Section
having an extensive cable system. Cables can be dam-
16.3.9).
aged by overvoltages, because of the reactive current
Automatic Reclosing generated by cable capacitance. Shunt reactors associ-
Transmission line automatic reclosing after line-to- ated with a cable circuit should be energized at the same
ground faults is frequently used, because lightning- time as the cable itself. There can still be substantial
related faults are frequently cleared after the fault arc is voltages rises, however, depending upon reactor size and
extinguished. However, careful consideration must be ability of the system to absorb reactive current.
given to automatic reclosing for combination over-
head/underground lines. Automatic reclosing should Energizing a cable without sufficient shunt compensa-
not be permitted on hybrid lines with HPFF or SCFF tion, or before the system is able to absorb the reactive
cable sections unless the protective system is designed to current from the cable without excessive overvoltages,
determine that the fault is not in the underground por- can result in cable damage and collapse of the system
tion of the line (IEEE 1996). If the fault is in a pipe-type that has been restored up to that time. If reactors are
cable section, and the fault arc did not burn a hole switched for voltage control, they must be easily and
through the pipe wall initially, there is a significant risk properly applied for system restoration.
that a second fault would cause pipe burn-through. A
pipe burn-through in a buried section would result in a HPFF cable systems require special consideration
significant fluid leak and possibly additional cable dam- related to system restoration. This is because more than
age due to water ingress. With XLPE transmission 24 hours is required to repressurize the cable system
cables, there is less chance of significant additional dam- after a total loss of dielectric fluid pressure. The pot-
age caused by reclosing into a fault. However, there is a heads must be vented to expel gases, which collect at
risk of causing a manhole fire if the fault is in a splice. high points during the loss of pipe pressure. Also, the
For SCFF and extruded-dielectric cable systems, auto- cable insulation may be wrinkled if the fluid pressure is
matic high-speed reclosing can also damage sheath volt- reapplied too quickly. These restoration problems may
age limiters (SVL) if the MCOV (maximum continuous be avoided if dielectric fluid pressurizing units are
operating voltage) of the SVLs is too close to the rms equipped with a backup source of power (e.g., com-
(root-mean-square) magnitude of the fault-induced pressed gas-driven pump or emergency generator) to
sheath voltage. Some utilities do use automatic reclosing keep the pressurizing pumps running even during total
on hybrid OH/UG lines with short sections of extruded- loss of electrical power.
dielectric cable without differentiation of overhead ver-
sus underground faults. However, most utilities do not 16.3.6 System Reliability
use automatic reclosing without differentiation of over- Although transmission cables are in general very reli-
head versus underground faults if the underground seg- able, their repair times are much longer than those for
ments are more than 15 to 25% of the total line length. overhead lines. For example, typical forced-outage rates
for 138-kV overhead and 138-kV HPFF underground
Relaying and instrument transformers can be installed transmission lines are 0.00468 outages/year/mile (0.0029
at transition stations to detect if the fault is in the outages/year/km) and 0.00165 outages/year/mile (0.0010
underground portion of the line and send a signal to the outages/year/km), respectively (Mallard 1966). (This is
remote substations to permit or prohibit reclosing. the fair-weather forced-outage rate. The adverse-weather
forced-outage rate for the same overhead line is 0.01 out-
ages/year/mile. The fair- and adverse-weather forced-
16-23
outage rates are the same for the cable circuit.) Typical the desired reliability. However, in the latter case, the loss
outage durations on the 138-kV overhead and under- of a single transmission cable between two substations
ground lines are 9 hours and 3 weeks, respectively. might result in long-term loss of load. Because of this
and the long repair times for underground transmission
As a result of the much longer repair times for under- lines, it may be necessary to design cable circuits with
ground transmission lines, it is common practice to 100% redundancy to meet the single-contingency design
design cable circuits with 100% redundancy. That is, two criteria. That is, two parallel cable circuits are installed,
parallel cable circuits are often installed with each of the with each of the two cables having a sufficient capacity to
two cables having sufficient capacity to carry the rated carry the rated load of the circuit alone for the assumed
load of the circuit alone for the duration of a contingency. duration of a contingency. For example, if a utility’s sys-
tem planning department determines that a 250-MVA
Forced outages for underground transmission cables power transfer capability is required for a circuit to meet
may be grouped into two categories. The first category peak demand, then two circuits with a long-term (i.e.,
consists of forced outages that do not involve dielectric 300 hours) emergency rating approaching 250 MVA
breakdown in the cable system and that may be repaired would be required if underground transmission cables
in less than 12 hours. An example of this type of forced are used. The length of the assumed outage duration
outage is an event that results in the temporary loss of varies from one utility to another. Another variation in
dielectric fluid pressure. The line may trip on low pres- design practice is whether to use cable long-term emer-
sure, but it may be restored quickly if pressure does not gency temperature ratings, which could result in loss of
drop too low. In this type of situation, the same emer- cable life, or to take a more conservative approach and
gency measures to temporarily supply loads or to use only the cable’s normal continuous rating. It appears
restore service may be used as is the case for overhead that the trend is toward the former as utilities search for
lines. However, the second type of forced outage, caused ways to make better use of existing facilities.
by dielectric breakdown in the cable system, either
through mechanical damage or a defect in the affected Another decision that must be made by utility planning
device, results in much longer outage times. This second departments is whether to choose the ampacity of
type of forced outage requires different design practices underground transmission cables to match that of adja-
for overhead and underground transmission circuits. cent overhead lines when the cable or cables are in series
with the overhead line. Most utilities recognize that
The design practices vary among utilities; however, they design considerations for overhead lines differ from
may be generally grouped into what are referred to as underground lines. For example, the conductors on an
single- or double-contingency design criteria. The overhead line may be determined by economic consider-
single-contingency (n-1) design criterion, which is the ations (i.e., cost of losses), corona, or audible noise con-
most common, means that the system must be designed siderations rather than the required current-carrying
such that an outage of any single transmission line in a capacity. In either case, the capacity of a cable circuit
given area will not result in loss of load. Similarly, the should be based on projected peak-load conditions for
double-contingency (n-2) design criterion means that the planned circuit and a single- or double-contingency
the system is designed such that there will be no loss of reliability criteria, as determined by system load-flow
load as a result of the loss of any two transmission lines studies.
at the same time. Actually, utilities that use the double-
contingency design criteria usually limit its application 16.3.7 Harmonic Resonance
to areas of the network, such as high-rise office areas, In some cases, the high capacitance of cable circuits can
where the total loss of load has particularly adverse cause harmonic resonance problems. Most problems of
consequences. harmonic resonance are due to parallel resonance
between shunt capacitance and the inductive imped-
These design criteria have different consequences ances of the power system at lower harmonics (Kundur
depending on whether the transmission system is a dense 1994). More specifically, when cable circuits are
urban network or a more widespread or sparse transmis- installed in an area where the cable system equivalent
sion system. In the former case, the load that is carried source impedance is relatively high (i.e., weak power
by a transmission circuit can generally be rerouted even system), the shunt capacitive reactance of the cable cir-
if the circuit is an underground cable with a three- or cuits may resonate with the power system source induc-
four-week repair time. In this type of system, there are tance at frequencies close to the second, third, or fourth
many instances where it is not necessary to put in multi- harmonic frequencies. The addition of shunt reactors
ple cable circuits between two substations to maintain generally does not solve harmonic resonance problems
16-24
because the value of their impedance increases with har- If the insulation coordination process is being per-
monic frequency. System planning studies should formed for a new type of cable system or for higher sys-
include the calculation of harmonic resonant frequen- tem voltages where industry-standard overvoltage
cies to avoid low-order harmonic resonance. If parallel withstand capabilities have not been established, then a
resonance frequencies are too close to the second, third, basic impulse insulation level (BIL) and possibly a basic
or fourth harmonic frequencies, the magnitude of tran- switching impulse insulation level (BSL) must be speci-
sient overvoltages may become problematic (Colla et al. fied after the most economical overvoltage protection
2005). methods have been investigated.
Overvoltages
16.3.8 Insulation Coordination
The three categories of overvoltages to which the cable
The objective of insulation coordination is to ensure system may be subjected are lightning overvoltages,
that the cable and its accessories are not exposed to switching overvoltages, and temporary ac overvoltages.
overvoltages that are high enough to cause a breakdown
or degrade their performance. An insulation coordina- Lightning Overvoltages
tion study includes the following steps: The cable is subjected to lightning strokes that strike
adjacent overhead lines. The magnitudes of the light-
1. Tentatively select a surge arrester to protect the cable
ning overvoltages are determined, therefore, by the
circuit. This selection is based on the highest normal
lightning performance of the overhead lines and the
voltage to be experienced by the arrester and on the
attenuation of the surges as they are conducted into the
maximum temporary overvoltage (TOV) expected at
cable system.
the cable circuit. The maximum TOV is usually
caused by fault- related overvoltages, but may result
In theory, lightning could strike the cable terminations,
from Ferranti rise and/or resonant conditions.
but substations and transition stations that contain the
2. Tentatively select the number and location of the cable terminations are typically built with overhead
surge arresters with respect to the equipment. If space shield wires, which effectively prevent the cable termina-
permits, it is recommended that surge arresters be tions from experiencing direct strokes. If a detailed insu-
placed at both ends of the cable circuits very close to lation coordination study is being performed, the
the cable terminations. maximum lightning overvoltage that may be conducted
3. Identify the magnitude and characteristics of over- into the substation or underground transition is deter-
voltages to which the cable system will be subjected. mined by an analysis of the nature of the strokes that
A detailed analysis of lightning and switching surge can enter the substation via the overhead transmission
overvoltages requires the use of a digital computer lines, and then an EMTP computer program is used to
program such as the EMTP (Electromagnetic Tran- calculate the overvoltage that actually appears along the
sients Program) (Dommel 1969; Long et al. 1990). length of the cable. An EMTP program simulates the
4. Check that the overvoltages are held by the arrester lightning surges that travel along transmission lines near
to values that will not damage the cable system and the speed of light, similar to a traveling wave in water.
other adjacent equipment. This step usually involves
checking the transient overvoltages against the indus- In the case of a lightning surge, the electrical charge
try-standard voltage-withstand levels for the cable, injected into the power system by the lightning stroke
joints, and terminations. travels along a transmission line, and the current corre-
sponding to the electrical charge is related to the voltage
5. Check the energy that surge arresters must absorb
surge by a quantity called the surge impedance. This
when they limit switching-surge overvoltages. The
surge impedance, Zc, is determined by the following for-
surge arrester must be able to absorb this energy
mula and is normally in the range of 20-60 ohms for
without damage, and this determination should be
cables.
made with an EMTP program for accurate results.
6. Check that the current through the arrester is within 60 Di
limits in terms of magnitude and duration. High- Zc = log e Ω
ε Dc 16-42
magnitude arrester currents are often associated with
switching of cables (low surge impedance) and can Where:
result in arrester damage. Di = diameter over cable insulation, in.
(or mm).
7. Repeat steps one through six, if necessary, to deter-
Dc = diameter over conductor, in. (or mm).
mine proper protection of the cable system with an
ε = insulation dielectric constant, dimension-
adequate safety margin.
less.
16-25
The surge impedance for most overhead lines is from

300 to 400Ω..
power frequency voltage, ( 2 )
3 Vl-l (rms) up to 3 to
3.5 times the crest value.
The relative surge impedance of overhead lines and
Temporary Overvoltages
underground lines is important because it determines
Temporary ac overvoltages (TOV) refer to overvoltages
the magnitude of the voltage which initially enters the
that may last from several cycles of the power frequency
cable from the overhead line. The voltage is indicated by
voltage to several seconds. The most common cause of
the following equation:
TOV is voltage rise on unfaulted phases during a line-
2Z c to-ground fault. This short-term increase in voltage is
VT = VI typically 10-30% above normal voltage. For longer
Z c + Z oh 16-43 cables, Ferranti rise may be important, and, especially
Where: for transformer-terminated cables, resonant conditions
Zc = traveling wave surge impedance of cable, Ω. may be significant.
Zoh = traveling wave surge impedance of over-
Limiting Overvoltages
head line, Ω.
The most effective measure of limiting overvoltages on a
VI = magnitude of incident traveling wave, V.
transmission cable circuit is to install surge arresters
VT = magnitude of voltage at the cable terminal,
close to the cable potheads at one or both ends of the
V.
cable circuit. In most cases, surge arresters are specified
for both ends of the cable circuit without a detailed
Thus, if the surge impedance of the cable is 30Ω, and
analysis of the overvoltages that may enter the cable cir-
that of an overhead line is 375Ω, the magnitude of the
cuit. This is because it is a conservative approach and
initial voltage wave that travels into the cable will be
because the cost of performing a detailed overvoltage
only 15% of the incident voltage from the line. The mag-
analysis may not be warranted, considering the cost of
nitude of the voltage on the cable eventually reaches the
the arresters versus the cost of the analysis. There are
same level as the voltage on the line, after reflected
some instances where substation space limitations make
waves have had time to travel back and forth several
it difficult to install surge arresters at both ends of the
times on the cable.
cable. A detailed overvoltage analysis may then be nec-
essary to determine if adequate protection can be
Another important characteristic of traveling voltage
obtained with an arrester at only one end of the cable.
waves is that they will double in magnitude at an open
One relatively common example of this is the case where
circuit if there is no surge arrester to limit the voltage.
the cable is terminated in a transformer without an oil-
Therefore, surge arresters should be connected to both
to-air pothead (Short et al. 1958). Since the effectiveness
ends of a cable circuit wherever possible.
of a surge arrester decreases with distance from the
device that it is supposed to protect, an analysis is neces-
Naturally occurring lightning overvoltages vary consid-
sary to see if the transformers, the terminations, and the
erably in both their duration and the rise time (time that
first sections of cable will be subjected to dangerously
it takes to increase from zero to approximately 90% of
high overvoltages.
their peak value). However, in general, they have rise
times of several microseconds and durations of tens to
Even though surge arresters may be placed on both ends
hundreds of microseconds.
of a cable circuit, it is possible for transient overvoltages
Switching Overvoltages to occur along the cable, which are higher than the pro-
Switching overvoltages are produced both by opening tective or discharge level of the arresters. The phenome-
and closing switching devices at various points in the non that makes this possible is related to the reflections
power system as well as by electrical faults. They also of traveling waves at the cable terminals and the super-
vary considerably in magnitude and voltage waveshape, position of the waves elsewhere along the cable. Refer-
but they are approximately 100 times slower than light- ence (Hileman 1982) indicates that a voltage of up to 1.5
ning overvoltages. That is, they generally have rise times times the arrester protective level may occur at locations
in the hundreds of microseconds and durations that away from the arresters. The user must verify that cable
range from hundreds of microseconds to several milli- BIL provides an adequate safety margin.
seconds. Their magnitudes and waveshapes can be pre-
dicted by computer programs such as EMTP or by a Arrester Selection
TNA (transient network analyzer). Their magnitude Arrester selection is a trade-off between using an
may range from slightly above the crest value of the arrester with a low protective voltage level and making
sure that the arrester will not fail due to the current and
16-26
energy that it must dissipate from surges and whether it determined by simulation and compared with arrester
can survive the transient overvoltages (TOV). The pro- capability. Note that for metal-oxide arresters, many
cess for estimating the maximum TOV is: manufacturers define energy absorption capability on
1. Calculate the temporary overvoltages during fault the basis of multiple operations over a 1-minute period.
conditions using R0/X1 vs. X0/X1 curves in industry For energy absorbed over a shorter period, such as
standards and reference books (IEEE Std. C62.22 might occur during a cable-energizing operation, it is
1997; Westinghouse 1963). R0, X0, and X1 are the prudent to reduce the rated energy capability by about
zero-sequence resistance, zero-sequence reactance, 15% from that given by the manufacturer.
and positive-sequence reactance of the power system
at the arrester location, which is normally determined A third important aspect of applying arresters on cable
from a short-circuit computer program. circuits is the magnitude and duration of the current.
Arrester current is normally much higher in magnitude
IEEE C62.22-1997 contains curves for coefficient of when limiting surges on cable circuits versus overhead
grounding (COG), which is defined as the ratio circuits, because cable surge impedance is far lower than
(expressed in %) of the highest line-to-ground voltage that of overhead lines. For silicon-carbide-type gapped
(rms) of the two unfaulted phases at a location dur- arresters, the gap structure in particular can be damaged
ing a line-to-ground fault to the normal line-to-line by high-magnitude, long-duration currents (lightning
voltage at that location. currents, while higher in magnitude, do not have the
2. Check for the possibility of other sources of TOVs same long duration). If the arrester current exceeds
that would cause higher voltages than calculated in 2 kA, energy in coulombs through the gaps (I2 t) should
step one. TOV phenomena that should be checked be summed until the current drops below about 1.5 kA.
are load rejection, switching, and resonance. The coulombs-per-operation should not exceed about
3. Check for prolonged system overvoltages that may one coulomb if the arrester is properly applied.
occur as a result of light-load conditions, Ferranti
voltage rise, or other abnormal conditions. For metal-oxide arresters, the problem is that at high
currents, the current tends to concentrate in specific
For modern metal-oxide surge arresters, the maximum areas of the disc. Although the total energy capability
continuous operating voltage (MCOV) of the arrester may not be exceeded, localized heating occurs, and
must be higher than the highest continuous line-to- block cracking and failure can result. One of the major
ground system voltage. At a minimum, the highest arrester suppliers recommends the values shown in Fig-
continuous system voltage is often assumed to be 5% ure 16-19 for determining arrester capability, but it
(345-kV system and below) or 10% (500-kV system and should be understood that the criteria may not apply to
higher) higher than the nominal system voltage (i.e., other manufacturers’ arresters.
[138 kV] (1.05)/( 3 ) = 84 kV for a 138-kV system), but
the result of Item 3 above may reveal higher power-
frequency voltage conditions.
The TOV capability of a metal-oxide arrester also has to

be above the system TOV. Arrester TOV capability is a
function of the magnitude of the overvoltage, its dura-
tion, and the amount of energy (from surges) absorbed
by the arrester immediately prior to being subjected to
the system TOV. For example, one manufacturer sug-
gests that its arrester can be subjected to a TOV of
about 1.38 times the MCOV (1.13 times arrester rating)
for a period of 1 second (assuming prior energy absorp-
tion) without sustaining damage. Figure 16-18 illus-
trates TOV capability as a function of voltage and time.
A second surge arrester parameter that should be con-

sidered is the energy absorption capability of the
arrester. This is not a problem with lightning surges, but
Figure 16-18 TOV capability following absorption of full
is important for switching surges, especially for cable rated energy.
application. Arrester energy requirements should be
16-27
To apply the criteria, divide the current waveform into Worked Example 4: Surge Arrester Selection for
small time elements (e.g., 0.05 microseconds). For each 138-kV Cable
time segment, determine the current magnitude, and Determine an appropriate metal-oxide arrester rating
from Figure 16-19, determine the “allowable time.” and estimate the protective ratio for a 10-mile, 138-kV
Divide the actual time element (they need not be all the extruded-dielectric cable system. The TOV capability of
same) by the allowable time. The sum over the entire the type of arrester to be used is shown in Figure 16-18.
wave must be less than or equal to 1.0 for proper appli- A short-circuit study yielded the following ratios of
cation. If the sum exceeds 1.0, either a higher-rated sequence impedances at the arrester locations:
arrester or a special arrester designed with additional
columns must be used. R0/X1=0.75
X0/X1=2.0
Protective Ratio
Once the arrester is selected and transient overvoltages
The cables have 1500-kcmil (760 mm 2 ) conductors,
are determined, the protective level of the surge arrester
which are 1.412 in. (35.9 mm) in diameter, and have an
should be compared to the voltage withstand levels of
insulation thickness of 850 mils (21.6 mm). The first
the cable system components.
step is to estimate the overvoltages that will result on the
unfaulted phases when a single line-to-ground fault
The ratio of these two numbers, equipment insulation-
occurs near the cable. Using the above ratios and the
withstand level to arrester protective level, is called the
grounding coefficients curves from IEEE Std. C62.22, it
protective ratio. Accepted industry practice is to design
is estimated that the voltage on the unfaulted phases will
for a protective ratio of at least 1.2 to provide an ade-
be approximately 70% of the normal line-to-line voltage.
quate safety margin (IEEE 62.22 1997). In most practi-
Assuming that the system is operating at maximum volt-
cal cases, the margin will be (and probably should be)
age, this would result in the following TOV on the
considerably higher than 20%.
unfaulted phases:
For system voltages below 345 kV, the only insulation
Fault TOV = 0.7(138) 1.05 = 102 kV
withstand to be considered for transient overvoltage
protection is the basic impulse insulation level (BIL) of
The continuous line-to-ground voltage may be as high
the cable, joints, and terminals. The basic switching
as 1.05(138)/ 3 = 84 kV.
impulse insulation level (BSL) for terminations must be
considered for system voltages of 345 and 500 kV.
Thus, the arrester must have an MCOV of at least 84 kV.
If no other temporary overvoltages were identified that
would exceed the above voltage of 102 kV, and the TOV
would not be expected to last more than 1 second, then,
assuming the arrester has immediately previously
absorbed its energy limit, the required arrester TOV
capability would be 107 kV (102/0.955) as determined
by the arrester characteristics in Figure 16-18.
A metal-oxide arrester with a rating of 108 kV (87 kV

MCOV) would meet the above requirements, but the
MCOV would be marginal. The next higher ratings of
120 kV (98 kV MCOV) would provide additional mar-
gin and should be selected as the tentative choice. The
120-kV arrester has a TOV capability of 142 x 0.955 kV,
or 136 kV, and this would be more than sufficient.
BIL is coordinated against the 10-kA discharge level of

the arrester, and the 120-kV arrester has a 284-kV dis-
charge level. If arresters are placed at both ends of the
cable circuit, the maximum voltage surge at the cable
Figure 16-19 Application for high-current metal-oxide terminals would be 284 kV. A traveling wave transients
arresters rated 54-360 kV (GE Arresters). program would be required to calculate the maximum
voltage that can occur in the cable as a result of subse-
16-28
quent traveling wave reflections; however, a conservative Fault Protection Schemes

assumption is that the voltage will not exceed 1.5 times Protective relaying schemes for underground cable sys-
the incident or arrester discharge voltage. Thus, the tems generally use some form of a communications
maximum lightning overvoltage in the cable should not channel between the two ends of the cable circuit. Since
exceed 1.5 x 284 kV, or 426 kV. almost all cable faults are line-to-ground faults, ground-
fault sensitivity is of utmost importance. Therefore,
With a BIL of 650 kV for 138-kV cables and accesso- high-speed pilot-relaying systems are the most common
ries, the protective ratio for cable system components relaying schemes applied for transmission cable systems.
not immediately protected by arresters would be The most commonly used relaying schemes for trans-
650/426 = 1.53. mission cable circuits are:
1. Directional Comparison. This protection scheme
Since the protective ratio of 1.53 is well in excess of the combines the use of a directional phase and ground
minimum recommended value of 1.2, the choice of the protective relays at both ends of the line with a com-
120-kV arrester rather than the 108-kV arrester is justi- munications link between the two ends to give fast
fied. tripping for all internal line faults and avoid tripping
for faults external to the protected line.
16.3.9 System Protection
2. Current Differential. This protective relaying scheme
In general, the schemes used for fault protection of compares quantities derived from the currents at one
transmission cables are the same as those used for over- end of the circuit with those derived from the oppo-
head transmission lines (IEEE 1996). However, special site (remote) end of the line. Communications
attention must be given to the following considerations. between the two ends of the cable circuit may be by
1. The circuit impedance per unit of length of cables is ac pilot wire, microwave, or optical fiber cables. Cur-
low compared to that of overhead lines. This affects rent differential protection schemes are frequently
the impedance settings and the ranges of relays that applied to underground transmission lines because
may be used for circuit protection. they are less dependent on the cable characteristics
2. Faults on underground cables are almost always per- that have an impact on distance relays. The relaying
manent; whereas, faults on air-insulated overhead settings for this relaying scheme are small and easy to
lines will clear in many cases after the fault current is compute.
interrupted. Consequently, reclosing on a faulted 3. Phase Comparison. Phase comparison relaying
cable should be blocked by protective relaying. schemes compare the phase angle of the phase cur-
3. The zero-sequence impedance of pipe-type cables rents at the two ends of the cable circuit to determine
varies as a function of the fault current magnitude. if the fault is on the protected line. Close agreement
The settings of ground distance relays must take this of the phase angle of the primary currents indicates
impedance variation into account over the expected that the fault is external to the line, while phase
range of fault currents. angles that are out of phase indicate that the fault is
in the protected line. A communication channel is
4. The angle of the zero-sequence impedance values for
necessary to provide remote end information to per-
cable is low compared to that of overhead lines, and
mit the phase comparison scheme to operate prop-
the positive-sequence angles are relatively high. This
erly. Cable charging current magnitude is generally
may result in the need to use a relay with special char-
small compared to the magnitude of fault current,
acteristics to better match the cable impedance char-
thus phase angle difference due to charging current is
acteristics.
usually not a problem in fault detection. However, on
5. The charging current in cables is high and may cause longer cable circuits with weak sources, phase angle
the phase distance relays to overreach if this is not displacement due to cable charging current should be
considered in selecting the relay setting. studied to prevent incorrect tripping due to faults
6. The ratio of zero-sequence to positive-sequence external to the protected cable.
impedance of cable circuits is considerably different
from that of overhead lines. This difference must be Distance relays are also used for protecting under-
considered when setting ground distance relays. ground cables (Kasztenny et al. 2004), but setting the
7. Cable circuit energization and de-energization result distance relays for underground transmission lines can
in high transient currents. The frequency and magni- be quite challenging. This is because of items 3 through
tude of these currents depend on the capacitance, 6 listed in the introduction to this section (16.3.9). Con-
inductance, and resistance of the circuit. sidering all of these complexities in selecting the proper
relay settings, most relay engineers prefer to protect
16-29
transmission cables using line-current differential pro- typical zone 1 reach, which is 80% of the total cable plus
tection systems, or phase comparison relays. overhead line positive-sequence impedance. For faults
in this zone, the relay sends a direct transfer trip signal
Hybrid Line Protection to trip the remote circuit breaker and allows high-speed
Fault protection schemes for hybrid transmission lines reclosing at the remote end for single-line-to-ground
must take into account the different reclosing practices faults. This scheme also permits high-speed reclosing for
for overhead and underground lines. Many utilities will single-line-to-ground faults for the previous condition at
allow high-speed reclosing if the overhead portion of the local end near the cable. The distance relay at the
the line length is much greater than the underground end of the line opposite from the cable has only one
cable. Lines where the cable length is less than 15 to 25% zone 1 element. The reach of this element is set at 80%
of the total circuit length frequently permit autoreclos- of the overhead line positive-sequence impedance.
ing for hybrid lines with XLPE transmission cables. If Faults detected in this zone trip the local breaker, send a
the underground cable is pipe-type or SCFF cable, direct transfer trip signal to trip the remote breaker, and
reclosing is always prohibited unless a line-current dif- allow high-speed reclosing. Faults detected in an over-
ferential relay system is used to differentiate between reaching zone 2 do not permit high-speed reclosing.
faults in the overhead and underground sections.
A differential current relaying scheme (Figure 16-20 B)
Another important factor in selecting the relay scheme is commonly used when the cable segment is in between
for hybrid lines is whether the cable segment is at the two overhead line segments. In this case, the differential
beginning of either end of the line (Figure 16-20 A) or if current relay that is monitoring the currents at the two
it is between two overhead line segments (Figure 16-20 ends of the cable positively identifies when the fault is in
B). If the underground transmission line segment is at the cable segment and blocks autoreclosing at the two
one end of the line, and the line length is much longer ends of the circuit via a digital communications (e.g.,
than the cable segment, two instantaneous zone 1 ele- fiber optic cable).
ments are set at the relay near the cable end of the line to
discriminate between faults in the cable and the over- Low Cable System Pressure Protection
head line section and to block autoreclosing for cable Some transmission cable systems include pressure
faults. The instantaneous zone 1 element (zone 1-1) for switches to trip the cable circuit to avoid cable or termi-
the relay near the cable end is set at 120 to 150% of the nation damage due to unacceptably low fluid or gas
cable positive-sequence impedance. Operation of this pressures in the cable or terminations.
zone trips the local breaker and sends direct transfer trip • Extruded Dielectric Cable Systems. Some extruded-
to trap and block high-speed reclosing of the remote cir-
dielectric cable terminations are filled with pressur-
cuit breaker. The second instantaneous zone 1 (zone 1-
ized SF6 gas at a pressure of approximately 45 psig
2) element of the relay near the cable end is set at the
Figure 16-20 Relaying schematics for hybrid transmission lines.
16-30
(315 kPa). In this case, pressure switches are provided 16.4.1 Cables without Ferromagnetic
with alarm contacts if the pressure drops below Components
15 psig (105 kPa) and to trip the circuit if the pressure Two-dimension or three-dimension magnetic field cal-
drops to atmospheric pressure. The majority of culations in the vicinity of underground cable systems
extruded-dielectric cable terminations are filled with without ferromagnetic components are performed using
dielectric fluid (polybutene or silicone oil) at close to Biot-Savart’s law and the principle of superposition.
atmospheric pressure. Some utilities require equip-
ment to detect low liquid levels in this type of termi- Two-Dimension Magnetic Field Calculations
nation; however, this is not common practice. The magnetic field may be calculated by applying Biot-
Savart’s law and superposition to all of the current-
• HPFF Cable Systems. On some HPFF cable systems,
carrying conductors in a cable circuit. In most cases, the
two dedicated pressure switches are wired in series to
transmission line is assumed to be straight and very
trip the circuit for low hydraulic system pressure. This
long compared to the distance from the cables where the
provides protection from cable system electrical
magnetic field is to be calculated. Using these assump-
breakdown for the case of an energized system that is
tions, Biot-Savart’s law reduces to the following equa-
rapidly losing pipe dielectric fluid or an inadvertent
tion (Equation 16-44), which may be used to calculate
closing of valves that supply pressure to the cable
the magnitude of the magnetic field density, B, in teslas
pipe. These pressure switches should have a setting of
(or gauss) at a distance, r, from an infinitely long line
approximately 100 psig (690 kPa). Normally the
carrying a current, I. The direction of the magnetic field
practice is not to trip HPGF cable systems on low gas
is perpendicular to a line connecting the conductor and
pressure.
the calculation point as determined by the right-hand
• SCFF Cable Systems. SCFF cable systems usually screw rule of basic electromagnetic field theory (see
have an alarm system to monitor either pressure or Figure 16-21).
volume and notify the system operator of any abnor-
mal conditions. These systems are usually not set to Î
trip the cable circuit. B = μ 0 H = 4 π x 10 −7 tesla
2πr
Î
16.4 MAGNETIC FIELDS = 4 π x 10 −3 gauss
Applications for construction permits for transmission 2πr
16-44
lines in North America usually include calculations for Where:
above-ground electric and magnetic fields. For transmis- μ0 = permeability of free space, 4π x 10-7 hen-
sion cables, the electric field is of no concern, because ries/m (or volt-sec/A).
the electric field lines from the high-voltage conductor H = magnetic field, A/m (or oersteds).
terminate on the grounded cable metallic shield or r = distance from line, m (or in.).
sheath. There is little or no electric field outside of Î = current in line, A.
shielded power cables. The above-ground magnetic field
values for most underground transmission lines and If the current, I, is a sinusoidal current, then both I and
overhead transmission lines differ primarily in the rate B are phasor quantities. Equation 16-44 assumes that
at which the magnetic field values decrease as a function the current-carrying conductor is surrounded by a
of distance from the transmission lines. The maximum medium that has a relative magnetic permeability of 1.0.
magnetic field values at the center of OH and UG trans- This assumption is valid for single-conductor transmis-
mission lines are typically similar for equal load cur- sion cables in most soils. As discussed later, this assump-
rents; however, the magnitude of the magnetic field for tion is not applicable to pipe-type cables. The use of
underground lines decreases much more rapidly than for Equation 16-44 also makes the assumption that any
overhead lines. This is because the high-voltage conduc- zero-sequence component of the phase currents or
tors are much closer together for underground transmis- induced sheath currents return in a path far enough
sion lines. The soils that surround the underground away from the cables that they do not appreciably affect
cables have little or no effect on the above-ground mag- the field near the cable. This is usually a good approxi-
netic field values, because they have a relative magnetic mation, because the earth return current flows several
permeability of unity. thousand feet (hundreds of meters) below the surface of
the earth for typical soil conditions.
The gauss or milligauss (mG) is a common unit of mea-

sure for magnetic field density in the United States when
16-31
making calculations for transmission lines. Microteslas calculate the value of induced shield/sheath currents, if
(μT) is the unit for magnetic flux density used in most multipoint bonding is used. The components of the
other countries. Using the conversion factor of 1 x 104 total magnetic field produced by these induced
gauss per tesla (or weber/m2), the horizontal, (x), and shield/sheath currents must be included in the summa-
vertical, (y), components of the magnetic field density at tion shown in Equations 16-47 and 16-48.
an arbitrary point (xc, yc in Figure 16-21) are given by
the following equations. ^ n ^
B xTotal = ∑ B xi mG (or μT) 16-47
i =1
^ ⎡y +d⎤
B x1 = −2 × Î ⎢ c 2 ⎥ mG ^ n ^
⎣ rc ⎦ B yTotal = ∑ B yi mG (or μT) 16-48

i =1
⎡y +d⎤
= −0.2 × Î ⎢ c 2 ⎥ μT Since the currents in Equations 16-47 and 16-48 are
⎣ rc ⎦ 16-45 phasor quantities, the x and y components of the total
magnetic field are also phasor quantities. This, in turn,
^ ⎡y +d⎤ means that the magnetic field vector at any point in
B y1 = 2 × Î ⎢ c 2 ⎥ mG space traces out an ellipse as the vector rotates (EPRI
⎣ rc ⎦ 2005). As the total magnetic field vector rotates, its pro-
jections on the horizontal and vertical axes describe the
⎡y +d⎤
= 0.2 × Î ⎢ c 2 ⎥ μT horizontal and vertical components of Equations 16-47
⎣ rc ⎦ 16-46
and 16-48. It should be noted that the maximum values
of these components do not coincide with the instanta-
In Equations 16-45 and 16-46, the “^” notation is used neous vectors but with the projections of the rotating
to indicate phasor quantities. The dimensions yc, d, h, vectors at some instants during the cycle. A method to
and rc shown in Figure 16-21 have the units of meters. calculate the magnitude and direction of the major axis
The negative sign preceding the number 2 in Equation of the magnetic field ellipse from the real and imaginary
16-45 assumes that the current is flowing out of the components of the x and y components of the field is
figure. For multiple current-carrying conductors, includ- contained in the EPRI Red Book (EPRI 2005). It is the
ing any induced sheath currents, the total x and y magnitude of the major axis of the magnetic field ellipse
components of the magnetic field are the summation of that would be measured if a single-axis magnetic-field
the horizontal and vertical components of the magnetic probe were rotated to obtain a maximum reading at the
field produced by each of the currents. The following calculation point. Three-axis gaussmeters measure the
equations show the total horizontal and vertical compo- resultant value of the magnetic field (Equation 16-49).
nent of the magnetic field where there are n current-
^ 2 ^ 2
carrying conductors in the cable system. The normal mG (or μT) 16-49
procedure in calculating cable magnetic fields is to Br = BxTotal + B yTotal
Three-Dimension Magnetic Field Calculations
A three-dimension (3-D) calculation procedure must be
used to calculate the magnetic field for locations where
the cables are not straight and parallel to each other.
Splice vaults and bends in the cable alignment are exam-
ples where 3-D magnetic field calculations must be per-
formed to give accurate results.
Biot-Savart finite length current segment modeling is

used to calculate the magnetic flux density produced by
the currents flowing in the underground cable system in
areas where the cables are not straight and parallel to
each other. This calculation procedure models the three-
Figure 16-21 Buried conductor and geometric
dimension magnetic field produced by the segments of
parameters for magnetic field calculations. the transmission line by means of numerous straight
line, finite lengths, current-carrying segments or
“sticks.”
16-32
Using the Biot-Savart law of electromagnetic field the- Worked Example 5: Magnetic Field Calculations for
ory, the magnetic field produced by a finite-length cur- Cable System without Ferromagnetic Components
rent-carrying conductor is given by Equation 16-50 This example shows the procedure for calculating the
(Haus and Melcher 1989). Figure 16-22 shows the phys- above-ground magnetic flux density for the single-point
ical meaning of the variables in Equation 16-50 for the bonded 230-kV XLPE cable system in Example 1 (Sec-
3-D magnetic field. tion 16.2.5). Calculate the resultant magnetic flux den-
sity at 1 m (3 ft) above ground directly above the cable
G μ Î c × a ⎛ a • c a • b ⎞ trench and at a distance of 25 ft (7.5 m) from the center
B= 0 ⎜ − ⎟ teslas 16-50 of the cable trench. The cables are carrying 1000 A (rms)
4 π | c × a |2 ⎝ | c | | b | ⎠
of positive-sequence current.
Where:
B = 3-D magnetic flux density vector, teslas. The first step is to calculate the current that is induced
× = Indicates vector cross product. in the ground continuity conductor. This is accom-
• = Indicates vector dot product. plished by adding the voltages induced in the ground
P = Point where magnetic field is to be calcu- continuity conductor by the currents in each of the cable
lated. high-voltage conductors and dividing by the self-imped-
a = Current segment vector with current I, ance of the ground continuity conductor.
amperes.
c = Vector from point P to end of current car- Using Equation 16-20, the mutual impedance between the
rying segment, a. Phase A cable and the ground continuity conductor is:
b = Vector from point P to start of current car- 33500
rying segment, a. Zag = 0.01807 + j0.0529 log10
(10) 2 + (10) 2
The total magnetic field at any point in the vicinity of
= 0.01807 + j 0.1785 Ω /1000 ft
the cable system is then the vector sum of the magnetic
field produced by each of the finite-length current-car-
rying segments. The magnetic field value that is calcu- The Phase B and Phase C mutual impedances to the
lated is the rms value of the resultant magnetic field ground continuity conductor are shown in Table 16-8.
given by Equation 16-51, and the units are milligauss
(mG) or microteslas (μT). The rms value of the resultant
Table 16-8 Mutual Impedances and Induced Voltages for
magnetic field is the magnetic field value measured by Example 4
most commercially available three-axis gauss meters.
Mutual Impedance Voltage Induced in
The magnetic field values that are reported are at 1 m Current with GCC GCC
(3 ft) above ground level. Phase (A) (ohms/1000 ft) (V/1000 ft)
A 1000 + j 0 0.01807 +j 0.1785 18.07+j 178.51
Br = B2x + B2y + Bz2 mG (or μT) B -500 + j 866.6 0.01807 +j 0.1865 -170.53 – j 77.59
16-51 C -500 – j 866.6 0.01807 +j 0.1785 145.56 – j 104.91
Total -6.90 – j 3.98
Calculation of the magnetic field values should include
the impact of all induced currents (i.e., cable shields Using Equation 16-18, to calculate the self-impedance
and/or ground continuity conductors). of the ground continuity conductor with earth return is:
⎛ 33500 ⎞
Zgg = 0.01807 + R gg + j0.0529 log10 ⎜ ⎟
⎝ 0.221 ⎠
= 0.067 + j 0.274 Ω /1000 ft
Figure 16-22 Finite-length current-carrying segment or

“stick.”
16-33
The current induced in the ground continuity conductor Some transmission cables are surrounded by ferromag-
is the total induced voltage divided by Zgg. netic metals. For example, pipe-type cables are
contained in a carbon-steel pipe, a ferromagnetic metal.
Vinduced
Igcc = In some transmission cable installations, ferromagnetic
Zgg shields have been installed in the vicinity of single-
conductor cables to reduce the magnitude of the above-
−6.90 − j3.98
= ground magnetic field values.
0.067 + j0.274
The presence of ferromagnetic metals in the vicinity of
= −19.51 + j20.39 A transmission cables significantly complicates the com-
putation of the magnetic field values. Since the magnetic
The next step is to calculate the horizontal and vertical permeability of the ferromagnetic metal is much higher
components of the magnetic field produced by each of than nonferromagnetic materials (i.e., soil, concrete, air,
the currents in the cable system using Equation 16-45 etc.), the magnetic flux produced by the cable concen-
and 16-46. trates in the ferromagnetic components. The nonlinear
magnetic properties of the ferromagnetic metals mean
Starting with the Phase A current at 1 m above ground that the split of magnetic field between the ferromag-
at the center of the cable trench, the horizontal and ver- netic and nonferromagnetic components varies with the
tical components of the magnetic field are shown in magnitude of the cable currents. Another complicating
Table 16-9. factor is that eddy currents are induced in the ferromag-
Table 16-9 Horizontal and Vertical Magnetic Field netic components, which must be taken into account.
Components, One Meter above Ground at Center of Cable
Trench (x=0) As shown in Figure 16-23, the magnetic flux density
Horizontal Compo- Vertical Compo- (B) increases rapidly in ferromagnetic metals as the
Current nent nent magnetic field intensity (H) is increased. However,
Phase (A) Bx (mG) By (mG) there is a material-dependent saturation point above
A 1000 + j 0 -747.45 71.62 which there is very little increase in magnetic flux den-
B -500 + j 866.6 341.56- j 591.61 -29.87+ j 51.73 sity with increasing magnetic field intensity. The slope
C -500 – j 866.6 314.47 + j 544.67 -25.29 – j 43.80 of the B-H curve (ΔB/ΔH) at a specific point is the mag-
-19.51+ netic permeability.
GCC 13.43 – j 14.04 0
j 20.39
Totals -77.99 – j 60.97 79.62 + j 61.37
The resultant magnetic field is the square root of the

sum of the squares of the horizontal and vertical mag-
netic field components.
Br = B2x + B2y
= (−77.99 − j60.97) 2 + (79.62 + j61.37) 2

= 100.5 mG
=10.05 μT
16.4.2 Cables with Ferromagnetic Components

Ferromagnetic metals are metals that have magnetic
permeabilities that are much greater than the permeabil-
ity of free space and nonferromagnetic materials (Bozo-
rth 1993). A second characteristic of ferromagnetic
metals is that their magnetic permeability varies in a
nonlinear fashion with the applied magnetic field
strength. Figure 16-23 shows the B (magnetic flux den- Figure 16-23 B-H characteristics of commonly used
ferromagnetic materials.
sity) versus H (magnetic field intensity) characteristics
of several ferromagnetic metals.
16-34
Figure 16-24 shows the relative magnetic permeability shielding; however, they generally are based on the
(μ/μ0) for samples of two different pipe-type cable pipes. assumption that the magnetic permeability of the
Although all pipe that is used for pipe-type cable cir- shielding material is constant. An analytical description
cuits is relatively low-carbon steel, the magnetic proper- (Wait and Hill 1977; Hoburg 1994) of a pipe with high
ties of the cable pipe vary with the amount of carbon magnetic permeability and high electrical conductivity,
and the manufacturing process (i.e., annealing rate, with a dipolar source inside of the pipe (Figure 16-25) is
mechanical stresses, etc.). The magnetic permeability given in Equation 16-52. The shielding factor is a com-
(actually B-H curve) of a certain supplier or production plex number, because the phase angle of the magnetic
run can be measured in the laboratory using ring sam- field outside of the pipe is shifted compared to the mag-
ples of the pipe. The magnetic permeability of steel pipe netic field with no magnetic shield. In this case, the
is rarely specified or measured by the manufacturer. shielding factor, SF, is constant for all locations outside
of the pipe. Equation 16-52 contains the modified Bessel
Accurate calculation of the magnetic field for cable sys- functions K and I of the first order and their derivatives,
tems with ferromagnetic components requires the appli- K’ and I’. Tables of values of the modified Bessel func-
cation of complex computation methods that are tions, as well as methods for calculating these values,
capable of taking into account the nonlinear character- may be found in several mathematical handbooks
istics of the B-H curve of the ferromagnetic material. (Abramowitz and Stegun 1964).
The finite element computation method is the most
common method (CIGRE 1997) for calculating the SF = N / M (unitless)
magnetic field of cable systems with ferromagnetic com- b
ponents. Finite difference and boundary element com- N = 2 μ r γb ⎡⎣ K1 ( γb) I1' ( γb) − K1' ( γb) I1 ( γb) ⎤⎦
puter programs have also been used for this type of a
magnetic field calculation. Most of these computer pro- M = ⎡⎣μ r K1 ( γa) − γaK1' ( γa) ⎤⎦ ⎡⎣μ r I1 ( γb) + γb I1' ( γb) ⎤⎦
grams require an operator who is very skilled in their
use, and the programs can be very sensitive to the many − ⎡⎣μ r I1 ( γa) − γaI1' ( γa) ⎤⎦ ⎡⎣μ r K1 ( γb) + γb K1' ( γb) ⎤⎦
subtleties involved in problem setup and solution con-
16-52
vergence (Clairmont and Lordan 1999).
Where:
Simplified calculation methods have been developed for
μ = magnetic permeability of the material, H/m.
some ferromagnetic shields that give approximate values b = Do/2 the pipe outside radius, m.
of the magnetic field shielding factor. These calculation a = (Do/2) the pipe inside radius, m.
methods model both flux shunting and eddy-current γ = (1 + j) / δ complex wave number, m-1.
δ = 2 ,m
ωμ σ
ω = power frequency, Hz.
σ = electrical conductivity, S/m.
Figure 16-24 Relative magnetic permeability versus

flux density for two samples of two different steel pipes Figure 16-25 End view of steel pipe with dimensions in
(Bahder et al. 1979). Equation 16-50.
16-35
Worked Example 6: Calculate the Magnetic Installation Geometry

Shielding Factor of an 8-in. Carbon Steel Pipe at It is well known (Haus and Melcher 1989) that the value
60 Hz of the magnetic field at any location is a linear function
The dimensions of an 8-in. (200 mm), Schedule 20 steel of the instantaneous currents in all of the cable system
pipe are: components, provided that there are no ferromagnetic
Pipe inside diameter: 8.125 in. (206.4 mm) components. Consequently, the above-ground magnetic
Pipe outside diameter: 8.625 in. (219.1 mm) field can be reduced by optimum placement of the
Pipe wall thickness: 0.25 in. (6.35 mm) cables to maximize cancellation of the magnetic field
between the individual currents in the cables. This is fre-
The electrical conductivity of carbon steel pipe is 6.76 x quently accomplished in three-phase power systems by
106 S/m. placing the cables close together and by optimum phase
placement of cables when there are multiple three-phase
The relative magnetic permeability of carbon steel pipe groups of cables. The maximum above-ground magnetic
at low flux densities is 180. field value can also be decreased by increased cable
burial depth.
The skin depth, δ, calculated using the above, parame-
ters is: Figure 16-26 shows the maximum magnetic field that
would occur 3 ft. (1 m) above ground as a function of
δ= 2 phase spacing for several different cable installation
(377 ∗180 ∗ 4 π 10 −7 ∗ 6.76 × 10 6 configurations with 700 amperes flowing in the cables
and single-point bonding or cross-bonding of the
= 0.00186 m (0.0732 in.)
sheaths. As can be seen from this figure, the maximum
γ = 536.8 + j 536.8 m -1 magnetic field, which generally occurs directly above the
centerline of the circuit, doubles when the phase spacing
is increased from 6 to 12 in. (15 to 30 cm). It should be
Using Equation 16-52, the numerator, N, and denomi-
noted, however, that the current-carrying capacity of the
nator, M, of the shielding factor are:
cables decreases as the phase spacing is decreased
N = 2 ∗ (0.110 / 0.103) ∗180 ∗ (58.82 + j 58.82) [(-0.00425 + j 0.00421) because of the mutual heating between the cables, as dis-
- (0.00425 - j 0.004286)] cussed in Chapter 11. The figure also shows that for a
= - 382.15 + j 0.0025 phase spacing of 9 in. (23 cm), the magnetic field varies
M = [1.672 x 10-23 - j 2.447 x 10-23 ] [-1.983 x 10 26 + j 3.198 x 10 26 ] almost two to one between a horizontal and triangular
- [2.596 x 10 24 - j 6.65 x 10 24 ] [5.263 x 10 -25 + j 9.97 x 10-27 ] installation configuration. Although the triangular con-
= 11141 - j 492.9 figuration gives the minimum magnetic field, its use is
SF = (-382.15 + j 0.0025) /(11141 - j 492.9) normally limited to situations where the cables can be
= - 0.0342 - j 0.00151 directly buried and the three cables are touching.
= 0.03425∠182°
Since the magnetic field near long parallel conductors

The shielding factor calculations indicate that the steel with balanced three-phase currents decreases approxi-
pipe will attenuate the magnetic field produced by cables mately with the square of the distance from the center-
inside of the pipe by a factor of 29.2 (i.e., 1/0.03425), line of the conductors, the maximum magnetic field
and the phase angle of the magnetic field on the outside
and inside of the pipe will be almost 180° out of phase
with the phase A current inside of the pipe.
16.4.3 Magnetic Field Management Methods

A number of methods have been devised by electric util-
ities (Bascom et al. 2006; Zemyan 1996) and research
organizations (Cooper 1993) to manage power fre-
quency magnetic field levels in the vicinity of under-
ground cables. Most practical magnetic field
management methods generally fall into one of the fol-
lowing categories. Figure 16-26 Magnetic field as a function of phase
spacing and installation configuration.
16-36
directly above the cables can be reduced by increasing When the cable shield/sheaths are multipoint grounded,
the depth of burial. Figure 16-27 shows a plot of the the induced shield/sheath current reduces the magnitude
maximum magnetic field 3 ft. (1 m) above the surface of of the magnetic field external to the cables. Figure 16-30
the ground for different depths of burial. Again, this shows the amount of magnetic field reduction that
magnetic field reduction method would reduce the occurs when the cable shield/sheaths are multipoint
ampacity of a cable circuit, as shown in Chapter 11. grounded, as a function of shield/sheath resistance. This
method of magnetic field reduction is not as practical as
Passive Conducting Loops adjusting the phase relationships, because the induced
The above-ground magnetic field values can also be shield/sheath currents decrease the ampacity of the cir-
reduced by the installation of conducting loops in the cuit. Also, the low-level magnetic field away from the
proximity of the cables (Bascom et al. 2006; Vavra and cable can actually be increased by the induced sheath
Wanda 2006). The magnetic field produced by the currents for horizontal or vertical installation configura-
power cables induces a current in the conducting loops, tions, because unequal currents are induced in the three
which, in turn, opposes the magnetic field produced by cable sheaths, and there is a net effective current that
the power cables, as stated by Lenz’s law. This method affects the magnetic field.
includes the use of ground continuity conductors or spe-
cial external shielding loops. For example, if two ground Cancellation with Multiple Cables per Phase
continuity conductors are placed above the cables One of the most effective methods of reducing the mag-
shown in Figure 16-28, there is a significant reduction in netic field in the vicinity of transmission cables is to
the above-ground magnetic field values, as shown by the alter the phase relationship of the cables when there are
magnetic field plots in Figure 16-29. two, three-phase cable circuits in close proximity or a
single circuit with two cables per phase. Figure 16-31
Figure 16-27 Magnetic field as a function of burial depth.
Figure 16-29 Magnetic field (1 m above) ground with and

without passive loop shielding (see Figure 16-28).
Figure 16-28 Transmission cables with passive loop Figure 16-30 Magnetic field as a function of shield/sheath
shielding. resistance.
16-37
shows the reduction in magnetic field that is obtained shunting and eddy currents induced in the metal. These
when reverse phasing is used for two cable circuits two mechanisms are described below.
installed in vertical configurations in the same duct
Flux Shunting
bank. The primary disadvantages of this reversed-
Materials with a relative magnetic permeability signifi-
phasing installation are the problems and possible haz-
cantly higher than one (i.e., ferromagnetic materials)
ards that may result during maintenance and repair of
conduct magnetic fields much more easily than nonfer-
the cables. Many utilities prefer to use the same phase
romagnetic materials. This is analogous to the conduc-
relationship for all cable circuits to avoid confusion in
tion of electric currents in metals versus semiconducting
identifying the cables in manholes. It should also be
materials. Figure 16-32 shows the magnetic field lines of
noted that the ten-to-one reduction in magnetic field
a cross section of a cylindrical shield of a high magnetic
produced by reverse phasing of the cables occurs only
permeability when it is placed in a uniform dc or low-
when the load current in the two different cable circuits
frequency magnetic field. The area inside of the cylinder
is exactly the same. The effectiveness of the magnetic
is partially shielded from the magnetic field.
field reduction decreases as the difference in current
magnitude between the two circuits increases.
The cylindrical shield gathers flux over a region whose
Metallic Shielding size is determined by the diameter of the shield and
A quantitative measure of the effectiveness of a mag- shunts it through the thickness of the shield. The flux
netic shield in reducing the magnetic field magnitude at density within the shield material is amplified from the
a given point is the shielding factor. The shielding factor, external uniform flux density by a factor determined by
SF, is defined as the ratio of the magnetic field magni- the ratio of shield diameter to thickness. The flux that
tude (rms value) at a given point when the shield is leaks into the shielded region (inside of the cylinder) is
present to the magnetic field magnitude (rms value) at attenuated from that within the shield material by the
the same point when the shield is absent. Thus, the ratio of material permeability to air permeability. These
smaller the shielding factor, the better the shielding effi- effects combine to produce a shielding factor that can be
ciency. A shielding factor of 1.0 represents no shielding. decreased (improved shielding) either by increasing the
material permeability relative to air or by increasing the
In general, the shielding factor is a function of the posi- material thickness relative to the shield diameter. This
tion at which it is measured. If the properties of a shield flux-shunting mechanism works for constant (dc) mag-
are independent of the magnetic field magnitudes, the netic fields as well as low-frequency magnetic fields. It is
shielding factor is independent of the magnitude of the important to note that, for a closed magnetic shield, the
magnetic field produced by the source. However, if the source needs to be entirely inside or outside the shielded
magnetic permeability of a shield material depends enclosure for it to be effective. For example, if a cylinder
upon the flux density within the material, the shielding is used to shield an area from the magnetic field pro-
factor becomes dependent upon the magnitude of mag- duced by currents in electric cables, then the cylinder
netic field from the source. should enclose all current-carrying conductors to shield
the large area external to the cylinder, or the cylindrical
Magnetic field shielding with metal structures occurs shield needs to enclose the area to be shielded with no
due to two different physical mechanisms termed flux current-carrying conductors inside of the shielded area.
Figure 16-32 High magnetic permeability cylinder placed

Figure 16-31 Magnetic field for like and reverse phase in a uniform magnetic field (example of flux shunting).
placement of cables.
16-38
It should be noted that placing a single current-carrying frequency. Equation 16-53 shows how the skin depth, δ,
conductor inside of a ferromagnetic pipe (with the return is defined in terms of the metal conductivity and perme-
current path outside of the pipe) does not reduce the ability.
magnetic field outside of the pipe.
2
Eddy-current Shielding δ= skin depth, m 16-53
When a metallic structure is exposed to a time-varying ω μσ
magnetic flux density, an electric field is induced in the μ = magnetic permeability of the pipe material,
material. If the material has electrical conductivity, σ, H/m.
the induced electric field creates an induced current σ = electrical conductivity of the pipe material,
within the material. This induced current constitutes an S/m.
additional magnetic field source that opposes the
change in the external magnetic flux. The superposition Shields fabricated from nonferromagnetic, conducting
of external and induced magnetic fields tends to “buck materials typically have a skin depth that is significantly
out” the imposed field, diverting the flux away from a less that the thickness of the shield. For example, at
shielded region in closed shield (Hasselgren and Luoni 60 Hz, the skin depths in copper and aluminum are 0.33
1995), while in open shields it produces a compression in. (8.5 mm) and 0.46 in. (11.8 mm), respectively. The
of the flux lines on the side of the source and a reduc- skin depth is significantly reduced in ferromagnetic
tion of the magnetic flux density beyond the shield (see materials, which usually show a lower electrical conduc-
Figure 16-33). tivity, but a much greater magnetic permeability. When
the skin depth is much less than the thickness of the
The induced current mechanism arises only when the metal, the induced currents flows uniformly over the
external magnetic field is time-varying, so that eddy- shield thickness. The metals used for conductive shield-
current shielding is not effective for shielding dc ing usually have electrical conductivity higher than 10
magnetic fields, and its efficiency increases with the MS/m. Copper and aluminum in the form of thin sheets
supply frequency. In conductive materials, the resulting (thickness of some mills or millimeters) are the most fre-
shielding factor is a complex number at each point in quently usually used nonferromagnetic shielding materi-
space, with a magnitude describing the ratio of magni- als.
tudes of shielded to unshielded fields at the point, and a
phase describing the difference in phase between Metals that have both good electrical conductivity and
shielded and unshielded fields. high magnetic permeability are often selected for mag-
netic shields because both shielding mechanisms (shunt-
Eddy-current shielding occurs in any electrically con- ing and eddy-current shielding) work together. Typical
ducting material, which may or may not be ferromag- ferromagnetic shielding metals are:
netic. In an ac field with angular frequency ω = 2π f, the • Carbon Steel. Common carbon steel, an alloy of iron
induced current density and total flux density decay
and carbon, is often used for magnetic shields
exponentially into the material, away from the surface
because it has a relatively high magnetic permeability,
of the shield that faces the source. A material-dependent
typically several hundred, and a low flux density to a
parameter, called skin depth, quantifies how deeply the
maximum of around 1000, and it has reasonably
magnetic field penetrates the metallic shield at a given
good electrical conductivity (i.e., around an eighth
that of copper). Although its shielding characteristics
are not as good as many other ferromagnetic metals,
carbon steel is relatively low cost, readily available,
and unaffected by mechanical stresses during instal-
lation. Low carbon steel (ASTM 1010), with 0.08 to
0.1% carbon, is preferred for magnetic shielding
because the permeability is higher than other carbon
steel alloys.
• Silicon Steel. Silicon steel or transformer grain-ori-
ented steel is an alloy of silicon and iron (3/97%) that
was developed to reduce losses in power transform-
Figure 16-33 High-conductivity, unity magnetic ers. It has significantly better shielding performance
permeability, cylinder placed in a uniform magnetic field than carbon steel. However, it is only available in rolls
(example of eddy-current shielding). of thin (i.e., 20 mils or 0.5 mm) sheets with a maxi-
16-39
mum width of several feet. So, multiple sheets of sili- They are effective in reducing the magnetic field pro-
con steel must be bundled together and preassembled duced by cables installed inside of them at all distances
to make useful magnetic-shielding structures. from the cable trench.
• Mumetal and Permalloy. Mumetal and Permalloy are
nickel-iron alloys (approximately 80/20%) that have The most commonly used open shield structures are flat
very high magnetic permeabilities. One variation of plates and inverted U-shaped shields placed on top of
mumetal, called supermalloy, has relative magnetic the cables. Both of these open shields are effective in
permeability values of up to 100,000. Mumetals typi- reducing the magnetic field directly above the cables;
cally are annealed in an oven with a hydrogen envi- however, they have little effect on the above-ground
ronment to maximize the magnetic permeability. The magnetic fields at distances greater than two or three
two primary disadvantages of mumetals are their times the width of the magnetic shield.
high cost and the fact that their magnetic permeabil- Practical Considerations
ity may degrade significantly during field fabrication Many magnetic field shields have the potential of signif-
if they are mechanically stressed or welded. Their use icantly reducing the above-ground magnetic fields pro-
is generally limited to shielding of laboratory equip- duced by transmission cables (Bucea and Kent 1998);
ment, electrical switchgear vaults, and room-size however, many design considerations must be addressed
areas. before magnetic field shielding can be applied in practi-
• Copper and Aluminum. Both copper and aluminum cal installations. Some of the practical design consider-
have been used for magnetic shielding of power sys- ations are:
tems equipment, because their high electrical conduc- 1. The most effective shields in reducing the magnitude
tivities make up for their lack of high magnetic of above-ground fields generate eddy currents in the
permeability and they are readily available. shield structure. However, these eddy currents gener-
ate joule heating in the metallic shield, which
Table 16-10 is a summary of the shielding parameters of increases the transmission line losses and reduces the
the commonly used shielding materials. The geometry ampacity of the cables because of the heat associated
and size of magnetic shielding structures have a signifi- with the losses. Hysteresis losses in ferromagnetic
cant impact on their magnetic-shielding performance. metals generate significant additional losses. For
example, the eddy-current and hysteresis losses in the
The various geometries used for magnetic shielding fall steel pipe for pipe-type cable systems often exceed
into the categories of open and closed shields. Closed 30% of the joule losses in the cable high-voltage con-
shields are cylinders (pipes), boxes, and spheres. Most ductors (see Chapter 11).
other types of shields (i.e., flat plates, channels, etc.) are
2. Many of the metals that are used for magnetic shields
open shield structures.
(i.e., carbon steel, silicon steel, aluminum, etc.) are
susceptible to corrosion. Therefore, some form of
Cylinders (pipes) and box-shaped shields are the most
corrosion protection should be considered when
efficient shielding geometries, because these geometries
designing the shield design.
completely enclose the source of the magnetic field (the
cables), and the closed paths inherent to these geome- 3. Most magnetic shields must have good electrical and
tries enhance flux shunting and eddy-current shielding. magnetic contact between sections that are assembled
in the field. If the electrical contact between adjacent
Table 16-10 Electrical Parameters of Commonly Used shield sections is poor, then eddy currents cannot
Magnetic Shielding Materials flow in these areas. If the magnetic contact is poor
between adjacent shield sections, then the flux shunt-
Relative Electrical
Metal Permeability Conductivity (S/m) ing field reduction mechanism is adversely affected.
For example, consider the inverted U-shape shield
Aluminum 1 3.05 x 107
shown in Figure 16-34. Due to thermal consider-
Copper 1 5.59 x 107
ations, the cables must be surrounded by compacted
granular backfill or Fluidized Thermal BackfillTM
Low Carbon Steel 200 to 1400 6.67 x 106 after they are laid in the cable trench. Consequently,
the top part of the magnetic shield must be installed
Silicon Steel 900 to 10,000 1.67 x 106 on top of the backfill material and then welded to the
side parts of the shield installed prior to the backfill.
Mumetal 40,000 1.64 x 106 Also, adjacent sections of the top shield should be
welded or have overlapping pieces to achieve good
16-40
shielding performance (Bottauscio et al. 2001). These 16.5 INDUCTIVE COORDINATION

welding operations are expensive due to the field Currents in underground transmission lines produce a
labor required and because this increases the time to magnetic field that surrounds the underground trans-
complete installation of the cable. mission line. This magnetic field, in turn, will induce
4. Installation of welded shields makes access to cables voltages on nearby electrical conductors such as tele-
for repairs more time consuming and difficult. On the communications lines, railroad communications facili-
positive side, steel shields can add significant ties, metallic pipelines, and fences. If the nearby
mechanical protection from dig-ins. conductors are continuous and grounded at more than
one location, then the induced voltages will produce a
Because of the preceding practical construction consid- current in the loop formed by the nearby conductors
erations, the most commonly used magnetic shields and earth. The magnitudes of the voltages induced on
applied to transmission cable installations in North nearby electrical conductors primarily depend on the
America have been flat steel plates (Durkin et al. 1994) following factors:
and carbon steel pipes (Zemyan 1996). Aluminum 1. The magnitude and phase angles of the currents flow-
plates, approximately 0.12 in. (3 mm) thick, have been ing in the underground transmission cables
used in some underground cable installations in Europe
2. The method of grounding the cable metallic shields
to reduce above-ground magnetic field levels. Studies
(i.e., single-point or cross-bonding). Designs that
(Canova et al. 2002) have shown that the parameter R
provide a return path for line-to-ground fault current
(see Figure 16-35) should be greater than six times the
that is close to the cable high-voltage conductors (i.e.,
distance between the cables and the plates to achieve
multi-point or cross-bonded sheaths) produce lower
good magnetic-shielding efficiency and to avoid
induced voltages compared to cable systems where
increased magnetic field levels near the ends of the alu-
the fault current return path is further away (e.g.,
minum plate. Adjacent plates should be welded together,
single-point with ground continuity conductor).
or additional aluminum plates should be placed on top
of the butt gaps to span the electrical discontinuity. Lab- 3. The separation between the transmission cables
oratory and field tests indicate that a shielding factor of 4. The radial distance (separation) between the cable
0.5 to 0.25 can be achieved with 0.12-in. (3 mm) thick trenches and the nearby electrical conductors
aluminum plates. 5. The longitudinal distance that nearby electrical con-
ductors are in close proximity to the underground
transmission line
6. The angles between the cable duct banks and the
nearby electrical conductors
The induced voltage per unit length (i.e., volts/1000 ft.

or volts/km) of the cable duct bank alignment at a given
transmission line current is highest when nearby con-
ductor(s) are approximately parallel to the underground
transmission line. The induced voltage per unit of length
decreases when the nearby metallic objects are oblique
Figure 16-34 Cross section of inverted “U” magnetic to the underground transmission line, and the induced
shield. voltage is zero if the nearby metallic objects are perpen-
dicular to the underground transmission line.
The voltages per unit of length that are induced in electri-

cal conductors in close proximity to the underground
transmission lines are relatively small compared to those
produced by overhead lines and some other underground
transmission lines because of the following factors:
1. The relatively close spacing between the cables (com-
pared to overhead lines) results in magnetic field val-
Figure 16-35 Flat plate (plane) magnetic shield geometry. ues that diminish rapidly with distance from the cable
trenches.
16-41
2. The relative phase placement of the cables for the (Table 16-12). No induced voltage limits for transient
underground lines are frequently selected to minimize conditions, such as fault current induced voltages, are
above-ground magnetic field values. The reduced contained in NACE RP0177-2000.
magnetic field values, in turn, result in relatively low 3. IEEE Standard 80 (IEEE 2000) indicates that electric
induced voltages. body currents below 116 mA / t , where t is the
duration of the current in seconds, can be tolerated
16.5.1 Industry Standards and Guides for by the average person.
Inductive Coordination
4. IEEE Standard 776 (IEEE 2000) is a document that
There are a number of different industry guides and provides calculation and measurement procedures to
standards that relate to inductive coordination between predict the exposure levels and susceptibility of com-
electric transmission lines and other utilities. These munications circuits to inductive coupling with power
guides and standards generally fall into the following lines. It also contains recommendations concerning
categories. threshold levels based on safety and telecommunica-
• Pipelines tions circuit degradation issues.
• Telecommunications lines 5. Canadian Standards Association C22.3 No. 3-98 is a
document that contains inductive coordination
• Railroads requirements for power lines and communications
facilities including railroad communications facilities.
Following is a summary of recommendations and
requirements in industry standards that relate to power 6. Canadian Standards Association C22.3 No. 6-91 is a
line inductive coordination. Most of these standards do document that is similar in scope to NACE Standard
not differentiate between requirements for overhead and RP0177-2000. It also requires that voltages induced
underground power lines. on pipelines shall be 15 V or less during normal oper-
ating conditions.
1. The 2002 National Electrical Safety Code’s (NESC’s)
(IEEE 2001) General Requirements Applying to
Underground Lines contains the following in Section Table 16-11 Human Resistances to Electrical Current
31.316 (Induced Voltage): “Cooperative procedures (NACE Std.RP0177 2000)
are recommended to minimize steady-state voltages Dry skin 100,000 to 600,000 ohms
from proximate facilities. Therefore, reasonable
Wet skin 1000 ohms
advance notice should be given to owners or opera-
tors of other known proximate facilities that may be Internal body-hand to foot 400 to 600 ohms
adversely affected by new construction or changes in
Ear to ear (about) 100 ohms
existing facilities.”
Section 09.095B2 of the NESC requires that “Made Table 16-12 Continuous 60-Hz Alternating Current Values
grounds or grounded structures should be separated by Affecting Human Beings (NACE Std. RP0177 2000)
3.0 m (10 ft) or more from pipelines used for the trans- Current Effects
mission of flammable liquids or gases operating at high
1 mA or less No sensation—Not felt
pressure (1030 kPa [150 lb/in2] or greater) unless they
Sensation of shock—Not painful; individual can
are electrically interconnected and cathodically pro- 1 to 8 mA
let go at will; muscular control not lost.
tected as a single unit. Grounds within 3.0 m (10 ft) of Painful shock—Individual can let go at will;
such pipelines should be avoided or shall be coordinated 8 to 15 mA
muscular control not lost.
so that hazardous ac conditions will not exist and Painful shock—Muscular control lost;
15 to 20 mA
cathodic protection of the pipeline will not be nullified.” cannot let go.
Painful shock—Severe muscular contractions;
2. NACE Standard RP0177-2000 (NACE 2000) is a 20 to 50 mA
breathing difficult.
document that addresses ac induced voltages on Ventricular fibrillation—Death will result if
metallic structures and corrosion control systems. 50 to 100 mA
prompt cardiac massage not administered.
Paragraph 5.2 of this document states, “For the pur- 100 to 200 Defibrillator shock must be applied to restore
pose of this standard, 15 V AC (rms) open circuit or mA normal heartbeat. Breathing probably stopped.
a source current capacity of 5 mA or more are con- Severe burns—Severe muscular contractions;
chest muscles clamp heart and stop it during
sidered to constitute an anticipated shock hazard.” 200 mA and
shock (ventricular fibrillation if prevented).
over
This requirement is based on the electrical resistance Breathing stopped—heart may start following
shock, or cardiac massage may be required.
values for typical human bodies (Table 16-11) and 60-
Hz alternating current values affecting human beings
16-42
16.5.2 Calculation of Induced Voltages and The mutual impedance, Z’ m, is given by the following
Currents equation
The magnetic field produced by the currents flowing in ⎛D ⎞

Ẑ'm = 0.01807 + j 0.0529 Log 10 ⎜ e ⎟ Ω /1000 ft
the cables induce voltages or currents in proximate (near ⎜ di j ⎟
⎝ ⎠
by) metallic conductors as shown in Figures 16-36 and
⎛ De ⎞
16-37. = 0.0289 + j 0.0846 Log 10 ⎜ Ω / km
⎜ d i j ⎟⎟
⎝ ⎠ 16-55
As indicated in Figure 16-36, the currents in the trans-
Where:
mission cables induce a voltage with respect to local
De = depth of earth return current (see Equa-
ground in parallel metallic objects. A typical example is
tion 16-15), in. (or mm)
a steel pipeline that is grounded at one end at a valve or
dij = distance between cable and pipe center-
cathodic protection location (i.e., right-hand end of Fig-
lines, in. (or mm)
ure 16-36) and is open at the opposite end (i.e., an insu-
lated pipe flange).
For three-phase cable systems, the total voltage induced
The voltage per unit of length (1000 ft or km) that is on the parallel metallic object is simply the vector sum
induced in a parallel metal pipeline (or other metallic of the induced voltage produced by each of the currents
object) is given by Equation 16-54. flowing in the three cables (Equation 16-56).
Vôc' = Iˆc Zˆ m′ volts/1000 ft (or volts/km) 16-54 VˆTotal

'
= IˆPh A Zˆ m′ A
Where:
+ IˆPh B Zˆ m′ B
Vôc
' = Induced open circuit voltage, volts
+ IˆPh C Zˆ m′ C V / 1000 ft (or V / km )
(rms)/1000 (or volts (rms)/km). 16-56
Where:
Iˆc = Current in transmission cable, A (rms).
VˆTotal
' =Total pipeline induced voltage, V/1000 ft (or
Zˆ m′ = Mutual impedance between transmission v/km).
cable and pipeline, Ω/1000’ (or Ω/km).
IˆPh A , IˆPh B ,IˆPhC = Transmission cable phase cur-
The carat (^) notation in Equation 16-54 indicates that rents A, B, and C, A.
the quantities are complex numbers (i.e., they have real
and imaginary components). The prime notation (’) indi- Zˆ m′ A ,Zˆ m′ B ,Zˆ mC
′ = Mutual impedances between
cates that the values are unit values (i.e., per unit of cable phases A, B, C, and pipeline,
length). Ω/1000 ft (or Ω/km).
When balanced three-phase currents are flowing in the

transmission cables, then the voltage induced in the par-
allel pipeline is less than the voltage produced by any
one of the three currents (i.e., the voltages partially off-
set each other).
When zero-sequence currents (i.e., in-phase currents)

flow in the underground line, then the pipeline induced
voltages do not cancel each other, and the induced volt-
age is significantly higher per unit of length compared
to positive-sequence currents. Single-line-to-ground
faults in the underground line (or through faults), which
are primarily zero-sequence current, in general, produce
the highest induced voltages on parallel metallic objects.
The mutual impedance shown in Equation 16-55 assumes

Figure 16-36 Voltage induced by transmission cables in that the underground line and the metallic object are par-
nearby pipeline. allel to each other. If the cables and parallel metallic
16-43
object are perpendicular to each other, then the mutual 16.5.3 Inductive Coordination with Pipelines
impedance and induced voltage is equal to zero. When underground transmission lines are installed in
close proximity to metallic pipelines, voltages and or
If both ends of the parallel metal object are grounded currents will be induced on the pipeline, depending on
(e.g., Figure 16-37), then the magnetic field produced by how the pipeline is grounded. The voltage or current
the transmission cable will induce a current in parallel induced on the pipeline will be a maximum when the
metallic pipelines. pipeline is parallel to the underground transmission line.
Conversely, the induced voltage or current will be a min-
The induced current flowing in a parallel metallic imum (theoretically zero) when the pipeline is perpen-
object, such as a pipeline, is given by Equation 16-57. dicular to the underground transmission line.
Vˆ ′ Pipeline Inductive Coordination Requirements

Iˆ = oc A Applicable industry standards (NACE 2000; CSA 2003)
Zˆ self
′ recommend or require:
16-57
Where: 1. The maximum induced voltage at all points along
Vôc′ = Open circuit induced voltage, volts/1000 ft. proximate pipelines should be less than 15 V (rms)
during normal operating conditions (NACE Std.
Zôc
′ = Self-impedance of the parallel metallic RP0177-2000 and CSA C22.3).
2. Grounded structures for underground transmission
object with earth return, ohms/1000 ft.
lines should be located at least 3 m from gas or flam-
mable liquid pipelines (NESC).
The self impedance, Z’self, is given by the following equa-
tion. 3. Below-ground fault current discharge facilities
should be located at least 10 m from metallic pipe-
Ẑ′self = R ′pipe + 0.01807 lines (CSA C22.3).
4. In areas of ac influence, any measured ac voltages
⎛ De ⎞ between a structure and ground (or some other adja-
+ j 0.0529 Log 10 ⎜ Ω /1000 ft
⎜ GMR pipe ⎟⎟ cent structure) shall be considered an indication that
⎝ ⎠ further study is required (CSA C22.3).
= R ′pipe + 0.0289 5. Permanent gradient control mats shall be used for all
above-ground metal pipeline appurtenances, except
⎛ De ⎞ test lead stations, if the touch voltage exceeds 15 V.
+ j 0.0846 Log 10 ⎜ Ω / km
⎜ GMR ⎟⎟
⎝ pipe ⎠ 16-58
6. Each utility should be aware of the other utilities’
facilities and cooperate in the mitigation of electrical
effects of one installation on the other. The mitiga-
As in the case of induced voltages, the induced current
tion required for a specific situation must be based on
for multi-point grounded parallel objects is the vector
safety considerations with good engineering judg-
sum of the currents induced by the currents flowing in
ment (NACE Std. RP0177-2000).
each of the three cables.
Based on the preceding recommendations, it is generally
necessary to perform induced voltage calculations for
underground transmission lines when they run parallel
to metallic pipelines for any appreciable distance. Some
states require that inductive coordination studies be per-
formed as a condition for placing transmission lines in
service. The general procedure for performing the
induced voltage calculations are summarized in the fol-
lowing section.
Calculation of Normal Operating Condition
Induced Voltages
The following parameters are necessary in order to cal-
culate voltages induced on parallel metallic pipelines by
an underground transmission line during normal oper-
Figure 16-37 Current induced in nearby pipeline. ating conditions.
16-44
1. Maximum positive and zero-sequence currents that the maximum induced voltage on sections of the pipe-
will flow on the underground transmission line or line between insulating flanges.
lines during normal and emergency operating condi- 1. Identify the location where a section of pipe is
tions. If there are multiple three-phase cable circuits grounded. A section of pipe for the purpose of
in the cable duct bank (or trench), then the direction induced voltage calculations is defined as the distance
of the currents for each of the circuits must also be between grounding point and the adjacent insulating
determined. The direction of current flow is impor- pipe flange or termination point. If it is not known
tant for duct banks with multiple cable circuits where the pipe is grounded between insulating pipe
because the induced pipeline voltages for currents flanges (or between a pipe flange and pipe end point),
flowing in opposite directions will partially cancel then it is assumed that either end of the pipe section
each other. is grounded.
The maximum positive-sequence current for induced 2. The plan and profile drawings are used to segregate
voltage calculations is typically the ampacity of the the pipe section into subsections where the distance
cable for continuous operation. Depending on local between the underground line and the pipeline may
requirements and emergency operating conditions, it be considered to be constant. The parallel (i.e., longi-
may also be necessary to perform induced voltage tudinal) length for each subsection, as well as the sep-
calculations using the short-term emergency ampac- aration distance (UG line to pipeline), must be
ity for the underground line. tabulated for each of the subsections. It is generally
In most cases, the zero-sequence current flowing in acceptable to use the horizontal distance between the
transmission lines does not exceed 2-4% of the posi- underground line and pipeline for induced voltage
tive-sequence current. calculations. However, if the lines are close together
2. The locations of the transmission cables in the cable and there is a significant difference in burial depth,
trench (i.e., duct bank cross-section). This informa- the distance between the pipeline and UG line should
tion is typically obtained from duct bank cross- be calculated from the horizontal and vertical separa-
section drawings at locations where the underground tions (i.e., the hypotenuse of the triangle).
line is parallel to the pipeline in question. 3. For each subsection, the unit induced voltage (i.e.,
3. Plan and profile drawings showing the distances volts/1000 ft or volts/km) is calculated using the sepa-
between the duct bank and the pipeline along the ration distances from the previous step, Equation
route of the underground transmission line. 16-55 (mutual impedance), Equation 16-56, and the
complex values of the positive-sequence currents
4. Locations where the pipeline is grounded and loca-
flowing in each of the three phases. The positive-
tions where there are insulating flanges in the pipe-
sequence induced voltage for the subsection is then
line. The pipeline may be directly bonded to
calculated by multiplying the subsection length by the
sacrificial anodes, rectifiers, polarization cells, etc.
unit induced voltage.
5. Information about whether or not the pipe has corro-
4. The previous step is repeated for the zero-sequence
sion protection. Most metallic pipelines (with the
current. For example, if the positive-sequence current
possible exception of cast iron pipelines) are installed
is 1000 A, and it is determined that the zero-sequence
with some type of corrosion-protection coating,
current does not exceed 2% of the positive-sequence
which electrically insulates the pipe from the sur-
voltage, then the induced voltage calculation for each
rounding soil. Although most pipe coatings have
subsection is performed with 1000 A * 2% /3 (e.g.,
some defects (called holidays), it is generally assumed
6.67 A) flowing in each of the three cable phases.
that the pipe is electrically insulated from ground,
except at locations where the pipe is intentionally 5. The total positive-sequence and zero-sequence volt-
grounded. If there is no pipe coating, then it is gener- ages for each of the pipeline sections in the pipeline is
ally assumed that there is no induced voltage (with calculated by adding up the induced voltage for each
respect to local ground) during normal operating of the subsections.
conditions. 6. The total induced voltage for each of the pipeline sec-
tions is then equal to the square root of the sum of
The principle of superposition and repeated application the squares of the positive and zero-sequence cur-
of Equations 16-55 and 16-56 are used to calculate the rents. For example, if the positive-sequence induced
positive- and zero-sequence induced voltages for loca- voltage for a pipeline section is 3 volts and the zero
tions where the distance between the underground line sequence induced voltage is 4 volts, then the maxi-
and pipeline may be considered to be constant. More mum induced voltage for that pipeline section would
specifically, the following procedure is used to calculate be 5 volts.
16-45
If the normal operating condition induced voltage for the transmission cable with earth return. Conse-
any of the pipeline sections exceeds 15 volts (rms), then quently, the majority of the single-line-to-ground
a method must be identified to reduce the induced volt- fault current would flow on the ground continuity
age for that section, or permanent gradient control mats conductors. For cross-bonded cable systems, the
should be used for all above-ground metal pipeline return path for line-to-ground faults will be a combi-
appurtenances, except test lead stations. nation of the cable sheaths and the ground continuity
conductors, if present.
Methods to reduce the pipeline induced voltage (NACE The impedances of the transmission cable high-
Std. RP0177 [NACE 2000]) are: voltage conductor with earth return is:
• Increase the separation distances between the pipe-
line and the underground line.
Ẑ′ce = R ′cond + 0.01807
• Ground the pipeline at the location(s) of the peak ⎛ De ⎞
induced voltage. This is typically accomplished by
+ j 0.0529 Log 10 ⎜ ⎟ Ω /1000 ft
means of polarization cells or solid-state isolators ⎝ GMR cond ⎠
such as the Dairyland Electric Isolation Surge Pro- 16-59
tection (ISP) device. Where electrolytic grounding Where:
cells, polarization cells, or similar devices are used, R’cond = resistance of the high-voltage conduc-
they should be properly sized, spaced, and physically tor, ohms/1000 ft.
secured in a manner that will safely conduct the max- GMRcond = geometric mean radius of the high volt-
imum amount of anticipated fault current. age conductor, in.
The impedance of the transmission cable high-volt-
Calculation of Fault-Induced Voltages age conductor with ground continuity conductor
In general, line-to-ground fault currents in underground return is:
lines generate the highest induced voltages in nearby
pipelines. The fault currents may be due either to a fault Ẑ′c −gcc = R ′cond + R ′gcc
in the underground line or to through faults. Single-line- ⎛ Sep ⎞
to-ground fault induced voltages are generally the most + j 0.0529 Log 10 ⎜ ⎟ Ω /1000 ft
⎜ GMR cond GMR gcc ⎟
severe due to the following reasons: ⎝ ⎠
• The currents during fault conditions are significantly 16-60
Where:
higher than normal and emergency operation condi-
R’gcc = resistance of the ground continuity
tions.
conductor, ohms/1000 ft.
• Three-phase fault currents are generally higher than GMRgcc = geometric mean radius of the ground
single-line-to-ground fault currents. However, the continuity conductor, in.
currents produced by single-line-to-ground faults are Sep = separation between center of transmis-
primarily zero-sequence currents, and there is no can- sion cable and ground conductor, in.
cellation with zero-sequence (in-phase) currents, as is
The impedance of a transmission cable high-voltage
the case for positive-sequence currents.
conductor with the three cable sheaths as a return
path is:
The following procedure is used to calculate the maxi-
mum induced voltage on sections of the pipeline R ′s
between insulating flanges. Ẑc −s = R ′cond +
3
1. The direction and magnitude of the currents for each
⎛ GMD ⎞
of the circuits in a duct bank must be determined by + j 0.0529 Log 10 ⎜ ⎟ Ω /1000 ft
performing short-circuit studies for the general loca- ⎝ Dsm ⎠
tion where the underground line and pipeline are
16-61
located. Where:
2. The division of fault current return between the cable R’s = resistance of the cable sheath, ohms/1000 ft.
sheaths, ground continuity conductor(s), and earth GMD = geometric mean distance between trans-
return must be determined. This division of the fault mission cables, in.
current between the available return paths is deter- Dsm = cable sheath mean diameter, in.
mined by the relative impedances of each of the 3. The pipeline ground points and subsections lengths
return paths. For example, the electrical impedance determined for normal operating condition induced
of the transmission cable with ground continuity con- voltages are used for the fault condition induced volt-
ductor return is much lower than the impedance of age calculations.
16-46
4. The tabulation of pipeline-to-cable duct bank dis- Figure 16-39 shows cross sections of the cable and pipe
tances from the normal operating condition induced trenches. The sheaths of the underground line are single-
voltages are used for the fault condition induced volt- point bonded, and a 4/0 AWG insulated ground conti-
age calculations.
5. The fault current induced voltage (per 1000 ft or km)
for each pipeline subsection is calculated from the
zero-sequence currents flowing in the transmission
cables and from the zero-sequence return currents
flowing in the opposite direction in each of the return
paths. The zero-sequence currents flowing in the
return paths are determined from step 2 above (divi-
sion of fault current between return paths). The total
induced voltage for each subsection is then equal to
the product of the subsection length and the induced
volts/1000 ft (or volts/km).
6. The total fault current induced voltage for each of
the pipeline sections is determined by performing a
complex number summation of all of the subsection
induced voltages. The magnitude of the fault current
induced voltage is then the magnitude of the complex
number induced voltage.
There are no specific guidelines or recommendations for

the maximum allowable pipeline induced voltages dur-
ing fault conditions. However, it is prudent to compare Figure 16-38 Plan view of 115-kV underground
transmission line and 8-in. gas pipeline.
the magnitude of the fault current induced voltage
obtained from the preceding calculation procedure to
the maximum value from the following relationship
(Meliopoulos 1988; IEEE Std. 80 [IEEE 2000]).
116 mA
I max = mA 16-62
t
Where:
Imax = maximum rms current, mA.
t = duration of fault current, sec.
The value of t in Equation 16-62 should be interpreted

as the maximum clearing time that is expected for
single-line-to-ground faults at the location of the under-
ground line.
Worked Example 7: Pipeline Inductive Coordination
This example demonstrates the method to perform
induced voltage calculations for a pipeline parallel to a
115-kV XLPE underground transmission line using the
equations and procedures outlined in Section 16.5.2
(“Calculation of Induced Voltages and Currents”). Fig-
ure 16-38 is a plan view of the underground transmis-
sion line and a gas pipeline. A cross section of the cable
duct bank and pipeline are shown in Figure 16-38. The
soil electrical resistivity is 100 ohm-m.
The 115-kV XLPE cables in Figure 16-38 have the

parameters:
Conductor Size: 2000 kcmil (1016 mm2), Cu Figure 16-39 115-kV cable and trench cross sections.
Sheath Mean Diameter: 3.5 in. (88.9 mm)
16-47
nuity conductor is installed in the cable trench, as quently, the voltage induced on the pipeline for the
shown in Figure 16-39 (a). duration of the fault would be 1000 * 1.106 V = 1106 V.
The maximum balanced, three-phase (i.e., positive- 16.5.4 Inductive Coordination with Fences
sequence) current that will flow in a 115-kV under- Voltages will be induced on nearby, parallel fences by
ground transmission line is 1000 A, and the maximum underground lines provided that there are metallic com-
zero-sequence current that is anticipated in the 115-kV ponents of the fence that are ungrounded or sections
underground line is 2% (20 A) of the positive-sequence that have only one ground point. Significant voltages
current. The impedances of the earth return paths (i.e., can be induced on barbed wire fences with wooden
remote earth and the ground continuity conductor) for fence posts, for example. The obvious induced voltage
the zero-sequence current result in 76% and 24% of the mitigation is to ground the components of the fence at
zero-sequence current to flow in the ground continuity regular intervals. There are no industry-recognized lim-
conductor and remote earth, respectively. Figure 16-40 its for voltages induced on metallic fences by under-
shows the positive- and zero-sequence currents flowing ground lines. However, the 15 V (rms) limit for
in the transmission cables. maximum pipeline voltage is considered to be a reason-
able limit based on pipeline induced voltage limits.
The distances between each of the conductors in the
cable trench and the pipeline are shown in Table 16-13. Table 16-13 Distances between Transmission Line
Conductors and Pipeline.
The mutual impedances between each of the transmis-
Distance
sion cable and the pipe line are calculated using Equa- Between
tion 16-12. Likewise, the mutual impedance between the Section Conductor
ground continuity conductor and the pipeline is calcu- (Figure and Pipe Mutual Impedance
16-38) Conductor (in.) (ohms/1000 ft)
lated using Equation 16-55.
1 Phase A 111.46 0.01807 +j 0.13108
1 Phase B 110.29 0.01807 +j 0.13132
The voltage induced on the pipeline by the positive- and
1 Phase C 110.02 0.01807 +j 0.13138
zero-sequence components of the currents in the under-
Ground
ground transmission line are shown in Table 16-14 for 1 120.27 0.01807 +j 0.12927
Conductor
pipeline Sections 1 and 2 (see Figure 16-38). As indi- 2 Phase A 370.44 0.01807 +j 0.10349
cated by the induced voltages in Table 16-14, the maxi- 2 Phase B 370.09 0.01807 +j 0.10351
mum induced voltage on the pipeline (1.1 V) is well 2 Phase C 370.00 0.01807 +j 0.10520
below the maximum value in NACE Std. RP0177-2000 Ground
2 360.20 0.01807 +j 0.10382
and CSA C22.3. It is also evident that the most of the Conductor
pipeline induced voltage is caused by the zero-sequence
component of the current. Table 16-14 Voltages Induced by Positive and Zero-
Sequence Currents (V rms)
If the ground continuity conductor were not installed in Voltage from
the duct bank, then all of the zero-sequence current Positive-Sequence Voltage from
Section Current Zero-Sequence
would return in remote earth, and the pipeline induced (Figure 16-38) (V) Current (V)
voltage would increase significantly (from 1.11 to 1 -0.003 – j 0.019 0.269 + j 0.502
4.22 V). If the line-to-ground fault current is 20 kA for 4 2 -0.004 – j 0.024 0.288 + j 0.454
cycles, then the zero-sequence current will be 1000 times Total (Magnitude) 0.044 1.106
the zero-sequence current in the above example. Conse-
Figure 16-40 Paths for positive- and zero-sequence currents.
16-48
The equations for calculating induced voltages on pipe- Wire” interface method (Figure 16-41) to determine
lines (Section 16.5.2) are applicable to voltage induced whether or not induction from electric supply lines are
in the components of fences that are insulated from likely to have adverse effects on communication lines.
ground.
IEEE Standard 776 contains recommendations for com-
16.5.5 Inductive Coordination with munications lines induced voltage threshold levels based
Communications Lines on the following considerations.
The operation of communication circuits can be 1. The open-circuit voltage that will be induced on a
adversely affected by currents flowing in nearby power 100-ft length of insulated 22 gauge “probe-wire”
lines. The magnetic field produced by the currents in placed at ground level, parallel to the power line
underground transmission cables and ground continuity either 50 ft or 75 ft from the geometric mean location
conductors induce voltages and currents on communi- of the power-line phase conductor distance (Figure
cation circuits that use metallic (copper) conductor 16-41). The 50-ft separation between the probe-wire
cables. Fiber optic cable communications circuits are and the power line applies for power lines rated 69 kV
generally immune to inductive interference from power and less. A separation distance of 75 ft applies to
lines. power lines with rated voltages above 69 kV.
Adverse Effects Caused by Induced Voltage 2. The length that the communications line is in close
The possible adverse effects of transmission line induced proximity to the power line. The lengths of close
voltages on communications lines are: proximity are categorized into Zones 1, 2, and 3.
1. Induced voltages can present a safety hazard to com- IEEE Standard 776 defines the distances for Zones 1,
munications company service personnel. 50 V rms 2, and 3 as follows:
continuously induced with respect to ground at 60 Hz Zone 1: 0 – 15 kft (0 - 24 km)
on telecommunications facilities has historically been Zone 2: 15 – 50 kft (24 - 80 km)
considered an upper threshold in North America for
Zone 3: >50 kft (> 80 km)
communications lines (IEEE Std. 776-1992 [IEEE
2000]). 3. The accessibility of the communications circuit to a
customer.
2. The induced voltage may degrade the quality of com-
munications. For example, 60 Hz hum from power- The term “inure” is defined as those facilities where
frequency induced voltages is a relatively common customers will not have access to conductors, and
problem. where the telecommunications utility is willing to
condition those conductors to operate safely at
3. The induced voltage may cause temporary outages in
induced voltage levels of 50-100 V rms to ground.
communications lines. For example, a phone line pro-
tective device may temporarily disconnect a phone The term “customer accessible” designates those
line as a result of fault current induced voltages. facilities that are accessible to the customer and
induced voltages are limited to less than 50 V rms to
4. Induced voltages may cause permanent damage to ground.
communications cables. If the induced voltage is high
enough, the electrical insulation around individual 4. The interruption tolerance of the communications
conductors may puncture. The induced voltages may circuit.
also damage equipment attached to the communica- 5. The categories “A” (will not tolerate interruptions)
tions cables. and “B” (will tolerate fault-caused interruptions).
IEEE Std. 776 (Recommended Practice for Inductive

Coordination of Electric Supply and Communications
Lines) (IEEE 2000) addresses the above inductive coor-
dination issues.
Threshold Levels for Induced Voltages
It is difficult to quantify the effect of electric supply lines
on each nearby communication circuit because the
interference susceptibility of the communications cir-
cuits depends on numerous parameters that may not
available to the power systems engineer. Consequently, Figure 16-41 IEEE Std 776 probe-wire interface.
IEEE Standard 776 uses what is called the “Probe-
16-49
Table 16-15 shows IEEE Standard 776 recommenda- Worked Example 8: Telecommunications Line
tions for the threshold 60 Hz rms voltages on a 22-gauge Inductive Coordination
probe-wire. This worked example demonstrates how to perform
probe-wire induced voltage calculations and application
Similar threshold induced voltages for fault conditions of the results to determine if voltage induction from an
are shown in Table 16-16. underground transmission line will adversely affect a
specific type of communications circuit. Figure 16-42 is a
Fault Induced Voltages profile view of the 115-kV underground transmission
As discussed earlier for pipelines, the highest induced line, and a cross section of the cable duct bank. Figure
voltages are caused by power system fault conditions. 16-43 shows the positive- and zero-sequence currents
flowing in the transmission cables. The soil electrical
Energy dissipation is dependent upon fault duration or resistivity is 100 ohm-m.
clearing time. Usually the fault duration at a particular
location is the time required for the nearest line side pro- The underground transmission line will operate at a
tective equipment to operate and isolate a fault, assum- maximum current of 1000 A during normal operating
ing the maximum calculated symmetrical fault current conditions, and the line will have a 24-hr emergency rat-
at the fault location. ing of 1200 A. The line-to-ground fault current will not
Experience (IEEE Std 776 2000) has shown that damage exceed 20 kA for 4 cycles (0.0667 s).
can occur to communications equipment if the product
of the square of the induced current and time (I2 t) falls The underground transmission line is being constructed
in the range of 16 to 80 A2 s. IEEE Std. 776 contains the parallel to an existing telephone line cable for a distance
following (Table 16-16) probe-wire induced voltage lim- of approximately 2000 ft. The objective of the analysis is
its for various fault current durations based on the 16 to to determine if the planned underground transmission
80 A2 s limits. The probe-wire induced voltage ranges line will adversely affect the operation and maintenance
for the Table 16-16 fault current durations were derived of the telephone line. The telephone line falls in the
from the fact that 100 ft of #22 AWG copper wire has a IEEE Std. 776 classifications of “customer accessible”
resistance of 1.6 ohms. and Class B (may tolerate interruption).
Table 16-15 Acceptable 60 Hz RMS Voltage Thresholds on As discussed in Section 16.5.5 (“Inductive Coordination
Probe-wire (Based on Safety and Circuit Degradation with Communications Lines”), the procedure is to cal-
Considerations) culate the “probe-wire” induced voltage and then com-
Class A A A B B B pare the probe-wire induced voltage to the IEEE Std.
1 2 3 1 2 3 776 threshold values that are applicable to a Zone 1,
Zone (V) (V) (V) (V) (V) (V) Class B, Customer Accessible communications circuit.
Steady c 0.3333 0.1000 0.0379 0.3333 0.1000 0.0379 The #22 AWG probe-wire is located at ground level and
State i 0.6667 0.2000 0.0758 0.6667 0.2000 0.0758 a horizontal distance of 75 ft from the closest phase of
i = inure, c = customer accessible the underground transmission line (see Figure 16-42).
Class A = may not tolerate interruption. Class B = may toler-
ate interruption (self restoring). The maximum balanced, three-phase (i.e., positive-
sequence) current that will flow in a 115-kV under-
Table 16-16 Probe-Wire Induced Voltage Threshold Values ground transmission line is 1000 A, and the maximum
for Equipment Destruction (IEEE Std. 776) zero-sequence current that is anticipated in the 115-kV
underground line is 2% (20 A) of the positive-sequence
Energy Energy
Dissipation Dissipation current. The impedances of the earth return paths (i.e.,
I2 ta I2 t remote earth and the ground continuity conductor) for
16 80 the zero-sequence current result in 76% and 24% of the
Fault
Duration Time
(Cycles) (s) (V) (V)
5 0.0833 22.17 49.57
10 0.1667 15.68 35.05
20 0.3333 11.09 24.79
40 0.6667 7.84 17.53
Figure 16-42 Elevation view of 115-kV underground line
80 1.3330 5.54 12.39
and probe-wire.
a. I2 t is the same for all zones and classes
16-50
Figure 16-43 Paths for positive and zero-sequence currents.
zero-sequence current to flow in the ground continuity The IEEE Std. 776 probe-wire threshold voltage based
conductor and remote earth, respectively. on safety and circuit degradation (Table 16-14) is 0.3333
V for Class B, Zone 1, customer accessible communica-
Figure 16-43 shows the positive and zero-sequence cur- tions lines.
rents flowing in the transmission cables.
As indicated by the induced voltages in Table 16-18, the
The distances between the probe-wire and the current- maximum induced voltage on the probe-wire (0.044 V
carrying conductors in the cable trench are summarized rms) is significantly below the applicable IEEE Std. 776
in Table 16-17. The mutual impedances between each of probe-wire threshold voltage.
the transmission cables and the probe-wire are calcu-
lated using Equation 16-55. Likewise, the mutual imped- If the line-to-ground fault current is 20 kA for 4 cycles,
ance between the ground continuity conductor and the then the zero-sequence current will be 1000 times the
probe-wire is calculated using Equation 16-55. zero-sequence current in the above example. Conse-
quently, the voltage induced on the probe-wire for the
The open-circuit voltage induced on the probe-wire by duration of the fault would be 1000 * 0.044 V = 44 V.
the positive- and zero-sequence components of the cur- The short-circuit current induced in the probe-wire
rents in the underground transmission line are shown in would be 27.5 amps, and the I 2 t value would be
Table 16-18. The short-circuit current (28 mA) induced 50.4 A 2 s. IEEE Std. 776 indicates that equipment
in the probe-wire by the maximum normal operating destruction will occur for I2 t values, ranging between 16
condition current would be the resistance of the #22 and 80 A2 s. Consequently, line-to-ground faults in the
AWG wire (1.6 ohms) divided by the open circuit volt- underground transmission line in this example may
age (0.044 V) divided by the resistance of the #22 AWG cause equipment destruction for the Class B, Zone 1
wire (1.6 ohms). communications line.
16.5.6 Railway System Inductive Coordination

Table 16-17 Distances and Mutual Impedances between The alternating currents in underground transmission
Conductors in Cable Trench and Probe-wire lines can interfere with nearby railway signals and com-
Distance Between munications circuits in the same manner as they would
Conductor and interfere with nearby telecommunications circuits. In
Pipe Mutual Impedance
Conductor (in.) (ohms/1000 ft) both cases, the source and mechanism of induction are
Phase A 900.98 0.01807 + j 0.08307
the same. Railway signals and communications systems
are, in addition, unique in regard to their operation,
Phase B 901.50 0.01807 + j 0.08306
safety, and reliability objectives. The signals and com-
Phase C 902.13 0.01807 + j 0.08304 munications systems must operate reliably and safely for
Ground Conductor 902.88 0.01807 + j 0.08302 economical train movement and train protection. Train
protection includes:
Table 16-18 Probe-Wire Voltages Induced by Positive- • Prevention of conflicting train movements
and Zero-sequence Currents (V rms)
• Maintenance of safe distances between trains travel-
Voltage From Positive- Voltage From ing in the same direction
Sequence Current Zero-Sequence Current
0.003 0.044 • Broken rail detection
• Open track switch detection
16-51
A railway communications system typically includes the The acceptable levels (CSA C22.3 No 3 1998) for longi-
following components subsystems: tudinally induced voltages in railway signaling and com-
• Track circuits munications circuits are as follows:
• Signal control circuits • Normal operating conditions. 50 V ac rms is the maxi-

mum acceptable induced voltage during normal
• Automatic highway grade-crossing warning systems power system operating conditions. 150 V may be
• Signal power supply circuits acceptable under special conditions.
• Supervisory control circuits • Fault conditions. The acceptable level for longitudinal
• Dispatcher and maintenance voice circuits induced voltage in railway signaling and communica-
tions circuits is 430 V rms under power line fault con-
• Hot box detector and similar special circuits ditions. This level applies to usual power line
• Train radio equipment and maintenance. Higher voltages may be
• Data circuits acceptable under special conditions, such as high reli-
ability power lines with high-speed relaying and fault
Railway signal systems involve the interconnection of clearing.
track circuits, signal control circuits, and signal appli-
ances to ensure the safe control of train movements. The EPRI Power System and Railroad Electromagnetic
Compatibility Handbook (EPRI 2006) is a source of
Track circuits detect the presence of trains. The basic technology and data for preventing and mitigating ac
principle of the track circuit is the shunting condition electrical interference problems on railroads. The book
produced by the train wheels passing along the electri- includes all aspects of electromagnetic compatibility
cally energized running rails. A basic direct current (EMC) where railroad systems are the receptors. This
track circuit is shown in Figure 16-44. The signal con- includes well-understood areas such as magnetic induc-
trol circuits use the information from signals, track tion from transmission lines as well as less understood
switches, and track circuits to cause the display of areas such as conducted interference from distribution
appropriate signal aspects to approaching trains. There systems and effects of harmonics. Chapters examine all
may also be supervisory control circuits from a central known effects of ac interference, including personnel
dispatching office. safety, operation of railroad equipment and systems,
and damage to railroad equipment. Direction is pro-
vided for studying the effects of proposed new and
upgraded installations as well as root cause analysis of
problems on existing installations.
Figure 16-44 Basic track circuit.
16-52
16.6 NOMENCLATURE
Units
A = shield cross-sectional area circular mils (mm2)
a = pipe inside diameter in. (m)
= complex operator
a -
(-0.5 + j 0.866)
B = magnetic flux density mG or (μT)
b = pipe outside diameter in. (m)
Br = resultant magnetic flux density mG or (μT)
= magnitude of magnetic flux density along a direction defined
Bα W/m2
by angle α, use SI
Bx1 = horizontal component of magnetic flux density gauss
By1 = vertical component of magnetic flux density gauss
= capacitance between each high-voltage conductor and the
C μF/ 1000 ft (μF/km)
grounded insulation shield
ds = sheath diameter in. (cm)
dan = distance from conductor a to conductor n in. (cm)
Dc = diameter over conductor semiconducting shield in. (cm)
De = equivalent depth of earth return current in. (cm)
Di = diameter over cable insulation in. (cm)
Dp = pipe inside diameter in. (cm)
Ds = diameter over skidwires in. (cm)
Ea, Eb, Eb = A, B, C phase voltages kV
Eo = operating line-to-ground voltage kV
[Ec] = column matrix of voltage drops across the cable conductors V
f = power frequency Hz
GMDφ = geometric mean distance between cable centers in. (cm)
GMRφ = geometric mean radius of conductor in. (cm)
GMRg = geometric mean radius of the ground conductor in. (cm)
H = magnetic field strength oersteds
I = current in line A
Ia, Ib, Ic = phase currents A
Ic = charging current A/l000 ft (A/km)
[Ic] = column matrix of currents in the high-voltage conductors A
[Is] = column matrix of the currents in the cable sheaths A
Isc = maximum short-circuit current A
k = constant for conductor shield/sheath temperature rise –
P12 = power flow between points 1 and 2 MVA
r = distance from line m
r1, r2 = radius to outside and inside of cylinder mm
Rφ = conductor resistance Ω/1000 ft (Ω/km)
Rc = conductor ac resistance μΩ/ft (μΩ/m)
Rg = resistance of the earth return Ω/1000 ft (Ω/km)
Rn = sheath or shield resistance Ω/1000 ft (Ω/km)
R’p = pipe effective resistance μΩ/ft (μΩ/m)
Rs = sheath/shield resistance Ω/1000 ft (Ω/km)
Rskidwire = skidwire resistance μΩ/ft (μΩ/m)
16-53
Units
rsm = mean radius of cable sheath in. (cm)
S = axial spacing between phases in. (cm)
Sag, Sbg, Scg = axial spacing between the centers of cables a, b, c
= magnetic field with shield/magnetic field w/o shield and the (unitless)
SF
ground conductor in. (cm)
V1 = magnitude of voltage at point 1 kV
V2 = magnitude of voltage at point 2 kV
VI = magnitude of incident traveling wave V
VT = magnitude of voltage at cable terminal V
X12 = reactance between points 1 and 2 Ω
X0 = zero-sequence reactance Ω/1000 ft (Ω/km)
X1 = positive-sequence reactance Ω/1000 ft (Ω/km)
Xc = conductor ac reactance μΩ/ft (μΩ/m)
Xm = mutual reactance Ω/1000 ft (Ω/km)
X’p = pipe effective reactance μΩ/ft (μΩ/m)
YC = conductor ac loss increment –
YP = pipe ac loss increment –
YS = shield ac loss increment –
ZO = zero-sequence impedance μΩ/ft (μΩ/m)
Z1 = positive-sequence impedance μΩ/ft (μΩ/m)
Zaa, Zbb, Zcc,
= self-impedances Ω/1000 ft (Ω/km)
Zss
Zc = traveling wave surge impedance of cable Ω
= square matrix of the self and mutual impedances for the high-
[Zcc] Ω/1000 ft (Ω/km)
voltage conductors
[Zcs] = square matrix of the conductor-to-sheath mutual impedances Ω/1000 ft (Ω/km)
Zmn = mutual impedances, m ≠ n Ω/1000 ft (Ω/km)
Zoh = traveling wave surge impedance of overhead line Ω
= square matrix of the self and mutual impedances of the cable
[Zss] Ω/1000 ft (Ω/km)
sheaths
δ = power angle between the voltage phasors at two locations degrees
δ = magnetic field skin depth of a type of metal m
σ = metal electrical conductivity S/m
ε = relative dielectric constant –
μ = relative magnetic permeability –
μ0 = magnetic permeability of free space H/m
r = earth electrical resistivity Ω-m
16-54
REFERENCES Cowley, D. J., and R.G. Foxall. 1983. “Planning, Design

Abramowitz, M., and J. E. Stegun. 1964. Handbook of and Construction of a 500 kV Transmission System to
Mathematical Functions. U.S. Department of Com- Vancouver Island Including Long Submarine Cable
merce. National Bureau of Standards. Links.” Transactions of the Canadian Electrical Associa-
tion, Engineering and Operating Division. Vol. 22. part 5.
Bahder, G., C. Katz, and G. W. Seman. 1979. Determi-
nation of AC Conductor and Pipe Loss in Pipe-Type CSA. 2003. Can/CSA-C22.3 No. 6-M91. “Principles
Cable Systems. EPRI Report EL-1125. and Practices of Electrical Coordination Between Pipe-
lines and Electric Supply Lines.” September 1991 (Reaf-
Bascom, E. C. III, J. H. Cooper, W. Banker, R. Piteo, A. firmed 2003).
M. Regan, and S. A. Boggs. 2006. “Magnetic Field
Management Considerations for Underground Cable Dommel, H. W. 1969. “Digital Computer Solution of
Duct Bank.” IEEE T&D Conference. Paper TD2005- Electromagnetic Transients in Single and Multiphase
000399. Networks,” IEEE Transactions on Power Apparatus and
Systems. Vol. PAS 88. pp. 388-399.
Bottauscio, O., M. Chiampi, D. Chiarabaglio, F. Fior-
illo, L. Rocchino, and M. Zucca. 2001. “Role of Mag- Dougherty, J. J., and C. S. Shifreen. 1962. “Long Cable
netic Materials in Power Frequency Shielding: Lines—AC with Reactor Compensation or DC.” AIEE
Numerical Analysis and Experiments.” IEEE Proceed- Transactions. Vol. 81. No. III. pp. 169-178.
ings on General Transmission and Distribution. Vol.
148. No 2. Durkin, C. H., R. P. Fogarty, T. M. Halleran, D. A.
Mark, and A. Mukhopadhyay. 1994. “Five Years of
Bozorth, R. M. 1993. Ferromagnetism. IEEE Press. Pis- Magnetic Field Management.” IEEE PWRD Transac-
cataway, NY. tions. 94 SM 392-1.
Bucea, G., and H. Kent. 1998. “Shielding Techniques to Edison Electric Institute. 1957. Underground Systems
Reduce Magnetic Fields Associated with Underground Reference Book. New York.
Power Cables.” CIGRE paper 21-201.
EPRI. 2006. Power System and Railroad Electromag-
Calabrese, G. O. 1959. Symmetrical Components Applied netic Compatibility Handbook. Revised First Edition.
to Power Systems. The Ronald Press. Report No. 1012652.
Canova, A., A. Manzin, and M. Tartaglia. 2002. “Eval- EPRI. 2005. Transmission Line Reference Book—345
uation of Different Analytical and Semi-Analytical kV and Above. 2nd Ed. Electric Power Research Insti-
Methods for the Design of ELF Magnetic Field tute. Palo Alto, CA.
Shields.” IEEE Transactions on Industry Applications.
Vol. 38. No 3. Glover, J. D., and M. S. Sarma. 2002. Power Systems
Analysis and Design. Third Edition. Brooks/Cole.
CIGRE. 1997. CIGRE JTF 36-01/21. “Magnetic Field
Calculation in Underground Cable Systems with Ferro- Hasselgren, L. and J. Luoni. 1995. “Geometrical
magnetic Components.” ELECTRA. No. 174. Aspects of Magnetic Shielding at Extremely Low Fre-
quencies.” IEEE Transaction on Electromagnetic Com-
Clairmont, B. A., and R. J. Lordan. 1999. “3-D Model- patibility. Vol. 37. No 3.
ing of Thin Conductive Sheets for Magnetic Field
Shielding: Calculations and Measurements.” IEEE Haus, H. A., and J. R. Melcher. 1989. Electromagnetic
Transactions on Power Delivery. Vol. 14. No. 4. Fields and Energy. Prentice Hall. Englewood Cliffs, New
Jersey.
Colla, L., F. M. Gatta, and F. Iliceto. 2005. “Design and
Operation of EHV Transmission Lines Including Long Hileman, A. R., R.W. Flugum, and T. F. Garrity, 1982.
Insulated Cable and Overhead Sections.” 7th Interna- “Lightning Insulation Coordination for a 600 kV DC
tional Power Engineering Conference. Gas Insulated Cable.” IEEE Transactions on Power
Apparatus and Systems.
Cooper, J. H. 1993. Transmission Cable Magnetic Field Vol. PAS 101. No. 11.
Management. EPRI Report TR-102003.
16-55
Hoburg, J. F. 1994. “Principles of Materials-Based Mag- Long, W., D. Catcher, D. Ruiu, P. Adam, S. Lee, and R.
netic Shielding.” IEEE/PES Transmission and Distribu- Adapa, 1990. “EMTP-A Powerful Tool for Analyzing
tion Conference. Power System Transients.” IEEE Computer Applications
in Power. Vol. 3, No. 3.
IEEE. 1992. Institute of Electrical and Electronic Engi-
neers. IEEE Std. 776-1992 (R1998). “IEEE Recom- Mallard, S. A., and V.C. Thomas. 1966. “A Method for
mended Practice for Inductive Coordination of Electric Calculating Transmission System Reliability.” IEEE
Supply and Communications Lines.” 2000. Transactions on Power Apparatus and Systems. Vol. PAS
87. pp. 824-834.
IEEE. 1993. Institute of Electrical and Electronic Engi-
neers. C37.015. “Application Guide for Shunt Reactor Meliopoulos, S. 1988. Power System Grounding and
Switching.” Transients. Marcel Dekker.
IEEE. 1996. Institute of Electrical and Electronic Engi- NACE. 2000. NACE Standard RP0177-2000 (Item No.
neers. Power Systems Relaying Committee. “Protective 21021), “Mitigation of Alternating Current and Light-
Relaying Considerations for Transmission Lines with ning Effects on Metallic Structures and Corrosion Con-
High Voltage AC Cables.” 96 WM 018-2 PWRD. trol Systems.” NACE International. September.
IEEE. 1997. Institute of Electrical and Electronic Engi- Neher, J. H. 1964. “The Phase Sequence Impedance of
neers. “IEEE Std. C62.22. “Guide for the Application Pipe-Type Cables.” IEEE Transactions on Power Appa-
of Metal-Oxide Surge Arresters for Alternating-Current ratus and Systems. Vol. PAS 83. pp. 795-804. August.
Systems.” December.
Short, C. M., T. J. Bliss, and M. K. Enns. 1958. “Surge
IEEE. 2000. Institute of Electrical and Electronic Engi- Protection of Cable-Connected Power Transformers at
neers. “IEEE Std. 80-2000. “Guide for Safety in AC Scattergood Steam Plant.” American Institute of Elec-
Substation Grounding.” January 30. trical Engineers Power Apparatus and Systems.
IEEE. 2002. Institute of Electrical and Electronic Engi- Stewart, J. R., D. D. Wilson, and J. A. Williams. 1986.
neers. C2-2002. “National Electric Safety Code.” “System Implications of Underground Cables.” IEEE T
August 1, 2001. & D Conference Proceedings. Anaheim, CA.
Kasztenny, B., I. Voloh, and J. G. Hubertus. 2004. Vavra, J., and M. Wanda. 2006. “400 kV Vienna The
“Applying Distance Protection to Cable Circuits.” 57th 400 kV Vienna North Input.” CIGRE Paper B1-101.
Annual Conference for Protective Relay Engineers.
Wait, J. R., and D. A. Hill. 1977. “Electromagnetic
Kundur, P. 1994. Power System Stability and Control. Shielding of Sources within a Metal-cased Bore Hole.”
McGraw Hill. IEEE Transactions on Geoscience Electronics. Vol.GE-
15. No. 2.
Lewis, W. A., and G. D. Allen. 1978 “Symmetrical Com-
ponent Circuit Constants and Neutral Circulating Cur- Westinghouse Electric Corporation. 1964. Electrical
rents for Concentric-Neutral Circuits.” IEEE Transmission and Distribution Reference Book. Fourth
Transactions on Power Apparatus and Systems. Vol. PAS Edition. East Pittsburgh, PA.
91. No. 3.
Zemyan, N. 1996. “138 kV Solid Dielectric Cable
Installed in Steel Pipes and Its Effect on EMF.” IEEE
Insulated Conductors Committee Fall Minutes.
16-56
CHAPTER 17 The Future of Underground

Transmission
Authors: Mohammad A. Pasha, The United Illuminating Company
Steve Eckroad, EPRI
Reviewer: Reza Ghafurian, Consolidated Edison Company of New York, Inc.
A continued increase is expected in the percentage of underground cable as part of total

transmission lines planned and built, and a higher percentage of solid dielectric cables is
predicted, making it possible to construct more underground transmission lines in urban
and suburban areas. Underground cable innovations are being driven by increased diffi-
culties in acquiring ROWs (rights-of-way) for overhead transmission lines, worldwide
demand on reliability of transmission lines, public desire for environmentally friendly
underground systems. Innovations include technological developments in insulating
materials and production gains in manufacturing processes.
Mohammad Anwar Pasha is a lead engineer with The United Illumi-

nating Company (UI). He holds a B.S. (1966) in physics from Punjab
University, Lahore, Pakistan, a B.S. (1970) in electrical engineering
(power) from the University of Engineering and Technology, Lahore,
an M.S. (1974) in nuclear technology from Islamabad University,
Islamabad, Pakistan, and an M.S.E.E (1982) from Syracuse University,
Syracuse, New York. Mr. Pasha began his career by working for the
Electric Company of Pakistan, Lahore, following which, he worked for
the Water and Power Development Authority, also in Lahore. Before immigrating to the
United States, he worked as a senior engineer with the National Engineering Services of
Pakistan, Lahore, Pakistan and was responsible for the engineering of the 205-mile,
500-kV Guddu-Jamshoro overhead transmission line. Presently he is working as project
transmission engineer for UI's 345-kV Middletown-Norwalk project. Mr. Pasha has been
chairman of EPRI's Underground Transmission Line Program since 2004. He is a past
chairman of the North East Transmission Group. He is a Member of IEEE, Insulated
Conductors Committee, and a licensed professional engineer in the state of Connecticut.
Mr. Pasha has authored numerous IEEE and EPRI publications and technical papers.
Steve Eckroad is a senior technical manager in the Power Delivery and

Markets sector of the Electric Power Research Institute (EPRI), where
he has been employed since 1992. He recently assumed program
management responsibilities for EPRI's Underground Transmission
Program. Mr. Eckroad has managed projects for the development of
commercial software and hardware, handbooks, guidelines, and
training courses in a wide variety of transmission and substation
equipment applications, as well as techno-economic assessments and
field demonstrations of emerging technologies, including energy storage and supercon-
ductivity. Prior to joining EPRI in 1992, Mr. Eckroad was an independent consultant to
the utility industry (1990-92). From 1979 to 1990, Mr. Eckroad was a project manager in
the Research and Development Division of Bechtel Group, Inc., where he performed
technology assessments and developed designs for advanced energy storage systems. Mr.
Eckroad is a Senior Member of the IEEE, and is a registered professional electrical
17-1
Chapter 17: The Future of Underground Transmission EPRI Underground Transmission Systems Reference Book
engineer in California. He is the author or co-author of in 1965, and carried out postgraduate studies in
a number of publications in the fields of energy storage electrical engineering at the University of Missouri-
and electric power delivery technology. Mr. Eckroad Rolla in 1990-1992.
received his BA Degree in physics from Antioch College
17-2
EPRI Underground Transmission Systems Reference Book Chapter 17: The Future of Underground Transmission
17.1 INTRODUCTION 17.2 CURRENT RESEARCH IN CABLE

Over the last five years, an extensive growth in installing TECHNOLOGY
underground transmission lines worldwide has been
observed. Nevertheless, predicting the need for and 17.2.1 Dielectrics
shape of future developments in underground transmis- Nanotechnology for Insulation
sion, defined here as cable systems rated 69 kV and The crosslinked polyethylene (XLPE) presently used for
higher, is uncertain at best. The only safe prediction is insulating HV and EHV cables has certain weaknesses.
that the requirement for underground transmission will These weaknesses include low electrical strength (there-
not decrease as a percentage of installed transmission fore the need for thicker insulating materials), intoler-
and may increase. ance to voids and contaminants, and a relatively high
coefficient of thermal expansion.
Underground transmission cannot compete head-on
with overhead transmission in many scenarios, simply EPRI has successfully developed and evaluated a new
because the dielectric system is much more complex dielectric material that has shown an outstanding
and expensive, and the cost of the installation can be increased performance (EPRI 2005). The material has
s eve ra l t i m e s t h e c o s t o f t h e ove r h e a d s y s t e m . been developed by adding an inorganic, nanosized filler
The suburban environment is another matter alto- to the insulating material commonly used for HV and
gether. This will become a “Next Frontier” for under- EHV extruded-dielectric cables. The new compound
ground transmission. material was chosen for its dielectric strength, dielectric
losses, cost, thermal conductivity, flexibility, and process
Communities often prefer to install underground trans- ability. The resulting insulating material has exhibited
mission lines as compared to overhead lines in their the following characteristics:
neighborhoods, because they believe that overhead lines
could devalue their property values or may have detri- • Breakdown strength is two to three times greater
mental environmental effects. Aesthetic issues, environ- than that of XLPE.
mental concerns, weather-related outages, and the • Voltage endurance is almost three orders of magni-
increasing cost of rights-of-way will all be factors in the tude higher than that of XLPE.
future that may make an underground cable an attrac- • Losses are lower, and the permittivity is equal to or
tive alternative to overhead. This situation is expected to
lower than that of XLPE.
continue while the cost difference between overhead and
underground transmission lines continues to narrow.
Once nano-based polymer insulation becomes commer-
Certain legislation, such as that in the State of Connect-
cially available, it will provide superior thin-wall insu-
icut, United States, has been monitoring the life-cycle
lated underground transmission cables. A reduction in
costs of the two options so that laws could be passed at
cable insulation thickness will mean longer lengths for
any time to make underground transmission lines man-
cable reels, which will result in fewer splicing vaults,
datory for all future transmission line construction. This
fewer splices, smaller duct sizes, and easier handling. All
trend is predicted to continue around the world. One
of these factors will reduce material, construction, and
point must be made clearly: simple cost ratios of under-
installation costs. Also, fewer splices will likely mean
ground to overhead options do not accurately reflect the
improved operation reliability.
benefits of underground vs. overhead. Each potential
scenario must be independently evaluated for all of its Development of Polypropylene Cable
advantages and disadvantages, some with costs associ- A polypropylene-type thermoplastic polymer could
ated and others that are not measured by cost. someday replace XLPE. This development would give
manufacturing advantages in not having to crosslink
This chapter looks ahead to the future of underground and degas the extruded cables. These advantages will
transmission—from developments on the verge of com- greatly increase the output of cable manufacturing
mercial application to those further out in the research plants since this has been a major bottleneck of the
frontier—and the achievements and benefits that may XLPE extrusion process.
one day be possible. Section 17.2 discusses current
research in cable technology. Section 17.3 reviews areas Dc Application of Thermoplastic Polymer Cable
of design and operation that are the subject of research As of today, the thermoplastic polymer has been used
and development, and Section 17.4 describes research in for dc applications up to +/- 150 kV. Many cable manu-
installation design. Section 17.5 speculates on future factures are conducting research and development of
cable systems employing high-voltage dc cables. thermoplastic polymer for dc applications at higher
voltage levels. Efforts have been made to replace
17-3
dicumyl peroxide and some of the antioxidants in order ductor cable system will be an integral part of the
to reduce the polar species that cause space-charge LIPA grid. The groundbreaking for this project was
problems when using XLPE with direct current. held on August 2, 2006, and commissioning of the
cable is expected in 2007. LIPA has stated that it
17.2.2 Dry-type Terminations and High-Stress expects to make the cable a permanent part of its
Splices transmission system “backbone” on Long Island.
Research is taking place by cable manufacturers to • The National Grid Cable Project is a 1148-ft (350 m)
develop true dry-type terminations as utilities demand HTS cable connecting two substations along the
fluid-free accessories due to safety and environmental Hudson River in Albany, New York and operates at
concerns. In addition, new types of XLPE cable accesso- 34.5 kV. This project will also install the first
ries are being developed that are more reliable to manu- “second-generation” HTS wire used in a utility grid
facture (some reject rates run at 30 to 60%) and can with a splice installed. The cable was put into opera-
withstand higher stresses. tion on July 20, 2006, and will operate for two years
as a demonstration before being considered for per-
17.2.3 Superconducting Cables manent use by National Grid.
The development of superconductors that have effec- • The American Electric Power (AEP) Cable Project is
tively near-zero resistance to ac currents at liquid nitro- a 657-ft (200 m) 13.2 kV HTS cable installed in an
gen temperatures (75 Kelvin, or about –200º C) has AEP substation in Columbus, Ohio. The cable is
progressed so substantially over the last 20 years that designed to carry 3000 amps and incorporates all the
superconducting power equipment is becoming a near- three phases in a single cable through three concen-
term commercial proposition. Superconducting cables tric layers of HTS wires. Due to the unique “triaxial”
and fault current limiters will likely be the first commer- design, the cable is very small—about 6 in. in diame-
cially viable technologies. The first generation of high- ter—yet is rated to carry 69 MVA. The cable was
temperature superconductor (HTS) material to be commissioned on August 8, 2006. AEP is investigat-
developed into a wire—known as “1G”—utilized a sub- ing other uses for this cable in its grid.
stantial amount of silver, making it relatively expensive.
A second-generation material, “2G,” now being pro- Other superconducting projects have been undertaken,
duced in small quantities will ultimately be cheaper to are under way, or are planned around the world—includ-
manufacture and have better performance characteris- ing China (35 kV, 33.5 m, warm dielectric), Korea (two
tics. Companies around the world are exploiting the projects, both 22.9 kV, 100 m, cold dielectric), Japan
possibilities for applying these materials to power deliv- (77 kV, 500 m, cold dielectric), Mexico, and Russia.
ery applications. (For a more detailed discussion of
superconducting cables, see Chapter 7, Section 7.4.) Superconducting cables offer a number of attractive
benefits, including:
As an indication of the increased interest in and
research on superconducting power cables, a number of
• Electricity savings (i.e., via lower losses) through
projects worldwide are demonstrating HTS cables. In near-zero resistance
the United States, in order to promote superconducting • Elimination of electromagnetic fields and thermal
power equipment commercialization efforts, the U.S. emissions to surrounding environments
Department of Energy (DOE) is currently co-funding • Three to five times higher ampacity ratings for cir-
three HTS cable demonstration projects. Each of these cuits, potentially reducing voltage levels in certain
is hosted by a utility company, and each one demon- applications
strates insertion of a load-carrying HTS cable into its
grid: Long Island Power Authority (in Holbrook, New
• Environmentally-friendly cooling medium and elimi-
York); National Grid (in Albany, New York); and nation of hydrocarbon fluids
American Electric Power (in Columbus, Ohio). These • Compact design, which allows for retrofit of existing
utilities have teamed with major U.S. and international underground duct infrastructure
companies that produce superconducting power equip-
ment components. Of these benefits, the market drivers for future imple-
mentation of this technology are anticipated to be
• The Long Island Power Authority (LIPA) Cable
increased power density and electricity savings. The
Project will connect two substations on Long Island,
technology will allow superconducting cables to move
New York, about 2000 ft apart, and will be the first
more power through the same underground duct cross-
application on a utility grid at transmission voltage.
section as normal copper cables. Also, more of the
With a capacity of 574 MW, the 138-kV supercon-
17-4
power produced by generators will be available to the • Greater understanding of cable design theory and
consumer. That is because, with traditional copper or practice
aluminum power cable, some of the generated power is
lost due to heat generated by ohmic losses in the con- Underground Transmission Workstation (UTW)
ductor, shield/sheath, and other metallic layers, along EPRI intends to rewrite its popular UTW during the
with dielectric losses generated in the insulation; losses coming years. A beta version for UTW 6.0 will be devel-
are usually of the order of 10% or more. oped and is expected to be made available by the end
2007. This software is expected to provide similar bene-
HTS cables will mean a reduction in right-of-way fits to those of the above mentioned Underground
(ROW) needs for building new transmission lines Transmission Systems Reference Book. The program
because superconducting cables offer high power capac- will provide an easy means to calculate ampacity, mag-
ity at lower voltage compared to conventional alumi- netic and electric field effects, pulling tensions, and other
num or copper wires. Existing ROW will also be utilized, practical calculations required for the design and instal-
since retrofitting cables in the existing pipe, conduits, lation of underground transmission systems. The UTW
and ducts will be possible. This will allow the utility to will have its examples based on worked examples in the
avoid digging up the streets and highways and the Green Book. Thus a user could read and understand the
accompanying disruption to traffic and local public. concept through worked examples in the Green Book
The ability to move greater amounts of power at and then run the same examples using UTW.
lower—even distribution—voltages will also reduce the End-of-life Criteria
need for siting and operating high-voltage substations in As utilities’ underground systems age, it will become
dense urban networks. essential to determine the remaining life of cable sys-
tems and to establish end-of-life criteria in order to
Also, the absence of electromagnetic fields, heat, and retire an underground system that may be vulnerable to
fluid will help to avoid potential environmental effects. sudden failures. Several questions must be asked: Is the
Further, because the cable is not affected by depth of system really at the end of its life, or does it have more
installation or thermal quality of surroundings (since life left due to low load levels in the past? If the system
the cable is self-cooled), installation at any depth will really is nearing the end of its life, is there any way to
be possible. extend it? It is essential that the condition of a cable,
splices, and terminations, and the quality of dielectric
17.3 IMPROVEMENTS IN DESIGN AND are determined. Based on such knowledge, a rational
OPERATION approach to extending the life of a cable system could
then be set.
17.3.1 Educational Tools
“End-of-life” criteria for paper-insulated cables have
Underground Transmission Systems Reference
been studied extensively. However, as XLPE cables
Book
installed in the 1980s start to approach 30 years of their
The current book represents a complete rewriting of the
operating life, the aging of these cable system types will
1992 edition of EPRI’s Underground Transmission Sys-
become an increasingly important topic.
tems Reference Book (“Green Book”) (EPRI 1992),
which was an outgrowth of the 1957 edition of the Knowledge Capture
underground reference book published by Edison Elec- As more and more underground engineers retire from
tric Institute. utilities, capturing knowledge from the outgoing engi-
neers will become necessary. It is anticipated that tools
This book provides reference material for new engineers will be developed that will help in capturing this
interested in learning underground transmission as well knowledge.
as a ready reference for expert underground engineers.
Benefits of providing a reference book like this include: 17.3.2 Thermomechanical Bending in Pipe-type
• Increased efficiency of engineering staff and reduced Cables
training costs for new engineers EPRI is engaged in determining whether there is a new
systemic failure mode for EHV HPFF cable systems
• Accurate and up-to-date design tools
where thermo mechanical bending (TMB) causes fail-
• Simplified design and construction practices ures within the pipe runs. The EPRI project intends to
• Reduced project costs evaluate whether these failures are due to age and/or
17-5
operating conditions. Procedures to mitigate TMB fail- trating radar system that could be further improved.
ures in pipes will be developed. The device consists of multiple arrays of radars that
send signals penetrating the ground. The signals that are
17.3.3 Recabling Upgrades reflected back are analyzed to indicate location and
Recabling existing pipe systems with higher voltage such presence of obstacles. The existing systems do not pos-
as 345 kV will be possible, because the new cables being sess satisfactory resolution. New and improved systems
developed will allow higher stress values and hence will be devised to help utilities make use of nonintrusive
reduction in cable diameter. Although we are presently methods for locating subsurface obstacles.
cautious with stresses of approximately 12-13 kV/mm,
stresses of 15-17 kV/mm should be possible with the use 17.5 FUTURE CABLE SYSTEMS—HV/EHV DC
of new insulating materials that are in the development CABLES
phase. The present problem of cable quality and joints Economical ac/dc converters are expected to be devel-
could be solved by proper quality control during the oped as use of dc transmission becomes more popular.
cable manufacture and splice installation in the field. It is, of course, essential to have high-temperature
XLPE/polymer cables available in conjunction with this
17.4 RESEARCH IN CABLE INSTALLATION type of transmission, since present cables are generally
DESIGN temperature restricted. Japan has been making sound
and continuous progress over the last 20 years and may
17.4.1 Construction Cost Reduction take their 500-kV dc cables out of the long-term test
What can be done to reduce underground transmission sites into commercial applications.
costs? Advanced methods are being explored to reduce
costs. Horizontal directional drilling, jack and bore, and The new generation of ac/dc converters would allow
microtunneling are very commonly employed for cross- tapping off radial lines, thus making expansion to an
ing rivers, railroads, highways, and other duct banks or existing system possible.
obstacles. This capability has made environmental agen-
cies more flexible in giving permission to build new 17.5.1 The SuperGrid Vision
underground transmission lines without fear of damag- Dc superconducting cables are envisioned as a future
ing the natural habitat and causing environmental con- long-distance high-power transmission application.
cerns. Long directional drilling up to 7200 ft (2200 m) Extremely high power links for transcontinental trans-
and pulling without a splicing chamber have now been mission of electric power may be possible using dc
made possible. superconducting cables operating at relatively low volt-
ages (50 to 100 kV) and tens of thousands of amperes.
17.4.2 Subsurface Structures Detection These HTS cables may be incorporated into a Super-
Tools for detecting existing subsurface structures are Grid that would also include transport of hydrogen for a
being developed. EPRI has developed a ground pene- future hydrogen economy.
Figure 17-1 Artist’s concept of SuperGrid long-distance underground transmission corridor.
17-6
The U.S Department of Energy (DOE) and the Office DOE and the OE, advanced materials such as high-
of Electric Transmission and Distribution (OE) (now temperature diamond materials, would be applied to the
called the Office of Electricity Delivery and Energy transmission, distribution, and control of electricity.
Reliability) has created their vision of Grid 2030. This Diamond technology could replace silicon and yield
came, under direction from President Bush, on revolutionary improvements in current density.
February 6, 2003, to “… modernize our electric delivery
system … for economic security … and for national The SuperGrid, originally conceived by EPRI’s founder,
security.” According to these agencies, “Grid 2030 Chauncey Starr, is envisioned with underground nuclear
energizes a competitive North American marketplace generation as a preferred option (though wind farms or
for electricity. It connects everyone to abundant, afford- large hydroelectric stations distant from load centers
able, clean, efficient, and reliable electric power may also prove attractive as sources of bulk power).
anytime, anywhere. It provides the best and most secure These plants would produce hydrogen for storage or fuel
electric services available in the world.” (EPRI 2003). with subsequent national distribution through a
"hydrogen-electricity" pipeline system. Superconduct-
Creation of a national electricity backbone is one of the ing dc cables would be placed inside the hydrogen
three major elements of the Grid 2030, the other two cooled pipes; the combined system would transmit large
being regional interconnections and local distribution. quantities of electrical power and/or hydrogen fuel
This electricity backbone would consist of high-capacity through the pipeline to dense urban areas.
transmission corridors linking the east and west coasts,
as well as Canada and Mexico, and would comprise of a 17.6 SUMMARY
variety of technologies. These technologies would
In conclusion, the future of underground cable systems
include controllable, very-low-impedance ac supercon-
looks promising as advancements in material, design,
ducting cables and transformers operating within the
operation, construction, and manufacturing processes
synchronous ac environment; high-ampacity, long-
result in cost reductions. This will make underground
distance dc superconducting cables and devices forming
transmission competitive with overhead transmission
connections between or around regions; and other types
lines in an increasing number of locations.
of advanced electricity conductors. According to the
17-7
REFERENCES EPRI, 1999. Underground Cable Installation and System

DOE. 2004. U.S. Department of Energy. Office of Cost Reduction Phase 2. Report No. TR-114457, Palo
Transmission and Distribution Superconductivity Part- Alto, CA. December.
nership with Industry “A National Effort to Introduce
New Technology into the Power Delivery Infrastructure”, EPRI. 2003. Transforming the Grid to Revolutionize Elec-
Washington, DC. May. tric Power in North America, “Grid 2030”: A National
Vision for Electricity’s Second 100 Years. Report No.
EPRI. 1992. Underground Transmission Systems Refer- 1013089. Palo Alto, CA. July.
ence Book. First Edition. Report No. TR-101670. Palo
Alto, CA. EPRI. 2005. An Examination of the Potential for Nano-
Composites in the Formulation of HV Cable Insulation.
EPRI, 1997. Underground Cable Installation and System Report No. 1010498. Palo Alto, CA. November.
Cost Reduction, Phase 1. Report No. TR-109150, Palo
Alto, CA. November. EPRI. 2005. Capturing Undocumented Expert Knowl-
edge; Electric Delivery Topics for Public Service Electric
and Gas. Report No. 1012127. Palo Alto, CA. August.
17-8
APPENDIX 1 EPRI Reports Related

to Underground
Transmission Systems
1977
Development of Extruded Dielectric Underground Transmission Cables Rated 138 kV,
230 kV and 345 kV, Volume 1. EL-428-V1. May.
Free and Forced Convective Cooling of Pipe-Type Electric Cables. EL-147. August.
1978
Research to Determine the Acceptable Emergency Operating Temperatures for Extruded
Dielectric Cables. EL-938. November.
1981
Development of Extruded Dielectric Underground Transmission Cables Rated 138 kV,
230 kV, and 345 kV, Volume 2. EL-428-V2. December.
1984
Designer’s Handbook for Forced-Cooled High-Pressure Oil-Filled Pipe-Type Cable Sys-
tems. EL-3624. July.
The handbook provides utilities with design procedures and specifications to optimize
their systems to reduce capital and operating costs.
1985
Flexible Gas-Insulated Cable Design for 500-kV Transmission. EL-3819. January.
By extrapolating from the 230-kV and 345-kV cable designs, researchers established the
characteristics and limits of higher-voltage (500-kV) cables, produced two system designs,
and explored their fabrication.
Calculating AC/DC Resistance Ratios for High-Pressure Oil-filled Cable Designs,

Volumes 1 and 2. EL-3977-V1. April.
With the new technique developed in this study, engineers can compute power losses
more precisely for cables of various configurations and voltage ratings.
Soil Resistivity Tests Using Modeling Techniques Copyright. EL-3983. May.
1987
Characterization of Arc By-Products of Sulfur Hexafluoride and Polymeric Construction
Materials. EL-5089. April.
A1-1
Appendix 1: EPRI Reports Related to Underground Transmission Systems EPRI Underground Transmission Systems Reference Book
Economics of PPP-Insulated Pipe-Type Cable. Transmission Cable Magnetic Field Research.

EL-5279. October. EL-6759-D. April.
This report establishes the economic range of applica- This report provides magnetic field data for one cable
bility for transmission cable systems insulated with configuration and documents the need for additional
paper-polypropylene-paper (PPP). research.
1988 Development of Moisture Barriers for Extruded Dielec-

tric Transmission Cables. EL-6857. June.
Thermal Overload Characteristics of Extruded Dielectric
This report presents the results of an EPRI project to
Cables. EL-5757. June.
develop moisture barriers for extruded dielectric cables,
and to develop an expansion layer between the cable
Thermomechanical Bending Effects on Extra-High
and the moisture barrier so that the construction can
Voltage Pipe-Type Cables. EL-5880. July.
withstand thermal cycling.
The report summarizes the work to identify the extent
of TMB in pipe-type cables, determine the effects of
Development of 230-kV High-Pressure, Gas-Filled,
cable construction variables on TMB damage resis-
Pipe-Type Cable System: Model Test Program Phase.
tance, and to develop a test that determines the proba-
EL-6933. September.
ble performance of new cable designs.
This report presents the results of an EPRI project to
establish a design and verify the technical and manufac-
1989 turing feasibility of a 230-kV HPGF pipe-type cable
Evaluation of Diagnostic Techniques for Cable Charac- system.
terization. EL-6207. February.
This study explores a variety of novel physical and Overload Characteristics of Paper-Polypropylene-Paper
chemical diagnostic techniques for assessing cable con- Cable. EL-6929. September.
dition and suggests a methodology for completely char- This report presents the short-time ratings of (Paper-
acterizing cable insulation. Polypropylene-Paper) PPP-insulated cables and com-
pares them with paper-insulated cables.
AC Losses in the New High-Temperature Superconduc-
tors. EL-6277. March. Space Charge in Polyethylene-Ionomer Blends for DC
This report addresses the properties of high-tempera- Cable Insulation. EL-6977. October.
ture ceramic oxide superconductors in low magnetic This research studied blends of ionomers (charged poly-
fields; it discusses ac losses in the superconducting and mers) with polyethylene to determine whether they
normal states, the influence of anisotropy, and a data- would prevent space charge buildup and be useful as
base for monitoring advances in superconductivity. insulation for extruded dc cable insulation.
1990 Evaluation of Sensitive Diagnostic Techniques for Cable

Characterization: Volume 1: Nine Diagnostic Tools.
Evaluation of Pipe-Type Cable Joint Restraint Systems.
EL-7076-V1. December.
EL-6760. March.
This study reviews a host of physical and chemical
This report details an investigation of several joint-
methods and discusses techniques to provide quantita-
restraining designs for pipe-type cables and recom-
tive information on cable insulation volatiles, including
mends one design.
water, ionic contaminants, and structural changes such
as water trees.
The Effect of Polymer Properties on Water Tree Growth.
EL-6807. April.
Proceedings: JICABLE/EPRI/CEA Workshop
Researchers studied the effect of polymer properties on
“Cable 89”: The Aging of Extruded Dielectric Cables.
water tree growth, using a simpler material, polystyrene,
EL-7090. December.
as a model to define the effects of parameters such as
This workshop addressed means of improving both dis-
crystallization, stiffness, and crosslinking on water tree
tribution and transmission-class cable lifetime and per-
initiation and growth.
formance reliability at a reasonable cost.
A1-2
EPRI Underground Transmission Systems Reference Book Appendix 1: EPRI Reports Related to Underground Transmission Systems
1991 Cross-Linked Polyethylene Cable Insulation:

Space Charge Measurement in DC Cable Materials. Electrically Stimulated Acoustic Wave Measurements.
EL-7301. May. EL-7402. August.
This research studied blends of ionomers (charged poly- This research evaluated the electrically stimulated
mers) with polyethylene to determine whether they acoustic wave nondestructive diagnostic technique to
would prevent space charge buildup, and facilitated the study charge distribution in crosslinked polyethylene.
development of the nondestructive thermal-pulse tech-
nique for measurement of space charge. Waltz Mill Testing of 345-kV PPP Cable.
EL-7429. October.
Accessories for Specially Bonded Extruded-Dielectric This report provides details of the installation, test
Transmission Cable Systems. EL-7259. July. program, and results of a prototype 345-kV PPP-
This project addressed the use of special bonding for (paper-polypropylene-paper) insulated pipe-type cable
extruded-dielectric cable systems to reduce cable shield subjected to a two-year accelerated life test program at
circulating currents and the resulting power losses. EPRI’s Waltz Mill Cable Test Facility.
Dynamic Rating and Underground Monitoring System Development of an Oil Deterioration Test Method to
(DRUMS). EL-7341. July. Monitor the Condition of High-Pressure Fluid-Filled
This study indicates that the various functions and Paper Cable. EL-7488-L. November.
communications systems that constitute the DRUMS This report summarizes the research on using a dis-
concept are technically feasible with currently solved-gas analysis (DGA) service for “fingerprinting”
available technology. as well as periodic analyses of HPFF cable to alert a
utility to potential catastrophic events and to thermal or
Terminations for Extruded-Dielectric Cables Rated electrical changes in the insulating fluid.
69 kV and 138 kV. EL-7331. July.
This report identifies the principal causes of failure in 1992
terminals of domestic manufacture and suggests modifi- Design of a Robotic Mouse and Other Techniques
cations to overcome these problems. for Instrumenting HPFF Pipe-Type Cable.
TR-100303. March.
Underground Transmission Computer Program: This report summarizes the ERPI project to design a
Debugging and Field Testing of the ICO Pipe Coupler; device or system that can place a fiber-optic sensor wire
Volumes 1 and 2. EL-7299-V1, V2. July. in the occupied steel pipe of HPFF cable systems; and
EPRI research confirms that ICO, Inc.’s interference fit to develop designs for a robotic device that can move
weldless pipe coupler is mechanically and electrically through small interstitial regions inside pipe-type cables
equivalent or superior to welded pipe joints. In related to inspect and install sensors.
work, EPRI’s HEATER code, a soil thermal profiler,
accurately analyzed nonrectangular trenches and Development of a Lightweight Composite Trench Cover.
determined the effect of pipe temperature on surround- TR-100296. March.
ing areas. This report presents the results of developing and test-
ing a lightweight composite plate that meets all Occupa-
Characterization of Cable Insulation Materials: tional Safety and Health Association (OSHA) as well as
Volume 2: Dynamic Mechanical Spectroscopy Studies: bridge deck/road requirements and also is comparable
Volume 2: Dynamic Mechanical Spectroscopy Studies. in strength and unit cost to steel plate.
EL-7076-V2. August.
This exploratory study showed that thermal aging in the Study of Trenching Versus Trenchless Construction
presence and absence of voltage stress yields different Methods for Installing Underground HPFF Cable Sys-
results for crosslinked polyethylene (XLPE) and high- tems. TR-100302. March.
molecular-weight polyethylene (HMWPE) and that This report summarizes a paper study of trenchless con-
service aging caused greater changes in HMWPE than struction techniques based on an actual 3.7-mile case.
in XLPE.
Waltz Mill Testing of 138-kV Factory-Molded Splice
Assemblies. TR-100415. April.
This report summarizes the research to determine the
integrity, reliability, usefulness, and life expectancy of
prototype 138-kV factory-molded splices.
A1-3
Computerized Database on Dielectric Materials. Field Measurement of Cable Dissipation Factor.

TR-100619. May. TR-102449. May.
The report presents the results of the development of an This report describes the development of a diagnostic
easy-to-use, interactive, computerized, dielectric materi- tool for determining the condition of transmission
als database that is readily accessible by telephone lines cables with oil-impregnated-paper insulation.
from a user’s computer terminal.
Transmission Cable Magnetic Field Management.
Development of 500/600-kV Solid-Type Nonpressurized TR-102003. June.
Oil-Paper DC Cable. TR-100621. May. This report provides detailed information on how to
This report summarizes the development of nonpressur- reduce magnetic fields caused by single-conductor
ized, solid-type, oil-paper dc cables rated 500/600 kV. transmission cables and what trade-offs are necessary.
Development of Optimized PPP Insulated Pipe-Cable Applications of Oxide Superconductors. TR-102780.

Systems in the Commercial Voltage Range. September.
TR-100416. May. Research was conducted to understand the basic physi-
This report presents the details of the design, manufac- cal properties and the materials characteristics of high-
ture, and successful laboratory testing of PPP insulated temperature superconductors that control their electri-
cable in the 138 to 345-kV range. cal and mechanical properties, and determine their
potential usefulness.
High-Voltage Solid-Dielectric Cable Test Termination.
TR-100620. May. Interfacial Bond Strength of Paper-Polypropylene-
This project attempted to resolve the difficulty of Paper Laminates. TR-102777. September.
achieving electrical breakdowns in extruded dielectric This research determined adhesion characteristics of
transmission cable through development of a hybrid ter- PPP under cyclic heating in oil and under tensile or
minal design, incorporating dielectric gases and fluids compressive stress, performed optical and electron
as well as a molded stress-control cone. microscopy, and developed an understanding of the
paper-polypropylene adhesion after the aging treatment.
Fundamental Studies of Tl-Ba-Ca-Cu-O Single
Crystals and New High-Temperature Superconducting Investigation of New Families of HTSC Compounds.
Compounds. TR-100731. June. TR-102778. September.
The research team developed new Tl-based high-tem- This project investigated the doping of copper chloride
perature superconductors and measured the properties to achieve superconductivity.
of the new materials.
1994
Waltz Mill Testing of 765-kV Paper-Polypropylene-
Diagnostics for the Detection and Measurement of
Paper (PPP) Cable. TR-100422. June.
Charge Motion in Water Trees in Polymers.
This report provides details of the installation, test
TR-103892. May.
program, and results of a prototype 765-kV PPP-insu-
The research was to fabricate a noninvasive diagnostic
lated pipe-type cable that was subjected to a two-year
system for detecting water treeing and to make that sys-
accelerated test program at EPRI’s Waltz Mill Cable
tem sensitive to the Johnson noise limit.
Test Center.
Evaluation of Trenchless Technology for Installation of
Microscopic Probes of High-Temperature Superconduc-
Pipe-Type Cable: Phase 1. TR-103219. May.
tivity. TR-100862. July.
Research was conducted to analyze guided boring and
microtunneling/pipe-jacking technologies with respect
1993 to the installation of power cable in city streets, to per-
Underground Transmission Systems Reference Book. form detailed economic studies for determining the
TR-101670. March. Research ended December 1992. most feasible installation technology, and to develop a
The Underground Transmission Systems Reference Book design method for cable installation.
covers all stages of cable system design and operation,
from initial planning studies to failure analysis. It
contains contributions from many of the industry’s
experts and represents practices from all parts of the
United States.
A1-4
Superconducting Low Voltage Direct Current (LVDC) Guidelines for the Design and Installation of Transmis-
Networks. TR-103636. April. sion and Distribution Cables Using Guided Drilling Sys-
This report presents a scheme for a superconducting, tems. TR-105850. December.
parallel-connected, multiterminal dc transmission This design guide will assist utility engineers as well as
system. the construction contractors by providing information
on the engineering plan, permitting, site analysis, and
Design Concepts for a Superconducting Cable. specifying equipment needed to perform guided drilling.
TR-103631. September.
This report discusses previous approaches to designing Superconducting Magnetic Energy Storage in Trapped
superconducting cables, describes the technical issues Field Magnets of High Temperature Superconductors.
that must be considered when designing a high-temper- TR-105751. December.
ature superconductor (HTS) cable, and presents several
design concepts for an HTS cable. 1996
Cable Pushing and Calculation of Pulling Tensions
New Superconducting Materials. TR-104414. October.
for Distribution and Transmission Cables. TR-106500.
This report presents criteria to guide the search for
September.
higher transition temperature materials.
This report presents the research of analyzing pulling
forces for several types of conduit and cable, and devel-
1995 oping revised pulling tension equations that take into
Basic Breakdown Study of HPFF and Extruded-Dielec- account the effect of cable pushing.
tric Cables. TR-105301. August.
This report describes a major project to develop a test 1997
plan and conduct a large number of tests on representa-
Guided Horizontal Drilling: A Primer for Electric Utili-
tive HPFF and extruded-dielectric transmission cables.
ties. TR-107687. February 1997, Research ended Janu-
ary 1997.
DRUMS Leak Detection for HPFF Pipe-Type Cable
This document is intended to be an introduction to
Systems. TR-105250. August.
guided horizontal drilling, also termed horizontal direc-
This project developed a leak detection system that lim-
tional drilling (HDD), as an alternative construction
its fluid loss to 0.05% of the total system volume of the
method to open trenching for the installation of under-
HPFF cable system.
ground power cables, pipes, ducts, or conduits.
DRUMS Dynamic Feeder Rating for Pipe-Type Cable
Testing of XLPE Transmission Cable Terminations at
Systems. TR-105251. October.
Three Utilities: Southern California Edison, Public
Research was conducted to develop an efficient method-
Service Company of Colorado, and PECO Energy
ology to calculate normal and transient ratings, and to
Company. TR-108073. May 1997. Research ended Sep-
validate the models and methods on full-scale cooling
tember 1996.
equipment and cable.
This report describes the testing procedures and results
obtained from performing very-high-frequency partial-
Effects of Intense Fluid Pumping on Forced-Cooled
discharge measurements of 14 crosslinked polyethylene
Cables. TR-105534. November.
(XLPE) cable terminations at Southern California Edi-
This study investigated the possible breakdown of
son, Public Service Company of Colorado, and PECO
dielectric fluid due to forced cooling and intense pump-
Energy Company.
ing of an HPFF cable system.
Computing the Magnetic Fields of High-Pressure Fluid-
Underground Cable Fault Location Reference Manual.
Filled Cables. TR-108269. June 1997, Research ended
TR-105502. November.
September 1994.
This fault location reference manual provides practical
This report describes a Fourier series expansion tech-
technical material on the art and science of locating
nique and a hybrid finite element and unimoment
cable faults, including a description of common fault
method for investigating the magnetic fields generated
location instruments and principles of advanced fault
by high-pressure fluid-filled (HPFF) underground
location techniques used by utilities throughout the
transmission cables.
United States.
A1-5
Reliability Centered Maintenance (RCM) Technical A Feasibility Study of an Application Using Supercon-
Reference for Power Delivery. TR-108068. October 1997, ducting Cable. TR-109654. December 1997, Research
Research ended April 1997. ended October 1997.
EPRI’s RCM Technical Reference for Power Delivery In this project, EPRI, Pirelli Cable Corporation, and
provides utilities with essential technical data and guid- South Carolina Electric and Gas (SCE&G) performed a
ance for conducting RCM studies of power delivery joint study to assess the technical and economic feasibil-
equipment and systems. ity of installing an HTS cable system on a highly loaded
SCE&G underground circuit.
HPFF Cable Leak Location Using Perfluorocarbon
Tracers. TR-109086. November 1997, Research ended 1998
September 1997.
Assessment of Extruded 345-kV Cable Technology.
In this project, researchers developed methods using
TR-110906. June 1998, Research ended December 1997.
perfluorocarbon tracers (PFT) to quickly and accu-
This report presents the prequalification tests results of
rately locate underground leaks.
345-kV XLPE cable technology with premolded joints.
Soil Thermal Properties Manual for Underground Power
Distributed Fiber Optic Temperature Monitoring and
Transmission: Soil Thermal Property Measurements,
Ampacity Analysis for XLPE Transmission Cables.
Soil Thermal Stability, and the Use of Corrective Ther-
TR-110630. June 1998, Research ended December 1997.
mal Backfills. TR-108919. November 1997, Research
Project identified appropriate instrumentation, modi-
ended December 1996.
fied it as necessary for operation on underground cables,
This guide provides comprehensive information on soil
and conducted a demonstration on a utility system.
thermal property measurement, surveying, interpreta-
tion, and improvement for underground transmission
Evaluation of Service-Aged 69- and 115-kV XLPE
cables.
Cables. TR-110882. June 1998, Research Ended
May 1998.
Underground Cable Installation and System Cost Reduc-
This report evaluates retention of ac voltage breakdown
tion: Phase 1. TR-109150. November 1997, Research
and other characteristics for extruded transmission
ended August 1997.
cables aged in service.
The first phase of a comprehensive study on cost reduc-
tion, this report examines current utility practices in
Assessment of Insulation Quality for 115-kV XLPE
underground cable design and installation.
Cables. TE-113557. September.
Cable Oil Monitor and Tester (COMAT). TR-109071.
Behavior of Paper-Polypropylene-Paper Laminate Under
December 1997, Research ended November 1996.
Thermal and Electrical Stresses. TR-111321. September
This report describes efforts to develop an on-line mon-
1998, Research ended December 1997.
itoring and diagnostic system, using dissolved gas anal-
Research was conducted to assess the gassing behavior
ysis (DGA), to provide more accurate evaluations of
and mechanical characteristics of PPP as affected by
HPFF cable conditions.
thermal and electrical aging conditions.
Deep Cable Ampacities: Guidelines for Calculating
Dissolved Gas Analysis (DGA) by EPRI Disposable Oil
Ampacities of Cables Installed by Guided Boring.
Sampling System (EDOSS). TR-111322. September
TR-109205. December 1997, Research ended
1998, Research ended April 1997.
October 1997.
This report summarizes the research to design an easy-
This project determined whether basic principles and
to use, inexpensive, but equally accurate version of
equations for standard-depth installations apply to
EPOSS, using a disposable glass vial for fluid collection
greater depths of cable installation.
and analysis as well as a commercially available head-
space analyzer to automate analysis.
A1-6
High Temperature Superconductivity—A Joint Feasibil- Evaluation of Field Diagnostic Techniques for Transmis-
ity Study for a Power Application with High-Temperature sion Cable Accessories. TR-112676. October 1999,
Superconducting Cable by South Carolina Electric and Research ended February 1999.
Gas Company. TR-110891. November 1998, Research The research provides diagnostic tools that will give util-
ended March 1998. ities and other system operators time to make necessary
This report documents a feasibility study co-sponsored repairs or schedule maintenance before disruptions to
by South Carolina Electric and Gas Company customer loads occur.
(SCE&G) to assess a real-world underground transmis-
sion application of this technology. Review of After-Laying Tests for Extruded Transmission
Cables. WP-114431. November.
High Temperature Superconductivity—A Joint Feasibil-
ity Study for a Power Application with High-Temperature Dissolved Gas Analysis (DGA) for Fluid Filled Termina-
Superconducting Cable by Peco Energy. TR-110892. tions of Extruded Transmission Cables. TR-114197.
November 1998, Research ended March 1998. December 1999, Research ended June 1999.
This report documents a feasibility study co-sponsored This report describes the equipment and procedure
by PECO Energy Company (PECO) to assess a real- required to sample fluid from extruded cable termina-
world underground transmission application of this tions for assessing the condition of these vital compo-
technology. nents through dissolved gas analysis (DGA).
Options for Reducing Environmental-Related Utility Underground Cable Installation and System Cost Reduc-
Costs Associated With Dielectric Fluids Employed in tion: Phase 2: Design Tradeoffs. TR-114457. December.
Cables and Transformers. TR-111722. December 1998, This report addresses a number of technical issues asso-
Research ended October 1998. ciated with cable installation, including circuit rating,
This report presents results of a literature review and trench configuration, and trench construction, and it
technical workshop on environmental management of evaluates the cost-saving installation approaches from
dielectric fluids, with emphasis on those properties that pipe-lining technologies to fast microtunneling and
strongly influence transport, fate, impacts, and costs of plowing-in cables.
a dielectric fluid release into the environment.
2000
Transmission Cable Life Evaluation and Management.
RCM Workstation for Power Delivery Version 1.1.
TR-111712. December 1998, Research ended Septem-
February 2000, Research ended October 1997.
ber 1998.
RCM Workstation for Power Delivery is a tool for
This report describes procedures and diagnostic tools
completion of PM (preventive maintenance) optimiza-
for assessing the condition and remaining life of these
tion studies using reliability-centered maintenance
high-value assets and effectively managing their opera-
(RCM) methods.
tion.
Industry Survey and Assessment of Available Corrosion
1999 Mitigation Technologies: 20th Century State of the Art.
Distributed Fiber-Optic Measurements on Distribution TE-114896. March.
Cable Systems. TE-114897. January.
Guidelines for the Interpretation of Dissolved Gas Analy-
Development of Improved Extrudable Dielectrics for sis (DGA) for Paper-Insulated Underground Transmis-
Transmission Cables. TE-113971. September. sion Cable Systems. 1000275. September.
138-kV Maintenance Hole Restraining System Application of Fiber-Optic Temperature Monitoring to

Testing. TR-113556. September 1999, Research ended Solid Dielectric Cable: DFOTS Installation at Con
June 1999. Edison. 1000469. November.
This report summarizes the testing of a restraining sys-
tem as designed by Los Angeles Department of Water Sagnac Interferometer for Electric Discharge Detection.
and Power (LADWP). 1000715. November.
Corrosion and Cathodic Protection of High Pressure

Fluid Filled Cable Circuits. WP-113987. October.
A1-7
Superconducting Cable Construction and Testing. Guide on Selection, Handling and Maintenance of Cable
1000160. November 2000, Research ended June 2000. Fluids. 1001924. December.
This report describes the development, construction, This guide provides an authoritative but easy-to-use
and testing of a superconducting cable system. source for information on the selection, storage, trans-
portation, maintenance, and disposal of cable fluids.
Aging Model Estimating the Life of Dry Aged XLPE
Cables: Progress Report. 1000277. December. Methods for Cleaning and Evaluating Pipe-Type
Cable Pipes for Retrofit with Extruded Dielectric Cable
Characteristics of High Pressure Fluid Filled Pipe-Type Systems. 1001937. December.
Cable Dielectric Fluids: Progress Report. 1000716. This report summarizes studies of pipe cleaning, “pig-
December. ging,” inspecting, and testing pipe-type cable pipes for
re-conditioning as conduits compatible with extruded
Guide for Operation & Maintenance of Paper-Insulated dielectric cables.
Transmission Cables: Part 1. 1000458. December.
Sagnac System for Detecting Defects in Cable Insulation
Reliability Parameters of XLPE Underground Transmis- and Jackets: Theory, Calibration, and Field Tests.
sion Systems Based on EDF Experience. 1000473. 1001935. December.
December. The research was conducted to design and build a Sag-
nac Interferometer system, calibrate the system for its
Software Requirements Specification for Underground position detection sensitivity, and demonstrate proof of
Line Inspection System. 1000459. December. principle in the field.
Testing of Ultrasonic Guided Waves for Pipe Evaluation. Ultrasonic Guided Waves for Pipe Ducts. 1001929.
1000445. December. December.
2001 Ultrasonic Guided Waves for Pipe Ducts. 1001930.

December.
Cost Reduction Activities in France for Installing Cable:
Mechanical Laying Techniques. 1000519. January.
Underground Transmission Workstation (UTW) 3.0. CD
Installer Software and User’s Guide, Version 1.0.
Development of Diagnostic Methods for Polyethylene
1006723. December.
Insulations in Transmission Cables. 1001393. February.
Underground Transmission WorkStation (UTW) 3.0. CD
Summary of Diagnostic Evaluation for Submarine Cable
Installer: Version 1.0 User’s Guide. 1006724. December.
Cathodic Protection System Performance: Feasibility
Assessment of Technologies. 1001931. July.
UTWorkstation ACE Version 4.0 User’s Manual.
1006688. December.
Underground Line Inspection System Status Report.
The UTWorkstation Alternative Cable Evaluation
1006263. August.
(ACE) program provides users with the capability to
evaluate ampacity and costs of major types of under-
CONQUEST—Construction Quick Estimator Code,
ground transmission cable systems: pipe-type (high
Version 1.0. 1006560. October.
pressure fluid-filled and high pressure gas-filled), self-
contained fluid-filled, and extruded dielectric.
Dry Electrical Aging and Breakdown of Extruded Cable
Insulation. 1001922. November.
Field Demonstration of a 24-kV Warm Dielectric HTS
This report presents a new aging model that has been
Cable at Detroit Edison: 2000 Annual Progress. 1001337.
validated against experimental data obtained by differ-
ent researchers at various temperatures.
2002
ACE—Underground Transmission Workstation Alter- Field Demonstration of a 24-KV Warm Dielectric Cable
nate Cable Evaluation Code, Version 4.0. 1006687. at Detroit Edison: FY2001 Annual Progress Report.
December. 1001934. February.
A1-8
Characteristics of Pipe-Type Cable Fluids and Develop- Condition and Power Transfer Assessment of Center-
ment of Risk-Based Cleanup Goals: Interim Report. Point Energy’s Polk-Garrott Pipe-Type Cable Circuit.
1001932. March. 1007539. December.
This report presents interim results of a project to char- This report summarizes distributed fiber optic tempera-
acterize the physical and chemical properties of HPFF ture sensing (DFOTS) on pipe-type cable circuits and
cable fluids, evaluate their fate and transport in the the results of modifying and implementing EPRI’s
environment, and develop soil cleanup levels for the Dynamic Thermal Circuit Rating (DTCR) system on
fluids that are protective of human health and the envi- CenterPoint’s Polk-Garrott pipe-type cable circuit in
ronment. Houston, Texas.
Application of Fiber-Optic Distributed Temperature Extruded Dielectrics for Transmission Cables: New
Sensing to Power Transmission Cables at BC Hydro. Developments and Aging Models 1001860. December.
1000443. May. This report summarizes the results of the initial work
This report describes applications of distributed temper- relating to the further development of existing aging
ature sensing methods to underground and submarine models and the review of materials for extruded insula-
cables at BC Hydro over a five-year period. tion systems.
Demonstration of Commissioning Tests for Extra-High Impact of Emergency Operating Temperatures on the
Voltage Cross-Linked Polyethylene Cable Systems at Integrity of XLPE Transmission Cable Systems—
Los Angeles Department of Water and Power. 1007298. Interim Report 1001858. December.
September. This interim report describes the results of the first four
months of work on a project to systematically analyze
Demonstration of Commissioning Tests for Extra-High the characteristics of XLPE transmission cables and
Voltage Cross-Linked Polyethylene Cable Systems at develop recommended limits for emergency operating
Los Angeles Department of Water and Power. 1001856. temperatures for 100- and 300-hour durations at 75 and
September. 100% load factor.
Workshop: Operation and Maintenance of Paper-Insu- Mechanical Effects on Extruded Dielectric Cables and
lated Transmission Cables. 1007446. October. Joints Installed in Underground Transmission Systems in
North America. 1001848. December.
Ampacity Evaluation and Distributed Fiber Optic Test- This report discusses findings and progress to date on a
ing on Pipe-Type Cables Under Bridgeport Harbor. project involving the modeling of mechanical effects on
1007534. December. pipe cable routes.
This report documents the results of case evaluations
and a detailed ampacity study to determine the actual Sagnac System for Detecting Electrical Shorts II: New
capacity of the lines under both normal and emergency Robust Techniques and Field Tests. 1001851. December.
ampacity conditions. This report summarizes the results of laboratory and
field tests that were undertaken for the purpose of
Cable System Technology Review of XLPE EHV Cables: assessing a Sagnac interferometer for detecting the posi-
220 kV to 500 kV. 1001846. December. tion of an electrical short.
This report reviews the worldwide state-of-the-art, gives
methods of calculating insulation design stress, and 2003
recommends critical factors to consider for reliable
Characteristics of Pipe-Type Cable Fluids: Environmen-
operation when specifying long-length EHV XLPE
tal Data Protocol and Comprehensive Test Matrix for
cable systems.
Fluid Evaluation. 1007615. January.
This report presents a decision protocol for conducting
a site investigation of cable fluid releases to soil and
groundwater, and a comprehensive testing and accep-
tance matrix for evaluation of new cable fluids.
A1-9
Field Demonstration of a 24-kV Warm Dielectric Cable Mechanical Effects on Extruded Dielectric Cables and
at Detroit Edison: FY2002 Annual Progress Report for Joints Installed in Underground Transmission Systems in
the Detroit Edison HTS Cable Field Demonstration North America. 1001849. March.
Project. 1001847. February. This report provides information on the mechanical
This report summarizes the progress of the demonstra- effects on extruded dielectric cables in underground
tion of a HTS cable system between the dates of 2 transmission systems.
December 2001 and 1 December 2002.
Novel Applications of Fiber Optic Sensor Technology for
Extruded Dielectrics for Transmission Cables: Evalua- Diagnostics of Underground Cables. 1008712. October.
tion of Aging Models 1002036. December. This report provides an overview of the state of the art
This report summarizes the evaluation of the three in the use of fiber optics in measurement applications,
models: (1) the Thermodynamic Model of Aging of with particular regard to applications suitable for mea-
Extruded Insulation by J. P. Crine; (2) the Space Charge suring or monitoring significant parameters for under-
Aging Model by L. A. Dissado, G. Mazzanti, and G. C. ground power cables.
Montanari (DMM); and (3) the Electromechanical
Aging Model by T. J. Lewis. DC Cable Systems with Extruded Dielectrics. 1008720.
December.
Impact of Emergency Operating Temperatures on the This report reviews the current state-of-the-art in
Integrity of XLPE Transmission Cable Systems. extruded-dielectric dc cable system technology.
1001859. December.
This report describes a research project that was carried Dissipation Factor Measurement of High-Pressure Fluid-
out to develop an understanding of the mechanical Filled Pipe-Type Cables Subjected to Prolonged Opera-
properties of XLPE transmission cables at high temper- tion at Low Hydraulic Pressure. 1011366. December.
atures. This report presents the results and conclusion of dissi-
pation factor measurement of HPFF pipe-type cables
NYPA Avoids Costly Cable Replacement and Improves subjected to prolonged operation at low hydraulic pres-
Reliability with Advanced Diagnostics to Assess Cable sure at PSE&G.
Condition and Power Transfer Capability. 1009232.
December. An Examination of the Potential for Nano-Composites
in the Formulation of HV Cable Insulation. 1002038.
Review of Materials and Materials Technologies for December.
Transmission Cable Applications. 1002037. December. This interim technical update describes the potential
This technical update reviews new materials and tech- for nano-composites in the formulation of HV cable
nologies, additives, impregnants, and nanofillers with insulation.
the objective of determining a path and direction of fur-
ther research work needed to support the growth of reli- Functional Description Preparation to Update the Under-
able, cost-effective, and high-performance transmission ground Transmission Workstation. 1008717. December.
cable systems. This report discusses the collection of information to
prepare a Functional Description document that will
UTDT Software, Version 1.0: Underground Transmission be used to direct the development of EPRI’s Under-
Design Tools Software. 1002035. December. ground Transmission Workstation (UTWorkstation)
The Underground Transmission Design Tools (UTDT) software system.
Software Version 1.0 enables users to search for relevant
EPRI products, including reports, computer programs, Lower Cost Underground Transmission Cable. 1008719.
videos, and web links relating to the design and installa- December.
tion of underground transmission cable systems. This report presents the potential methods to reduce
the installed cost of underground transmission cable
2004 circuits.
Field Demonstration of a 24-kV Warm Dielectric Super-
Review of the “Underground Transmission Systems
conducting Cable at Detroit Edison: FY2003 Annual
Reference Book.” 1008718. December.
Progress Report. 1002040. March.
This report summarizes a critical and editorial review of
This report summarizes the progress of the demonstra-
the Underground Transmission Systems Reference Book.
tion of a HTS cable system between the dates of 2
December 2002 and 1 December 2003.
A1-10
Simplified Leak Detection System for HPFF Cable Sys- Formation of Nanovoids in Extruded Dielectrics Caused
tems: Evaluation of Novel Leak Detection Technologies. by Mechanical Fatigue and Fracture. 1010497. August
1009809. December.
This document describes the evaluation of novel leak Functional Description for UTWorkstation 5.0. 1010499.
detection technologies. September.
Simplified Leak Detection System for HPFF Cable Sys- An Examination of the Potential for Nano-Composites in
tems: Evaluation of a New Leak Detection Technology. the Formulation of HV Cable Insulation. 1010498.
1008713. December. November.
This document presents the work that has been per-
formed in 2004 for developing a simplified leak detec- Updating the EPRI Underground Transmission Systems
tion system for HPFF cables, including a general Reference Book. 1010500. November.
description of the model and a detailed experiment plan
for the evaluation of the model. Simplified Leak Detection System for HPFF Cable Sys-
tems. 1010876. December.
Thermo-Mechanical Performance of 230-345kV High
Pressure Pipe Type Cable Systems: Selection of Full Size Simplified Leak Detection System for HPFF Cable Sys-
Test Rig. 1008714. December. tems. 1010503. December.
This report describes the front-end project to the wider
experimental study of the thermo-mechanical perfor- Unified Thermo-Mechanical and Thermal Model for the
mance of high-pressure fluid-filled (HPFF) transmis- UT Extruded Dielectric Systems. 1010496. December.
sion class cables in pipe systems.
2006
Using Guided Waves to Detect and Locate Corrosion in
EPRI Cable Testing Network (ECTN). 1008969. May.
Coated Piping. 1008715. December.
This report documents the results of laboratory and
Find Hot Spots, Verify Ratings, and Increase Capacity
field experiments designed to evaluate the use of ultra-
on Underground Power Cables Using Distributed Fiber
sonic guided waves for the detection of corrosion under
Optic Temperature Sensing (DFOTS)—Field Demon-
tar-coated pipes, for detection of delamination of such a
stration. 1012943. May.
coating from pipes, and for coated pipe weld inspection.
Near-Term Projected Transmission Circuit Ratings
2005 Based On Weather Prediction—Up to a Week in
Best Practices for HPFF Pipe-Type Cable Assessment, Advance. 1013057. May
Maintenance, and Testing: Blenheim-Gilboa 345-kV
Cable Systems. 1011489. March. Optimize Maintenance through Integrated Monitoring &
This report presents the results and conclusions of an Diagnostics (IMD). 1012853. May.
extensive series of tests on the NYPA Blenheim-Gilboa
345-kV high-pressure fluid-filled (HPFF) transmission Design Tools for Underground Transmission. 1012334.
cable systems to determine their condition. December.
Evaluation of Soft Spots in HPFF Circuits Installed in Guide for Partial Discharge Measurements of Transmis-
the St. Lawrence Power Plant. 1012009. June. sion Cables in the Field. 1012338. December.
An Assessment of ESR Spectroscopy as a Means of 2007

Detecting Electrical Aging in Polyethylene. 1012088.
Highly Reliable Extruded Dielectric Insulation Systems.
August.
1012337.
Capturing Undocumented Expert Knowledge. 1012127.
Thermal and Mechanical Stresses in Extruded Dielectric
August.
Underground Transmission Systems Installed in Ducts or
Pipes. 1012335.
A1-11
Glossary
ac/dc Resistance Ratio. Ratio of the actual alternating Annular Conductor. Conductor with an expanded diam-
current resistance of a conductor to its direct current eter to reduce skin effect and/or to provide a core for oil
resistance, taking into account the increments of ac pressurization and various other properties.
resistance due to skin effect and proximity effect.
ANSI. American National Standards Institute.
Accelerated Aging. Test in which certain factors, such as
voltage, temperature, and so forth, are increased or ASTM. American Society for Testing and Materials.
decreased beyond normal operating values to obtain
observable deterioration in a reasonable period of time, Beta Coefficient. Stress coefficient of electrical resistivity
and thereby afford some conditions or some measure of of a direct current insulating material.
the durability of the equipment when exposed to the
factors being aggravated. BIL (Basic lightning impulse insulation level). Electrical
strength of insulation expressed in terms of the crest
AEIC. Association of Edison Illuminating Companies. value of a standard lightning impulse under standard
atmospheric conditions.
Aggregate. Any of several hard inert materials such as
sand or gravel used for mixing with a cementing mate- Bird Caging. Condition in stranded conductors in which
rial to form concrete. the inner layer comes through the outer layer, backing
up and bulging out. This condition sometimes occurs in
Alkylbenzene. Family of synthetic insulation liquids concentric stranded conductors (or cables) when the lay
characterized by excellent hydrogen gas absorption is too long or assembly tensions are incorrect.
properties.
Butt Space, Butt Gap. The distance between adjacent
Alpha Coefficient. Temperature coefficient of electrical turns of a helically applied tape.
resistivity of a direct current insulating material.
Cable Clearance. Free space between cable and duct.
Aluminum. Material with 61% the conductivity and
0.3027 times the weight of copper and used in power Cable Configuration. Physical configuration that three
cables, overhead conductors (ACSR, AAC, ACAR, cables take when pulled into a common conduit.
AAAC), and some building wire.
Cable Entrance . Duct section closest to cable-feeding
Alumoweld. Thin coating of aluminum fused to a steel equipment. It is usually the lowest tension point of the
core. Used in line wire and cable messengers. cable in the duct section under examination.
Annealing. Process in which the conductor is heated Cable Exit . Duct section closest to the cable-pulling
about 700°F for copper, somewhat lower for aluminum equipment. This is usually the highest tension point of
and allowed to cool. Used to allow conductors to be the cable in the duct section under examination.
bent without breaking, to remove stiffness and facilitate
compacting. Cable Pusher. Device that allows cable to be pushed
into a duct. Also, can be used in tandem with cable pull-
ing to reduce installation tensions.
G-1
Glossary EPRI Underground Transmission Systems Reference Book
Cable Vibrator. Vibration cylinder used to reduce the Coefficient of Friction (Dynamic). Used to calculate the
tension required to remove a cable. tension required to start a cable in motion.
Capacitance . Property of a system of conductors, Coefficient of Friction (Effective). The coefficient of fric-
shields, and dielectrics that permits storage of electri- tion that takes into account the cable configuration for
cally separated charges when potential differences exist three cable pulls.
between the conductors and shields or sheaths.
Coefficient of Friction (High) . Measure of friction
Capacitive Voltage Stress. Alternating current cables between the cable(s) and conduit, which is used for pull-
(and dc cables under transient and ripple conditions) ing tension calculations when the cable is being pulled
have their voltage stress distribution governed by capac- around bends where the SWBP is less than 150 lb/ft.
itive effects. The capacitive gradient for a cable of con-
ductor shield radius, r1 and outer shield radius, r2 is Coefficient of Friction (Low) . Measure of friction
given as follows for an applied voltage V: between the cable(s) and conduit, which is used for pull-
Gx = V / (rx Ln (r2/r1) ) ing tension calculations at the start of the pull or when
the cable is being pulled around bends where the SWBP
Direct current cables will be subjected, under steady- is 150 lb/ft or greater.
state conditions, to an average electrical gradient equal
to the voltage divided by the electrode spacing, and a Coefficient of Friction (Static). Used to calculate the ten-
gradient at specific radial locations in the electric field sion required to start a cable in motion.
determined by the insulation resistivity and electrode
geometries. See also Gradient, Electrical. COG (Coefficient of Grounding). Ratio, line-to-ground
voltage (ELG)/line-to-line voltage (ELL), of the highest
Capacitively Graded Pothead . For ac operation, pot- root-mean-square (rms) line-to-ground power-fre-
heads at 230 kV and above generally have longitudinal quency voltage on a sound phase, at a selected location,
voltage grading controlled by capacitor elements during a fault to ground affecting one or more phases to
installed between the porcelain body of the pothead and the line-to-line power-frequency voltage, ELL, that
a phenolic or epoxy tube inside the pothead. Resistive would be obtained at the selected location with the fault
grading is desired for dc operation; it may therefore be removed.
necessary to change potheads when converting an ac
line to dc. See also Pothead. Cold Dielectric Superconducting Cable. Superconduct-
ing cable where the electrical insulation (dielectric) is
Capillary. Held by the surface tension of water in the located inside the cold space of the cable and must oper-
soil. ating at cryogenic temperatures. The dielectric is typi-
cally a tape-type dielectric wound in multiple layers on
Centerline Radius . Radius of curvature of the bend top of the phase HTS wires.
extending to the center of the duct. Manufacturers
speak in terms of centerline radius; however, the side- Compact Round Conductor (ASTM B-8) . Concentric
wall bearing pressure calculations use the inside radius. stranded conductor with the layers of individual wires
all laid in the same direction and rolled or die com-
CM (Circular Mil). System for specifying wire size by con- pacted by layer to a predetermined size. This results in
ductor area. For solid conductors, circular mils are an extremely smooth conductor surface approaching
obtained by multiplying the conductor diameter in that of a solid conductor at the same time preserving the
inches by 1000 and squaring the result. For concentric flexibility of a stranded conductor. Advantages are min-
stranded conductors, the area of the individual wires is imum overall diameter for a given conductor area and
determined and summated. The areas of compact and freedom from “bird caging” of strand under bending,
compressed strands are determined by weighing a unit and the practical elimination of space between strands.
length of a conductor, then calculating the area using Used in medium and high voltage power cables.
the known specific gravity (1/density) of the metal.
Compaction (1). Term for describing the amount of free
Coefficient of Friction (Basic). Coefficient of friction space in a stranded conductor.
based on the two contacting surface materials and the
lubricant used. Compaction (2). Process of increasing the density of
soils using a variety of devices.
G-2
EPRI Underground Transmission Systems Reference Book Glossary
Compressed Gas-Insulated Transmission Line (CGI) . smooth surface (copper) or compressed and welded
Transmission system consisting of a large-diameter con- (aluminum).
ductor, typically tubular aluminum, centered via epoxy
or plastic spacers in a larger-diameter shield, also typi- Copper. Principal electrical conductor. This material has
cally aluminum. Sulfur hexafluoride (SF6) gas, either by an electrical resistivity of 10.371 ohms-cmil/ft.
itself or in mixture with nitrogen and operating at a
pressure of 2-6 atmospheres, is the insulating fluid. Copperclad. Thin coating of copper fused to an alumi-
Some dc development has been done as part of a com- num core. Used in some building wires (#12 and larger).
pact substation. This system is also known as SF6 bus,
GITL, or gas-spacer system. See also Gas-Insulated Copperweld . Thin coating of copper fused to a steel
Transmission Line. core. Used in line wire and cable messengers, and
stranded with copper for strength or extending flex life.
Compressed Round Conductor (ASTM B-8). Concentric
lay conductor that has been reduced 3% in diameter in Cradled Configuration . Configuration where center
comparison to its ASTM-specified nonsized structure cable rests between the other two cables.
by light rolling or die-drawing. Used in high-voltage
extruded dielectric power cables. Critical Current. Maximum current that a superconduct-
ing wire may transmit at a given temperature and mag-
Compressive Strength. Compressive force or stress at n e t i c fi e l d wh i l e m a i n t a i n i n g s u p e rc o n d u c t i n g
which a material is permanently deformed. properties.
Concentric Strand (ASTM B-8). Conductor formed in a Critical Length. The capacitance charging current of an
layer pattern generally with 7, 19, 37, 61, or 91 wires. ac cable increases linearly with the cable length until, at
Normally adjacent layers have reversed direction of lay. the so-called critical length, the magnitude of the charg-
Control cables favor 7 or 19 wires, and power cables 7 ing current equals that of the thermal current rating of
through 91 wires depending on size and flexibility the cable so that no real power can be transmitted.
desired.
Critical Temperature. Maximum temperature at which a
Conci Conductor. Conci conductors are generally made superconducting wire exhibits superconducting proper-
of high-conductivity copper and consist of a number of ties.
annular layers of “conci,” or keystone segments laid up
around a central copper rod. Each layer is applied with Cross-Bonding. Technique of eliminating or minimizing
a helical lay opposite to that of the underlying layer to the I 2 R losses in sheaths of single-core cables which
improve torsional stability. This type of conductor is a accrue from voltages induced in the sheaths by the load
de facto standard for HVDC submarine cables where currents flowing in the conductors when the sheaths are
the absence of ac skin and proximity effects means that bonded so as to permit current flow. Cross-bonding uti-
the use of stranded-wire Milliken conductors is not lizes the phaser relationships among the induced volt-
required. ages, electrically transposing the connections between
three equal length sections so that the net vector voltage
Conduction Current. Conduction current of an imper- at the ends is equal, obviating current flow. Eddy cur-
fect dielectric material is proportional to the applied rent losses remain but in general are of a much lower
voltage gradient and is independent of time of voltage magnitude than the circulating current losses. An elec-
application for fixed conditions of test. Conduction cur- tric connection across a joint in the armor or lead
rent is strongly temperature-dependent and mildly volt- sheath of a cable, or between the armor or lead sheath
age-dependent for cable insulations. and the earth, or between the armor or sheath of adja-
cent cables.
Conductor Stress. Conductor stress is the maximum
stress that the cable being pulled can withstand. The Crosslinked Polyethylene (XLP, XLPE, X-LINK PE). HV
allowable stress is a function of the conductor type. and control cable insulations (90ºC). Special appliance
wiring material (125ºC). Extruded polyethylene that is
Connector. At splicing locations where two cable sec- rendered thermosetting (rubberlike) by crosslinking
tions are joined, the copper or aluminum conductors are (“vulcanizing”, “knitting”) the PE molecules usually
placed in a tight-fitting sleeve that is compressed with chemically (heat activated catalyst) or for special,
hydraulically-actuated dies and often machined to give a smaller cables by electron beam irradiation.
G-3
Cryogenic System. Cooling or refrigeration system used ing electrostatic field of the polar (lossy) molecules of
to deliver subcooled and pressurized liquid nitrogen to the dielectric. In dc cables or dielectrics, the dielectric
superconducting cables. loss is simply the square of leakage current times the
leakage resistance.
Cryostat. Thermal insulating pipe used in a supercon-
ducting cable. The cryostat is double-walled, stainless Dielectric Loss Factor. For ac cables, the product of the
steel, corrugated, vacuum-jacketed pipe. A high-level dielectric constant times the dissipation factor; common
vacuum is pulled between the two walls of the pipe with values for HV paper cables are dielectric constant of 3.5
multilayer thin film metallized mylar insulation (MLI) and dissipation factor of 0.002, for a dielectric loss fac-
wrapped around the inner pipe. The MLI and vacuum tor of 0.007. Dielectric loss factor is often used as a fig-
together provide thermal insulation on the order of 1.0- ure of merit for comparing the losses of alternative cable
1.5 W/m heat leak at 77 K operating temperature. insulations.
CSA. Canadian Standards Association. Dielectric Polarization. Appearance of surface charges

of opposite sign on the opposing faces of a dielectric
DDB (Dodecylbenzene). Impregnant of the alkylbenzene due to displacement of positive and negative charges
type commonly used as the impregnant for SCFF contained in its molecular structure. The displacement
cables. is proportional to the applied electric field.
Degree of Soil Saturation . Fraction of pore space Dielectric Strength. Potential gradient at which electric
between soil particles that contain water. failure or breakdown occurs. To obtain the true dielec-
tric strength, the actual maximum gradient must be con-
Deionized Water. In the paper-making process for cellu- sidered, or the test piece and electrodes must be
losic insulation, the final rinses of the paper pulp are designed so that uniform gradient is obtained. The
often done with deionized water, to reduce the amount value obtained for the dielectric strength in practical
of conductive material left in the paper after drying. tests will usually depend on the thickness of the material
This is especially important for dc operation, where and on the method and conditions of the test.
paper losses are primarily due to conduction currents.
Dielectric Strength, Intrinsic . Ultimate dielectric
Dielectric Absorption. When dc voltage is applied to a strength of a material in a completely uniform field, in
dielectric, the charge continues to flow with a decreasing the absence of any external effects such as voids, con-
amount over a long time. When the dielectric is shorted, taminants, etc.
the charge is not dissipated instantaneously; the dielec-
tric material may accumulate a substantial charge at its Direct Burial. In rural and suburban areas, and occa-
electrodes. For high-voltage cables, especially extruded sionally in urban areas, it is possible to install self-con-
cables with their larger volume resistivity, care should be tained oil-filled or extruded dielectric cables directly in
taken to have the conductor and shield shorted for the ground, without having to install ducts and pull the
many hours before handling the cable. cables through the duct. A controlled backfill is usually
installed around the cables to optimize heat dissipation
Dielectric Constant. Physical quantity describing elec- and to prevent damage from rocks, etc. and care is taken
tric field effects. That property of a dielectric material to maintain proper spacing to reduce mutual heating
that determines how much electrostatic energy can be effects. Because of the more direct heat transfer to the
stored per unit of volume when unit voltage is applied. earth, directly buried cables have a higher current rating
It is determined by the ability of a material to polarize than the same cable installed in ducts. For important
in response to the field, and thereby reduce the field circuits, the cables are protected against damage by con-
inside the material. Also called “relative permittivity.” crete slabs or troughs.
An older, but seldom used, term for dielectric constant
is “specific inductive capacity (SIC).” Direction of Lay. Direction of lay of strand (or insulated
wire) is determined by the direction of the rotation of
Dielectric Loss. Heating in and of a dielectric caused by the top strand (or insulated wire) when looking away
radial flow of capacitive charging current through the from the viewer along the axis of the conductor (or
resistance represented by the loss component of the cable). For example, a right-hand lay is indicated when
shunt admittance. Physically the phenomenon consists the top strand (or insulated wire) goes away from and to
of frictional heat generated by rotation in the alternat-
G-4
the right of the viewer looking along the axis of the con- Feed Joint. A joint or splice used in SCFF cable systems
ductor (or cable)—i.e., clockwise. to enable fluid to be fed from the feed tanks to the cable
central duct without presenting a hydraulic barrier
Dodecylbenzene (DDB). An impregnant used in SCFF between adjacent cable sections.
Cables. See also Alkylbenzene
Feed Tanks. Static pressurized fluid reservoirs con-
Duct Bank. Arrangement of conduit providing one or nected to SCFF cable systems designed to accommo-
more continuous ducts between two points. In urban date the thermal expansion and contraction of the
areas where it is not possible to have long street open- impregnating fluid.
ings, or in any area where there may be a need to
remove/reinstall cables without disturbing the street sur- Ferromagnetic. Materials that have magnetic permeabil-
face, ducts are installed and cables are later pulled into ities that are much greater than the permeability of free
the ducts. Use of ducts generally results in a lower cable space. They are also characterized by nonlinear mag-
rating and the distance between splices is shorter netic properties.
because of the relatively high coefficient of friction
between the cable sheath and the duct. Filled Strand. “Glop” wiped into all interstices of the
strand, which when cured, forms a rubbery mass serving
EHV. Extra-High-Voltage. to block the flow of liquids or gases through and along
the strand. Used in some extruded power cables, pump
Emanuelli Cables. Design of HV and EHV transmission cables, shipboard cables, and cables in hazardous location.
cable utilizing three paper-insulated oil-impregnated
cables, each with its own factory-applied metal sheath Fluid-Filled Cables (SCFF, LPFF, FF). Design of HV and
that contains the pressure medium, a low-viscosity oil at EHV transmission cable utilizing three paper-insulated
1/2 to 15 atmospheres, fed to the cable insulation via a fluid-impregnated cables, each with its own factory-
duct in the center of the conductors. See also Fluid- applied metal sheath that contains the pressure medium,
Filled Cable and Self-Contained Fluid-Filled Cable. a low-viscosity impregnant at low pressure, fed to the
cable insulation via a duct in the center of the conduc-
Ethylene Propylene Rubber (EPR, EPM, EPDM). Syn- tors. See also Emanuelli Cable and Self-Contained Fluid-
thetic HV insulation (90ºC) with excellent weathering Filled Cable.
and ozone resistance. Control cable insulation (105ºC)
and jacket. Forced-Cooling. Means to artificially remove heat from
the cable system. For HPOF cables, the filling oil is peri-
Extruded Cable Jacket. Cable jacket that substantially odically removed from the pipe, cooled (oil-to-air or oil-
encapsulates the metallic shield. It is also known as an to water exchangers, or refrigeration exchangers), and
encapsulating cable jacket. returned to the cable pipe at a separate location. Self-
contained cables can have external (lateral) cooling (a
Extruded Cables. Cables having a rubber or polymeric water pipe laid in the trench with the cables), integral
material that is extruded onto the conductor to give a cooling (the cables installed in a water pipe), or internal
monolithic insulation. Insulation material is commonly (core) cooling (the oil in the cable core periodically
crosslinked polyethylene, non-crosslinked polyethylene, removed, cooled, and reintroduced). Extruded cables
or ethylene propylene rubber. can have external (lateral) or integral cooling.
Feed-in Section. Short section of duct, either a straight or Freezing. Technique for temporary hydraulic isolation
bend section, that allows the installer to reach the cable for repairs or to facilitate installation, where the cable's
entrance to the duct run. If a feed in tube is used, the impregnate and its filling oil are frozen using various
feed-in section could be used to address the tension drop liquified gases or low temperature fluids to form a solid
in the feed-in tube. Alternatively, the reel back tension plug.
could be increased and the feed-in section eliminated.
Gage. System for specifying wire size. The American
Feed-in Tube. Flexible tube that is placed through the Wire Gage (AWG), also known as Brown & Sharpe
manhole cover and is attached to the duct. Cable is Gage, is used for copper. An increase of three gage num-
pulled through the tube, which protects the cable. bers (e.g., from #1 to #4) halves area and weight and
doubles dc resistance.
G-5
Gas-Insulated Transmission Line (GITL). Transmission 138 kV and below. Also known as Oil-O-Static, POF,
system consisting of a large-diameter conductor, typi- and HPGF cable.
cally tubular aluminum, centered via epoxy or plastic
spacers in a larger-diameter shield, also typically alumi- High-Temperature Superconductors (HTS). Those con-
num. Sulfur hexafluoride (SF6) gas, either by itself or in ductors that exhibit superconductivity at or above the
mixture with nitrogen and operating at a pressure of 2-6 temperature of liquid nitrogen, or -196°C (77 K).
atmospheres, is the insulating fluid. Some dc develop-
ment has been done as part of a compact substation. Horizontal Bend. In the field, some portions of a hori-
This system is also known as SF 6 bus, CGI, and gas- zontal bend actually have both horizontal and vertical
spacer system. See also Compressed Gas-Insulated components. When this occurs, the section is treated as
Transmission Line. a straight pull, and any elevation change is made up in
the section preceding or following the bend using a
Geometric Mean Radius (GMR). The geometric mean slope up or slope down pull.
radius of a conductor is the radius of a fictitious thin,
hollow conductor that produces the same magnetic flux Horizontal Directional Drilling (HDD). Installation of
as the actual conductor for the same current flowing in pipes and utilities beneath obstacles such as waterways,
the two conductors. railroad tracks, or roadways. It is a trenchless technol-
ogy that provides numerous benefits over a traditional
GITL. See Gas-Insulated Transmission Line. open-cut trenching.
GMR. See Geometric Mean Radius. HPGF. See High-Pressure Gas-Filled Cable.
Gradient, Electrical. Electrical stress in an insulating HPOF. See High-Pressure Oil-Filled Cable.
material, determined by dividing the applied voltage by
the electrode spacing, with suitable geometric factors for HTS. See High-Temperature Superconductors.
electrode geometries. Where the insulating material is
intrinsically or intentionally made nonhomogeneous (as HV. High Voltage.
for stress-grading purposes), the stress distribution must
be calculated for the actual dielectric constants (ac HVDC Integral Return Cable (HVDC-IRC). HVDC cable
cables) or using the alpha and beta coefficients (dc systems require a current return path. This can be pro-
cables). vided in at least three different ways, namely: a
ground/sea return using an anode/cathode electrode sys-
Harmonic Resonance. Resonance of an electrical circuit tem, a separate return cable, or a metallic return con-
at frequencies that are integer multiples of the power ductor integrated into the HVDC cable. To date, the
system frequency (i.e., 60, 120, 180, 240…Hz for a Moyle HVDC interconnection between Scotland and
60 Hz power system). Northern Ireland is the only project to employ HVDC
integrated return cables.
High-Pressure Gas-Filled (HPGF) Cable. Design of HV
and EHV transmission cable utilizing three paper- ICEA. Insulated Cable Engineers Association.
insulated, mass-impregnated cables without individual
metal sheaths, installed in a common steel pipe, which Impedance Bonding . Bonding method consisting of
contains the pressure medium, usually oil, at nominally grounding the cable sheaths through an impedance.
15 atmospheres; occasionally nitrogen gas at voltage
138 kV and below. Also known as HPOF, Oil-O-Static, Impermeability. Also known as Air Impermeability. A
and POF cable. See also High-Pressure Oil-filled Cable, measure of the resistance of insulating paper to the flow
Oil-O-Static, and Pressurized Oil-Filled Cable. of air through its fiber structure and thus to the amount
the pulp was beaten in the manufacturing process.
High-Pressure Oil-Filled (HPOF) Cable. Design of HV
and EHV transmission cable utilizing three paper- Impregnant (Impregnating Fluid). Fluid used to saturate
insulated, mass-impregnated cables without individual the paper and fill all the spaces between taped insulation
metal sheaths, installed in a common steel pipe, which layers and between conductor strands. A suitable sub-
contains the pressure medium, usually oil, at nominally stance is of low viscosity at the impregnating tempera-
15 atmospheres; occasionally nitrogen gas at voltage ture and fills the spaces between the fibers. To be
considered suitable, the impregnating substance must
G-6
have good insulating properties, gas-absorbing proper- (remote operated vehicles) are preferred burial equip-
ties, and resistance to oxidation. ment where seabed conditions permit. The working
principle of the Jet Plow is to fluidize the seabed materi-
Impregnation Coefficient. Measure of the degree of als in a narrow path and to a predetermined depth
completeness of impregnation of an SCFF cable. (down to ~ 8 ft). The Jet Plow is towed by a support or
the cable-laying vessel itself, and cable laying and burial
Inductance. Property of an electric circuit by virtue of take place simultaneously.
which a varying current induces an electromotive force
in the circuit or in a neighboring circuit. Kraft Paper. See Paper (Kraft Paper).
Insulation. Material applied over conductors for electri- Laminated Polypropylene (LPP). Also known as PPL,
cal isolation between conductors. The part that is relied PPP, and PPLP. A taped insulation material used in
upon to insulate the conductor from other conductors SCFF and HPFF cables and manufactured by extrud-
or conducting parts or from ground. ing a thin layer of polypropylene between two layers of
paper. It has significantly improved electrical properties
Insulation Coordination. Selection of insulation strength compared to kraft paper.
consistent with expected overvoltages to obtain an
acceptable risk of failure. Lapping Head. Piece of cable manufacturing machinery
used to apply tapes in a helical manner to the cable.
Insulation Resistance. Ratio of applied dc voltage of an
insulating material to the total current flowing though Lay. Distance between successive points where the same
the material for prescribed conditions of test. strand (or insulated conductor) when twisted with one
or more strands (or insulated conductors) presents itself
Interlock Armoring. A 0.020-in. or 0.030-in. thick strip in the same position. The helical arrangements formed
of mild steel or aluminum is preformed and spiralled by twisting together the individual elements of a cable.
around the cable core, each turn being “interlocked”
with the preceding turn ala “BX” wiring. Laying Up. Process of bringing the individual cores of a
multicore cable together without imparting a twist to
Ionization Factor Test. Test performed on laminar the cores.
dielectric cable to detect ionization (partial discharges)
in the cable high voltage insulation. The ionization fac- Leach. To dissolve out by the action of percolating
tor test is the difference in insulation dissipation factor liquid.
values at two different electrical stress levels.
Leakage Surface. Applied to longitudinal stress as on
Jack and Bore. Method of simultaneously jacking cas- joints or potheads—the total length of the insulating
ing through the earth while removing the spoil inside the surface between the two electrodes.
encasement by a rotating auger. Also known as “auger
boring.” Link-Box. Waterproof enclosure housing the bus bars
and connection hardware for interconnecting the cable
Jacket. Self-contained fluid-filled and extruded cables sheaths in a cross-bonded scheme, together with the
often have a thermoplastic or thermosetting jacket surge protection devices when used to control the volt-
applied over the lead or aluminum cable sheath, or over age from sheath to earth across the protective jacket.
the cable armoring. This jacket provides additional
mechanical protection, especially during the cable-pull- Load Factor. In cable rating or loss evaluation; the ratio
ing operations. The jacket also electrically isolates the of the sum of the hourly currents to 24 times the peak
sheath from the earth, permitting special sheath bonding. hourly current. Alternatively, the daily energy usage
divided by 24 times the daily peak load. (Annual or sea-
Jam Ratio. Calculated constant that warns if a cable is sonal load factors can be calculated as desired for evalu-
likely to become stuck or jammed. ating equipment utilization.)
Jet Plow. Submarine cables are generally buried to pro- Loss Factor. As used in cable rating, the ratio of the sum
vide protection against damage due to third-party of squares of the hourly current readings divided by 24
marine activities. Jet Plows and Water-Jetting ROVs times the square of the maximum hourly reading. An
G-7
empirical relationship to the load factor at transmission each other in order to reduce proximity and skin effects
voltages has been found to be: in alternating current systems.
Loss factor = 0.3 Load Factor + 0.7 (Load Factor)2
Minimum Bending Radius. Minimum radius (in feet) in
(Annual or seasonal loss factors can be calculated as which a cable can be bent without structure damage.
desired when evaluating equipment losses.)
Minimum Bending Radius Pulling Tension. Tightest
LPP. See Laminated Polypropylene. allowable radius that a cable can be pulled around with-
out damage.
Lubricant. Slippery material used to reduce the coeffi-
cient of friction between cable and conduit. Moisture Content. Soil moisture has a strong influence
on the soil's thermal resistivity. The moisture collects at
Maintenance Test. Field test made during the operating the contact points between soil grains, aiding in conduc-
life of a cable system. It is intended to detect deterioration from soil grain to soil grain resulting in overall
tion of the system and to check the serviceability so that lower thermal resistivities. Due to the loss of water at
suitable maintenance procedures can be initiated. the grain to grain interface, the thermal resistivity of a
dry soil can be three to six times higher than the thermal
Mandrel. Device that has a diameter slightly smaller resistivity of a wet soil. The moisture content of the soil
than the duct. This device is pulled through the duct to is calculated according to the following relationship:
ensure that the run is not blocked. Percent Moisture = ((Wet soil weight - Dry soil
weight) / Dry Soil Weight) ∗100
Mass Impregnation. Process used in the manufacture of
paper and LPP insulated cables where, after the cable The moisture content is said to be calculated on a dry
has been dried, the entire drying and impregnating ves- basis when this equation is used because the soil weight
sel is filled with impregnating fluid. The cable is then in the denominator does not include moisture.
metal-sheathed directly from the vessel via a tube filled
with impregnant. Mouse . Any device attached to a small rope that
restricts air flow through a duct. A compressor is used
Maximum Bending Radius. Highest tension that can be to blow the mouse and the rope through the duct.
placed on a cable without risk of damage.
Multipoint Bonding. Bonding method consisting of
Maximum Pulling Tension. Maximum tension (in grounding of the cable sheaths at multiple locations,
pounds) that can be placed on the cable(s) at the bend typically at all terminations and splices.
exit.
Nanotechnology. Science of engineering that deals with
Maximum Sidewall Bearing Pressure. Maximum pres- things smaller than 100 nanometers.
sure (in pounds/foot) that the side of a cable can with-
stand as it is being pulled through a bend. NEMA. National Electrical Manufacturers Association.
Maximum Stress. In ac cables and uniform temperature Normal Joint. Splice or joint not incorporating any fea-
dc cables, the highest dielectric stress always occurs at tures to provide hydraulic isolation between adjacent
the conductor screen surface. In dc cables, the location cable sections.
of maximum stress varies with temperature and stress
characteristics. Oil. Fluid(s) used to impregnate the insulating tapes of
pipe-type cables, and to maintain a positive pressure in
MCOV (Maximum Continuous Operating Voltage). Max- the cable system to increase dielectric strength and sup-
imum designated root-mean-square (rms) value of press ionization. Pipe-type cables usually contain a min-
power-frequency voltage that may be applied continu- eral oil or a polybutene oil at pressures of 15-30
ously between the terminals of a surge arrester. atmospheres. Some of the synthetic and semi-synthetic
insulations under development may require a synthetic
Milliken Conductor. Conductor construction where the oil such as a silicone oil.
total cross sectional area comprises a number, typically
four or six, of smaller segments lightly insulated from Oil-O-Static Cable. Design of HV and EHV transmis-
sion cable utilizing three paper-insulated, mass-impreg-
G-8
nated cables without individual metal sheaths, installed Pipe-Type Cable. Design of HV and EHV transmission
in a common steel pipe, which contains the pressure cable utilizing three paper-insulated, mass-impregnated
medium, usually oil, at nominally 15 atmospheres; occa- cables without individual metal sheaths, installed in a
sionally nitrogen gas at voltage 138 kV and below. Also common steel pipe, which contains the pressure
known as HPOF, POF, and HPGF cable. See also High- medium, usually oil, at nominally 15 atmospheres; occa-
Pressure Oil-Filled Cable, High-Pressure Gas-Filled sionally nitrogen gas at voltage 138 kV and below. Also
Cable, and Pressurized Oil-Filled Cable. known as HPOF, Oil-O-Static, POF, and HPGF cable.
Oversheath. See Jacket. Polarity Reversal. Switching operation to change the

polarity of the voltage applied to a conductor—e.g.,
Paper (Kraft Paper). Insulating material in almost uni- from positive to negative. Of concern to insulated dc
versal use for pipe-type and self-contained cable. cable systems because of the dielectric absorption of the
dielectric. See also Dielectric Absorption.
Paper Lead Cable. Paper lead cable is an oil-impreg-
nated paper-insulated cable with a copper conductor Polarization Index. Method of evaluating dielectric
and a lead covering. absorption of an insulating material by determining its
insulation resistance after one minute and comparing
Paper-Polypropylene-Paper (PPP). Trade name for that value to its insulation resistance after 10 minutes.
Laminated Polypropylene (LPP). The ratio is the Polarization Index.
Parkway Taping. Flat metal tape applied with an open Polybutene. Synthetic oil commonly used for pipe-type
lap over the cable core or inner jacket. A duplicate sec- cables, in a viscosity as an impregnating oil, and in lower
ond tape is applied centered over the opening of the first viscosities as pipe-filling oils. Polybutene is obtained by
tape. polymerizing the isobutylene gases that are produced in
petroleum refining.
Partial Discharge (PD). Localized electrical discharge
that only partially bridges the insulation between con- Polyethylene. Extrudable polymeric insulation charac-
ductors and that may or may not occur adjacent to a terized by very low dielectric loss, very high intrinsic
conductor. Partial discharges are a consequence of local dielectric strength but high vulnerability to degradation
breakdown either as a result of (a) an electric field by corona discharge. Due to its thermoplasticity and rel-
enhancement within or on the surface of the insulation, atively low fusion temperature, the short-time emergency
or (b) a region of low breakdown field. PDs appear as loading capability of polyethylene cables is limited.
individual events of very short duration, are always
accompanied by emissions of light, sound, and heat, as Polyvinyl Chloride (PVC). Extrudable thermoplastic
well as electromagnetic pulses, and often result in chem- polymeric material compounded from base polymer,
ical reactions. plasticizer, stabilizer, lubricant, filler, and color. Used as
an extruded jacket (oversheath) on supertension SCFF
Particle Size Analysis. Particle size analysis (or sieve and XLPE cables and as insulation on LV cables.
analysis) shows the grain size gradation of soils. This
test can be performed on all soils and is used to deter- Pothead. Device providing both the required electrical
mine if the soil is predominantly sand, silt, or clay. The stress grading and insulating hydraulic seal to atmo-
particles are segregated by running the soil through sphere or other medium (e.g., SF6) where the cable ter-
sieves as outlined in ASTM Standard D 422. Soils with minates. Electrical stress grading is effected by flared
good thermal properties are generally characterized as build-ups of insulation-paper rolls or molded epoxy ele-
having a wide gradation of particle sizes. The thermal ments typically, by distributed capacitors or combina-
resistivity of a soil is decreased as the number of contact tions thereof. The internal pressure is contained, and
points is increased. Consequently, evaluation of the external creepage and weathershed (for in- air termina-
sieve analysis of a soil gives an indication of the poten- tors) provided by petti-coated porcelain insulators; in
tial of a soil to be a good thermal conductor. certain designs with an internal pressure-support tube
of ceramic or fiberglass. For SF6 or oil-immersed termi-
PD. See Partial Discharge. nations, various designs exist using cast epoxy or
ceramic seals, usually of a smooth conical shape. A
Permeability. Ability of a liquid to flow through pores device that seals the end of a cable and provides insu-
or openings.
G-9
lated egress for the conductor or conductors. See also conductor. In effect, the current density is increased in
Terminator and Sealing-End. sections of conductors nearer each other and decreased
in the sections farthest from each other.
PPP. See Paper-Polypropylene-Paper.
Pulling Eye Stress. Pulling eye stress is the stress on the
Pre-qualification Test. Series of long-term electrical tests cable when pulling eyes are used.
required for approval to supply a prospective design and
manufacturing technique of EHV XLPE cable and Pulling Eyes. Device that is attached to cable conductor
accessories. for subsequent cable pulling.
Pressurized Oil-Filled (Type POF) Cable. Design of HV Pulling Grip Tension. Tension on the cable when pulling
and EHV transmission cable utilizing three paper- grips are used.
insulated, mass-impregnated cables without individual
metal sheaths, installed in a common steel pipe, which Pulling Grips. Woven steel basket grips that slide over a
contains the pressure medium, usually oil, at nominally cable for subsequent cable pulling.
15 atmospheres; occasionally nitrogen gas at voltage
138 kV and below. Also known as HPOF, Oil-O-Static, Pulling Section Dip Depth. Distance in feet that a verti-
and HPGF cable. See also High-Pressure Oil-filled cal dip bends below the horizontal. Thus, it is only rele-
Cable, High-Pressure Gas-Filled Cable, and Oil-O-Static vant to vertical dip section types.
Cable.
Pulling Section Length. Length in feet of a pulled sec-
Pressurizing Plant. Unit consisting of a pressurizing tion.
pump and oil storage tank (typical capacities 4000-
200,000 liters) with controls and alarms. Its purpose is Pulling Section Offset Angle. Angle from the vertical, in
to maintain the designed pressure in the cable system, degrees, that a section travels as it is pulled through a
and to provide a means for accommodating oil volume bend.
changes due to temperature variations in the cable. All
pipe-type oil-filled circuits have such a plant, and a few Pulling Section Radius. Centerline radius of curvature,
long self-contained circuits have pressurizing plants in feet, of a section that is pulled through a bend.
instead of reservoirs.
Pulling Section Total Angle. Angle in degrees that a sec-
Proctor Curve. The Proctor Curve provides valuable tion travels as it is pulled through a bend.
data in predicting the thermal behavior of soil. The
Proctor Curve is a plot of the maximum dry density that Pulling Tension. Tension on a cable at the cable exit
can be obtained as a function of moisture content. For which results from pulling the cable through a run. You
most soils, the maximum obtainable density increases as might want to add jumps to the pulling grips and pull-
the moisture content increases until a maximum dry ing eyes here since these are two primary methods used
density is achieved. At higher moisture contents, the to pull the cables through a run.
water begins to displace soil particles resulting in lower
Proctor dry densities. The Proctor Curve is useful in Pumping Plant. Unit consisting of a pressurizing pump
selecting the moisture content that the maximum Proc- and oil storage tank (typical capacities 4000-200,000
tor density is obtained. Compaction of backfill soils at liters) with controls and alarms. Its purpose is to main-
the moisture content that corresponds to the maximum tain the designed pressure in the cable system, and to
dry density is desirable. It provides conditions that max- provide a means for accommodating oil volume changes
imize heat transfer in soils in the vicinity of under- due to temperature variations in the cable. All pipe-type
ground power cables. Field densities of 95% maximum oil-filled circuits have such a plant, and a few long self-
Proctor density are commonly specified for backfills. It contained circuits have pressurizing plants instead of
is likely that densities of 85 to 90% of the Proctor den- reservoirs.
sity will be obtained unless strict quality control is
maintained during compaction. PVC. See Polyvinyl Chloride.
Proximity Effect. Phenomenon of nonuniform current Qualification Test. Type tests or qualification tests are
distribution over the cross section of a conductor caused performed to demonstrate the adequacy of designs,
by the time variation of the current in a neighboring manufacturing, and materials of a cable system or com-
G-10
ponent to achieve the desired performance characteris- Rod. A 5/16-in. diameter rod of copper drawn from wire
tics. The term “qualification test” is commonly used in bars. Used in making solid conductors and individual
Association of Edison Illuminating Companies (AEIC) strands.
and Insulated Cable Engineers Association (ICEA)
cable specifications and standards. IEC cable standards Rope. Rope used to pull power cables. Typically made of
contain similar tests, but they are called “type tests.” steel or of polymers.
Quality Factor. For an inductor, the ratio of the reac- Round Wire Armoring. High tensile strength round steel
tance in ohms divided by the resistance in ohms. wires are spiralled, close spaced, around the cable for
mechanical protection or strength. Steel wire sizes vary
Reel Back Tension. Tension required to unspool cable from #24 through #8 solid.
off a reel.
Routine Test. Series of electrical and mechanical tests
Registration (of tapes). Measurement of the percentage carried out on every reel of completed cable, or acces-
of the width of one tape that overlaps the tape beneath sory, before shipping.
it. Thus 35/65 registration has 35% of the tape lying in
front of and 65% of the tape lying behind the trailing Saturated (1). To be full of moisture, thoroughly wet.
edge of the tape beneath it.
Saturated (2). Alternative term for impregnated.
Reinforcement. Layer or layers of metal tapes wound
with a short lay around a SCFF cable lead alloy sheath SCFF. See Self-Contained Fluid-Filled Cable.
to contain the internal fluid pressure.
Screen. Conducting or semiconducting layer(s) over or
Relative Permittivity. Physical quantity describing elec- under the cable dielectric used to define and confine the
tric field effects. That property of a dielectric material electrostatic field between the energized and grounded
that determines how much electrostatic energy can be earthed) electrodes, in a completely radial pattern. The
stored per unit of volume when unit voltage is applied. screening layers also serve to smooth out irregularities
It is determined by the ability of a material to polarize in conductors and are designed with thermal character-
in response to the field, and thereby reduce the field istics to obviate void formation between materials of
inside the material. Also called “dielectric constant.” An varying thermal expansion coefficients, viz. 13, conduc-
older, but seldom used, term for relative permittivity is tors and insulation.
“specific inductive capacity (SIC).”
Screening. Material separated out by passage through
Reservoir. Tank of typical capacity 100 liters, used to or retention on a screen.
maintain oil pressure and allow for oil expansion and
contraction on self-contained circuits. The reservoirs are Sealing End. Device providing both the required electri-
usually placed in splicing manholes, at spacings from cal stress grading and insulating hydraulic seal to atmo-
1000 m to several thousand meters, depending upon ele- sphere or other medium (e.g., SF 6 ) where the cable
vation changes. terminates. Electrical stress grading is effected by flared
build-ups of insulation-paper rolls or molded epoxy ele-
Residual Gas Pressure. Measure of the amount of air ments typically, by distributed capacitors or combina-
or other gas dissolved in the impregnating fluid of an tions thereof. The internal pressure is contained, and
SCFF cable. external creepage and weathershed (for in- air termina-
tors) provided by petti-coated porcelain insulators; in
Resistivity. Volume resistivity in ohm-cm of a dielectric certain designs with an internal pressure-support tube
is the ratio of the dc potential (volts/cm paralleling the of ceramic or fiberglass. For SF6 or oil immersed termi-
flow of current in the material) to the current density nations, various designs exist using cast epoxy or
(amperes/cm) after a prescribed time of voltage appli- ceramic seals, usually of a smooth conical shape.
cation.
Sector Conductor. Stranded conductor whose cross sec-
Rho. Characteristic of insulations and thermal environ- tion is approximately the shape of a sector of a circle. A
ments that determine temperature rise as a function of multi-conductor insulated cable with sector conductors
watts of heat dissipated. has a smaller diameter than the corresponding cable
with round conductors.
G-11
Segmental Conductor. Stranded conductor consisting Such tapes are commonly used for conductor shielding
of three or more stranded conducting elements, each and in conjunction with metallic shielding over the insu-
element having approximately the shape of a circle, lation.
assembled to give a substantially circular cross section.
The sectors are usually lightly insulated from each other Semi-Stop Joint. A splice or joint in HPFF cable with
and, in service, are connected in parallel. The light insu- features to effect only a partial or temporary hydraulic
lation is in order to reduce proximity and skin effects in isolation between adjacent cable sections usually for
alternating current systems. maintenance or leak location or rate reduction pur-
poses. A joint designed to restrict movement of the
Self-Contained Fluid-Filled (SCFF) Cable. Design of HV dielectric fluid between cables being joined.
and EHV transmission cable utilizing three paper or
LPP insulated fluid-impregnated cables, each with its Sheath. Metal tube, usually lead, lead alloy, or alumi-
own factory-applied metal sheath that contains a low num formed, almost universally by extrusion, over the
viscosity fluid at low pressure, fed to the cable insulation cable core as a hermetic seal and pressure containing
via a duct in the center of the conductors. Three core vessel. The overall protective covering for the insulated
types with all phases inside one metal sheath are avail- cable.
able up to a certain voltage level. Also known as LPFF,
FF, and Emanuelli cable. (In North America, it is more Shield. Conducting or semi-conducting layer(s) over or
common to use the term “Self-Contained Fluid-Filled under the cable dielectric used to define and confine the
(SCFF) Cable”). A pressure cable in which the con- electrostatic field between the energized and grounded
tainer for the pressure medium is an impervious flexible (earthed) electrodes, in a completely radial pattern. The
metal sheath, reinforced if necessary, that is factory screening layers also serve to smooth out irregularities
assembled with the cable core. A self-contained pressure in conductors and are designed with thermal character-
cable in which the pressure medium is low-viscosity istics to obviate void formation between materials of
fluid having access to the insulation. varying thermal expansion coefficients, viz. 13, conduc-
tors and insulation.
Self-Contained Mass-Impregnated (SCMI) Cable. In
common with self-contained fluid-filled cables, SCMI Shielding Factor. Ratio of the magnetic field magnitude
cables have insulation consisting of paper tapes impreg- (rms value) at a given point when a magnetic shield is
nated with a dielectric fluid. In the case of SCFF cables, present to the magnetic field magnitude (rms value) at
a low-viscosity synthetic fluid—generally an alkylben- the same point when the shield is absent.
zene—is maintained under pressure by a pressurizing
system. SCMI cables, on the other hand, employ a high- Sidewall Bearing Pressure (SWBP). Radial pressure
viscosity fluid that does not require a pressurizing sys- experienced by the cable as it is pulled through a curved
tem. SCMI cables can in theory be used for ac transmis- section. The pressure is caused by the tension and
sion at voltages up to 69 kV, though their major field of weight of the cable, which tends to force it against the
application is in HVDC transmission, where they are conduit wall. The parameters that influence the SWBP
used at voltages up to and including 500 kVdc. The are cable tension and the inside radius of curvature of
absence of the pressurizing system requirement means the bend.
that there is no length limitation as in the case of SCFF
cables. Sieve. Device with meshes or perforations allowing finer
particles of a mixture to pass, while retaining coarser
Semiconducting Layer. A conducting or semi-conduct- materials.
ing layer(s) over or under the cable dielectric used to
define and confine the electrostatic field between the Single-point Bonding . Bonding method consisting of
energized and grounded (earthed) electrodes, in a com- grounding the cable sheaths at only one end along a seg-
pletely radial pattern. The screening layers also serve to ment of cable.
smooth out irregularities in conductors and are
designed with thermal characteristics to obviate void Skid Wires. Nonmagnetic metal wires, usually D-shaped
formation between materials of varying thermal expan- applied helically over the single core of pipe-type cables
sion coefficients, viz. 13, conductors and insulation. A to serve as protective runners to the core during pulling
tape or such resistance that, when applied between two into the steel pipe. Wire or wires, usually D-shaped,
elements of a cable, the adjacent surfaces of the two ele- applied open spiral with curved side outward with a
ments will maintain substantially the same potential. suitable spacing between turns (lay) over the outside
G-12
surface of the cable. Its purpose is to facilitate cable Spacer-Gas Cable. Transmission system consisting of a
pulling and to provide mechanical protection during large-diameter conductor, typically tubular aluminum,
installation. centered via epoxy or plastic spacers in a larger-diame-
ter shield, also typically aluminum. Sulfur hexafluoride
Skin Effect. Phenomenon of charge of current distribu- (SF6) gas, either by itself or in mixture with nitrogen and
tion over the cross section of a conductor with conse- operating at a pressure of 2-6 atmospheres, is the insu-
quent increase in ac resistance caused by the alternating lating fluid. Some dc development has been done as part
flux created by the ac current carrying conductor itself. of a compact substation. This system is also known as
The tendency of an alternating current to concentrate in SF6 bus, GITL, and CGI. See also Compressed Gas-
the areas of lowest impedance. Insulated Transmission Line and Gas-Insulated Trans-
mission Line.
Slump. To fall or sink.
Specific Inductive Capacity (SIC). An older, but seldom
Slurry. Watery mixture of insoluble matter. used, term for relative permittivity or dielectric con-
stant.
Soil Density. Density of a soil affects thermal stability.
As the dry density of a soil increases, the void space Spiral Duct. Component of an SCFF cable designed to
between soil particles decreases and the paths for vapor support other components of the cable and thus provide
outward flow are reduced. At increased dry densities, a flow path for impregnating fluid along its length. It is
the total soil contact area is increased, resulting in a normally manufactured from metal strip wound around
lower soil thermal resistivity. At an applied heat input, into an open spiral.
the increased conduction in a dense soil results in lower
temperature gradients across the soil. The driving Splice. Process and hardware components used to make
potential for the movement of water vapor is reduced. the electrical and hydraulic connections between adja-
Therefore, at a given soil moisture content, the soil is cent lengths of cable. It entails connecting the conduc-
thermally stable as the dry density of a soil is increased. tors usually by compression or swaged fittings for
copper conductors, welding for aluminum, restoration
Soil Thermal Stability. A cable whose ampacity was cal- of the insulation almost universally by hand taping, and
culated based on the thermal resistivity of a moist soil restoration of the hydraulic pressure enclosure pipe or
will reach a higher temperature if moisture is driven sheath by welding, brazing or soldering. The physical
away from the cable by the thermal gradient in the soil. connection of two or more conductors to provide elec-
As the moisture content of the soil decreases further, the trical continuity.
thermal resistivity of the soil increases, resulting in
higher cable temperatures compounding the problem of Stop Joint. Splice or joint with features to effect a com-
overheating even further. This type of behavior is called plete hydraulic isolation between adjacent cable sec-
“thermal runaway.” Soil under the existing conditions tions, as may be required to limit pressures due to
would be thermally unstable. If a soil is thermally stable profile changes or fluid loss in the event of sheath dam-
under a given cable heat input, the movement of soil age. A joint that is designed to prevent any transfer of
moisture (in a vapor state) away from the cable surface dielectric fluid between the cables being joined. Many
is balanced by the movement of liquid moisture back designs will also allow feeding of the individual sections
toward the cable, due to capillary action. from static feed tanks.
Solid Dielectric Cable. Cables having a rubber or poly- Stop Mark . A circumferential discontinuity in the
meric material that is extruded onto the conductor to smooth extruded surface of a lead alloy sheath caused
give a monolithic insulation, which is highly resistant to by the need to raise the ram of the extrusion press to
the flow of electric current. Insulation material is com- recharge the extruder with metal.
monly cross-linked polyethylene, non-cross-linked poly-
ethylene, or ethylene propylene rubber. Stranding. The bringing together by twisting or laying
of smaller wires to form a single larger conductor. Used
Spaced. Installation configuration wherein the three to provide, flexibility and ease of handling.
cables are spaced (almost always laterally) from each
other in order to improve heat dissipation. Stress Cone. Shaped insulating device and related
shields to control the electrostatic stress concentration
at a termination.
G-13
Superconductor. Conductor used in specialized cable Thermal Backfill. Thermal backfills are required where
systems that when cooled to very low temperatures higher cable ampacities are expected for both transmis-
develops a near-zero conductor resistance. Practical sion and distribution cables. Backfills are used around
cable systems using superconductors operate at near-liq- pipe type cables, cables in duct, or direct buried cables.
uid nitrogen temperatures (77 kelvin, or about -200° C). Thermal backfill materials can be used at selected loca-
tions along a cable route where there is poor native soil.
Surge Arrester. Protective device for limiting surge volt- This will reduce the chances of developing local hot
ages on equipment by discharging or bypassing surge spots. If the native soil thermal resistivity is higher than
current; it limits the flow of power follow current to desired or if the soil is considered to be thermally unsta-
ground, and is capable of repeating these functions as ble or the soil is unsuitable for compaction, a thermal
specified. backfill can be placed along the entire length of the
cable. The overall earth thermal resistivity is affected by
Surging. Resonant phenomenon where tensions fluctu- the native soil thermal properties and by the properties
ate rapidly during pulling. of the thermal backfill. The thermal resistivity of the
backfill has a stronger influence on the heat dissipating
Symmetrical Components. Analytical method for trans- capacity of the system because it is in the immediate
forming three-phase electrical circuits into three single- vicinity of the cables. The primary requirement of a
phase circuits (zero-sequence, positive-sequence, and thermal backfill is to dissipate heat away from the cable
negative-sequence circuits). This transformation is per- so the design temperature of the cable is not exceeded.
formed to simplify the solutions for currents and volt-
ages in the three-phase circuit. Thermal Conductivity. Inherent quality of a substance
that indicates its ability to conduct heat.
Symmetrical Component Impedance. Impedance of
power system components to zero-, positive- and nega- Thermal Probe. Thermal probes are used to measure the
tive-sequence circuit currents. thermal properties of soils. Thermal probes are con-
structed from stainless steel tubing. The diameters and
Synthetic Insulation Cable. Replacement of the cellu- lengths of the probes are selected depending on the
lose insulation on a pipe-type or self-contained cable desired application. An insulated heater element made
with synthetic or semi-synthetic (e.g., polypropylene/cel- of nichrome or constantan wire is run down and back
lulose laminate) insulation to reduce ac dielectric losses the active length of the thermal probe. These heater ele-
and allow application of a higher voltage stress. ments are preferred, because the resistance of these
materials does not change significantly with tempera-
Temporary Overvoltage (TOV). An oscillatory overvolt- ture. Thermocouples are slipped down the steel tubing
age, associated with switching or faults and/or nonlin- and placed at selected locations in the heated section of
earities of relatively long duration, which is undamped the probe. Copper constantan is a preferred thermocou-
or slightly damped. ple material.
Tension. Pulling force required to install cable in a duct. Thermal Resistivity (Rho). Characteristic of insulations
and thermal environments that determine temperature
Terminator. Device providing both the required electri- rise as a function of watts of heat dissipated.
cal stress grading and insulating hydraulic seal to atmo-
sphere or other medium (e.g., SF 6 ) where the cable Thermal Sands. Thermal sand contains a wide grada-
terminates. Electrical stress grading is effected by flared tion of particle sizes as well as small amounts of silt and
build-ups of insulation-paper rolls or molded epoxy ele- clay. This soil type will have lower liquid and vapor per-
ments typically, by distributed capacitors or combina- meabilities than a uniformly grained sandy soil. This is
tions thereof. The internal pressure is contained and because the smaller silt and clay particles tend to fill the
external creepage and weathershed (for in-air termina- voids between larger sized sand grains, reducing the
tors) provided by petti-coated porcelain insulators; in path for either vapor or liquid flow. Used for mixing
certain designs with an internal pressure-support tube with a cementing material to form concrete. Thermal
of ceramic or fiberglass. For SF6 or oil-immersed termi- sands are more likely to dry out than weak mix backfills.
nations, various designs exist using cast epoxy or Thermal sands are considered to be more thermally sta-
ceramic seals, usually of a smooth conical shape. See ble than most sands, because they have better heat-car-
also Pothead and Sealing-End. rying capacities and they contain silt and clay, which
reduces the permeability of the sand.
G-14
Thermistor. Electrical resistor made of semiconducting Vertical Dip. Most vertical bends are close to 90º and are
material whose resistance changes dramatically with used to bring cable into an enclosure or up a riser pole.
temperature. Therefore, the designer can treat the section as a bend
section having a section total angle of 90º and a section
Thermocouple. Special matched alloy conductors. Used offset angle of 0º. The possible bend sections are: con-
in temperature-measuring applications. A pair of dis- cave up, concave down, convex up, and convex down.
similar conductors so joined at two points that an elec-
tromotive force is developed by the thermoelectric Voltage Source Converter (VSC). The fundamental pro-
effects when the junctions are at different temperatures. cess that occurs in an HVDC transmission line is the
conversion of ac to dc at the transmitting end and of dc
Thermoplastic (PVC, Polyethylene, Nylon, etc.). Term to ac at the receiving end. The use of Voltage Source
used to describe the extruded materials which after pro- Converters is one of three ways of achieving this conver-
cessing, can be repeatedly made to flow under the appli- sion. Voltage Source Converters are unique in that the
cation of heat. Having the quality of softening when employ semiconductors that can be switched on and
heated above a certain temperature range and of return- switched off. Stated advantages for this type of conver-
ing to its original state when cooling below that range. sion are independent control of active and reactive
power and improved power quality.
Thermosetting (rubber, neoprene, hvpaion, etc.). Term
used to describe extrudable materials which are cross- Voltage Stress. Applied voltage divided by the electrode
linked (vulcanized chemical agents) using steam and spacing.
heat or radiation. After processing, the materials cannot
be made to flow under the application of heat. Warm Dielectric Superconducting Cable. Supercon-
ducting cable where the electrical insulation (dielectric)
Total Bend Angle. Sum of the total angles of all the sec- is located outside the cold space of the cable. The dielec-
tions in the pull. tric is typically an extruded polyethylene or EPR on the
outside of the cable cryostat.
Trefoil. Installation configuration wherein the three
cables (in ac circuits) are in a close equilateral triangle Water-Jetting ROV. Submarine cables are generally bur-
with sheaths or jackets touching. ied to provide protection against damage due to third-
party marine activities. Water-Jetting ROVs (remote
Triangular Configuration. Configuration where the cen- operated vehicles) and Jet Plows are preferred burial
ter cable rests on top of the other two cables. equipment where seabed conditions permit. The Water-
Jetting ROV combines seabed fluidizing and hydrody-
Triplexed Cables. Three cables that have been twisted namic transport of the fluidized material. The fluidizing
together and then placed on a single reel. process creates a trench into which the cable sinks
before the fluidized material is allowed to settle and
Tubed Cable Jacket. Cable jacket loosely applied over start the trench-filling process. In this process, the cable
the metallic shield. Usually a separator tape separates is laid on the seabed by the cable ship prior to com-
the jacket and the metallic shield. Also known as a mencement of burial by the Water-Jetting ROV.
sleeved cable jacket.
Weight Correction Factor. Calculated correction factor
Type Approval Test. Series of mechanical and electrical that adjusts the effective coefficient of friction for three
tests carried out by a prospective manufacturer to gain cable pulls.
approval to supply the design of cable and accessories to
be tested. Wire Coating. Applied over conductor to prevent insula-
tion from attacking copper. Eliminates difficulties in
Type Test. See Qualification Test. soldering, and prevents deterioration of copper at high
temperatures. Coating (Maximum Condition Rating
ULI. Underwriters Laboratories Incorporated. Degrees Celsius): Tin (200), Lead Alloy (200), Nickel
(250), and Silver (250).
Vacuum Sheathing. Process used in the manufacture of
SCFF cables whereby the dried insulated core is metal XLPE. See Crosslinked Polyethylene.
sheathed under vacuum before impregnation.
G-15
Index
All references are to section or subsection numbers, unless otherwise noted.
A Bipole cable, 3.4.1

Ac cable. See Alternating current cable. Bird caging, 15.3.4
Accessories. See Cable accessories. Bonding, 10.2.1, 12.2.9
Ac voltage withstand and step-breakdown tests, 14.2.1 Bonding leads, 8.4.1
AEIC specifications, 2.11, 5.7.1 Bonding scheme, 3.2.7
After-installation tests, 13.14, 14.1.1, 14.4 Cross-bonding, 10.2.1, 12.2.9
Extruded-dielectric cables, 14.4.1 Impedance bonding, 10.2.1
High-pressure fluid-filled and high-pressure gas-filled Multipoint bonding, 10.2.1
cables, 14.4.2 Sheath bonding, 8.4.1, 10.2.1
Self-contained fluid-filled cables, 14.4.3 Single-point bonding, 10.2.1, 12.2.9
Series impedance measurements, 14.4.5 Bridges, 12.3.4
Sheath-bonding verification tests, 14.4.4 Butt gap, 2.8.3, 4.23.1
Alkylbenzene liquids, 5.5.2, 6.16 Butt space, 5.4.4, 6.5
Alternating current cable, 3.4.2
Ampacity, 11.1.1 C
Ampacity for trenchless installations, 11.2.7 Cable accessories, 2.9, 3.2.4, 8.2–8.4
Ampacity implications for crossing external heat Bonding leads, 8.4.1
sources, 11.2.8 Cable outer sheath/jacket, 8.4.1
Ampacity verification and uprating, 11.5 Cathodic protection and grounding, 2.9.5
Cable data, 11.2.3 Fluid-handling equipment, 2.9.4
Cables in air and tunnels, 11.2.6 Link boxes, 2.9.3, 8.4.1
Calculation principles, 11.2.1 Monitoring, 2.9.6
Calculation procedures, 11.2.5 Sectionalizing, 8.4.1, 10.2.1
Fault current calculations, 11.2.9 Sheath bonding, 8.4.1, 10.2.1
Forced cooling, 11.4 Splices, 2.9.1. 8.2 (See also Joints, Splices.)
Installation data, 11.2.4 Surge diverters, 8.4.1
Neher-McGrath paper, 11.2.1 Terminations, 2.9.2, 8.3 (See also Potheads,
Thermal ohm-meter, 11.2.1 Terminations.)
Transient rating and temperature calculations, 11.3 Cable as a two-port network, 3.7.1
Uprating, 11.5.2 Cable burial, 7.3.6
Utility system data, 11.2.2 Cable components, 3.2.3
Worked examples, 11.6 Conductor, 2.4.1
Ancillary equipment, 3.2.5 Ground continuity conductor, 3.2.3
Armoring, 2.4.4, 5.5.1, 6.10, 6.17 Ground return conductor, 3.2.3
ASTM standards, 5.7.2 Insulation, 2.4.3
Outer layers, shielding, and sheath, 2.4.4
B Semiconducting shield, 2.4.2
Backfilling, 13.7, 13.2.7 Shield conductor, 3.2.3
Basic lightning impulse insulation level. See BIL. Cable design, 3.6
Bedding tapes, 4.17.1 Cable length, 3.6.7
Bell and spigot conduits, 13.4.2, 13.6 Conductor, 3.6.1
Bending radius, 3.6.7, 4.11 Conductor shield, 3.6.2
BIL, 3.3.2, 3.4.2, 5.4.5, 16.3.8 Ground return conductor, 3.6.5
I-1
Index EPRI Underground Transmission Systems Reference Book
Insulation, 3.6.3 Field bends, 13.4.3

Insulation shield, 3.6.4 Joining sections of conduit, 13.4.2
Jacket, 3.6.6 Proof testing, 13.4.4
Cable electrical parameters, 3.7 Construction, cable system, 13.1–13.17
Cable fluid, 6.16 After installation tests, 13.14
Cable installation, 13.8 Backfilling, 13.7
Cable length, 3.6.7 Cable installation, 13.8
Cable maintenance, 15.3.4 Conduit installation for extruded-dielectric cables,
Cable Probabilistic Design Model (CPDM), 4.23.5 13.4
Cable system considerations Cost reduction research, 17.4.1
Impedance calculations, 16.2 Dielectric fluid installations, 13.11
Inductive coordination, 16.5 Direct-buried cable installation, 13.5
Integration of cables into the utility system, 16.3 Fiber-optic cables installation, 13.16
Magnetic fields, 16.4 High-pressure fluid-filled pipe and conduit
Cable System Endurance Model (CSEM), 4.23.6 installation, 13.6
Calculations, cable system designs, 12.4 Horizontal directional drill installations, 13.3.3
Capacitance, 2.3, 2.10.1, 3.6.3, 5.4.4, 5.4.5, 6.8, 11.1.1 Jack-and-bore installations, 13.3.2
Capacitance and reactance effects, 16.3.1 Nitrogen gas installations, 13.12
Casings, 10.3.11 Pipe freezes installation, 13.15
Cathodic protection and grounding, 2.9.5, 10.3 Pressurization plant installations, 13.13
Cathodic protection criteria, 10.3.4 Splicing, 13.9
Cathodic protection for pipe-type cable, 10.3.5 Submarine cable installation, 13.17
Field rectifiers, 10.3.5 Terminations, 13.10
Galvanic protection systems, 10.3.5 Trenching and manholes, 13.2
Polarization cell, 10.3.5 Tunneling, 13.3.1
Resistor-rectifier, 10.3.5 Continuous steady-state rating, 3.5.1
Solid-state isolator, 10.3.5 Continuous vulcanizing, 4.5.2
Voltage-breakdown gap, 10.3.5 Corrosion, 10.3
Charging currents, 2.4.4, 2.10.1, 4.1.2, 4.16.3, 5.5.1, Atmospheric corrosion, 10.3.10
16.3.2 Casings, 10.3.11
Chemical crosslinking process, 4.6.1 Cathodic protection criteria, 10.3.4
Chill or backer ring, 13.6.1 Cathodic protection for pipe-type cable, 10.3.5
Coatings, 10.3.8 Coatings, 10.3.8, 13.6.2
Damage, 15.3.2 Corrosion Coordinating Committees, 10.3.2
Disbonded coatings, 10.3.9 Corrosion prevention, 10.3.3
Coefficient of friction, 5.4.4, 5.4.7 Corrosion theory, 10.3.1
Coefficient of grounding, 16.3.8 Disbonded coating, 10.3.9
Cold dielectric. See Superconducting cable. Dynamic interference, 10.3.1
Conci conductors, 6.3 Electrolytic (stray current) corrosion, 10.3.1
Conduction current, 5.5.2 Galvanic corrosion, 10.3.1
Conductor Resistance R, 3.7.3 Lead-sheathed cable, 10.3.12
Conductor, sector, 3.6.1 Pipe to soil potential surveys, 10.3.7
Conductor, segmental, 3.6.1 Prevention, 10.3.3
Conductor shield, 3.6.2 Static interference, 10.3.1
Pipe-type cables, 5.3.1 Stray current mitigation, 10.3.6
Self-contained fluid-filled, 6.4 Theory, 10.3.1
Conductors, 2.4.1, 3.6.1 Cost considerations, 2.10.5
Conci, 6.3 Critical length, 7.3.1, 11.2.5
Conductor resistance, 3.6.1 Cross-bonding, 8.4.1, 10.2.1
Milliken, 6.2.1, 6.3 Crosslinked polyethylene, 2.5, 4.5–4.9
Pipe-type cables, 5.2 Benefits, 4.5.1
Self-contained fluid-filled cables, 6.3 Challenges, 4.5.2
Stranded round wire conductors, 3.6.1 Continuous vulcanizing, 4.5.2
Stranded shaped wire conductors, 3.6.1 Development of, 1.5
Conduit installation for extruded-dielectric cables, 13.4 Electrical breakdown mechanism, 4.9.1
Conduit materials and types, 13.4.1 Electrical design stresses, 4.8
I-2
EPRI Underground Transmission Systems Reference Book Index
Electrical properties, 4.9.2 Distributed temperature sensing (DTS), 2.9.6, 4.17.7,

Insulation, 4.6 11.5.1
Landmarks, 4.7 Dodecylbenzene cable fluid, 6.16
Physical properties, 4.9.3
Properties, 4.9 E
Thermal properties, 4.9.4 Educational tools, 17.3.1
Cryogenic system. See Superconducting cable. Electrical characteristics of underground cables, 2.3
Cryostat. See Superconducting cable. Electrical stress distribution, 3.6.3
Current rating, 3.5 Electromagnetic field, 3.3
CV. See Continuous vulcanizing. Electrostatic field, 3.3.1
CV lines, 4.20.2–4.20.6 EMF. See Electromagnetic field.
Emanueli cable, 1.2, 1.3
D Emanueli, Luigi, 1.2, 5.4.1
Dc cables. See Direct current cables. Environmental considerations, 2.10.4
Demand calculations, 9.3.2 EPR. See Ethylene propylene rubber.
Design considerations Ethylene propylene rubber, 2.5, 4.10–4.13
Ancillary equipment, 3.2.5 Applications, 4.11.3
Accessories, 3.2.4 Benefits, 4.10.1
Bonding scheme, 3.2.7 Challenges, 4.10.2
Cable, 3.2.3, 3.6 Electrical breakdown mechanism, 4.13.1
Cable electrical parameters, 3.7 Electrical properties, 4.13.2
Cable length, 3.6.7 First production, 1.5
Cable system, 3.2 Formulation, 4.11.1
Conductor, 3.6.1 General description, 4.11
Conductor shield, 3.6.2 Landmarks, 4.12
Current rating, 3.5 Physical properties, 4.13.3
Electromagnetic field, 3.3 Properties, 4.13
Ground return conductor, 3.6.5 Thermal properties, 4.13.4
Hydraulic and pneumatic scheme, 3.2.6 Excavating trenches, 13.2.2
Installation design, 3.2.10 Expulsion calculation, 9.3.1
Insulation, 3.6.3 Extruded cable core, 4.20
Insulation shield, 3.6.4 Extruded-dielectric cables, 2.5
Jacket, 3.6.6 Cable life and design stress, 4.23
Operation, maintenance, and testing, 3.2.11 Conductor construction, 4.14
Power rating, 3.4 Conductor and insulation shield construction, 4.15
Thermal design, 3.2.8 Crosslinked polyethylene, 4.5–4.9
Thermomechanical design, 3.2.9 Ethylene propylene rubber, 4.10–4.13
Dielectric constant. See Relative permittivity. Insulation construction, 4.16
Dielectric constants for insulation materials, Table 16-2 Jacket construction, 4.18
Dielectric fluid Joints, 8.2.2
Installation, 13.11 Manufacture of compounds, 4.19
Maintenance, 15.3.5 Manufacture of extruded cable core, 4.20
Research, 17.2.1 Outer layer construction, 4.17
Dielectric loss, 2.2, 3.3.1, 3.6.3, 4.1.2, 4.10.2, 4.11, 5.4.5, Polyethylene cable, 4.2–4.4
6.2.1 Quality control, 4.21
Dielectric loss factor, 3.6.3, 4.2.1, 5.4.5, Table 6-1 Shipping, 4.22
Dielectric strength, 3.2.6, 4.9.2, 5.4.2, 5.4.4 Terminations, 8.3.2
Direct-buried cable installation, 13.5 Extruders, 4.20.1
Direct current cables, 2.8.2, 3.4.1, 6.18, 7.3
Cable accessories—splices and terminations, 7.3.4 F
Cable design, 7.3.4 Factory production (routine) tests, 14.1.1, 14.3
HVdc underground cables, 7.3.3 Failure repair, 15.5
Insulation design, 7.3.4 Extruded-dielectric, 15.5.1
Selection of dc versus ac cables, 7.3.1 Pipe-type, 15.5.2
Submarine cables, 7.3.2 Self-contained fluid-filled, 15.5.3
Tests, 7.3.5
I-3
Failure repair material, 15.6 Structure grounding, 10.2.2

Extruded-dielectric, 15.6.1 Surge Voltage Limiters, 10.2.1
Pipe-type, 15.6.2 Gutta percha insulation, 1.2
Self-contained fluid-filled, 15.6.3
Fault location, 15.4 H
Features of underground cable systems, 2.1.1 Harmonic resonance, 16.3.7
Feed joint, 6.2.2, 8.2.1 HDD. See Horizontal directional drilling.
Feed tank, 6.15.2 HDPE. See High-density polyethylene insulation.
Ferranti cable, 1.2 High capacitance, 2.3
Ferromagnetic, 3.3.2 High-density polyethylene insulation, 2.5, 4.4.1
Fiber-optic cables, installation, 13.16 High-pressure fluid-filled cable, 2.6. 5.1.1
Field bends First U.S. installation, 1.3
Extruded-dielectric cables, 13.4.3 Pipe and conduit installation, 13.6
High-pressure fluid-filled cables, 13.6.5 High-pressure gas-filled cable, 2.6
Fluids, 6.16 First U.S. installation, 1.3
Fluid-handling equipment, 2.9.4 High-pressure liquid-filled cable, 2.6, 5.1.1
Fluidized Thermal BackfillTM, 12.2.6, 13.2.7 High-pressure oil-filled cable, 2.6, 5.1.1
Forced-cooling, 11.4 High-temperature superconducting cable, 2.8.3.
Axial thermal calculations, 11.4.5 See also Superconducting cable.
Calculating losses, 11.4.3 Cable design options, 2.8.3, 7.4.2
Cooling plant design, 9.5.3 Cold dielectric, 2.8.3, 7.4.2
Pressure-loss calculations, 11.4.6 Demonstration projects, 1.9
Thermal-hydraulic characterization, 11.4.4 Discovery of, 1.7
System components, 9.5.2 Warm dielectric, 2.8.3, 7.4.2
Systems design, 9.6.2 High-voltage direct-current cable
Freezing, 15.3.7 Long Island Sound installation, 1.9
FTBTM. See Fluidized Thermal BackfillTM Horizontal directional drilling, 12.3.3, 13.3.3, 13.2.8
HPFF. See High-pressure fluid-filled cable.
G HPGF. See High-pressure gas-filled cable.
Gas cabinets, 9.4.4 HPLF. See High-pressure liquid-filled cable.
Gas-insulated transmission line, 2.8.1, 7.2 HPOF. See High-pressure oil-filled cable.
Enclosures, 7.2.3 HTS. See High-temperature superconducting cable.
Epoxy spacers, 7.2.2 HTS. See Superconducting cable.
Factory assembled and tested GIL HVdc. See High-voltage direct current cable.
Field assembled and tested GIL Hybrid lines, 16.3.4
Field test, 7.2.7 Hydraulic calculations, 9.3
First U.S. installation, 1.6 Demand calculations, 9.3.2
High-voltage conductor, 7.2.4 Expulsion calculation, 9.3.1
Operation and maintenance, 7.2.8 Pressure drop due to flow, 9.3.3
Sulfur hexafluoride (SF6) gas, 7.2.1 Hydraulic design and equipment, 9.2–9.7
Geometric mean distance, 16.2.3 Engineering design principles, 9.2
Geometric mean radius, 16.2.3 Forced-cooling equipment, 9.5
GIL. See Gas-insulated transmission line. Hydraulic calculations, 9.3
GITL. See Gas-insulated transmission line. Pressurization equipment, 9.4
Gradient, electrical, 4.3, 8.2.5 System design considerations, 9.6
Ground continuity conductor, 3.2.3 Thermal and hydraulic driving forces, 9.2.3
Ground return conductor, 3.2.3 Worked example, 9.7
Grounding, 10.2, 12.2.9 Hydraulic equations, basic, 9.2.2
Alternatives to sheath bonding, 10.2.1 Hydraulic and pneumatic scheme, 3.2.6
Cable jacket, 10.2.1
Design calculation, 10.2.3 I
Link boxes, 10.2.1 ICEA specifications, 2.11
Sectionalizing—shield/sheath interruptions, 10.2.1 IEEE specifications, 2.11, 5.7.3
Sheath bonding, 10.2.1 Impedance, 3.7.4, 3.7.5
I-4
Impedance bonding, 10.2.1 Splices, 12.3.1

Impedance calculations, 16.2 Trenching, 12.3.2
Calculation method for symmetric component series Trenchless installations, 12.3.3
impedance, 16.2.4 Underwater installations, 12.3.5
Series symmetrical component impedances, 16.2.2 Insulation, 2.4.3, 3.6.3
Shunt impedance, 16.2.1 Crosslinked polyethylene, 4.6
Single-conductor cable symmetrical component Electrical stress distribution, 3.6.3
series impedance, 16.2.3 Extruded-dielectric, 4.16
Symmetrical component series impedances, 16.2.6 HVdc, 7.3.4
Worked examples, 16.2.5 Pipe-type, 5.4
Impermeability, 6.5 Polyethylene, 4.3
Impregnant, 1.3, 2.4.4, 5.4, 6.5 Self-contained fluid-filled, 6.5
Impregnated paper, 5.4.2 Insulation coordination, 16.3.8
Induced voltages, 10.2.3 Insulation resistance, 3.6.3, 4.21.4, 5.4.5
Inductance L, 3.7.2 Insulation shield, 3.6.4
Inductive coordination, 16.5 Pipe-type cables, 5.3.2
Calculation of induced voltages and currents, 16.5.2 Self-contained fluid-filled cables, 6.6
Inductive coordination with communication lines, Integration of cables into the utility system, 16.3
16.5.5 Capacitance and reactance effects, 16.3.1
Inductive coordination with fences, 16.5.4 Harmonic resonance, 16.3.7
Inductive coordination with pipelines, 16.5.3 Hybrid lines, 16.3.4
Industry standards and guides, 16.5.1 Insulation coordination, 16.3.8
Railway system inductive coordination, 16.5.7 Load sharing and power flows, 16.3.3
Installation considerations, 2.10.2 Reactive compensation, 16.3.2
Installation System protection, 16.3.9
Cable installation, 13.8 System reliability, 16.3.6
Conduit installation for extruded-dielectric cables, System restoration, 16.3.5
13.4 Ionization factor, 6.12
Dielectric fluid, 13.11
Direct-buried cable installation, 13.5 J
Extruded-dielectric cable installation, 13.8.2 Jack-and-bore, 12.3.3, 13.3.2
Fiber optic cables, 13.16 Jacket, 3.6.6, 4.18. 6.11, 10.2.1
High-pressure fluid-filled pipe and conduit Jam ratio, 12.4.2
installation, 13.6 Jet plow, 7.3.6
Nitrogen gas, 13.12 Joint, semi-stop, 8.2.3
Pipe freezes, 13.15 Joint, stop, 2.9.1, 8.2.1, 8.2.3, 15.5.3
Pipe-type cable installation, 13.8.3 Joints, 2.9.1, 8.2 (See also Splices.)
Pressurization plants, 13.13 Anticorrosion protection, 8.2.1
Research, 17.4 Connectors, 8.2.1
Self-contained fluid-filled cable installation, 13.8.4 Electrical design, 8.2.1
Submarine cable, 13.17 Extruded-dielectric joints, 8.2.2
Installation design, 3.2.10 Function, 8.2.1
Common design considerations, 12.2 General description, 8.2.1
Cable-type selection, 12.2.1 Maintenance, 15.3.6
Grounding/cathodic protection, 12.2.9 Mechanical/hydraulic design, 8.2.1
Mechanical considerations, 12.2.7 Metallic sheath closures, 8.2.1
Monitoring, 12.2.8 Pipe-type joints, 8.2.3
Overhead-to-underground transition, 12.2.5 Self-contained fluid-filled joints, 8.2.4
Rerouting existing lines, 12.2.4 Thermal design, 8.2.1
Route considerations, 12.2.2
Soil thermal properties and special backfill, 12.2.6 K
Surveying, 12.2.3 Kraft paper, 5.1.1, 5.4.2, 6.5
Thermomechanical bending, 12.2.7 Capacitance, 5.4.5
Installation modes, 12.3 Dielectric loss, 5.4.5
Bridges, 12.3.4 Dielectric strength, 5.4.5
I-5
Insulation resistance, 5.4.5 Maintenance or proof tests, 14.1.1, 14.5

Physical properties, 5.4.6 Mandrel, 13.4.2
Manholes, 13.2
L Maintenance, 15.3.3
Laminated paper polypropylene, 5.1, 5.4.3 Manhole installation, 13.2.9
Development of 1.6 Manhole racking, 12.2.7
Electrical properties, 5.4.5 Mass impregnation, 5.4.4
History, 5.4.1 Medium-pressure gas-filled cable
Physical properties, 5.4.6 First U.S. installation, 1.3
Self-contained fluid-filled cables, 6.5 Metallic sheath, 3.6.5, 4.17.3, 4.17.6, 6.9
LDPE. See Low-density polyethylene insulation Microtunneling, 12.3.3
Lead-sheathed cable, 10.3.12 Milliken conductor, 6.2.1, 6.3
Leak management, 15.7 Mineral oils, 5.5.2
Causes of leaks, 15.7.2 Moisture content, 4.17.2, 5.4.4, 6.5
Leak detection techniques, 15.7.3 Monitoring
Leak location techniques, 15.7.4 Distributed temperature monitoring, 12.2.8
Leak prevention, 15.7.8 Leak detection, 12.2.8
PFT leak location method, 15.7.6 Leak location, 12.2.8
Pinpointing methods, 15.7.5 Temperature monitoring, 12.2.8
Link boxes, 2.9.3, 8.4.1, 10.2.1 Monopole cable, 3.4.1
Load factor, 3.5.1, 11.2.1 MPGF. See Medium-pressure gas-filled cable.
Load sharing and power flows, 16.3.3 Multipoint bonding, 10.2.1
Loss factor, Table 5-1, Table 6-1, 11.2.1, 11.2.4
Low-density polyethylene insulation, 2.5, 4.2–4.4 N
Low impedance, 2.3 Nitrogen gas installation, 13.12
Low inductance, 2.3
Low-pressure gas-filled cable O
First U.S. installation, 1.3 Operation of cables, considerations for, 2.10.3
Low-pressure oil-filled cable. See Self-contained fluid- Operation, maintenance, and testing, 3.2.11
filled cable. Operations procedures, 15.2
Low-temperature superconductors. See Overhead-to-underground transition, 12.2.5
Superconducting cable. Oversheathing. See Jacket.
LPGF. Low-pressure gas-filled cable.
First U.S. installation, 1.3 P
LPOF. Low-pressure oil-filled cable. See Self-contained Paper-insulated cable, first uses, 1.3
fluid-filled cable. Partial discharge, 2.9.6, 3.2.5, 4.3
LPP. See Laminated paper polypropylene. PD. See Partial discharge.
LTS. Low-temperature superconductors. See PE cable. See Polyethylene cable.
Superconducting cable. PFT leak location, 15.7.6
Pipe freezes, 13.15, 15.5.2
M Pipe-type cables, 2.6
Magnetic field, 3.3.2, 16.4 Conductors, 5.2
Cables with ferromagnetic components, 16.4.2 General description, 5.1.1
Cables without ferromagnetic components, 16.4.1 Insulation and impregnant, 5.4
Magnetic field management methods, 16.4.3 Joints, 8.2.3
Maintenance, routine, 15.3 Outer layers and filling medium, 5.5
Cable, 15.3.4 Quality control, 5.6
Cable installed in pipe or duct, 15.3.1 Shields, 5.3
Dielectric fluid, 15.3.5 Specifications and standards, 5.7
Joints, 15.3.6 Terminations, 8.3.3
Manhole and joint casings, 15.3.3 Polarity reversal, 3.6.3, 6.18
Pipe coating damage, 15.3.2 Polybutene, 5.5.2
Pressurization system, 15.3.8 Polyethylene cable, 4.2–4.4
Terminations, 15.3.7 Benefits, 4.2.1
I-6
Challenges, 4.2.2 S
First U.S. use, 1.4 SCFF. See Self-contained fluid-filled cable.
Landmarks, 4.4 SCLF. Self-contained liquid-filled cable. See Self-
Polyethylene insulation, 4.3 contained fluid-filled cable.
Polyethylene insulation, 4.3 Screen, 2.8.3, 3.6.3, 4.8, 4.11.2, 5.4.2, 6.6
Water trees, 4.4.2 Sealing ends, 2.9.2
Potheads, 2.9.2 (See also Sealing ends and Sectionalizing, 8.4.1, 10.2.1
Terminations.) Selecting underground cable systems, considerations for,
Power rating, 3.4 2.2, 2.10
PPP. Paper polypropylene. See Laminated paper Cost considerations, 2.10.5
polypropylene. Environmental considerations, 2.10.4
Pressure drop due to flow, 9.3.3 Installation considerations, 2.10.2
Pressure limits, of cable system, 9.2.1 Operations considerations, 2.10.3
Pressurization equipment, 9.4 System considerations, 2.10.1
Computing reservoir volume, 9.4.5 Self-contained fluid-filled cable, 2.7
Gas cabinets, 9.4.4 Cable fluid, 6.16
Installation, 13.13 Characteristics, 6.1
Maintenance, 15.3.8 Components, 6.2.2
Pressurization plants, 9.4.2, 9.6.1 Conductor, 6.3
Reservoir systems, 9.4.3 Conductor shield, 6.4
Protection, system, 16.3.9 Dc operation, 6.18
Proctor density, 12.2.6 Design description, 6.2.2
Pulling tension, 2.9.1 Developments, 6.2.1
Drying and impregnation, 6.8
Q Electrical tests, 6.14
Qualification tests, 3.2.11, 3.5.3, 3.6.3, 4.4.2, 4.7.3, First U.S. installation, 1.3
11.2.5, 14.1.1, 14.3 Historical landmarks, 6.2.1
Quality factor, 14.2.6 Hydraulic and fluid quality tests, 6.15
Insulation 6.5
R Insulation shield, 6.6
Radial water barrier, 3.6.5 Jacketing or oversheathing, 6.11
Rating Joints, 8.2.4
Continuous steady-state rating, 3.5.1 Laying up, 6.7
Current rating, 3.5 Metal sheathing, 6.9
Power rating, 3.4 Reinforcement, 6.10
Short-circuit rating, 3.5.2 Routine testing, 6.12
Reactive compensation, 16.3.2 Shipping, 6.13
Relative permittivity, 3.3.1, 3.6.3, 5.4.5, 5.4.6, 5.5.2, 6.5, Submarine use, 6.17
Table 5-1, Table 5-2, Table 6-1 Terminations, 8.3.4
Reliability, system, 16.3.6 Semiconducting layer, 2.9.1, 4.9.4
Rerouting existing lines, 12.2.4 SF6 gas. See Sulfur hexafluoride gas.
Research in cable technology, 17.2 Sheath, 3.2.7, 3.3.2, 3.6.4, 2.4.4, 4.9.4, 5.4.1, 6.2.1
Cable installation design, 17.4 Sheath bonding, 8.4.1, 10.2.1
Dielectrics, 17.2.1 Alternatives to sheath bonding, 10.2.1
Dry-type terminations and high-stress splices, 17.2.2 Cable jacket, 10.2.1
Superconducting cables, 17.2.3 Design of bonding methods, 10.2.1
SuperGrid, 17.5.1 Link boxes, 10.2.1
Reservoir systems, 9.4.3 Sectionalizing—shield/sheath interruptions, 10.2.1
Computing reservoir volume, 9.4.5 Surge Voltage Limiters, 10.2.1
Restoration, system, 16.3.5 Sheath or surge voltage limiters (SVLs), 2.10.3, 8.3.3,
Route considerations, 12.2.2 8.4.1
Shields, 2.1.1, 2.2, 3.2.3, 3.6.1, 3.6.2, 4.9.4, 5.3, 6.1
Shield conductor, 3.2.3, 4.17.4. 5.3.1, 6.4
I-7
Shipping Superconducting cable, 7.4

Extruded-dielectric cables, 4.22 Accessories, 7.4.3
Self-contained fluid-filled cables, 6.13 Applications, 7.4.1
Shoring/sheeting, 13.2.4 Cable projects, Table 7-5
Short-circuit rating, 3.5.2 Cold dielectric, 7.4.2
SIC. See Relative permittivity. Cryogenic systems, 7.4.3
Sidewall bearing pressure, 7.4.4 Cryostats, 7.4.3
Sieve, 12.2.6 Design, 7.4.2
Single-point bonding, 10.2.1 History, 7.4.1
Skid wires, 2.4.4, 3.2.3, 3.6.4, 5.4.4 Installation, 7.4.4
Skin effect, 3.3.2, 3.6.1, 3.7.3, 11.2.5 Materials, 7.4.1, 7.4.3
Sleeves, 13.6.1 Research, 17.2.3
Soil density, 12.2.6 Standards, 7.4.5
Soil thermal properties and special backfill, 12.2.6 Warm dielectric, 7.4.2
Fluidized Thermal BackfillTM (FTBTM), 12.2.6 SuperGrid, 17.5.1
Thermal Property Analyzer (TPA), 12.2.6 Surge arrester, 2.1, 2.10.1, 5.4.7
Thermal resistivity and diffusivity, 12.2.6 Surge diverters, 8.4.1
Thermal stability, 12.2.6 Surge Voltage Limiters, 10.2.1
Solid dielectric cables. See Extruded-dielectric cables. Surveying, 12.2.3, 13.2.1
Special application cables, 2.8 SVLs. See Surge Voltage Limiters.
Special-purpose tests, 14.1.1, 14.6
Specific inductive capacity. See Relative permittivity. T
Specifications and standards, 2.11 Temperature monitoring, 2.9.6
AEIC specifications, 2.11, 5.7.1 Temporary overvoltage, 16.3.8
ASTM standards, 5.7.2 Terminations, 2.9.2, 13.10
Contact information, 2.11.3 Components, 8.3.1
ICEA specifications, 2.11 Electrical design, 8.3.1
IEEE specifications, 2.11, 5.7.3 Extruded-dielectric cable terminations, 8.3.2, 13.10.1
Listing, 2.11.2 Failure, 15.5.1, 15.5.2
Pipe-type cables, 5.7 Function, 8.3.1
Splices, 2.9.1, 8.2, 12.3.1, 13.9 (See also Joints.) General description, 8.3.1
Extruded-dielectric cables, 13.9.1 High-pressure pipe-type termination, 13.10.2
High-pressure pipe-type cables. 13.9.2 HVdc, 7.3.4
HVdc, 7.3.4 Installation, 8.3.1
Self-contained fluid-filled cables, 13.9.3 Maintenance, 15.3.7
Stop joint. See joint, stop. Mechanical/hydraulic design, 8.3.1
Stray current mitigation, 10.3.6 Pipe-type cable terminations, 8.3.3
Stress cone, 8.2.1 Self-contained fluid-filled cable terminations, 8.3.4,
Submarine cables, 2.8.2, 7.3 13.10.3
Alternating-current cables, 2.8.2, 7.3.2 Specification and standards, 8.3.5
Cable accessories—splices and terminations, 7.3.4 Thermal design, 8.3.1
Cable burial, 7.3.6 Tests, 4.1–4.6
Cable design, 7.3.4 Ac voltage withstand and step-breakdown tests,
Direct-current cables, 2.8.2, 7.3.2 14.2.1
Installation, 13.17 After-installation tests, 13.14, 14.1.1, 14.4
Insulation design, 7.3.4 Compressed-gas-insulated cables, 14.3.6
Major ac cable links, Table 7-2 Dc Hi-Potential tests, 14.2.2
Major dc cable links, Table 7-3 Development tests, 14.1.1
Reliability, 7.3.6 Electrical tests for self-contained fluid-filled cables,
Self-contained fluid-filled cables, 6.17 6.14
Tests, 7.3.5 Extruded-dielectric cables, 14.3.2, 14.3.3
Worldwide experience, 7.3.2 Factory partial-discharge measurements, 14.2.5
Sulfur hexafluoride (SF6) gas, 7.2.1 Factory production (routine) tests, 14.1.1, 14.3
I-8
Field partial-discharge measurements, 14.2.6 Street and highway crossings, 13.2.8

HVdc and submarine cable system, 7.3.5 Survey of route, 13.2.1
Hydraulic and fluid quality tests for self-contained Traffic control, 13.2.5
fluid-filled cables, 6.15 Water removal/well pointing, 13.2.3
Impulse and switching surge tests, 14.2.3 Trenchless installations, 12.3.3
Insulation dissipation factor test, 14.2.4 Horizontal directional drilling, 12.3.3
Joints, 14.3.7 Jack-and-bore, 12.3.3
Longitudinal water penetration tests, 15.6.4 Microtunneling, 12.3.3
Long-term prequalification tests, 14.1.1 Triangular configuration, 16.2.5, 16.4.3
Maintenance or proof tests, 14.1.1, 14.5 Tunneling, 13.3.1
Mechanical bending tests, 16.6.2 Type or qualification tests, 14.1.1, 14.3
Pipe-type cables, 14.3.4, 14.3.5
Special-purpose tests, 14.1.1, 14.6 U
Standards organization, 14.1.3 Underground Transmission Systems Reference Book,
Terminations, 14.3.8 17.3.1
Test standards and guides, 14.1.2 Underground Transmission Workstation, 17.3.1
Thermomechanical bending tests, 14.6.3 Underwater installations, 12.3.5
Type or qualification tests, 14.1.1, 14.3 Uprating, 11.5.2
Thermal backfill, 2.9.1, 11.2.1, 11.2.4, 12.2.6, 16.4.3 Dielectric fluid circulation, 11.5.2
Thermal conductivity, 4.9.4, 5.5.2 Heat pipes, 11.5.2
Thermal design, 3.2.8 Quasi-dynamic, 11.5.2
Thermal probe, 15.7.4 Real-time dynamic ratings, 11.5.2
Thermal resistivity, 3.2.8, 4.9.4, 5.4.6, 11.2.1, 11.2.4,
11.2.5 V
Thermomechanical bending, 12.2.7, 17.3.2 Voltage source converter, 7.3.3
Thermomechanical design, 3.2.9 Voltage stress, 5.5.2, 6.5
TMB. See Thermomechanical bending.
Transient rating and temperature calculations, 11.3 W
Trefoil, 3.7.1, 6.17, 12.2.8 Warm dielectric. See Superconducting cable.
Trenching, 12.3.2, 13.2 Water blocking tapes, 4.17.2
Backfilling, 13.2.7 Water removal/well pointing, 13.2.3
Bed of trench, 13.2.6 Water trees, 4.4.2, 4.9.1
Excavating, 13.2.2 Weight correction, 12.4.2
Manhole installation, 13.2.9
Pavement restoration, 13.2.10 X
Shoring/sheeting, 13.2.4 XLPE. See Crosslinked polyethylene.
I-9
Export Control Restrictions The Electric Power Research Institute (EPRI), with major
Access to and use of EPRI Intellectual Property is granted with the locations in Palo Alto, California; Charlotte, North Carolina; and
specific understanding and requirement that responsibility for ensur- Knoxville, Tennessee, was established in 1973 as an independent,
ing full compliance with all applicable U.S. and foreign export laws nonprofit center for public interest energy and environmental
and regulations is being undertaken by you and your company. This research. EPRI brings together members, participants, the Institute’s
includes an obligation to ensure that any individual receiving access scientists and engineers, and other leading experts to work
hereunder who is not a U.S. citizen or permanent U.S. resident is collaboratively on solutions to the challenges of electric power. These
permitted access under applicable U.S. and foreign export laws and solutions span nearly every area of electricity generation, delivery,
regulations. In the event you are uncertain whether you or your com- and use, including health, safety, and environment. EPRI’s members
pany may lawfully obtain access to this EPRI Intellectual Property, you represent over 90% of the electricity generated in the United States.
acknowledge that it is your obligation to consult with your company’s International participation represents nearly 15% of EPRI’s total
legal counsel to determine whether this access is lawful. Although research, development, and demonstration program.
EPRI may make available on a case-by-case basis an informal as-
Together...Shaping the Future of Electricity
sessment of the applicable U.S. export classification for specific EPRI
Intellectual Property, you and your company acknowledge that this
assessment is solely for informational purposes and not for reliance
purposes. You and your company acknowledge that it is still the ob-
ligation of you and your company to make your own assessment
of the applicable U.S. export classification and ensure compliance
accordingly. You and your company understand and acknowledge
your obligations to make a prompt report to EPRI and the appropriate
authorities regarding any access to or use of EPRI Intellectual Prop-
erty hereunder that may be in violation of applicable U.S. or foreign
export laws or regulations.
Program:
Underground Transmission
© 2007 Electric Power Research Institute (EPRI), Inc. All rights reserved. Electric Power
Research Institute, EPRI, and TOGETHER...SHAPING THE FUTURE OF ELECTRICITY are
registered service marks of the Electric Power Research Institute, Inc.
Printed on recycled paper in the United States of America
1014840

3420 Hillview Avenue, Palo Alto, California 94304-1338 • PO Box 10412, Palo Alto, California 94303-0813 USA
800.313.3774 • 650.855.2121 • askepri@epri.com • www.epri.com

Libro Verde

Uploaded by

Copyright:

Available Formats

Libro Verde

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Libro Verde

Uploaded by

Copyright:

Available Formats

EPRI Underground Transmission Systems

Place Image Here

Final Report, March 2007

EPRI Project Manager

ELECTRIC POWER RESEARCH INSTITUTE

(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I)

(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER

ORGANIZATION(S) THAT PREPARED THIS DOCUMENT

Electric Power Research Institute

This report was prepared by

Electric Power Research Institute

Results & Findings

Challenges & Objectives

The Electric Power Research Institute is pleased to AUDIENCE AND SCOPE

ACKNOWLEDGMENTS significantly as both an author and a reviewer. Finally, I

Allen MacPhail, BC Hydro Howard Jones, American Electric Power (AEP)

Frank Musick, WA Chester, LLC (WAC) Takashi Kojima, BC Hydro

Deepak Parmar, Geotherm, Inc. Richard Raymond, CenterPoint Energy

Mohammad A. Pasha, The United Illuminating Erich F. Schoennagel, CenterPoint Energy

Mark R. Smith, Georgia Power Company Rachel Mosier, Northeast Utilities

Chapter 1 Historical Perspective 2.10 SELECTING CABLE TYPES . . . . . . . . . . . . . . . . . . . . 2-17

1.8 PROGRESS THROUGH THE 1990s . . . . . . . . . . . . . . 1-10

5.4 INSULATION AND IMPREGNANT . . . . . . . . . . . . . . . . .5-5 6.16 CABLE FLUID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-16

6.7 LAYING UP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-9

6.8 DRYING AND IMPREGNATION . . . . . . . . . . . . . . . . . .6-10 Chapter 8 Cable Systems Accessories

6.10 REINFORCEMENT. . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-11 8.2 JOINTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-3

Chapter 9 Hydraulic Design and Equipment Chapter 11 Ampacity

9.2 ENGINEERING DESIGN PRINCIPLES . . . . . . . . . . . . .9-2 Ampacity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2

Chapter 12 Installation Design 13.5 DIRECT-BURIED CABLE INSTALLATION . . . . . . . . . 13-16

Chapter 14 Testing Cable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-5

14.4 AFTER-INSTALLATION TESTS . . . . . . . . . . . . . . . . .14-24 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-24

Chapter 17 The Future of Underground Transmission

17.3 IMPROVEMENTS IN DESIGN AND OPERATION . . . . 17-5

CHAPTER 1 Historical Perspective

This chapter presents an overview of the history of underground transmission—from the

1.1 INTRODUCTION cation. In 1812, Baron Schilling, a Russian, fired a mine

The first full-size installation of low-pressure, gas-filled

REFERENCES EPRI. 1992. Underground Transmission Systems Refer-

CHAPTER 2 Cable System Types

Chapter 2 summarizes basic components of ac transmission cable systems, describes

This chapter is an overview of underground transmission cable systems; subsequent

G. Allen MacPhail graduated from the University of British Columbia,

2.1 INTRODUCTION by laying concrete cover slabs directly above buried

Extruded-dielectric cable installations in North 2.6 PIPE-TYPE CABLES

Figure 2-1 345-kV XLPE-insulated cable (courtesy of

The three phases making up a line are pulled together

Figure 2-4 Two-pipe trench installation.

Several names have been used to describe this cable: gas-

GIL cable was used for a 525-kV underground applica-

The GIL system is inherently simple. The conductor

Three types of submarine power cable can be considered

Self-contained, fluid-filled cables are very similar to the

Figure 2.9 shows a warm-dielectric HTS cable produced

All splices consist of a connector to join adjacent

Figure 2-10 Splice components.

Figure 2-11 Extruded-dielectric composite termination.

Table 2-1 General Comparison Among Cable Types

Transition requirements are significantly different for

Table 2-2 Typical Electrical Characteristics—345-kV